Section 4:

Technology, its Implications, and Inequality


The essays in this section consider the inter-relationship between technology, entrepreneurship, and income inequality. Philip E. Auerswald’s essay, “The Great Man-Machine Debate,” opens this section with a discussion of the implications of advances in automation for work and inequality. Technology, he explains, will displace humans in some areas of work but create new types of human work elsewhere, just as technological change has in the past. Entrepreneurial opportunities will be found in connectivity, training, localizing/individualizing, and creating. He argues that we need to rethink the (mis)match of education policy and social policy with the emerging new world of work, entrepreneurship, and automation, and we need to improve the way we measure productivity and innovation. Solutions to the problem of digital disruption, he concludes, involve expanding meaningful work rather than creating jobs involving digital skillsets.

In their essay, “Choosing a Future in the Platform Economy: The Implications and Consequences of Digital Platforms,” John Zysman and Martin Kenney focus on the repercussions of the new Platform Economy that is emerging from the Internet, algorithmic revolution, and cloud computing. This Platform Economy, they explain, has major implications for how value is created, how work is organized, and trends in entrepreneurship. We may see more innovative entrepreneurship, but precarious employment also may result. As our policy choices will influence the shape of the Platform Economy, we need to rethink our social policy, especially the provision of employee benefits and social insurance, as well as competition policy, taxation, and intellectual property.

While most of the other essays in this section look at the impact of technology on income inequality, Elisabeth Jacobs considers the reverse—the implications of economic inequality for entrepreneurship – in her essay, “What Do Trends in Economic Inequality Imply for Innovation? A Framework for Future Research and Policy.” Income inequality, she explains, is contributing to a slowdown in overall business creation, and the gap in entrepreneurial outcomes across income quintiles is caused by educational inequities. The middle-class has experienced rising time pressures, liquidity constraints, and economic insecurity, all of which have increased risk aversion. To mitigate these effects, we need to take advantage of opportunities to connect low-income students to mentors, employers, and skill-based programs. We need to see early childhood investment as a low-income, middle-class, and entrepreneurship policy priority. And, we need to look at tax reform, corporate governance reform, and social insurance policy that can reduce the risk inherent in entrepreneurship.

Phillip Longman’s paper, “Wealth and Generations,” focuses specifically on the widening generational wealth gap that is part of the broader rise in inequality. This gap, he argues, hurts entrepreneurship and has contributed to lower rates of entrepreneurship among people between the ages of twenty and thirty-four years old. To close this gap, he recommends that policy focus on asset-building: housing subsidies need to be redirected, antitrust enforcement needs to be changed, and higher education costs must be reduced.

Finally, in “Firms as Drivers of Growth and (In-)Equality,” Christian Moser considers the implications of differences among firms for income inequality in the United States and Brazil. While it appears that differences among firms might help explain rising income inequality in the United States, in Brazil it has been established that changes between firms help explain falling inequality. Characteristics of firms, he finds, can have important distributional functions and consequences, and policy needs to help firms adopt more productive technologies and practices, access export markets, and reduce barriers to worker mobility.

The Great Man-Machine Debate

By Philip E. Auerswald Author and Associate Professor at the School of Policy, Government, and International Affairs George Mason University
Philip E. Auerswald
Philip E. Auerswald


In June 2015, a group of leading academics, entrepreneurs, and technologists issued an Open Letter on the Digital Economy. The letter begins with the assertion that “the digital revolution is the best economic news on the planet,” but goes on to note that the emergence of the digital economy has not succeeded in solving old problems related to economic inclusion, and is likely to be creating new ones: “The majority of U.S. households have seen little, if any, income growth for more than twenty years, the percentage of national income that’s paid out in wages has declined sharply in the U.S. since 2000, and the American middle class, which is one of our country’s great creations, is being hollowed out.” The letter concludes that, “we can only create a society of shared prosperity if we update our policies, organizations, and research to seize the opportunities and address the challenges brought by these tools.”1 The letter proposed specific policies in five areas: education, infrastructure, entrepreneurship, immigration, trade, and basic research.

In this paper, I seek to explore the problem of digital disruption posed in the Open Letter from five perspectives: history, measurement, structure, solutions, and redefinition. I will argue that, while the framing of the problem as articulated in the Open Letter is correct, the set of proposed policy solutions is outdated. My conclusion will be that solutions to the problem of digital disruption lie not in creating “jobs” involving digital skillsets, but rather in expanding opportunities for meaningful work involving human skillsets.

History: “This Time Will Be Different” . . . ?

The Open Letter on the Digital Economy comes fifty-one years after a comparably distinguished group that dubbed itself the Ad Hoc Committee on the Triple Revolution—a group that included Linus Pauling (recipient of both the Nobel Prize in Chemistry and the Nobel Peace Prize) and Gunnar Myrdal (a future Nobel recipient)—published a similar, but more strongly worded, manifesto, warning of dangers associated with the emergence of the “Cybernation Revolution.” The “underlying cause of excessive unemployment is the fact that the capability of machines is rising more rapidly than the capacity of many human beings to keep pace,” the Ad Hoc Committee asserted. As a consequence, “a permanent impoverished and jobless class is established in the midst of potential abundance.” Echoing contemporary analysts, the Ad Hoc Committee noted that the unemployment rate of 5.5 percent at the time failed to capture the extent of under-employment and attrition of the labor force: “Even more serious is the fact that the number of people who have voluntarily removed themselves from the labor force is not constant but increases continuously. These people have decided to stop looking for employment and seem to have accepted the fact that they will never hold jobs again.”2 What followed, of course, was a decade in which GDP growth was not only strong by historical standards, but the poverty rate in the United States fell precipitously—from 15 percent in 1964 to 9.1 percent in 1978.

Going back further still, George Gordon Byron—the poet Lord Byron—described the adverse consequences of the introduction of new machinery in his first remarks before the House of Lords, delivered on February 27, 1812. The debate on that date concerned the Frame-Work Bill, which proposed that the destruction of mechanized looms known as “stocking frames” be deemed a capital felony and, hence, a crime punishable by death. “By the adoption of one species of frame in particular, one man performed the work of many, and the superfluous laborers were thrown out of employment,” Byron began. “Yet it is to be observed that that work thus executed was inferior in quality; not marketable at home, and merely hurried with a view to exportation.” While the domestic market could not absorb the significant increase in output enabled by the mechanized loom, overseas markets could. Byron made the claim that disruption of trade due to the Napoleonic Wars intensified the waste and suffering due to the mechanization of production: “Although the adoption of the enlarged machinery in that state of commerce which the country once boasted might have been beneficial to the master without being detrimental to the servant; yet, in the present situation of our manufactures, without a prospect of exportation, with the demand for work and workmen equally diminished, frames of this description tend materially to aggravate the distress and discontent of the disappointed sufferers.”3 What followed was the Industrial Revolution and, after roughly 1850, an unambiguous improvement in human wellbeing that has spanned the globe and redefined the human experience.

Lord Byron’s speech before the House of Lords on the subject of the Frame-Work Bill posed a fundamental question that has endured for two centuries in the core of economic theory as well as in politics: is the adoption of “enlarged machinery” in fact “beneficial to the master without being detrimental to the servant” in some, or even most, cases? Or is it instead the case, as Bryon asserted, that mechanization results only “in the enrichment of a few individuals” while it throws “the workman out of employment, and the laborer unworthy of his hire?” This question has, over time, engaged not only one of England’s greatest poets, but also some of the greatest contributors to the history of ideas: Gottfried Wilhelm Leibniz, David Ricardo, Ada Lovelace, John Maynard Keynes, Norbert Weiner, and Herbert Simon notable among them. All of these people grappled with the difficult balance the human societies must always find between the benefits of technologically enabled disruption and their short-terms costs—The Great Man-Machine Debate.

For all the complexities of The Great Man-Machine Debate, its dynamics have, over the past century at least, settled into a fairly straightforward rhythm: One side cites historical evidence demonstrating that, in the past, technology-enabled disruptions have always ended up boosting both employment and output, so we have little to fear about the future; the other side counters with some variant of “this time is different.” The Great Man-Machine Debate raged in the 1930s, involving not just Keynes but other greats of the economics profession, including Joseph Schumpeter, John Hicks, and Paul Douglas (of Cobb-Douglas fame); it raged again in the 1950s and ’60s, at the time the Ad Hoc Committee issued its manifesto, prompting scores of Congressional hearings and the appointment of multiple Presidential Commissions. And it is raging again today, prompted by a combination of macro-economic circumstance (the 2008 recession and its aftermath) and new scholarship. Every time The Great Man-Machine Debate has recurred, the same intellectual contest has taken place: “History Proves We’ll Be Fine” vs. “This Time is Different.”

Measurement: All Trends Are Not Created Equal

To update The Great Man-Machine Debate for the age of digital disruption—if not to endeavor to resolve it definitively—we first need to update evidence of the existence of a problem. That task is complicated by the multiplicity of economic indicators with which the process of digital disruption is associated. Focusing on the United States, these include:

  • GDP growth
  • Productivity growth
  • Labor’s share of national income
  • Stock market index levels
  • Income and wealth inequality
  • Share of the economy comprised of small business
  • Rates of firm entry and exit
  • Real wage growth

More important than the direction taken by the indicators in the past decade is their relevance to the specific societal challenges articulated by Lord Byron in 1812, the Ad Hoc Committee of the Three Changes in 1964, and the signatories to the Open Letter on the Digital Economy.

  • Gross Domestic Product (GDP) and its growth rate. GDP measures flows of goods and services in the aggregate. It cannot capture changes in how production occurs or in the distribution of returns from production. It is not relevant in this context.4
  • Labor’s share of income (in an aggregate production framework). More than a half century ago, Joan Robinson decisively won her debate with Paul Samuelson and Robert Solow, having opened it with the assertion that “The production function has been a powerful instrument of mis-education.”5 For some technical (but significant) reasons first articulated by Robinson (1953–1954), the measurement of capital was never very good; furthermore, for some commonsensical (and even more significant) reasons articulated by McCloskey (2014) and Weil (2015), absent good approaches for measuring human capital and intangibles, the measurement of capital is not getting any better. Indeed, the concern that motivated Robinson to begin with was that the reification of the production function (particularly in the unit-elasticity of substitution, Cobb-Douglas form favored by macroeconomists) distracted attention from issues related to technical change: “By concentrating upon the question of the proportions of factors [in a given state of technical knowledge, the production function] has distracted attention from the more difficult but more rewarding questions of the influences governing the supplies of the factors and of the causes and consequences of changes in technical knowledge.” While potentially relevant to the Great-Man Machine debate, measures of labor share of income are not ones that can be considered uncritically.6
  • Economy-wide productivity measures. Summarizing mostly well-known and accepted critiques, Amar Bhidé recently argued that aggregate productivity measures are “incorrigibly implausible proxies for the process one might think of as innovation. They are, in fact, bunk.”7 If Bhidé is correct (I believe he is), then the widely discussed mystery of the productivity slowdown is, in fact, a mystery of how indicators that lose meaning persist in use.8 Whatever innovation might be, it is not the same thing as productivity measured in the aggregate.
  • Stock market valuations. Equities are an asset class. As such, stock market valuations are determined in an equilibrium with other asset classes. While labor market dynamics play some role in the trends of stock market valuations, that role is minimal. Trends in stock market valuations, on their own, are, at best, minimally relevant in this context.
  • Income and wealth inequality. As Jones and Kim (2014) document, income and wealth inequality statistics are driven almost entirely by changes in the concentration of wealth among top income earners and wealthy families. Furthermore, as McCloskey (2014) has described, while it is indeed the case that wealth inequality is back to levels not seen since the Gilded Age, that fact alone is not sufficient to conclude that the wealth distribution is problematic. Looking at Figure 1, drawn from Piketty (2014), one might reasonably ask (as Piketty does) why wealth inequality has grown since the 1980s. But one might even more reasonably ask how wealth inequality attained such low levels in the 1950s–1980s to begin with. Piketty has an answer to that. On p. 237 of Capital in the Twenty-First Century,he states: “[T]he two world wars, and the public policies that followed from them, played a central role in reducing inequalities in the twentieth century. There was nothing natural or spontaneous about this process...” He leaves out the 1918–1919 Spanish Flu (which, by plausible estimates, killed more Americans than WWI and WWII combined) and the Great Depression (which had an effect on accumulated wealth in the United States). As Piketty states, there was “nothing natural or spontaneous” about the process of getting to the 1950s–1980s levels of inequality in the United States. As a consequence, there is no reason to hold that era up as the point of reference for the economy going forward.9 Furthermore—and perhaps more to the point for this workshop—to the extent that entrepreneurial opportunity driven by digital disruption is (as many observe) leading to a concentration of wealth among platform owners (Google, Facebook, Uber, and so forth), then digital disruption is as likely to be upsetting as reinforcing the persistence of inherited wealth. For these and other reasons, the relationship between income and wealth inequality on one side and the dynamics of digital disruption on the other is not a simple one.
  • Share of the economy comprised of small business and rates of firm entry and exit. While these indicators of economic activity are different, they have in common one fact: They are clearly and consistently defined and measured. Trends in the share of the economy comprised of small business and in rates of firm entry and exit are important economic indicators that relate directly to digital disruption.10
  • Real wages. Nominal wages are reported directly by employers to the federal government. The Bureau of Labor Statistics calculates the Consumer Price Index employing an inherently imperfect, but nonetheless rigorously considered and regularly updated, methodology. Real wages mean something, and their trend matters for discussions of the impact of digital disruption.

One could continue in this vein with other indicators, but the bottom line would be the same: all indicators are not created equal. As a general rule, aggregate indicators that derive from the aggregate production function and fail to account adequately for human capital and intangibles are of decreasing relevance in the digital age. Microeconomic indicators based directly on transactional reporting remain most meaningful.

Of course, we as economists should not take the decreasing relevance of treasured methodologies of aggregate measurement as a rebuke. As I argue with co-authors in Agwara, Auerswald, and Higgingotham (2014), “just as the advent of science-based innovation motivated an earlier generation of economists to create new theoretical frameworks and analytic techniques to understand the rate and direction of technical change, so the advance of the algorithmic frontier is challenging the current generation of economists to respond in a like manner.” If what we are measuring matters less, then we are presented with an opportunity to define new measures that matter more.

Structure: The Bifurcation is Near

To come up with measures of economic outcomes that are relevant for the digital age, we need to understand better the structure of change driven by digital technologies.

From an economic standpoint, the argument made by the cosignatories of the Open Letter on the Digital Economy can be summarized as follows:

  • The power of technology is growing at an exponential rate.
  • Technology nearly perfectly substitutes for human capabilities.
  • Therefore, the (relative) power of human capabilities is shrinking at an exponential rate.

If they are correct, we should be deeply worried indeed about the process of digital disruption.

In sharp contrast, Kurzweil (2005) famously argued that the exponentially increasing power of technology — particularly, though not exclusively, digital computing technologies — will trigger epochal discontinuity in the human experience. From an economic standpoint, Kurzweil’s argument is comparably straightforward:

  • The power of technology is growing at an exponential rate.
  • Technology nearly perfectly complements human capabilities.
  • Therefore the (absolute) power of human capabilities is growing at an exponential rate.

Like many others, Kurzweil argues that “only technology can provide the scale to overcome the challenges with which human society has struggled for generations.” But he goes further, tracing the arc of technologically enabled progress forward into the immediate future to sketch the outlines of “The Singularity,” which “will result from the merger of the vast knowledge embedded in our brains with the vastly greater capacity, speed, and knowledge-sharing ability of our technology, [enabling] our human-machine civilization to transcend the human brain’s limitations of a mere hundred trillion extremely slow connections.” When it comes to algorithmically empowered robots taking our jobs, Kurzweil’s prescription is straightforward: If you can’t beat ’em, join ’em — maybe even literally, in cyborg fashion.

These are not the only two options, however. In addition to the limiting-case arguments offered by Rifkin and Kurzweil, a third line of argument is possible:

  • The power of technology is growing at an exponential rate.
  • Technology only partially substitutes for human capabilities.
  • Therefore the (relative) power of human capabilities is shrinking at an exponential rate for those categories of work that can be performed by computers, and not in others.
The best evidence to support this line of argument comes from the labor market studies conducted by MIT economists David Autor, Daron Acemoglu, and Frank Levy, and Harvard economist Richard Murnane, in dozens of papers and one book written in various combinations and with co-authors over the past dozen years. In a seminal 2003 paper published in the Quarterly Journal of Economics, Autor, Levy, and Murnane summarize their findings as follows:11
We contend that computer capital (1) substitutes for a limited and well-defined set of human activities, those involving routine (repetitive) cognitive and manual tasks; and (2) complements activities involving non-routine problem solving and interactive tasks. Provided these tasks are imperfect substitutes, our model implies measurable changes in the task content of employment, which we explore using representative data on job task requirements over 1960–1998. Computerization is associated with declining relative industry demand for routine manual and cognitive tasks and increased relative demand for non-routine cognitive tasks.

From the Autor, Acemoglu, Levy, and Murnane perspective, the impact of digital disruption on the future of work depends critically on the nature of the work itself — in other words, the “how” of production and not just the “what.” Tasks that are routine and can be easily encoded will be performed by computers, where those that are not will continue to be performed by people.

Herbert Simon understood the fundamentals of this technology-induced bifurcation in work more than half a century ahead of the most recent writing on digital disruption.  In a paper written in 1960, he makes the following startling claim:
The fact that chess programs, theorem-proving programs, music-composing programs, and a factory-scheduling program now exist indicates that the conceptual mountains have been crossed that barred us from understanding how the human mind grapples with everyday affairs. It is my conviction that no new major ideas will have to be discovered to enable us to extend these early results to the whole of human thinking, problem solving, and decision-making activity.

Well ahead of his peers, Simon understood that “automation” would extend well beyond the factory floor:
We should not make the simple assumption that the higher status occupations, and those requiring the most education, are going to be the least automated. There are perhaps as good prospects technically and economically for automating completely the job of a physician, a corporate vice-president, or a college teacher, as for automating the job of a man who operates a piece of earth-moving equipment.

Yet, Simon also understood intuitively what research since has demonstrated: as the power of computers increased, work would bifurcate. Computers would perform those tasks most readily subject to algorithmic definition while humans would perform those tasks most resistant to algorithmic definition:
The change in the occupational profile depends on a well-known economic principle, the doctrine of comparative advantage. It may seem paradoxical to think we can increase the productivity of mechanized techniques in all processes without displacing men somewhere. Won’t a point be reached where men are less productive than machines in all processes, hence economically unemployable?

The paradox is dissolved by supplying a missing term. Whether man or machines will be employed in a particular process depends not simply on their relative productivity in physical terms, but on their cost, as well. And cost depends on price. Hence—so goes the traditional argument of economics—as technology changes and machines become more productive, the prices of labor and capital will adjust themselves as to clear the market of both. As much of each will be employed as offers itself at the market price, and the market price will be proportional to the marginal productivity of that factor.

While this description of comparative advantage derives directly from the classical economists,12 Simon advances upon the conventional story of comparative advantage very significantly with a very subtle decision to use the word “process,” rather than occupation, in this critical passage:
By operation of the market price, manpower will flow to those processes in which productivity is high relative to the productivity of machines; it will leave those processes in which its productivity is relatively low. . . .

This leads to a situation in which any technological unemployment that occurs is transient, and the economy is generally at full employment, but, importantly, where
full employment does not necessarily mean a forty-hour week, for the allocation of productive capacity between additional goods and services and additional leisure may continue to change as it has in the past. Full employment means that the opportunity to work will be available to virtually all adults in the society...

Simon also directly addresses the humanization of work (p. 27):
Automation does not mean “dehumanizing” work. On the contrary, in most actual instances of recent automation, jobs were made, on the whole, more pleasant and interesting, as judged by the employees themselves, than they had been before.

Simon foresees a process by which the processes of work will bifurcate, with humans focusing on those tasks in which their (our) skills are comparatively strong, and computers focusing on those tasks in which their skills are comparatively strong:
We conclude that human employment will become smaller relative to the total labor force in those kinds of occupations and activities in which automatic devices have the greatest advantage over humans; human employment will become relatively greater in those occupations and activities in which automatic devices have the least comparative advantage.

He continues (p. 38):
In the entire occupied population, a larger fraction of members than at present will be engaged in occupations where “personal service” involving face-to-face human interactions is an important part of the job. I am confident in stating this conclusion; far less confident in conjecturing what these occupations will be.

At every stage in this process, technology has not only partially substituted for human capabilities, but also generated new kinds of human requirements and capabilities.13

These disruptions follow a predictable pattern: The creation of a new, high-volume, low-price option creates a new market for a low-volume, high-price option. Every time this happens, the introduction of a new technology forces a bifurcation of markets and of work. Innovations begin as ideas, and end as algorithms. In each stage, the works associated with the innovation bifurcates, in the manner described by Simon.

In assessing the claims made by Simon a half century ago, we have one advantage: an additional fifty years of data. What do the data show us? The four occupations to have shown the greatest decline since 1850 include:

  • Blacksmiths
  • Shoemakers
  • Sailors and deckhands
  • Cabinetmakers

In other words, human employment has become smaller relative to the total labor force in those kinds of occupations and activities in which automatic devices have the greatest advantage over humans. (In the case of “blacksmiths,” the automatic devices in question were machine tools and, indirectly, the automobile; in the case of “sailors and deckhands,” the automatic device was ever-larger cargo ships, enabled by innovations in engineering, in particular that of containerized shipping.)

What employment categories have grown the most rapidly since 1850?

  • Cooks and bakers
  • Nurses
  • Cashiers
  • Actors, artists, and musicians14

Human employment has grown most rapidly in those occupations and activities in which automatic devices have the least comparative advantage, notably including “occupations where ‘personal service’ involving face-to-face human interactions is an important part of the job.

While hardly conclusive, these data points in combination with the AALM work strongly suggest that Simon was correct. The answer to the question, “What can humans do better than robots?” may turn out to be fairly straightforward: Humans are better at being human.

Solutions: From “Scale-Up” to “Scale Out” Innovation

For much of the last decade, the emphasis of entrepreneurship and innovation policy has been on “scale up” innovation: rapidly growing firms and the economic growth they generate. In the digital age, metrics of success must necessarily expand to focus (at least) equally on “scale out” innovation: innovation that creates opportunities for meaningful work and reliable livelihoods for as many people in the economy as seek such work.

Compressing a lengthy treatment of the solution space, I propose that the majority of “scale out” innovations will take one of four forms:

  • Connecting (overcoming information asymmetries that limit access of trained people to existing work opportunities—e.g., platforms like Uber and AirBnB; initiatives to facilitate veteran employment)
  • Training (skills mismatch issues—e.g., GeneralAssembly, Code Academy, Girls Who Code; initiatives to create pathways to employment for marginalized youth)
  • Localizing/Individualizing (farm-to-table, urban gardening, craft furniture and clothing; initiatives to shift the culture to value authenticity and local production)
  • Creating (fundamental novelty—e.g., health care to the home, commercial space travel, space mining; initiatives to increase the frequency of novelty in the economy that makes use of new human capabilities)

Tools available to get these solutions include:

  • Platforms (facilitating connections, fragmenting work)
  • Analytics (improving understanding)
  • Insurance (de-risking the twenty-first-century economy, e.g., freelancer’s union and, taking a different angle, guaranteed minimum income or an expanded earned income tax credit)
  • Space (creating density, ref cities, libraries, co-working spaces, virtual networks, etc.)
  • Imagination (getting beyond jobs)

The transition to the digital economy is, above all, defined by a shift from the “what” of human productive activity to its “how” and “why.”

Redefinition: From “Permanent Income” to “Dynamic Purpose”

The primary value of the foregoing analysis is to establish the context for a redefinition of the problem posed in The Great Man-Machine Debate. How do we get there?

Consider the following: In the immediate aftermath of the 2008 financial crisis in 2008, economists devoted a great deal of attention to the shortcomings of macroeconomic and financial models that, at best, failed to predict the breakdown, and, at worst, may have helped to bring it about. Hyman Minsky’s “financial instability hypothesis,” to which few previously had paid much attention, was newly celebrated; Eugene Fama’s “efficient-market hypothesis” was newly questioned. Yet, as events have unfolded, the profession has begun to take more seriously the structural factors that are shaping the twenty-first-century economy and driving economic outcomes that go beyond the business cycle. From the standpoint of the reconsideration of theory, this means shifting attention from macroeconomics to microeconomics and rethinking fundamental models of both consumption and production.

High on the list of models ripe for reconsideration is the “permanent income hypothesis,” introduced into the field of economics by Milton Friedman in 1957 in a book titled, A Theory of the Consumption Function. The idea, as we all know, is simple: Early in life, we as consumers optimize our lifetime earnings by going into debt to invest in education; education delivers the skills that form the foundation for a career. Early in our working lives, consumers stop investing in education and start to save. We keep saving increasing fractions of our income until, all at once, we retire. At that point, we spend down our savings, timing the depletion of savings to coincide perfectly with the depletion of … well, our lives.

The model Friedman developed of the arc of a human life is as technically sound today as it was in 1957. Furthermore—somewhat like the efficient markets hypothesis, which was also developed at the University of Chicago at about the same time—the permanent income hypothesis has become encoded in the operating system of the economy in such fundamental ways that we barely notice its influence. From the Pell Grants to the 401(k), the experience of consumers from youth to death remains framed by the notion that institutions are sufficiently slowly changing and we are sufficiently short-lived that we can invest (one time only) in education at the front of our lives to reap a reward that we ultimately enjoy at the end of life. Predictable and familiar policy prescriptions follow:

  • Too much student debt and too few quality jobs for recent graduates? . . . Need more and better education.
  • Too much unemployment and too little stability in the labor market? . . . Need to spend more on worker protections.
  • Too many retired people and too little saving? . . . Need more and better health insurance.

In a recent column for The New York Times, Robert Shiller wrote: “Most people complete the majority of their formal education by their early 20s and expect to draw on it for the better part of a century. But a computer can learn in seconds most of the factual information that people will get in high school and college, and there will be a great many generations of new computers and robots, improving at an exponential rate, before one long human lifetime has passed.” Colleges and universities have yet to respond adequately to these changes, Shiller concluded. “We will have to adapt as information technology advances . . . . We must continually re-evaluate what is inherently different between human and computer learning.”

Shiller is right: We need to update our thinking about the function of higher education. We also need to update our thinking about workforce training, retirement, and aging to fit the realities of the twenty-first century.

The counter-intuitive, but pervasive, reality is that technological advance is most likely to do in the future what it is has in the past: humanize work. Therefore, to adapt to a world where computers can think may be just the beginning. The hard part for us humans could be to understand—and to value—what it means to be human in the first place.

Conclusion: Entering Unfamiliar Territory

In one fundamental respect, our era is different from any before: population growth has slowed and, in many places, is now reversing, because the world is becoming more prosperous. More than 40 percent of the world’s population already lives in nations with sub-replacement fertility, defined as a fertility rate below 2.1 children per woman. This list includes all of North America, all of Europe, all of East Asia (notably including China), and additional populous countries including Brazil and Iran. The total population of Europe, including Russia and non-European Union countries, peaked in the year 2000. Japan is well into population decline, and many more countries are on the brink. And recent data from the National Center for Health Statistics show that the U.S. birthrate fell for the sixth straight year in 2013 to an all-time low.

There is no historical analogue to the coming era of global population decline. Why not? Because, in our era, population decline has been caused not by war, disease, and famine, but rather by prosperity. Countries are simultaneously experiencing population aging and population decline. In all of human history, this has never before happened. No books, plays, or poems exist to guide us in an era of prosperity-driven population decline. It is new.

John Stuart Mill predicted two centuries ago that human society ultimately would reach a stage in which a stagnant population earns high (relative) incomes and has steadily improving well-being, all sustained by continued investment in R&D. John Maynard Keynes offered a similar vision in a 1930 essay titled “Economic Possibilities for our Grandchildren,” in which he predicted that humanity was on a trajectory to solve the problem of economic scarcity within a matter of decades; humanity’s problem thereafter, according to Keynes, would be to solve the problem of leisure time, noting that “it is fearful for the ordinary person, with no special talents, to occupy himself, if he no longer has roots in the soil or in custom or in the beloved conventions of a traditional society.” In such a world, the promise of meaningful occupation ultimately comes to be valued above wealth for its own sake, and returns from income are correspondingly subject to diminishing returns.

Because of the uniqueness of the coming era of depopulation, it is not easy to analyze its implications. Nonetheless, as I argue with Joon Yun in Auerswald and Yun (2014) there is good reason to believe that the trajectory of human progress will continue over the century to come, even as global population begins to decline.

A combination of population decline and population aging suggests a powerful coincidence of needs: as the human population plateaus in the aggregate, continued improvements will require individuals to exhibit greater levels of creativity across the population, and across a human life, than has ever been the case in the past. Some of that creativity will contribute to aggregate growth through improved productivity and new innovations, while an even greater share will contribute to solving what Keynes referred to as the problem of leisure— what we might today think of as the challenge of living a meaningful, fulfilling life.

But what if the Singularity arrives in full force, and “computers” evolve into sensing, thinking, feeling, loving beings completely indistinguishable from humans — into full-fledged immigrants from the future? . . . 

Herbert Simon has something to say about that, as well:
I am confident that man will, as he has in the past, find a new way of describing his place in the universe — a way that will satisfy his needs for dignity and for purpose.

He has a point.

About the Author
Philip Auerswald is an associate professor at the School of Policy, Government, and International Affairs. His work is about entrepreneurship and innovation in a global context. He is most recently the author of The Coming Prosperity: How Entrepreneurs are Transforming the Global Economy (Oxford University Press, 2012). Since 2010 Professor Auerswald has served as an adviser to the Clinton Global Initiative on topics related to job creation, education, and market-based strategy. During 2011-2012 he was a senior fellow at the Kauffman Foundation. He is the co-founder and co-editor of Innovations: Technology | Governance | Globalization, a quarterly journal from MIT Press about entrepreneurial solutions to global challenges, and an associate at the Belfer Center for Science and International Affairs, Harvard University.


Autor, David H., Frank Levy, and Richard J. Murnane. 2003. “The Skill Content of Recent Technological Change: An Empirical Exploration.” The Quarterly Journal of Economics 118 (4): 1279–1333. Accessible at

Bhidé, Amar. 2008. The Venturesome Economy: How Innovation Sustains Prosperity in a More Connected World. Princeton, N.J.: Princeton University Press.

Bhidé, Amar. 2015. “The Demise of US Dynamism Is Vastly Exaggerated–But Not All Is Well,” unpublished manuscript.

Byron, George Noël Gordon [Lord Byron]. 1831. The complete works of Lord Byron, including his Lordship’s suppressed poems with others never before published. Volume 1. Paris: Galignani.

Brynjolfsson, Erik, Thomas W. Malone, Vijay Gurbaxani, and Ajit Kambil. 1994. “Does Information Technology Lead to Smaller Firms?” Management Science, 40(12), December: 1628–1644.

Coyle, Diana. 2014. GDP: A Brief But Affectionate History. Princeton, N.J.: Princeton University Press.

Fernald, John. 2014. “A Quarterly, Utilization-Adjusted Series on Total Factor Productivity.” Federal Reserve Bank of San Francisco Working Paper 2012-19, April.

Gollin, Douglas. 2002. “Getting Income Shares Right.” Journal of Political Economy, 110 (2): 458–474.

Karabarbounis, Loukas, and Brent Neiman. 2014. “The Global Decline of the Labor Share.” The Quarterly Journal of Economics 129 (1): 61–103.

Kurzweil, Ray. 2006. The Singularity is Near: When Humans Transcend Biology. New York: Penguin Press.

Jones, Charles I., and Jihee Kim. 2014. “A Schumpeterian Model of Top Income Inequality.” NBER Working Paper No. 20637, October.

McCloskey, Deirdre Nansen. 2014. “Measured, Unmeasured, Mismeasured, and Unjustified Pessimism: A Review Essay of Thomas Piketty’s Capital in the Twenty-First Century.”Forthcoming, Erasmus Journal of Philosophy and Economics.

Piketty, Thomas. 2014. Capital in the Twenty-First Century. Cambridge MA: Belknap Press.

Planet Money. 2015. “How Machines Destroy (And Create!) Jobs, In 4 Graphs.”, May 18.

Robinson, Joan. 1953–1954. “The Production Function and the Theory of Capital.” The Review of Economic Studies, 21 (2): 81–106.

Rowthorn, Robert. 2014. “A note on Piketty’s Capital in the Twenty-First Century.Cambridge Journal of Economics, 38 (5): 1275–1284.

Samuelson, Paul. 1966. “A Summing Up.” The Quarterly Journal of Economics, 80 (4): 568–583.

Simon, Herbert. 1960. “The Corporation: Will it be Managed by Machines?” In Management and the Corporations (M. L. Anshen and G. L. Bach, eds.). New York, NY: McGraw Hill: 17–55.

Shiller, Robert. 2015. “What to Learn in College to Stay One Step Ahead of Computers.” The New York Times, May 22.


Figure 1. Top 1 percent wealth share in the United States, 1913–2012. (Piketty 2014).

Figure 2. Trends in Routine and Nonroutine Task Input, 1960–1998 (Autor, Levy, and Murnane 2003).



  3. Byron (1831), 672–673.
  4. See Coyle (2014) for a thorough treatment of the interpretation of GDP.
  5. She continued: “The student of economic theory is taught to write Q = F (L, K), where L is a quantity of labor, K a quantity of capital and Q a rate of output of commodities. He is instructed to assume all workers alike, and to measure L in man-hours of labor; he is told something about the index-number problem in choosing a unit of output; and then he is hurried on to the next question, in the hope that he will forget to ask in what units K is measured. Before he ever does ask, he has become a professor, and so sloppy habits of thought are handed on from one generation to the next.”
  6. Karabarbounis and Neiman (2014) offer a careful empirical assessment of trends in labor shares that is an exception that proves the rule in this literature. See also Gollin (2002), who states: “It is common practice to use employee compensation as a measure of labor income. From a conceptual perspective, however, employee compensation excludes some important forms of non-wage compensation and may include rents accruing to particular skills, including returns to entrepreneurial ability. More important for the purposes of this paper, employee compensation omits the labor income of people who are not employees. In some countries, the self-employed account for huge fractions of the workforce. As a result, in these countries, labor income is badly understated by the employee compensation measure.”
  7. Talk given at the Cato Institute conference on The Future of U.S. Economic Growth, December 4, 2014. See also Bhidé (2008), chapter 13, and Bhidé (2015). See Fernald (2014) for a description of the current state of the art in constructing aggregate productivity measures. The extent of the corrections to “naïve” or “raw” total factor productivity (TFP) estimates into ones that plausibly might reflect the rate of technological changes are themselves an indicator that the underlying project of calculating TFP may have outlived its usefulness.
  8. To do justice to the topic addressed in this single bullet point clearly would require a paper on its own, if not a book. However, even if we were to allow a more direct correspondence between productivity statistics and actual innovation in the economy than is reasonable, trends in productivity would do anything but support the concern raised by the authors of the cosignatories of the Open Letter on the Digital Economy; productivity has been flat during the era the signatories of the Open Letter contend has one of rapid technological advance.
  9. For an understanding of the weakness of the arguments presented in Piketty, there is no substitute for reading McCloskey (2015) in full. See also Weil (2015).
  10. See Brynjolfsson et al. (1993) for important early work on the relationship between IT and firm size. In Auerswald (2009) and (2010), I make the connection between IT-related changes in transactions costs and shifts in the relative price of entry into new markets between new and incumbent firms.
  11. See Figure 2.
  12. In particular, David Ricardo.
  13. I thank Jordan Greenhall for this observation.
  14. Planet Money (2015). 

Choosing A Future in the Platform Economy:

Implications And Consequences Of Digital Platforms

By John Zysman Professor, Political Science and codirector of the Berkeley Roundtable on the International Economy University of California, Berkeley
John Zysman
John Zysman


We are entering a Platform Economy—one in which tools and frameworks based on the power of the Internet will frame and channel our economic and social lives. The algorithmic revolution, in which computable algorithms are applied to activities from consumption and leisure to services and manufacturing, is the foundation of this digital transformation.1 Algorithms now live in the cloud and form the basis of digital “platforms.” For our purposes, platforms are “frameworks that permit collaborators—users, peers, providers—to undertake a range of activities, often creating de facto standards, and forming entire ecosystems for value creation and capture.”2 

The cloud is at once infrastructure, marketplace, and ecosystem.3  The variety of platforms nearly defies categorization. To illustrate, Google and Facebook are digital platforms that provide search and social media, but also are platforms on which other platforms are built. Amazon is a marketplace, as are Etsy and eBay. Amazon Web Services provides infrastructure and tools with which others can build, while Airbnb and Uber are forcing deep change on quite different businesses. These diverse, cloud-residing platforms however we categorize them, are provoking a profound reorganization of markets, work arrangements, and, fundamentally, of value creation in the contemporary economy. 

Our basic premise is that the emergence of a platform-based economic reorganization will not dictate our future, although it will and already is beginning to frame the choices we make.4 How we choose to deploy platform-based tools will reflect corporate strategy and public policy, and will condition the society we are building. 

Knowledge to Shape the Future

Will the Platform Economy—and the reorganization of markets, enterprises, and social organization it portends—catalyze economic growth and a surge in productivity driven by a new generation of entrepreneurs? Or will the reorganization concentrate gains in the hands of those who generate the platforms, and possibly stifle future entrepreneurs? Will it spark a wave of entrepreneurial possibilities, or an avalanche of dispossessed workers trying to make their way with gigs and temporary contracts? Ultimately, what do we need to know and understand to shape this future?

How pervasive will the platform effect be? The word of the day is disruption—the sense that many traditional business models, organizations, and forms of value creation will be swept aside or radically transformed.  Although control or ownership of platforms is a separate matter, groups of peers that coordinate activities and transactions on platforms challenge existing business models. For example:

  • Taxi businesses are threatened by Uber.
  • The music industry is challenged by iTunes, Pandora, and Spotify.
  • The camera industry is disrupted by GoPro, which, through its website, is trying to organize itself as a platform firm, even as YouTube and app stores emerge.

We must consider the impact of platform strategies on competition in diverse segments of the economy, and what competitive strategies will be introduced. Long ago, online stores emerged in the retail sector. Consignment businesses have grown dramatically, and have opened the way for an array of newly minted entrepreneurs. And we know that many products have become parts of systems, the classic example being the iPhone, which provides access to a platform store that has a previously unimaginable variety of virtual products. The effects certainly will vary across sectors in very different ways.

While some will question whether productivity and growth will be accelerated, the more profound questions may be whether economic and social life will be transformed, and whether the outcome will lead to a very different distribution of wealth and power in global society. 

The platform is likely to effectively define the digital era, with the algorithm, Internet and cloud as the building blocks. We contend that we are seeing the beginning of a digital transformation, which will extend to the Internet of Things and beyond, and is likely to release both enormous creativity and wicked management problems.

What will the Platform Economy do to work, entrepreneurship, income, and inequality? Platforms are creating new opportunities for earning income, and are generating an array of entrepreneurial opportunities … but of what sort? Will we have an increasing number of laundromat entrepreneur equivalents who are viable in their own right but unlikely to generate sustained productivity and growth, thus reflecting the patterns of growth rather than generating them? Will platforms induce a set of new businesses that, in their turn, drive employment and growth? Or, put differently, will the platform transformation create a community of incipient entrepreneurs or vulnerable workers? 

Work is being reformatted.  For many, traditional employment—a single organization providing long-term engagement, usually with some form of social benefits—is giving way to gig and contract arrangements. Not surprisingly, business strategies shape job quality.  Even in low-margin, low-price businesses, there are “better” job strategies that can provide workers higher wages and benefits, and help strengthen a firm’s competitive position.  In the aggregate, the shifting place and character of entrepreneurship, and the reorganization of work, may powerfully alter the distribution of wealth and income in societies.

Addressing the Elements of a Platform Economy

Assuming we are moving to a Platform Economy, we must address four elements. At first glance, each is evident. But a closer look reveals that what to achieve and how to achieve it in each case are not as evident.

1. Infrastructure:  We obviously need appropriate infrastructure—but what kind? And how do we acquire or build it? Does it simply require ever-faster broadband access for the community as a whole? Or does effective infrastructure require tools and training for firms and consumers? If so, should they be provided as a public utility or by the market?

2. Training and skills:  Does the Platform Economy require a new set of skills, or only a recasting of emphasis? Needs certainly include widespread comfort with using platforms and apps. But does that require a heavier investment in STEM (science, technology, engineering and manufacturing), or in design and art?  The answer is less obvious than it might appear.

3. Social protections:  Will entrepreneurs and contract work arrangements be facilitated and encouraged by broadening social protections?  Or will those protections inhibit the flexibility required of the economy?

4. Regulatory transitions: Rules and regulations for the marketplace and labor markets will not adapt themselves to the needs and logic of the Platform Economy. Moreover, this transition will lead to debates and fights about the adaptation of those rules. There will be struggles about protections for communities, clients, workers, and the market itself. 
As communities, the evident instance is whether Airbnb represents a change in land use—do I want a virtual hotel in my neighborhood?  And should Airbnb hosts be able to discriminate against people they do not want to welcome, whereas a hotel is legally proscribed from discriminating? 
As clients, do we want assurance of health and safety in our Uber cars or Airbnb rentals? 
As workers, what risks should we bear, and what risks should be shouldered by the platform owner? Indeed, the matter of when workers are contractors and when they are employees is being adjudicated now in a number of states and nations. 
As the market itself, what protections and policies are required to preserve competition?  The European Union’s efforts to regulate Apple, Google, Facebook, and other globally dominant, Silicon Valley-domiciled platform leaders is an example.5 In fact, what are the competition issues when dissatisfaction with Google Search, for example, can lead to a costless, seamless, immediate switch to Microsoft Bing or the independent DuckDuckGo?

The Question of U.S. Leadership

Finally, we should consider whether U.S. entrepreneurs or, more specifically, Silicon Valley or Silicon Valley-inspired entrepreneurs, will retain their early and currently obvious dominance in a platform era. Throughout the digital era and this current wave of change, American policy initiatives and firms have led the way. Silicon Valley’s enormous success with finding and funding new products, firms, and disruptions makes it seem that the digital transformation is an American prerogative, and that others can only follow and whine.

While generations of digital change have issued forth from Silicon Valley, inevitability is truth only until things change. Henry Ford, we are often told, launched the era of mass production that represented American innovation and contributed to an era of American predominance. Indeed, policymakers and scholars alike referred to Fordism as an economic revolution.

Seventy years later, Toyota arrived, and American advantages became obstacles. As Cohen and Zysman observed in 1988, the U.S. faced decades of manufacturing decline. Comebacks were slow, and our manufacturing base and domestic skill sets eroded.6  Silicon Valley breakthroughs did not help the industrial heartland, as Florida and Kenney argued in 1990.7 Can this happen again? Let us not too quickly point gleefully to the mistaken overenthusiasm about Japan’s consumer-electronics and semiconductor successes, which faded as Silicon Valley-driven U.S. entrepreneurial firms reasserted themselves through success in software-driven electronics. Rather than dismiss the question, we should ask what might happen if a surge in platform leadership emerged from outside the United States.

Two possibilities establish the reality of this question:

  1. The huge size of the Chinese market suggests that platforms bred in that largely protected market could have the scale and financial sustainability to move abroad and establish standards. With the possible exception of Alibaba, there currently is little evidence that Chinese platforms are gaining traction, even in East Asia. But will this change? 
  2. German manufacturing firms are concerned about maintaining position in what is labeled Industrie 4.0, which is what they term the Internet of Things. They want to assure that their small, midsized, and major manufacturers capture the advantages of platforms and next-generation cloud-based computing, and define that trajectory globally. 

Asked generally, will European firms’ manufacturing, industrial and service strengths— as well as the Chinese operating behind protectionist barriers—create application domains that can outcompete U.S. entrepreneurial firms? 

For now, as a beginning, we return to our basic premise.  The Platform Economy is upon us.  What it means for our economy and society will be a choice, not an inevitable unfolding of the technology.

Labels Matter 

How we label this transformation matters.  Labels specify targets for policies, strategies, and studies.

The new, digitally founded economy has been given a variety of names based on its perceived attributes. Each name lends itself to the study of different outcomes and activities over others.  At various times, it has been called the creative economy8 and, by contrast, the gig economy/the precariat/1099 economy,9 focusing on the impact this emerging economy is having on work.   

Recently, “Sharing Economy” has been a popular label10. However, much activity labeled sharing—such as Uber and Airbnb—is far from the stated visions of:

  • Wikipedia—the shared construction of a knowledge tool.
  • Napster—sharing music, legally or not.
  • Open-source software—shared among all.

Despite the attractive label and entrepreneurial successes, there are drawbacks to calling this a Sharing Economy. Uber and Airbnb are entrepreneurial initiatives that facilitate the conversion of consumption goods, such as automobiles and apartments, into commercial offerings. Airbnb is closer to nineteenth-century, working-class women taking in boarders. Uber facilitates gig work for drivers providing their personal vehicles, or buying vehicles to join the game.

This ”sharing” can and often does resemble a putting-out economy from the early industrialization period before factories, but the putting out of work to individuals now is facilitated by digital platforms.

Simultaneously, the rapidly growing mobile-app and user-generated-content firms, such as YouTube or Instagram, are structured as digital consignment industries, borrowing from the compensation scheme used by artists working for galleries, for example. Airbnb also has some consignment aspects.

The “Platform Economy” is our preferred term for this emerging organizational model. Why? Digital platforms are the base upon which an increasing number of connection-based activities—marketplace, social, and political—are organized. If the Industrial Revolution was organized around the factory, today’s changes are organized around platforms that apply algorithms to enormous databases running in the cloud. The salience of these digital platforms suggests that we are in the midst of an economic reorganization in which platform firms are developing power roughly equivalent to that of Ford, General Motors, and General Electric during earlier eras.

Debating the Future

There is a ferocious debate about whether the future of the Platform Economy will be a utopia of abundance, or a dystopia of limited employment and stunning inequality.  

The early pioneers in the industry—particularly Bob Noyce, Steve Jobs, and Bill Gates on the West Coast—truly believed they were creating the future, and opening the world’s possibilities and prospects.11 And optimists still abound: San Francisco, for example, now has been called the new Hollywood, as visions of profitable disruptions mobilize entrepreneurs and data scientists.12

The question for investors, inherently optimists who search for profitable activities, is: How is value created and captured in the platform era? To answer, many point to the benefits of emerging platforms. Zipcar, by reducing the need for individual auto ownership —or at least potentially increasing access to auto transport for those who do not own vehicles—stands as a commercial vision of sharing, although it actually is a particular form of rental,13 alongside true sharing activities, such as Wikipedia.

However, as already noted, the optimistic, utopian vision of contract workers as simply proto-entrepreneurs who treasure their flexible schedules often collides with claims that they simply are a new precariat—dependent contractors in precarious roles within a modern putting out. Similarly, the utopian vision argues that platforms such as Uber and Lyft can unlock the commercial value in underused personal goods—apartment rooms or cars can become investment goods in commercial markets without consequences for the community. This is a large assumption. Similarly, the platform businesses that match workers and tasks may make labor markets more efficient but, at the same time, can generate fragmented work schedules, and increasing levels of part-time work without employment-related benefits that previously characterized much full-time work.

Will, then, the tools we have built turn on society? Even as the digital era unfolded in its utopian phase, there were skeptics, and perhaps most prescient was novelist Kurt Vonnegut. In his first novel, Player Piano,14 Vonnegut envisioned a digital future of abundance—albeit a digital future of machines built with tubes rather than semiconductors—with radical social division between a creatively employed and highly credentialed elite, and an underclass. His dystopian vision now is finding full expression in the fear that digital machines, artificial intelligence, robots, and the like will displace work for a vast swath of the population.

Bill Davidow, once at Intel and then at his own Silicon Valley venture capital company, expressed this in the business literature, “What Happens to Society When Robots Replace Workers?”15 An outpouring of popular books and more formal articles from the economics profession argue that jobs will be displaced by digital automation and robotics. The best known and most popular economic expression of this belief is The Second Machine Age.16

The question really is about what balance will exist between jobs created as the digital wave flows through our economy and society, and what jobs will be displaced. It is feasible to catalogue existing work, particularly routine work that can be fully characterized, as likely dislocated by digital tools, and perhaps estimate the numbers of such existing jobs that may become obsolete.17 More difficult to catalogue, and open to speculation, are new kinds of work being created now and that will be created in the future. Some early indicators can be enumerated, but certainly not exhaustively counted. Algorithms and databases are automating work but, as this occurs, new work is being created. There will be new products—goods and services, as well new production processes—that are likely to require intensive design, creativity, and technology.

Moreover, the character of some existing work—much or little we cannot know—will be reframed, but not eliminated by digital technology. Uber, TaskRabbit, Handy, and other platform firms are transforming industries by connecting "workers" with customers in new ways. In some cases, this is displacing or threatening existing, often regulated, service providers, such as taxis and hotels. In other cases, it is formalizing previously less organized or locally organized work. Still other platforms, such as app stores and YouTube, are creating entirely new occupations or occupational branches. Finally, existing organizations are creating digital and social media marketing departments and jobs. The question in these cases is not whether there will be jobs, but what system of control and value capture is in place. Having written this, our sense is that, across the board, "employment" appears to be more precarious than ever.

These changes are not likely to result in the "workerless" society. Rather, we risk a society within which the preponderance of the work and value creation is more dispersed than ever before, even as platform owners centralize transactions and capture value from activities on their platforms. Importantly, we can only speculate about the balance and character of firms and jobs destroyed, created, and transformed, and about the character of the work and organizations generated.18

Indeed, we would note that there is a classic dilemma in the use of digital automation: anything that can be characterized sufficiently to become computable can be copied.19 At that point, another round of innovation and imagination is required.20 

Can automation innovate itself?  Or will teams of people and digital tools be required to be competitive?  Note that the Turing Test might establish that a digital machine can imitate intelligence. But the test does not establish, or purport to establish, consciousness, nor consider whether human consciousness differs in fundamental ways from current algorithmic tools.21 The debate about jobs created or destroyed cannot be resolved. Importantly, we can only speculate about the balance and character of firms and jobs destroyed, created, and transformed.22 We can, we emphasize, only examine indicators and traces.

We are confident, though, that the outcomes—jobs created, and jobs evaporated or transformed—will depend not on the technology itself but on how the technology is deployed. Choices about deployment will turn on entrepreneurial initiative, corporate strategies, and public policies. We know, for example, that the consequences of deploying radio-frequency identification in retail differed dramatically in Denmark, France, and the United States. The outcomes, interestingly, were not a product of labor-management fights, but of conflicts between producers and distributors.23 

Similarly, in the discussion of the Internet of Things or the digitally based reorganization of manufacturing, we find significant differences among national emphasis and investments.24 Which communities, this leads us to ask, are most likely to be the sources and beneficiaries of the emerging Platform Economy? Which are most likely to be discomfited? The strategies for deployment are, of course, precisely the substance of our choices for a future in the Platform Economy.

The Algorithmic Revolution25 and Clouds:26
Technical Foundations of the Platform Economy

The algorithmic revolution and cloud computing are the foundations of the Platform Economy. Computing power in itself is only the beginning of the story. Computing power is converted into economic tools by algorithms operating on raw data. When aspects of activities can be converted into processes, which can be formalized and codified with clearly defined rules for their execution, they can be reduced to computable algorithms.

The software layer that stretches across and is interwoven with the economy is a fabric of algorithms.27 That software layer—that algorithmic fabric—encompasses manufacturing. It is the Internet of Things/Internet of Everything/Industrial Internet, with its implied webs of sensor networks. This includes services, which often employ those sensor networks, and covers other diverse social, political and economic activities. The software layer extends the availability and lowers the cost of access to digital tools, and to traditional tools accessed by and controlled by digital processes. Costs sometimes drop through open-source software—the race to zero in cloud computing—and/or by the ability to collectively provide tools through online platforms or physical commercial sites, such as TechShops. 

Cloud computing is about how computing is done. It is much less about geography, or where it is done.28 It rests on virtualization and abstraction of computing processes.31 While the details of how it works do not matter to our discussion, the consequences do.

We should note that the major providers of cloud services remain, at least for now, large American firms that developed the cloud paradigms and systems for their own internal use. For provision, scale does matter. For users—individuals, small- and medium-sized enterprises, startups, and corporations alike—the consequence is a radical reduction in the cost of computing resources, and information- and communication-technology tools. As important, as is now widely recognized, the terms of access to computing resources change, as well. Users can “rent” resources in units rather than having to own or build entire computing systems. Computing, and the applications and platforms it facilitates, are then available as an operating expense rather than a capital expense.

Let us link the story of algorithms and cloud computing to the emerging Platform Economy.

Algorithms live in the cloud and dramatically ease the creation of platforms.29 Digital platforms are, then, computing frameworks upon which users can undertake a range of activities, often forming entire ecosystems for value creation. Many of the current Internet platform firms use Amazon Web Services. So, indeed, platforms can grow on platforms as an array of applications.

Many of those platforms-on-platforms are what we would call complementors. Complementors include emerging actors such as:

  • AppAnnie, which ranks revenue generated by apps.
  • Advertising “agencies,” which analyze YouTube advert buying.
  • TubeMogul, which classifies YouTube “stars” and measures their reach.
  • Various agencies that cultivate new YouTubers.

These complementors are powerful allies in building and maintaining the lock-in for the master platform. Platforms are, to put this differently, algorithm-enabled cyberplaces where constituents—people or machines—can act or transact. Of course, building a platform is work, but platforms themselves then generate or organize the work of others by providing digital locations for connections that organize work and other activities.

Making Sense of the Diversity of Platforms in the Platform Economy

Digital platforms, based on algorithms and databases, are restructuring ever-greater parts of the global economy. In many cases, they have disrupted the existing organization of economic activity by resetting entry barriers, changing the logic of value creation and value capture, repackaging work, and/or repositioning power in the economic and network system.

Speculations aside, we currently have no real theory of the effect of these diverse platforms on the overall economy, nor a particularly dominant approach to categorizing platforms. The intent here is to begin to structure a discussion about how an economy increasingly organized to operate on platforms affects the organization and practice of competition, work, and entrepreneurship.

Traditionally, we have categorized by sector, but sectors are blurring. A smartphone is at once a communications device, a camera, and a music system, to start a list. We might ask whether the consequences of platforms differ by the original sector being transformed. But, ultimately, increasingly blurred sectors are not good organizing categories for understanding the Platform Economy.

Perhaps we should categorize platforms by function or business model. Here again, categories blur and overlap. As a result, an initial listing is partial and choppy, an awkward cut into the complexity.   

  1. Platforms for platforms. In a sense, the Internet itself is the foundation of the Platform Economy, but there are a series of businesses that then provide the infrastructure and tools for the rest. For example, Amazon Web Services facilitates the construction of cloud services, the tools with which other platforms are built.
  2. Platforms mediating work. In some platforms, this resembles the function of electronic headhunters or human resource departments. In other cases, it can be seen as a modern form of the putting out system of nineteenth-century industrialization. Mediating work itself has many versions, which include: 
    a. Globally biddable work. Examples include Odesk/eLance (since renamed Upwork), Innocentives, and Amazon Mechanical Turk.
    b. Occasional informal work. Facilitated by apps and, therefore, cyberformalized, Task Rabbit, Handy, and Homejoy are perfect examples.
  3. Platforms making tools become available online. Github is becoming the repository of all kinds of open-source software available to anyone. This dramatically reduces the cost of software tools and building blocks.
    a. Automated HR, Job Rooster, and Wonolo provide diverse HR functions.
    b. Zenefits provides an online marketplace of HR tools free to small businesses. In the process, it makes mediation by local insurance brokers unnecessary.
  4. Electronic goods markets for retail and business. These run an entire gamut.
    a. Virtual markets for physical goods. Etsy and eBay are distinct versions.
    b. Retail sales platforms, such as Amazon or hundreds of company-specific apps.
    c. Apple and Android stores, which are platforms that facilitate diffusion of the apps.
    d. Virtual consignment platforms, such as YouTube, Amazon self-publishing, and many others.
  5. Platforms intending to transform service industries. Airbnb and Uber are examples of platforms intending to convert consumer goods into investment goods. For example, rather than sharing, Uber connects drivers with customers algorithmically. In this case, drivers are treated as contractors, which puts them in a more precarious position.
  6. Shifting the place of intermediaries in finance.
    a. Platforms, such as Kickstarter or Indiegogo for project funding, can replace traditional intermediaries.
    b. Finance platforms displace traditional financial institutions, e.g., AngelsList for venture capital, or Zopa or Rate Setter for peer-to-peer lending support.
  7. Facilitating social and political organizations, including worker organizations.

Does beginning with the platform categories above provide insight into the character and ability to scale the entrepreneurial opportunity? Diverse as they are, all these platforms have created business ecosystems, repackaged work relationships, and/or transformed terms of competition.

These platforms represent a multiplicity of business models and functions, and raise questions that suggest wildly varied answers. As a place to start a discussion, we might ask these questions about each platform, or type of platform.

-    How is value created? The Platform Economy itself is a distinctly new set of economic relations resting on the Internet. The ecosystem created by each platform is a source of value, and sets the terms on which owners and platform users can participate.

-    Who captures the value?  There are a variety of mechanisms with varied implications for gains distribution.

  • Owners of some platforms “tax” all transactions, while others monetize their services through advertising.
  • Contractors, consigners, or quid pro quo workers do work previously reserved for traditional employees, or take on entirely new categories of work created by platforms.
  • Venture laborers, so named by Gina Neff, earn high wages working in platform firms. More significantly, if the firm is successful, the value of the platform is capitalized in the stock markets, resulting in remarkable wealth for the firm’s direct employees and entrepreneurs.30  If the firm falters or fails, these individuals must find new employment.
  • Mini-entrepreneurs or consignment workers—alternately positive and negative terms for the same individuals—provide platforms such as app stores, YouTube or Amazon self-publishing with goods that are, usually, but not necessarily, “virtual.” Many will be unsuccessful or marginally profitable, but some can be fabulously successful. While as yet unmeasured, it seems likely that this is creating many more opportunities for entrepreneurship. In certain cases, particularly in apps, products in the consignment economy may become so successful that venture capitalists will invest in the entrepreneur/firm, and the employees will become venture labor. Some of these apps can become platforms themselves. Put differently, the consignment level has significant upside for participants, but the upside is accompanied by high risk.

-   Who owns or controls the platform? The answer differs and makes a difference. For users and producers, for example, there are differences between Wikipedia (a network managed by consensus rules), the Danish agricultural cooperative platform (a consortium in which participant owners know each other, and set clear boundaries between themselves and others) and Uber (a tightly held firm funded by venture capitalists, whose value eventually will be capitalized by sale of control, either through acquisition or stock offering). Power may be centralized (controlled transaction and communication systems separate buyers/users from providers, and separate providers from buyers), or decentralized (such as how Wikipedia diffuses power over content).

-   How is work packaged and value created?  Some workers—including those employed by Microsoft, Google, LinkedIn, and Facebook—retain traditional employment relationships. These firms expect long but relatively flexible working hours, and offer free food, drinks, transportation, and a myriad of other benefits, which make them seem almost like corporate paradises. Those working on gigs, consignments, or contracts have radically different relations, though the hours are flexible and largely self-controlled.31 One important question, then, is what percentage of work now is organized in these radically new ways?

-   What is the distribution of risks and rewards for those in these various ecosystems?

The Consequences of Platforms for Entrepreneurship and Work

Entrepreneurship and the packaging of work are tightly interwoven issues.

Consider entrepreneurship. Media attention and much talk in the venture community focuses on “disruptions,” which appear to be where massive opportunities exist. Uber disrupts taxi companies. Airbnb challenges hotels. Zenefits threatens local insurance brokers.

But how many instances of disruption are there? Do these disruptions create a flood of viable entrepreneurial possibilities, or destroy the security of employment relations? Do they operate to create new sources of income and reasonably compensated work throughout the society?

It is evident that platforms open an array of entrepreneurial opportunities. Some entrepreneurs, such as Zipcar’s Robin Chase, envisioned not only an alternative economic model, but also an alternative social model: Own a car, or access occasional use of one through the Zipcar platform. If that model spreads widely, it would result in a drop in overall demand for auto production. This may or may not disrupt Hertz (Zipcar was sold to Avis).32 But could it dramatically affect automakers if fewer persons buy automobiles.33  In other words, such “sharing” solutions could have unforeseen ripple effects on entire market ecosystems, as encyclopedia producers discovered to their dismay.

Many platforms, by their very nature, prove to be winner-take-all models in which the owners of only one or two surviving platforms appropriate a portion of the entire value created by all the platform participants. More importantly, power is centralized to the platform owner who, after winning the initial competition, becomes a monopolist who can make decisions to maximize his/her own welfare. At the same time, the monopolist platform owner squeezes the platform community—the drivers on Uber, the content providers, the consigners—who are instrumental in producing the value in the first place.

Consider, by contrast, how platforms affect work. In these business models, what happens to the organizational forms of work, and to the form in which work is packaged? 

Conceptually, if not literally, Uber converts taxi company employees or former medallion owners into contractors who access income through the Uber platform. Are these contractors, mini-entrepreneurs, or extremely precarious workers relabeled as contractors?  Are some Airbnb offerings not just another form of rentals that transform apartments from long-term residences to short-stay offerings? 

Is this entrepreneurship in any significant way?  Moreover, how do we understand the business models of the winners among those who produce apps or YouTube videos, or self-publish books on Amazon?  For these individuals, there is a law of returns—a few big winners are remunerated by advertising, product placement fees, and personal appearances.  And there is a very long tail of producers who are creating the vast bulk of consigned content without compensation.

Considering disruptions, mini-entrepreneurs, contractors, and gig workers leads us to ask: Does the Platform Economy point to an even more unequal society?  Does the answer depend on the character of platforms, or on the policies and politics of the Platform Economy?

Policy and Politics

The policy objectives and concerns in a platform era seem evident. The late nineteenth century saw the emergence of the corporate organization as a means of orchestrating economic activity and organizing markets.34 In the twenty-first century, we might speculate, the platform in the cloud takes on a variety of these functions.

Hence take Google, the Platform Economy giant. It is now a large firm, and yet has only 50,000 employees. Uber has only about 1,500 employees, and already is a global business. What policy and political issues emerge when the orchestrators of economic activity are relatively small firms rather than organizations as large as Ford Motor Company, General Electric or, the behemoth of them all, General Motors were in their heyday?

From the American standpoint, we must ask how policy will influence entrepreneurship and work in a cloud-founded Platform Economy. Let us set aside, in this brief essay essential questions about whether cloud technologies, and the platform-driven economic reorganization they prompt, will drive productivity growth, and whether the reorganization, on balance, will destroy jobs or reduce the skill levels required.

Instead, we ask:

  • On balance, are the array of entrepreneurial, innovative opportunities widespread? Or will they hover around a few big winners, and an array of small–scale, highly vulnerable players?
  • Do we create a new source of productivity or a new form of putting out? 
  • Can Uber drivers be self-supporting contractors in a 1099 economy, rather than stable workers in an employment economy? Or are they extremely vulnerable gig workers? 
  • Do we generate labor market flexibility or a “precariat,” as some believe, that resembles a cyberized Downton Abbey, replete with a new and sizeable underclass.  

The policies we adopt may determine the balances achieved in the platform era. If we want an entrepreneurial spirit to infuse the platform world, then we want risk-taking entrepreneurs who form platforms or seek advantage as contractors/consigners within it.  What encourages risk? Is it a certainty that, if a gamble fails, one can always play again? 

Similarly, if we want workers to accept the new flexibility, how do we assure them that, if they accept the flexibility, they will be beneficiaries, not victims, of the greater social value and wealth that is being created? As victims, they will resist. As beneficiaries, they may help facilitate the shift. Stated simply, public policies will shape the gains, risks, and responsibilities both for work and entrepreneurship.

How then can we make the Platform Economy a vibrant source of growth?  We view two public policy domains as critical.

First, social policy, sometimes called welfare, shapes the risks workers and entrepreneurs take, and affects their evaluation of whether to pursue or resist change. In the United States, benefits such as pensions and medical coverage, until the emergence of ObamaCare, have been tightly tied to employment. Lose your employment, lose the protections. The American debate often assumes that expanded welfare protections mute initiative,35 pointing to Europe as investing in social protections at the cost of economic dynamism. Regardless of whether this was, in fact, ever the case, the question is whether social protection will inherently mute initiative now. The real issue was never the protections themselves, but how they were organized. The Danish flexible security model provides evidence that social protections can lubricate the engines of change.

The Platform Economy, with expanding contract work and gig employment, should lead us to look again at the Nordic social policy model. Simply put, it attaches many social benefits to citizenship rights. The notion of flexible security gives employers extensive rights to adjust their work forces as needed while still providing workers social securities, or protections, in the form of training, job placement, and basic income. This is no panacea. Struggles to sustain employment endure, and there are continuing fights about these programs. But we must consider that, in this environment, addressing the downside risk of entrepreneurial efforts, while providing worker flexibility with broader social safety nets, are social rights that may make a Platform Economy a source of sustainable growth.

What assurances of social safety do we want to provide to risk takers, to encourage those risks?  Could Uber get drivers in an environment with a greater social safety net and a reasonable basic standard of living guaranteed? Evidence from Europe suggests an emphatic “yes,” as governments are forced to control Uber drivers.

Second, what market rules are appropriate for a Platform Economy? There will be an array of political struggles about these rules, and we should welcome them as part of defining the market and society in a cloud/platform era. There will be political fights about how to protect communities, clients, and workers against market disruptions. Some of those fights will be about business models playing a game of policy arbitrage, while others may be about rules on consignment platforms. In this game, the advantage of platform-based companies often rests on an arbitrage between the practices they adopt versus established companies’ operating rules, which are intended to protect clients, communities, workers, and markets themselves.

These are classic issues, but it is worth working through examples and formally laying out the problem. Taxi companies pay insurance that protects clients and other drivers should their employees have accidents. What insurance protections should be required for Uber or Lyft, and who should pay for them? Airbnb effectively ignores, in many cases, land-use rules intended to preserve particular community values. Should that be permitted? A taxi driver is legally required to pick up anyone hailing the cab in any part of town, but an Uber driver can refuse. A hotel must lodge a boarder regardless of ethnicity, and cannot decide that a person’s religion or ethnicity is undesirable. Should the same apply to Airbnb hosts?  With these examples in mind, policy arbitrage is not the ideal basis of a new competitive business model.

The list of policy domains that will have to be rejiggered in a platform era must include competition policy, taxation rules, service provision requirements, and intellectual property rules. We may want to review existing public policies, recognizing that platform entrants playing the arbitrage game likely will press for policy revisions.

Importantly, it is not possible, or even appropriate, to simply declare that old rules and values will apply in this new era. For one thing, it is not always possible to transport the values of one era to the next. Clearly, platforms raise new issues about market dominance and the ability to extend position from one market into another. We only have to look to Europe’s struggles with Google, and look back at the battles around Microsoft. As seemingly settled fights in all these policy domains are reopened, issues will be refought, and new outcomes can be expected.

Conclusion: Takeaways to Consider

Technologies—the cloud, big data, algorithms, and platforms—will not in themselves dictate our future. Rather, the future is ours to choose. How we deploy the technologies, and the rules set for their deployment and use, will be critical. When we look at cases such as electric utility grids. call centers or radio-frequency identification in retail, we find that the market and social outcomes of new technologies vary across countries. Of course, a chosen technology path frames choices.36 Larry Lessig, what seems like long ago, wrote that Code is Law37, and code is increasingly West Coast law. If not solely the technology, what then explains Platform Economy variances between and within countries?

Deployments differ with corporate strategies and public policies. Consider corporate strategies. Will companies view workers as costs to be contained, or as assets to be developed and promoted, even in an era of algorithms, robots and automation?38 And, as important, will worker assets be directly tied to the firm, and what entity should bear the costs of their conservation and improvement? In other words, the old question of the boundaries of the firm are being reposed.

Consider policies. What balance do we seek between flexibility/adaptability and social protection? Do we recognize that flexibility can come with protections against risk?

The consequence is that we will be making choices in an inherently fluid environment shaped, to some degree, by unpredictable technical changes, and by social reaction to these changes. Ultimately, all of this will depend on how we believe markets should be structured, what we socially value, and how we will channel the enormous value that these socio-technical changes are creating.

About the Authors and Acknowledgements
Martin Kenney is professor of Community and Regional Development at the University of California, Davis, and a member of the Berkeley Roundtable on the International Economy.

John Zysman is codirector of the Berkeley Roundtable on the International Economy, and also is a professor, Political Science, at the University of California, Berkeley.

Each author contributed equally to the formulation and development of the ideas in this paper. They extend thanks to Ruth Collier, Lilly Irani, Bryan Pon, and Anne Visser for their comments about earlier work, and their contributions to the discussion from which this paper emerged.


  1. Zysman, J. 2006. “The algorithmic revolution — the fourth service transformation,” Communications of the ACM. 49 (7).
  2. Direct wording is borrowed and slightly modified from Mattila, J., and T. Seppälä. 2015. “Machines in a Cloud – or a Cloud in Machines? Emerging New Trends of the Digital Platforms in Industry and Society”. The foundational book describing platforms and their use is Gawer, A., and M. A. Cusumano. 2002. Platform Leadership: How Intel, Microsoft, and Cisco Drive Industry Innovation (Boston, Harvard Business School Press). Also, see the more recent Gawer, A., and M. A. Cusumano. 2014. “Industry Platforms and Ecosystem Innovation,” Journal of Product Innovation Management 31(3), 417-433.
  3. Kushida, K. E., J. Murray, and J. Zysman. 2012. "The Gathering Storm: Analyzing the Cloud Computing Ecosystem and Implications for Public Policy," Communications and Strategies 85:63-85.
  4. For a theoretical conceptualization of platforms as private market regulators, see Boudreau, K. J., and A. Hagiu. 2008. Platform rules: Multi-sided platforms as regulators. available at SSRN 1269966.
  5. For an early discussion of the issues, see Ballon, P., and E. Van Heesvelde. 2001. “ICT platforms and regulatory concerns in Europe,” Telecommunications Policy, 35(8), 702-714.
  6. Cohen, S., and J. Zysman. 1988. Manufacturing Matters: The Myth of the Post-Industrial Economy (New York: Basic Books).
  7. Florida, R., and M. Kenney. 1990. The Breakthrough Illusion: Corporate America’s Failure to Move from Innovation to Mass Production (New York: Basic Books).
  8. Florida, R. 2002. The Rise of the Creative Class (New York: Basic Books).
  9. Friedman, G. 2014. “Workers without employers: Shadow corporations and the rise of the gig economy,” Review of Keynesian Economics2(2): 171-188.; Standing, G. 2011. The Precariat: The New Dangerous Class (London: A&C Black).
  10. Yochai Benkler is an early proponent of this framing. See, for example, Benkler, Y. 2006. The Wealth of Networks: How Social Production Transforms Markets and Freedom (New Haven: Yale University Press). For a critique of Benkler, see Dijck, J. V. 2013. The Culture of Connectivity: A Critical History of Social Media (Oxford: Oxford University Press).
  11. Robert Noyce founded Intel, Steve Jobs established Apple, and Bill Gates founded Microsoft. Others in that early cohort, particularly the semiconductor industry folks who were driving the revolution, included Jerry Sanders of AMD, Charlie Sporck of National Semiconductor, and Bill Davidow, who re-emerged as a pessimistic commentator on our digital futures following a career at Intel and in venture capital.
  12. Gapper, J. 2015. “Silicon Valley has become a dream factory,” Financial Times, May.
  13. Unsurprisingly, Zipcar was purchased by Avis in 2013. It is important to note that Zipcar is an asset-heavy business model, as it needs to own the vehicles.
  14. Vonnegut, K. 1952. Player Piano (New York: Charles Scribner).
  15. Davidow, W. H., and M. S. Malone. 2014. “What Happens to Society When Robots Replace Workers?” Harvard Business Review.
  16. Brynjolfsson, E., and A. McAfee. 2014. The Second Machine Age: Work, Progress, and Prosperity in a Time of Brilliant Technologies (New York: Norton).
  17. See Frey, C. B., and M. A. Osborne. 2013. “The Future of Employment: How Susceptible Are Jobs to Computerization?,” September 17, 2013.
  18. From our own research, we would propose that today’s home-run firms create far fewer jobs than did GE, GM, or Ford in their era. And yet, today’s firm exert just as much organizing influence over the social world and over our thinking about the social world. Google employs fewer folks than Microsoft does, and maybe fewer than Microsoft did at a similar stage in its life. However, Google is far more omnipresent than Microsoft ever was.
  19. This point was introduced into our debates by Niels Christian Nielsen’s persistent arguments. See, for example, Nielsen, N. C., J. Murray, and J. Zysman. 2013. The Algorithmic Revolution and Empowered Human Value Creation (Danish Technology Institute Rosenthals/Book PartnerMedia).
  20. Ibid.
  21. Turing, A.M. 1950. “Computing machinery and intelligence,” Mind, 59, 433-460. See also John Searle’s Chinese Room discussion. “John Searle in his construction of the Chinese Room problem distinguishes likewise between the imitation of consciousness, in this case capacity to understand and use a language, and the ability to imitate and translate without understanding.”  For a simple depiction of the argument: John Searle, pages 21-24 (Continuum Books, Joshua Rust, 2009).
  22. This often will require case study research, and what Richard Nelson has termed “appreciative theorizing.” Appreciative theorizing is qualitative in nature as opposed to the formal modeling and quantitatively derived theorizing of economics.
  23. For example, Watson, B. C. 2013. “Platforms, Productivity, and Politics: Comparative Retail Services in a Digital Age,” The Third Globalization? Can Wealthy Nations Stay Rich in the Twenty-first Century?.; Breznitz, D., and J. Zysman, editors. 2013. (Oxford University Press). On call centers, Batt, R. 2002. “Managing customer services: Human resource practices, quit rates, and sales growth,” Academy of Management Journal, 45(3), 587-597.
  24. The German studies Industrie 4.0 and Smart Services Welt put a distinct emphasis on existing German strengths and how to preserve them. They are a very different flavor, in our view, from the American discussions.
  25. Zysman, J. 2006. “The algorithmic revolution — the fourth service transformation,” Communications of the ACM. 49 (7).
  26. Jonathan Murray, Kenji Kushida, and Patrick Scaglia have been essential to our understanding of these issues. See, for example, Kushida, K. E., J. Murray, and J. Zysman. 2011. "Diffusing the Fog: Cloud Computing and Public Policy." Journal of Industry, Competition and Trade 11(3):209-237.; Kushida, K. E., J. Murray, and J. Zysman. 2012. "The Gathering Storm: Analyzing the Cloud Computing Ecosystem and Implications for Public Policy." Communications and Strategies 85: 63-85.; Kushida, K. E., J. Murray, and J. Zysman. 2013. "Clouducopia: Into the Era of Abundance," CLSA Blue Book January.; Kushida, K. E., J. Murray, and J. Zysman. 2014. “The Next Epoch in Cloud Computing: Implications for Integrated Research and Innovation Strategy (BRIE Working Paper,  2014-4).; Kushida, K. E., J. Murray, and J. Zysman. 2015. "Cloud Computing: From Scarcity to Abundance," Journal of Industry, Competition and Trade 15(1): 5-19.
  27. Hatch, M. 2014. The Maker Movement Manifesto: Rules for Innovation in the New World of Crafters, Hackers, and Tinkerers (New York: McGraw Hill).
  28. Geography is not completely irrelevant. For many functions, speed matters; consider high-speed trading activities. Even search benefits from fast responses.
  29. It is important to note that fundamental insights regarding the importance of platforms come from work by Michael Cusumano and Annabelle Gawer that drew upon the history of Microsoft, Intel, and Cisco. Gawer, A., and M. A. Cusumano. 2002. Platform Leadership: How Intel, Microsoft, and Cisco drive industry innovation (Boston: Harvard Business School Press). There is now a large and fascinating body of literature about how firms should develop their platforms.
  30. Neff, G., 2012. Venture Labor: Work and the burden of risk in innovative industries (Cambridge: MIT Press).
  31. There are exceptions, as Uber and Lyft have quite stringent control over their service providers. Ultimately, the courts may decide that their drivers are, in fact, employees, thus challenging their current model. See, for example, the court ruling in San Francisco that a lawsuit about whether Uber drivers are employees could be tried by a jury. The lawsuit’s outcome will significantly affect the status of labor in a number of these firms. Wilson, M. 2015. “Juries to Decide Whether Uber, Lyft Drivers Are 'Employees” (Findlaw, March 12, 2015).
  32. Of interest is what happens to these “sharing” sites as they grow, and the owners receive venture capital and must monetize operations in preparation for an exit event.
  33. Today, the Prius seems to be the universal choice for Uber drivers.
  34. Fligstein, N. 1993. The transformation of corporate control (Cambridge: Harvard University Press).
  35. Many have noted that a number of young entrepreneurs who created platforms came from social backgrounds in which their parents were able to support them during their “hacking adventures.” If they failed, they could return to the university. The fact that these entrepreneurs were backstopped by family and social status does not diminish their accomplishments, but provides context.
  36. See, for example, Latour, B. 1990. “Technology is society made durable,” The Sociological Review, 38(S1), 103-131.
  37. Lessig, L. 1999. Code and Other Laws of Cyber Space, (New York: Basic Books).
  38. Ton, Z. 2014. The Good Jobs Strategy: How the smartest companies invest in employees to lower costs and boost profits (Amazon Publishing).

What Do Trends in Economic Inequality Imply for Innovation?

A Framework for Future Research and Policy

By Elisabeth Jacobs Senior Director, Policy + Academic Programs Washington Center for Equitable Growth
Elisabeth Jacobs
Elisabeth Jacobs


Most economists agree that contemporary levels of economic inequality in the United States are at near-record highs. And most economists also concur that innovation is a critical engine driving economic growth. In this essay, I develop a framework connecting the two—rising inequality on one hand and declining levels of innovation and dynamism on the other. The result is a set of research questions designed to stimulate a conversation about how the innovation channel may be a key mechanism through which rising levels of inequality are affecting the overall health of the American economy. Inequality’s impacts on innovation may work differently across the income distribution, and the framework presented in the pages to follow accordingly works through how inequality may be impeding innovation through its implications for the bottom, middle, and top echelons of the American economy. When viewed this way, the potential mechanisms through which inequality may be impacting economic growth come into clearer focus, as do a set of policy implications worthy of consideration.

Rising Inequality, Declining Dynamism

The rise in economic inequality over the last quarter century has been well documented. UC-Berkeley economist Emmanuel Saez finds that the share of pre-tax income going to the top 1 percent of American taxpayers has risen from 9 percent in 1970 to 20 percent in 2012.1 The non-partisan Congressional Budget Office finds that, between 1979 and 2007, post-tax-and-transfer income grew by 275 percent for the top 1 percent of households, as compared to less than 30 percent for those in the twenty-first to eightieth percentile.2 The concentration of wealth has been even more dramatic. Gabriel Zucman at the London School of Economics finds that wealth is highly concentrated at the pinnacle of the distribution, with the richest 0.1 percent of taxpayers holding 22 percent of the nation’s assets in 2012, up from 7 percent in 1979.3 In short, no matter how you slice the economic pie, a larger and larger piece has gone to those at the tippity-top.

Politicians and researchers long have worried about the implications of the contemporary distribution of economic resources—the most unequal since the Roaring Twenties, by most measures. Between 1947 and 1979, across the income distribution, average family incomes grew at a pace of just over 2 percent a year. In the period from 1979 to 2007, families in the bottom quintile experienced virtually no growth, while families in higher quintiles saw progressively greater annual income growth. Moreover, income growth regardless of quintile has sputtered over the last quarter century, with average growth rates well below the 2 percent of the post-World War II period.4 The economic lives of those in the middle and bottom of the income distribution have shifted in fundamental ways as more of the returns to growth have flowed to the top.

An emerging body of research is beginning to re-investigate what trends in inequality mean for the dynamism of the economy as a whole. For more than half a century, Simon Kuznets’s Nobel Prize-winning observation dominated conventional wisdom, holding that inequality was a transitional phase that would reverse itself once poor countries evolved into developed nations.5 Six decades later, and armed with far more sophisticated data and computational power than what Kuznets had available at Harvard in the 1950s, today’s researchers offer serious challenges to the U-shaped Kuznets Curve. For instance, IMF economists Andrew Berg and Jonathan Ostry made waves with their 2011 study demonstrating the negative relationship between sustained economic growth spells and high levels of income inequality.6 World Bank economist Roy van der Weide and City University of New York Graduate Center economist Branko Milanovic use state-level data for the United States and find that high levels of income inequality decrease income growth for those at the bottom of the distribution while simultaneously increasing income growth for those at the top.7 Harvard economist Nathanael Hendren calculates that income inequality reduced U.S. economic growth by 20 percent over the last four decades, adding up to a “social cost” of roughly $400 billion.8

While the empirical literature on the macro relationship between inequality and growth continues to evolve, a key set of questions remains urgently awaiting answers. Even if macroeconomists were to come to a universal consensus tomorrow that inequality was negatively correlated with economic growth, policymakers still would need evidence-backed guidance directing them to the appropriate levers for mitigating the problem of inequality and, in turn, jump-starting equitable growth. In order to gain traction—to make an abstract conceptual problem concrete and to give policymakers specific places to look for solutions—we need to shift the focus to include mechanisms. Specifically, how might widening inequality impact economic growth and stability? Through what channels might inequality’s impacts on the broader economy flow? When we turn the lens in this direction, the importance of innovation and dynamism begin to come into focus.

Innovation, Dynamism, and Growth

Innovation is a critical engine driving economic growth. As Council of Economics Advisers Chairman Jason Furman recently remarked, for an advanced economy like the United States, “Catching up to the productivity frontier is not possible when you are already there.”9 Borrowing from Simon Kuznets (via Kenneth Arrow), the concept of innovation is best understood broadly, as “a new combination of existing knowledge to create something useful (in some sense).”10 A long history of economics research puts innovation at the center of most growth models, with more recent work focusing on small, young firms—startups—as the locus of a disproportionately large share of innovation.11 Startups’ unique role in fostering innovation stems from their ability to generate market turbulence, competition, and industry renewal, and they can serve as particularly effective competitors in arenas that require flexibility and the ability to efficiently respond to niche markets.12 And startups generally require an individual actor, the business founder/owner—in other words, the entrepreneur.

Why are startups so important for generating innovation and, in turn, economic growth? The research points to two main channels: job creation and productivity. While the empirics on the productivity channel are less than clear, the evidence on the job-creating power of the startup is more persuasive.13 Business startups account for 20 percent of gross job creation in the United States, while high-growth businesses (which are disproportionately young) account for almost 50 percent of gross job creation. Taken together, startups and high-growth firms (which are typically young businesses) account for a whopping 70 percent of gross firm-level job creation annually. As University of Maryland economist Ryan Decker and his colleagues summarize, “Startups and young businesses are small, the underlying reason why many commenters describe small businesses as the engine of job growth.”14 It is worth noting that firm age is more important than firm size—once researchers control for firm age, a business’s size has no relationship to its growth trajectory.15

Also worth noting: most startups fail. The majority that last don’t grow substantially. But, among the minority that survive are the handful of firms that make an outsized contribution to net job creation. The lesson here is that supporting entrepreneurship is a high-risk, high-reward enterprise on multiple levels. On a micro level, the prospect of business failure for a given individual entrepreneur is high, and comes with substantial costs. On a macro-level, promoting an economy based on the dynamism inherent in startups requires substantial tolerance for failure and volatility. On both micro and macro levels, the risk inherent in an entrepreneurial economy requires a set of public policy solutions designed to incentivize “smart” risk-taking and to protect against the major downsides of that risk.

Given the importance of young firms for innovation, it is worth pausing here to review the contemporary empirical state of affairs for startups in the United States. Recent evidence suggests that many American entrepreneurs do not have aspirations for high growth or innovation, but rather have started their businesses for non-pecuniary reasons such as flexibility in work hours.16 These aspirations are consistent with the industry characteristics of most small businesses, which are concentrated among skilled craftsmen, lawyers, real estate agents, doctors, small shopkeepers, and restaurateurs.17 So it is perhaps unsurprising that the majority of startups exhibit low rates of job creation. And it remains an open question whether and how policy shifts could incentivize more of these “subsistence entrepreneurs,” whose self-employment plays a key role for the individuals and their families, to make the leap into “transformational entrepreneurs,” who play a key role in both innovation and job creation.18

Multiple studies have noted a marked decline in the startup rate. The annual startup rate declined from 12 percent in the 1980s to an average of 10.6 percent just before the Great Recession, when it plunged to below 8 percent. The rate of the decline has varied across sectors, but applies broadly across the economy.19 The slowing average startup rate has not been counterbalanced by an increase in the size of the average startup—depending on the data used, average startup size has either decreased or remained stable since the 1980s.20 The combined result of these two empirical trends is that startup businesses play a smaller role in the economy today than they did in the past. Firms five years old or younger comprised 47 percent of all firms in the late 1980s, but declined to 39 percent prior to the Great Recession, and their share has continued to fall since. Recent empirical work shows that startup slowdown accounts for 32 percent of the observed decline in job creation, 20 percent of the decline in job destruction, and 26 percent of the decline in job reallocation between the 1980s and 2000s.21

Given their role in fostering innovation and driving job creation, the startup slowdown has obvious implications for economic dynamism and growth. An open question, and the focus of the remainder of this essay, is how economic inequality might be playing a role in the startup slowdown as well as on other indicators of economic dynamism, especially innovation. For the sake of introducing some analytic clarity, I approach inequality slice-by-slice, suggesting ways that inequality’s impacts may flow through the economic realities for the bottom, middle, and top of the income distribution. Doing so necessarily requires a focus not just on firms, but on individuals—a departure for much of the research on growth, dynamism, and entrepreneurship, but a critical one. The bottom-middle-top framework below provides an analytic lens for structuring a research agenda with many critical, yet unanswered, questions.

Inequality-Innovation Channel I: Wasted Potential at the Bottom?

Entrepreneurship is at the heart of the rags-to-riches mythology, the lifeblood of the American Dream. The idea that a poor kid with gumption and a great idea can rise out of poverty and into the vaunted top echelons of society has been echoed time and time again, from Horatio Alger to Jay Gatsby to Mark Cuban. A wealth of new empirical work suggests that the intergenerational mobility embodied by the American Dream largely has been overstated.22 What remains an open question is whether the broken promise of the American Dream is linked to trends in entrepreneurship and innovation. Are low-income individuals more or less likely to succeed in entrepreneurship compared to their wealthier counterparts, all things being equal? How has increasing inequality affected the likelihood of the poor achieving entrepreneurial success?

New research from a team of economists from Harvard University, the U.S. Treasury Department, and the London School of Economics quantifies just how far out of reach entrepreneurial success is for the vast majority of children born into low-income families.23 The team matches data on U.S. patents with tax returns of the people receiving the patents and the tax returns of those patent recipients’ parents. This multi-generational match allows them to see not only how much these inventors earn as adults, but also their families’ economic resources during the inventor’s childhood.24 As a result, for the first time ever, we have the beginnings of a picture of the lifecycle of an inventor.

Unsurprisingly, children born to wealthy parents are far more likely than are poor children to obtain a patent later in life. More surprising is the sheer magnitude of that patent gap. For every 10,000 children born to families in the top 1 percent, 22.5 will receive a patent in adulthood. In contrast, just 2.2 of every 10,000 children born to families with incomes below the median will receive a patent in adulthood. Worth noting: just as income inequality is characterized by runaway rates at the very top of the distribution, so too are patent rates. Children born to families in the ninety-ninth percentile are twice as likely to obtain patents in adulthood as compared to children born into the ninetieth percentile.

Importantly, the data indicate that the vast majority of the patent gap is not explained by inherent skills differences between low- and high-income children. The researchers merge test score data for third graders in their study, and find that patent rates in adulthood are roughly equal for low- and high-income kids with similar test scores—except for the high-performers. Among high-performing children, those born into high-income families are about four times as likely to obtain patents as are their peers born into low-income families.

What explains the patent gap, then? The lion’s share of the gap is explained by the cumulative impact of educational disadvantage. The share of the patent gap explained by test scores grows by nearly 5 percent per grade from the third-grade baseline through eighth grade (the last year for which the researchers have testing data available). And, when the researchers look at the role of college quality in determining patent receipt, the importance of the human capital channel grows even starker: 90 percent of the income-innovation relationship is explained by whether or not an individual attended a high-quality college.25

In sum, the empirical evidence suggests that low-income children face major barriers to successful entrepreneurship, even when they show significant promise. Despite early measured ability, poor children are significantly less likely to patent than are their wealthy peers. Subsequent schooling essentially accounts for this difference, which suggests that human capital policy is hugely important for fostering innovation from across the income spectrum. But we need far more research quantifying the lifecycle of inventors and, more generally, entrepreneurs. The Harvard team’s cutting-edge creation of a dataset with multigenerational layers of administrative data represents the frontier for research that will allow us to better understand the implications of inequality for entrepreneurship at the bottom. Better understanding macro-level phenomena requires the application of novel, large-scale micro-level data to answer key questions, an exciting, promising endeavor that remains in its early days.

Inequality-Innovation Channel II: Risk Aversion in the Middle?

For millions of middle-class Americans, the idea of “striking it rich” is far from their imaginations. Middle-class incomes have stagnated, with rising women’s labor force participation largely keeping the middle-class family afloat over the last half century.26 The experience of the typical working American may mitigate against entrepreneurship for multiple reasons. First, the time crunch faced by millions of American families may crowd out the potential for developing successful business plans. Second, the increase in both perceived and real economic insecurity may mean the middle class is substantially more risk averse than in the past. Third, the middle class may face serious capital constraints in the face of low and eroding wealth, with implications for the rate of business startups.

Working families face a serious time crunch. Delayed marriage and child-bearing, more births outside of marriage, the increase in women’s labor force participation, and the aging of the population have altered family life and created new challenges for those with caregiving demands.27 American mothers have decreased the time they spend on housework as their time in the labor market has gone up, but they have also increased the time spent on childcare. Fathers have increased the time spent on childcare as well. Intensive child rearing practices are more common, perhaps in response to the rat race that comes along with rising inequality and higher perceived costs of failure.28 Workers increasingly receive non-standard schedules, with high levels of unpredictability adding additional stress for families balancing care responsibilities.29

Combined with dual-earner families and single parenting, these factors all translate into unprecedented levels of time pressure for many working American families. How is the time crunch impacting the entrepreneurial success rate of the American middle class? If potential entrepreneurs are collapsing at the end of a busy day, exhausted after a long day of work and “second-shift” responsibilities caring for children (and, increasingly, aging parents), what’s the likelihood that they have the energy to push their nascent big ideas into transformational ventures?

As noted earlier, entrepreneurship inherently involves risk. The vast majority of startups fail. Recent empirical work raises serious questions about whether the middle class may have grown more risk averse, even among those whose lives are financially stable enough to serve as fertile ground for potential entrepreneurship.

Even prior to the Great Recession, public opinion polls suggested high levels of economic anxiety. A 2007 poll revealed that more than half of all Americans felt they had not moved forward, while nearly a third say they have fallen back.30 Only 41 percent said they were better off than they were five years ago—the lowest level in nearly 50 years. Meanwhile, the share saying they were worse off than they were five years ago rose to 31 percent, the highest level in almost half a century. Note that these perceptions match the reality of the American experience fairly closely. Lower- and middle-income families’ income growth has been very slow over recent decades, and has declined somewhat over recent years. In contrast, higher income households’ income growth has been strong, much stronger than growth for everyone else over the last four decades. So, it is easy to see why many Americans feel they are falling further and further behind.

Survey data provide ample evidence of the economic precariousness of the American experience. For instance, the 2009 TNS Economic Crisis survey asked households about their capacity to come up with $2,000 in thirty days. About one-quarter of Americans reported that they would certainly not be able to come up with such funds, and an additional 19 percent reported they would do so by pawning or selling possessions or taking payday loans. Based on this finding, Dartmouth University economist Annamaria Lusardi and her colleagues determined that nearly one-half of Americans are financially fragile, including a sizeable fraction of seemingly “middle-class” Americans.31 Families simply don’t cope with risk using “precautionary savings,” either. While savings come first in the “pecking order” of coping methods, the typical household relies heavily on family and friends, formal and alternative credit, increased work hours, and selling items. Empirical studies of short-term income volatility—for instance, the likelihood of experiencing a large monthly or annual drop in income—suggest a rising risk of economic insecurity. More than one in seven families experienced a drop of income of at least 50 percent in a given four-month period, on average, between 1996–2004. The probability of not fully recovering from a substantial drop in income within a year rose sharply between 1996 and 2004, with 81.9 percent failing to recover in 1996 versus 92.4 percent in 2004.32

Stagnant incomes, financial fragility, and a relatively high probability of experiencing a substantial and irredeemable drop in income in the short term all combine to provide good reason to wonder whether risk tolerance among America’s middle class has indeed shifted. In the face of economic precariousness, what happens to the probability of entrepreneurial success? How, if at all, does the half-century-long shift in economic risk from institutions to individuals impact the entrepreneurial spirit?33

Finally, consider the impact of the erosion of middle-class wealth on the probability of entrepreneurial success. There are plenty of good reasons to imagine that liquidity constraints affect an individual’s likelihood of starting a business—simply put, starting a business costs money. Personal wealth can be seed capital or loan collateral. One way of quantifying liquidity constraints is through individual or household wealth. Early seminal work from Global Economics Group economist David Evans and his colleagues established a positive relationship between wealth and the probability of becoming an entrepreneur.34 More recent work from World Bank economist Camilo Mondragon-Velez extends on this analysis to show that most potential entrepreneurs in the economy—especially those below the top of the wealth distribution—face capital constraints when making the decision to start businesses.35 Moreover, given the role of liquidity constraints, differential access to financial markets can impact who becomes a successful entrepreneur. Increased financing costs and limited access to borrowing for lower-income individuals can hamper entrepreneurship.

The empirical trends in middle-class families’ asset portfolios provide reason to ask whether liquidity constraints are hampering successful entrepreneurship. Median wealth in America in 2013 was at its lowest rate since the early 1980s.36 The average family’s net worth plummeted between 2007 and 2010, mainly due to high leverage of the average American household prior to the recession and the prominence of housing in the average asset portfolio.37 The racial and ethnic wealth gap, largely stable from 1983 to 2007, widened dramatically over the course of the Great Recession. The wealth of households under the age of forty-five has taken an especially hard hit in recent years.38

Taken together, these three basic sets of trends outline an empirical reality for middle class America with potentially important implications for successful entrepreneurship. American families face serious time pressures, which may be constraining their ability to do the work necessary to move from the daydreaming stage to the reality of a startup. The economic precariousness that now characterizes millions of Americans’ everyday lives may be holding back potentially transformative ideas from taking flight, as Americans become more risk averse in the face of rising downside risk. And liquidity constraints stemming from eroding wealth and limited access to borrowing for credit-constrained individuals may be an additional roadblock. All three of these pathways are key channels through which inequality’s impacts on the middle of the economic distribution may be holding back innovation and, in turn, hampering broader economic growth and dynamism.

Inequality-Innovation Channel III: Perverse Incentives at the Top?

Top-end income and wealth have skyrocketed over the last half century. Twenty percent of pre-tax income in the United States now goes to the top 1 percent of American taxpayers.9 The wealthiest 0.1 percent of taxpayers held 22 percent of the nation’s assets in 2012.40 How does the pulling away of the top of the economic distribution impact levels and trends in innovation and entrepreneurship? Below, I work through two potential mechanisms worthy of further exploration.

First, consider the fact that an increasing fraction of top-end income earners in the United States are employed in the financial sector—investment bankers and institutional investors.41 How has financialization impacted innovation and entrepreneurship? The transition of the United States economy over the last three decades from manufacturing-dominated to finance-driven is well established.42 Corporate governance is increasingly more responsive to financial markets than to product markets.43 Financialization has reshaped managerial priorities away from market share and toward short-term profits, which, in turn, has implications for investments in research and development—and, in turn, potential consequences for innovation, as firms are more inclined to take a short-term approach to business rather than a long-term view.44 The research and development funding necessary for breakthrough innovations may have long-term payoffs but yield limited short-term results, thus taking a backseat in an economy characterized by high levels of financialization.

Relatedly, investing in workers’ skills and talents also may be disincentivized in a world where short-term shareholder value trumps a longer-term vision. Underinvesting in workers may have hidden negative spillover effects for innovation and entrepreneurship. Today’s large firms are the incubators of tomorrow’s high-growth startups.45 If financialization means a shift in focus from long-term goals to short-term goals, the role of the firm as a locus for human capital development may have shifted in fundamental ways. And, for reasons noted below, the public sector has not responded accordingly to the erosion of a key private-sector capacity.

Second, consider the potential political dynamics at play. The outsized influence of money in the political process means that a well-organized conservative movement deeply committed to reducing the size and scope of government plays a powerful role in shaping political debates and outcomes.46 Continued cuts in public investment in research and development—including cuts to the National Institutes of Health, the National Science Foundation, and elsewhere—are arguably a consequence of organized economic elites determined to rein in the role of government and, in turn, are starving the country of investments in critical research and development that could spur important innovations.

Likewise, the anemia of the social insurance system that protects against some of the risks faced by middle- and lower-class households also can plausibly be traced down the same path. Consider the organized fight against the Affordable Care Act. A growing body of empirical work shows that access to affordable, universal health insurance can play a critical role in stimulating entrepreneurship by untethering workers’ health care needs from employers.47 Yet, the push by a small but highly-organized, well-resourced economic elite jeopardizes a key-element social insurance policy with the potential to jump-start entrepreneurial activity.

From Research Frameworks to Policy Implications

The above overview is meant to provide a provocative set of questions, rather than definitive answers. More research is needed if we are to fully understand whether and how inequality impacts economic growth via the innovation channel. But the above analytic framework implies a constellation of policy areas worthy of consideration by those looking to rejuvenate a dynamic, growing economy.

First, human capital policies are critical. The research on the relationship between childhood economic status and adult patent receipt suggest that the American educational system is fundamentally failing students born into low-income families, especially those who show academic/intellectual promise in the early years. Gifted and talented programs for promising low-income youth, mentorship programs, and other creative investments in public schools are all worth putting on the table. The evidence suggests that patent receipt maybe be just as much about who an individual knows as it is about what that person knows. As a result, policies that more effectively connect talented low-income young adults with mentorship opportunities and connections to investors are worth considering. Finally, given the early age at which gaps open up, universal pre-kindergarten programs belong in the discussion—especially because of the important role they can play for families across the income distribution. Pre-kindergarten is an investment in children, but it is also an investment in parents, including middle-class families facing a major time crunch that may be impacting their ability to act on their entrepreneurial instincts.

Second, social insurance policies are a key element of a new entrepreneurial policy framework. Recent empirical work on the role of social insurance suggests the key role it plays in fostering entrepreneurship, especially for those in the lower and middle tiers of the income distribution. For instance, HEC Paris economist Johan Hombert and his colleagues find that extending unemployment benefits to individuals who start their own companies (and further extending those benefits if the new company failed) increased the rate of business startups by 10 percent across all industries.48 The startups facilitated by the unemployment benefits extension were just as high-quality as other startups, as measured by job creation, growth, and survival rates. And entrepreneurs who used the unemployment benefits to finance their new businesses reported higher levels of ambition than other entrepreneurs did. While Hombert’s study used French data, two recent studies from Harvard Business School economist Gareth Olds find similarly positive impacts of access to social safety-net programs in the United States, including the State Child Health Insurance Program (S-CHIP) and Supplemental Nutrition Assistance Program (SNAP, commonly known as food stamps).49 In short, reducing downside risks through robust social insurance programs can incentivize entrepreneurial entry and promote entrepreneurial success.

Policies designed to mitigate the potential impact of inequality on innovation cannot ignore the impact of top-end inequality. A constellation of tax reforms is almost certainly a necessary part of this solution, given the erosion of the top income tax rates over time. But tax policy alone cannot be the only answer. As noted above, financialization is likely a major culprit in the shift in the American way of doing business, with implications for innovation and entrepreneurship flowing through several channels. Undoing the impacts of financialization may require corporate governance reforms that tackle the problems associated with shareholder value theory—that is, the prevailing corporate ethos of maximizing short-term returns at the expense of longer-term growth, stability, and innovation, and, often, at the expense of the worker.50

Finally, and critically: The vast majority of Americans don’t work for startups, and they aren’t entrepreneurs. Half of private-sector employment in the United States is accounted for by the less than 1 percent of firms with more than 500 employees.51 This basic fact has critical implications for policy, namely that boosting economic dynamism and promoting entrepreneurship requires boosting the quality of life for all Americans. We simply don’t know where the next blockbuster idea might come from, and therefore should be promoting an economy where all people have the capacity to pursue those great ideas, regardless of their economic status.

The simple fact that we don’t know where the next great American breakthrough startup will come from means that bread-and-butter economic policy priorities issues are important, including policies that don’t immediately come into play in the entrepreneurship debate—or, if they do, they typically get push-back in the opposite direction. Consider policies such as the minimum wage, paid sick and parental leave, and expanded access to childcare. All have potential implications for promoting growth and dynamism in the United States and belong on the table. If the goal is to stimulate transformational entrepreneurship, it is not enough just to push narrow policies designed to “support entrepreneurs.” We need a comprehensive package of creative solutions that moves the economy toward broad-based health and growth for all.

About the Author
Elisabeth Jacobs is senior director for Policy and Academic Programs at the Washington Center for Equitable Growth. Her research focuses on economic inequality and mobility, family economic security, poverty, employment, social policy, social insurance, and the politics of inequality.


  1. Saez, Emmanuel. 2013. “Striking It Richer: The Evolution of Top Incomes in the United States.” Berkeley, CA: University of California.
  2. Congressional Budget Office. 2011. “Trends in the Distribution of Household Income Between 1979 and 2007.” Washington, DC: Congressional Budget Office.
  3. Zucman, Gabriel. 2014. “Wealth Inequality in the United States Since 1913: Evidence from Capitalized Income Tax Data.” NBER Working Paper No. 20625. Cambridge, MA: National Bureau of Economic Research.
  4. Mishel, Lawrence, Josh Bivens, and Heidi Shierholz. 2012. The State of Working America (12th Edition). Ithaca: Cornell University Press.
  5. Kuznets, Simon. 1955. “Economic Growth and Income Inequality.” American Economic Review XLV(1): 1–28.
  6. Berg, Andrew G., and Jonathan D. Ostry. 2011. “Inequality and Unsustainable Growth: Two Sides of the Same Coin?” 2011. IMF Staff Discussion Note. Washington, DC: International Monetary Fund.
  7. Van der Weide, Roy, and Branko Milanovic. 2014. “Inequality is Bad for Growth of the Poor (But Not for That of the Rich). World Bank Policy Research Working Paper 6963. Washington, DC: World Bank.
  8. Hendren, Nathanael. 2014. “The Inequality Deflator: Interpersonal Comparisons without a Social Welfare Function.” NBER Working Paper No. 20351. Cambridge, MA: National Bureau of Economic Research.
  9. Furman, Jason. 2014. “Patents, Innovation, and Productivity.” Speech to the Sixth Annual Patent Law and Policy Conference, Georgetown University Law Center and University of California, Berkeley Center for Law and Technology.
  10. Arrow, Kenneth. 2012. “The Economics of Inventive Activity over Fifty Years.” In The Rate and Direction of Inventive Activity Revisited, eds. Josh Lerner and Scott Stern. Chicago, IL: University of Chicago Press.
  11. Simon Kuznets observed in 1962 that the greatest challenge to understanding the role of innovation in economic processes has been the lack of meaningful measures of innovative inputs and outputs. Despite major advances in data collection capturing innovation (patents, research, and development, stock market values of inventive output), precise operationalization of innovation remains a slippery task. See, for instance, Acs, Zoltan J., and David Audretsh. 1988. “Innovations in Large and Small Firms: An Empirical Analysis.” American Economic Review 78(4): 678–690.
  12. Acs, Zoltan J., and David Audretsh. 1990. Innovation in Small Firms. Cambridge, MA: MIT Press.
  13. Decker, Ryan, et al. 2014. “The Role of Entrepreneurship in U.S. Job Creation and Economic Dynamism.” Journal of Economic Perspectives 28(3): 3–24.
  14. Decker et al. 2014.
  15. Haltiwanger, John, et al. 2013. “Who Creates Jobs? Small versus Large versus Young.” Review of Economics and Statistics XCV(2): 347–361.
  16. Hurst, Erik, and Benjamin Wild Pugsley. 2011. “What Do Small Businesses Do?” NBER Working Paper No. 17041. Cambridge, MA: National Bureau of Economic Research.
  17. Hurst and Pugsley. 2011.
  18. The distinction between “subsistence” and “transformational” entrepreneurs is from MIT economist Antoinette Schoar, who developed the concepts to apply to developing economies. Schoar, Antoinette. 2010. “The Divide Between Subsistence and Transformational Entrepreneurship.” In Innovation Policy and the Economy, eds. Joshua Lerner and Scott Stern. Chicago, IL: University of Chicago Press.
  19. Decker et al., 2014.
  20. Reedy, E.J., and Robert E. Litan. 2011. “Starting Smaller, Staying Smaller: America’s Slow Leak in Job Creation.” Kauffman Foundation Research Series: Firm Formation and Economic Growth. Kansas City, MO: Ewing Marion Kauffman Foundation.
  21. Decker et al. 2014.
  22. See, for example: Chetty, Raj, et. al. 2014. “Where is the Land of Opportunity? The Geography of Intergenerational Mobility in the United States.” NBER Working Paper 19843. Cambridge, MA: National Bureau of Economic Research; Corak, Miles. 2013. “Income Inequality, Equality of Opportunity, and Intergenerational Mobility.” Journal of Economic Perspectives 27(3): 79–102.
  23. Bell, Alex, et al. 2015. “Innovation Policy and the Lifecycle of Inventors.” Presentation of preliminary research. Brussels, Belgium. As Bell and his co-authors note, patents are a widely used, yet imperfect, metric for innovation.
  24. Note that the researchers are only able to match data for parents and children born after 1980, so the focus on “adult” patent recipients is perhaps more accurately described as young patent holders, i.e., those that receive a patent by age thirty-two. Still, given that 8.35 percent of patent holders in 2000 were below age thirty-two, their sample is still impressively large. See Bell et al., 2012, for more detail on the data.
  25. Bell, Alex, et al. 2015.
  26. White House Council of Economic Advisers. June 2014. “Nine Facts About American Families and Work.”
  27. Bianchi, Suzanne M. 2011. “Family Change and Time Allocation in American Families.” The Annals of the American Academy of Social and Political Science 636(1): 21–44.
  28. Cooper, Marianne. 2014. Cut Adrift: Families in Insecure Times. Berkeley, CA: University of California Press; Ramey, Garey, and Valerie Ramey. 2010. “The Rug Rat Race.” Brookings Papers on Economic Activity 41(1): 129–199. Washington, DC: Brookings Institution.
  29. Bianci, Suzanne M. 2011. “Changing Families, Changing Workplaces.” Future of Children 21(2): 15–36.
  30. Taylor, Paul, et al. 2008. “Inside the Middle Class: Bad Times Hit the Good Life.” Washington, DC: Pew Research Center.
  31. Lusardi, Annamaria, et al. 2011. “Financially Fragile Households: Evidence and Implications.” NBER Working Paper 17072. Cambridge, MA: National Bureau of Economic Research.
  32. Acs, Greg, and Austin Nichols. 2010. “America Insecure: Changes in the Economic Security of American Families.” Washington, DC: Urban Institute. See also Jacob S. Hacker and Elisabeth Jacobs. 2008. “The Rising Instability of American Family Income, 1968–2004: Evidence from the Panel Study of Income Dynamics.” Washington, DC: Economic Policy Institute.
  33. For more on the shift of economic risk from institutions to individuals, see Hacker, Jacob S. The Great Risk Shift: The New Economic Security and the American Dream. New York, NY: Oxford University Press.
  34. Evans, David, et al. 1989. “An Estimated Model of Entrepreneurial Choice Under Liquidity Constraints.” Journal of Political Economy 97:808–827.
  35. Mondragon-Velez, Camilo. 2009. “The Probability of Transitioning to Entrepreneurship Revisited: Wealth, Education, and Age.” Annals of Finance 5:421–441.
  36. Fry, Richard, and Rakesh Kochhar. 2014. “America’s Wealth Gap Between Middle-Income and Upper-Income Families Is Widest on Record.” Washington, DC: Pew Research Center.
  37. Wolff, Edward N. “The Asset Price Meltdown and the Wealth of the Middle Class.” NBER Working Paper 18559. Cambridge, MA: National Bureau of Economic Research.
  38. Wolff, Edward N. “The Asset Price Meltdown and the Wealth of the Middle Class.” NBER Working Paper 18559. Cambridge, MA: National Bureau of Economic Research.
  39. Saez, Emmanuel. 2013.
  40. Zucman, Gabriel. 2014.
  41. Kaplan, Steven N., and Joshua Rauh. 2010. “Wall Street and Main Street: What Contributes to the Rise in the Highest Incomes? Review of Financial Studies 23(3):1004–1050.
  42. Lin, Ken-Hou, and David Tomaskovic-Devey. 2013. “Financialization and U.S. Income Inequality, 1970–2008.” American Journal of Sociology 118(5): 1284–1329.
  43. Davis, Gerald. 2009. Managed by Markets: How Finance Reshaped America. New York, NY: Oxford University Press.
  44. Stockhammer, Engelbert. 2013. “Why Have Wage Shares Fallen? A Panel Analysis of the Determinants of Functional Income Distribution.” ILO Conditions of Work and Employment Working Paper Series Report No. 35. Geneva, Switzerland: International Labor Organization.
  45. Gompers, Paul, et al. 2003. “Entrepreneurial Spawning: Public Corporations and the Genesis of New Ventures, 1986–1999.” NBER Working Paper No. 9816. Cambridge, MA: National Bureau of Economic Research.
  46. Gilens, Martin, and Benjamin Page. 2014. “Testing Theories of American Politics: Elites, Interest Groups, and Average Citizens.” Perspectives on Politics, Fall (2014).
  47. Fairlie, Robert, et al. 2010. “Is Employer-Based Health Insurance a Barrier to Entrepreneurship?” Santa Monica, CA: Kauffman-RAND Institute for Entrepreneurship Public Policy.
  48. Hombert, Johan, et al. 2014. “Can Unemployment Insurance Spur Entrepreneurial Activity?” NBER Working Paper 20717. Cambridge, MA: National Bureau for Economic Research.
  49. Olds, Gareth. 2014. “Entrepreneurship and Health Insurance.” Manuscript. Cambridge, MA: Harvard Business School; Olds, Gareth. 2014. “Food Stamp Entrepreneurs.” Manuscript. Cambridge, MA: Harvard Business School.
  50. Stout, Lynn. “The Shareholder Value Myth.” European Financial Review, April 13, 2013.
  51. Decker et al., 2014.

Wealth and Generations

By Phillip Longman Policy Director and Managing Editor, Open Markets Program, Senior Editor at Washington Monthly and lecturer at Johns Hopkins University
Phillip Longman
Phillip Longman


Writing during 1965, Social Security Commissioner Robert Ball announced that the American economy was on its way to becoming so productive that there was no reason why government could not eliminate all want among the citizenry. “Poverty in the past has been basically the result of the fact that there was not enough to go around,” wrote Ball. “By contrast, today it can be taken as a fact that the abolition of want in the United States is no longer a problem of economic capacity.”

The next generation was bound to become rich beyond imagination, Bell explained, and so could well afford to pick up the cost of making programs such as Social Security far more generous. “Extremely conservative projections of what has been happening in industry lead to almost unbelievable conclusions," wrote Ball. "If we take not the rate of productivity increases that seems likely to result from the new [automated] approach to problems of production, but instead merely the average rate during the past 100 years, our grandchildren will be able to produce in one day as much as we do in a forty-hour week.”1

The American Dream may be a cliché, but faith in its promise has long defined our social contract and ideologies across the political spectrum. During many eras of U.S. history, liberals and progressives such as Ball have used the assumed upward mobility of future generations to bolster and legitimize their agendas, arguing during the 1960s, for example, that the cost of the War on Poverty, and of new or expanded entitlement programs, such as Medicare, Medicaid, and Social Security, easily could be borne by future taxpayers who inevitably would be many times richer.

The same assumption has been perhaps even more critical to the appeal of conservative ideas throughout American history. In explaining why socialism, communism, and other forms of class-based politics barely got a toehold in the United States, historians typically point to the broad faith most Americans have in their own and their children’s ability to rise up the economic ladder. Broad upward mobility across generations diminished the importance of inherited wealth, strengthened faith in the fairness of markets, and, as Richard Nixon proclaimed in his famous “kitchen debate” with Nikita Khrushchev at the height of the Cold War, made it at least superficially plausible to argue that America was a “classless society.”

Yet with the benefit of hindsight, we now can see that this underlying premise of the American creed has gradually been turned on its head. Until roughly the 1970s, inequality between generations was large and increasing, but for the happy reason that most members of each new generation far surpassed their parents’ material standard of living. Today, inequality between generations is increasing for the opposite reason. Most workers are, indeed, many times more productive than their counterparts in the past yet, by most measures, they are falling farther and farther behind their parents’ generation in economic well being.

The implications run deep. For younger Americans, the new normal of stagnant or falling living standards compared to the prior generation requires new life strategies, including much shrewder and more deliberative plans for building human capital and lifetime net wealth. At the same time, we need public policies that do not simply assume that each new generation will be richer than the last, but that give individuals and families the specific tools they need to pursue opportunity and upward mobility in the twenty-first century.

Downward Mobility Takes Hold

Defining and measuring the “standard of living” enjoyed by different generations is not straightforward. Among the complicating factors are changes in family structure, the role of women, and the ethnic and racial profile of the population. Other considerations include the true measure of inflation, the amount of financial and unemployment risk borne by individuals in different eras, and changes in educational attainment. Although no single metric is perfect, in combination they tell a dramatic and, by and large, depressing story.

The most straightforward “apples-to-apples” comparison is between the amount of income the typical (median) male with a specific level of education makes today compared with what his counterpart in the previous generation made. According to work done by economists Michael Greenstone of the University of Chicago and Adam Looney of the Brookings Institution, the steepest downward mobility has been among male high school dropouts who, during 2009, earned 66 percent less (adjusted for inflation) than their counterparts did during 1969,2 due to a combination of falling real wages and declining labor force participation rates. The slide for men with only a high school degree, who constitute the majority of men, was a staggering 47 percent. College-educated men did better, but only by falling not as far. For prime-aged male college graduates, real earnings during 2009 were 12 percent below those enjoyed by their counterparts forty years before. Even among college graduates who worked full time, real earnings were 2 percent below that of their counterparts during 1969.

These trends were well in place before the coming of the great recession. According to work by Jeff Madrick and Nikolaos Papanikolaou, between 1969 and 2005, real earnings for full-time male workers, ages 25-34 with only a high school degree, declined from $34,681 to $30,000 (in 2005 dollars).3 Meanwhile, full-time, college-educated male workers of the same age eked out hardly any gains compared with their counterparts in the previous generation, as real wage and salary income for this group increased at an annual growth rate of just 0.1 percent between 1969 and 2005.

A similar comparison between today’s working women and their counterparts a generation ago reveals an only slightly less dramatic story. For example, among full-time working women, ages 30-45, who lack a high school degree, real wages were 12 percent lower during 2013 than they were for their counterparts during 1990. For the typical woman in this age group who has a high school degree but never graduated from college, wage and salary increases have been hardly measurable from one generation to the next, rising by just 3 percent between 1990 and 2013. Only college-educated women who worked full time saw any substantial gains compared with their counterparts of 1990. This was mostly because of increasing numbers of women moving into managerial jobs rather than to any general increases in wages for the same work.4

These trends for men and women converge in the statistics about family income, which, especially for the young, had been falling consistently year after year, even before the great recession. The median income among families headed by someone younger than 35 was just $35,500 during 2013. Adjusted for changes in the Consumer Price Index, that is nearly 20 percent below what young families earned during 2001.5

Lifetime Income Gains Peak Earlier and Contract

Another measure to consider is the ever earlier age at which workers’ earnings peak. During nearly all previous eras, workers normally saw their incomes rise in their 20s, 30s, 40s and 50s as they gained education and experience, and as wage rates in general grew. Although their earnings might be interrupted by illness or temporary unemployment, most workers generally earned more each successive year until they retired, typically in their 60s. This pattern still held until 2000, after which Americans started seeing their earnings peak and then decline at younger and younger ages, even as the standard retirement age went up.  

The tipping point came with those born between 1946 and 1950. The median household income of these early-wave Baby Boomers rose steadily during their early working years. Adding to these gains in household income was a sharp increase in the number of working women, as the “two paycheck” family gradually became the middle-class norm. Yet despite this additional income from women, median earnings for these households started declining when their prime wage earners were still in their early 50s—a time of life when members of previous generations still typically were seeing gains in real income from year to year. For these early Baby Boomers, median household income peaked during 2000 at $78,458 at age 50-54 and fell each year thereafter, reaching an inflation-adjusted $50,834 during 2013.6

This pattern has grown progressively worse since then. For example, those born between 1953 and 1957 saw their median household income peak at $77,543 during 2002, when they were ages 45-49. For them, household income subsequently fell by 0.5 percent annually during the so-called economic recovery years of 2002 to 2007, and then declined much more during and after the great recession, falling to $60,100 by 2013. Financially speaking, age 50 turned out to be the new age 65 for these cohorts, even as they were expected to live longer.7 And the situation only got worse with younger generations. For example, among persons born between 1962 and 1966, median household income peaked during 2007, when they were still between the ages of 41 and 45, and has not yet recovered.

Among today’s newest workers, most already have missed out on the rapid increase in earnings that members of previous generations typically enjoyed in their 20s and 30s. This early-career earnings deficit has left them with fewer dollars to save while young, putting them even farther behind their parents in building long-term assets, such as adequate savings for retirement.

Middle-Class Children Who Grow Up To Be Poor

Contributing to this downward mobility trend are Americans who were raised in middle-class homes, but who have fallen down the economic ladder as adults. According to a Pew Charitable Trusts study of children born during the late 1970s, one-third of those raised in middle-class families—defined as families between the 30th and 70th percentiles of the income distribution —have fallen out of the middle class during adulthood. This phenomenon is particularly pronounced among members of minority groups. Among African Americans who were raised in middle-class families, for example, 37 percent were no longer middle class at middle age. For whites, 25 percent were now in the bottom tiers.8

How do these rates compare with the number of Americans who move up the income ladder? Recent research by Raj Chetty and others shows that, during the last two generations, fewer than one in ten children born to parents in the bottom one-fifth of the income distribution managed to rise to the top one-fifth as adults.9 This ratio has not changed appreciably since the 1970s. Yet overall income inequality has increased substantially since then, causing the rungs of the income ladder to stretch farther apart. This in turn makes the consequences of failing to rise up the ladder, or falling down it, harder to bare.  

Family Balance Sheets Deteriorate

Income alone does not define a standard of living. Getting ahead in life also requires accumulating assets, such as home equity and savings, that exceed one’s debts and other liabilities. Without at least some net wealth, it is impossible to finance a first home, pay for a child’s college education, enjoy financial security in old age, or leave behind an inheritance.

Until the present era, despite vast disparities and inequalities across different racial, ethnic, and other demographic groups, most American families enjoyed a rising net worth, both within and across generations. Today’s older Americans still exemplify this pattern. Americans who were 74 years old or older during 2010 had an average net worth that was 149 percent higher than that enjoyed by Americans who were the same age during 1983 (after adjusting for inflation).10 This pattern has since disappeared, however. The precise tipping point came among people born in 1952. They would become perhaps the first generation in American history to have less real net worth on the threshold of retirement than people born ten years earlier had at the same age. From there, the real net worth of subsequent birth cohorts has generally been stagnant or has declined compared to the lifecycle experience of Americans roughly ten to twenty years older.11 For example, after adjusting for inflation, the median net worth of families headed by a person 35-40 years old was 30 percent less in 2010 than it was for their counterparts during 1983.12

Retirement Looms

Because of the vast upward mobility of Americans born before the 1950s, and the downward mobility of Americans born later, the economic security of the next generation of elders will, on current course, be much less than that of today’s retirees—and their children are even less likely to be able to make up any shortfall. One study by the Pew Charitable Trusts found that the typical retiree couple born between 1936 and 1945 had enough net wealth to replace 100 percent of preretirement income when combined with annuitized assets, such as private pensions and Social Security. In contrast, a typical Gen-X couple born between 1966 and 1975 is on course to see income decline by half in retirement.13

To make matters worse, this 50 percent decline assumes that both members of such a couple are able to continue working until the previously normal retirement age, which may well not happen. Labor force participation rates for men younger than 65 have been declining sharply, owing to corporate downsizing, low wages, obsolete job skills, rising rates of chronic illness such as diabetes, long-term unemployment, and other factors.14 Since the 1960s, the share of prime-age men no longer in the workforce has rougly tripled.15 Taken together, these trends paint a picture of steadily declining mobility, and shorter and less secure attachment to the workforce for men.  

Adding to the difficulties facing the future elderly is the disappearance of windfall Social Security benefits. During the late 1970s, Social Security paid out benefits to retirees that exceeded the value of their contributions by between $250,000 and $300,000 in today’s money.16 Subsequent birth cohorts have paid a far higher share of their income into the system, but under current law, most members are promised back little more in benefits that they paid in taxes. Social Security payroll taxes remained below 2.5 percent through the 1950s and below 4 percent until the end of the 1960s. But workers born in the 1960s have paid 6.2 percent of their income into the system throughout most of their working lives, and really double that, because most economists agree that the employer contribution in payroll taxes ultimately is born by employees.17

Having effectively paid about one out of eight dollars they earned into Social Security, the ability of Americans born during and since the 1960s to save for their own retirement has been correspondingly reduced, even as the Social Security system’s rate of return has become progressively less for each new generation. The same diminishing rate of return is found in many private pension plans, as well, even as pension coverage itself also has fallen precipitously among today’s young and middle-aged workers.

The Mounting Cost of Living

The declining cost and increasing quality of digital technologies, as manifested by smartphones and their apps, gives many of today’s Americans access to goods and services that were beyond the reach of even the richest people a generation ago. Yet the cost of the goods and services Americans most need to help themselves and their children rise up the economic ladder has grown much faster than family income or general inflation. This is another large factor behind the stark increase in wealth inequality among the generations.

One major example is the inflation in higher education costs. During the past generation, graduating from college has become a near prerequisite to obtaining middle-class status or avoiding losing it. Yet even as paying for higher education became, for that reason, harder for families and individuals to avoid, the cost of attending a public or private college escalated 40 percentage points more than the Consumer Price Index between 2005 and 2015.18

Compounding the burden, the share of the higher education sector’s revenue paid by families and students rose from one-third during 1980 to one-half during 2012, reflecting not just rising tuition, but a sharp decline in needs-based financial aid during the past generation.19 Closing the gap has dumped a mountain of debt on household balance sheets. The share of young adults with student loans rose from 26 percent during 2001 to 40 percent during 2013.20 Total student debt now surpasses $1.1 trillion, and sadly, much of this debt is held by people who never finished college, and who often have been victimized by predatory lending practices. Among seniors graduating during 2013, the average borrower owned $28,400 in student loans.21

Meanwhile, the dramatic rise of health-care costs relative to family incomes has been, and will be, particularly burdensome for younger generations. As recently as the 1960s, health-care costs were an incidental expense for most young American families. During 1964, health-care spending was just $197 per person per year. This low cost meant that, with a mere seventy-eight hours of labor (or by the end of the second work week in January, for those working full-time), the average nonsupervisory worker earned enough to cover the per capita cost of health care, including that of all children and retirees.

By contrast, during 2012, such a worker had to put in 452 hours to cover the average per capita burden of medical expenses, which by then had risen to more than $8,915. Put another way, by 2012, it was nearly March before the typical American working a forty-hour week earned enough to pay the health-care sector’s growing claim.22 The total annual cost of health care for a typical family of four—even one covered by a typical employer-sponsored plan—reached $23,215 during 2014, or roughly the equivalent cost of buying a brand new Honda Accord LX every year.23 The growing burden of health-care costs is a major reason why employers are so reluctant to hire, and why wages remain stagnant.

Although some of the increase in health-care costs reflects genuine advances in medicine, most simply reflects rising prices for existing medical services combined with an increasing volume of redundant tests, unnecessary surgeries, and other forms of over-treatment that do not improve health.24 Peer countries achieve better population health and life expectancy while expending as little as half as much per person in health-care services. As such, most of the increasing cost of health care does not reflect improvement to the average Amerian’s standard of living.

Predatory Lending and the Housing Bust

Another factor behind the downward mobility of Americans is the growth of payday loans, subprime mortgage lending, and other wealth-destroying consumer finance products. Americans who came of age before the 1970s were largely protected from predatory lending by usury laws, for example, which capped fees and interest costs on loans. But starting during the 1980s, these consumer finance protections largely disappeared. At the same time, financial engineering, including securitizations, led to the growth of financial institutions with business models that allowed them to prosper—at least in the short term—by lending money to people who could not afford to repay.

These trends, combined with generally lagging or falling individual and household incomes and rapidly expanding access to credit, often on predatory terms, led to an explosion of borrowing. When this was followed, in turn, by a collapse in home prices, the result was devastion to the balance sheets of most Americans younger than age 50. By 2010, the average family of people ages 25-49 had a net worth that was 32 percent below that of their counterparts during 1989.25

This sequence of events particularly damaged members of Generation X, many of whom took out mortgages on predatory terms at or near the top of the housing bubble. Largely as a result, from 2007 to 2010, Gen-Xers as a whole lost nearly half (45 percent) of their wealth, or an average of about $33,000 subtracted from already low levels. Many were pushed into negative net worth, as their houses became worth less than their mortgages. By contrast, those born during the Great Depression (between 1926 and 1935) experienced zero loss of net wealth as a group during the great recession (2007-2010).26

Declining Assets and the Sharing Economy

Most millennials, whose oldest members still are in their mid-30s) were too young to be in the market for real estate during the housing bubble and, therefore, did not directly experience the evaporation of real-estate wealth caused by the great recession. While they may have dodged that bullet, however, the longer-term trend of declining asset ownership among today’s younger Americans has potentially very negative implications for their future net wealth. The rate of homeownership among households headed by a person younger than age 35 has fallen from 43 percent during 2005 to 35 percent during 2014. To be sure, not every millennial wants or needs to own a house. But homownership has been the major means by which most ordinary Americans in previous generations built their net wealth and financed their retirements. Moreover, home prices have been recovering since the bottom of the great recession, and in many places have escalated sharply. Thus, the continuing decline in the homeownership rate among young households has probably reduced what millennials’ aggregate net wealth would have otherwise been.27 And if the typical millennial winds up a renter for much if not all of his or her life, this certainly will require that the generation acquire some other major means for building assets during a lifetime.

A sharper decline in stock ownership among young adults does not bode well for that possibility. During 2001, 48 percent of persons ages 18 to 31 owned stock; by 2013, this share had dropped to 37 percent.28 This long-term decline in stock ownership among the young occurred during a period in which stocks, despite volatility, appreciated in value by several-fold. Younger Americans also are increasingly less likely to own their businesses. On a per capita basis, the rate of new business formation declined by 50 percent between 1977 and 2009, a trend that leaves more businesses failing each year than are started.29 As Federal Reserve Chair Janet Yellen has pointed out, the declining share of Americans who are business owners diminishes what historically has been “a vital source of opportunity for many households to improve their economic circumstances and position in the wealth distribution.”30

The trend now seems to be compounding among millennials, who, despite high aspirations to entrepreneurship, are having a difficult time starting successful businesses. A recent report by the Kauffman Foundation concludes that, although Millennials have higher levels of education than previous generations and lifelong exposure to information technology, their shaky finances mean most “can’t afford to become entrepreneurs.”31

Millennials also are less likely than young adults of the past to own other forms of assets, including cars and many durable consumer items. In some instances, this can be positive. If, for example, the growth of services such as Zipcar makes owning a depreciating asset such as an automobile unnecessary, this is at least potentially a gain to one’s net worth. Being able to “monetize” previously underused assets, such as by renting a spare bedroom through Airbnb, also can have the same positive effects on personal balance sheets.

Yet this “sharing” economy depends on and contributes to  the “gig” economy, in which more and more workers no longer are employees, but rather are freelancers responsible for the financial security, health, and retirement once provided by employers. The Uber driver, for example, is responsible for purchasing and maintaining the car he or she uses, just as the contract white-collar worker often must finance and maintain his or her own office space, IT systems, career training, and other hard and soft assets necessary for the job. Although difficult to measure, the increasing uncertainty and contingency of today’s employment has to be counted as a net negative for most workers’ standards of living.


To be sure, some of the factors behind generational downward mobility are difficult to address through public policy. For example, during the last several generations, the number and share of single-parent families has grown rapidly.32 Abundant social-science research documents that this is both a cause and a consequence of diminishing economic opportunity, yet there is no single policy lever that will reverse the trend.

But many of the major causes of downward mobility do rest squarely within the realm of political economy and public control. One example is the woeful inefficiency of the U.S. health-care system. A large body of research now pegs the amount of waste in this burgeoning sector at between 30 and 50 percent of all health-care spending. According to the National Academy of Medicine, eliminating this waste would be enough to provide every young person in America (ages 18-24) with the average annual tuition and fees of a four-year institution of higher learning for two years—to take but one example of its tremendous opportunity cost.33

The higher-education sector is also badly in need of systematic rethinking and overhaul. Individuals need to be cognizant of both the mounting cost of not acquiring an education and of the lifelong damage that can result from excessive student debt. At the same time, government and society at large need to attack inflating college costs, which seem to result primarily from growth in administrative spending and a lack of transparency about educational outcomes.34

Another priority should be redirecting the vast subsidies the federal government has long expended to help households accumulate financial and tangible assets. These subsidies currently total more than $350 billion a year, with the lion’s share going to already wealthy households and individuals. For example, American taxpayers annually spend roughly $70 billion to cover the cost of the home-mortgage deduction. Yet 70 percent of this money goes to households in the top 20 percent of the income distribution, while just 8 percent accrues to middle-income households, and almost nothing to the bottom 40 percent. Similar tax breaks nominally meant to encourage saving for college and retirement have similar “Robin-Hood-reverse” qualities.35 Much more can and should be done to target resources for asset building for those in, and struggling to reach, the middle class.

Let’s not forget another possible policy lever: the money supply. Moderate levels of price and wage inflation have always tended to benefit younger adults disproportionately, because younger households tend to have more debts and fewer assets than older households. Conversely, hard money tends to help older generations, who have fewer debts, less need to worry about unemployment, and more assets to protect from inflation. A big part of the reason that today’s 70-somethings did so comparatively well financially during their lives was that, while they were young, the general wage and price inflation of the 1960s and 1970s eroded the value of their mortgages even as it inflated the value of their homes.  Today’s young people, being particularly encumbered by debt, would benefit from modest levels of general inflation so long as wages kept pace.

More generally, we need policies that will allow today’s workers to retain more of the value of their increased productivity. In many sectors of the economy, workers do, indeed, produce as much in one day as their counterparts in the 1960s did in a forty-hour week—just as Robert Ball predicted. Yet the benefits of this increased efficiency have gone overwhelmingly to already-established owners of assets, rather than to each new generation of workers.

The reasons behind this shift are varied, but hardly inevitable or unalterable. Since the 1980s, for example, the United States has radically reduced enforcement of antitrust and fair trade policies. The resulting trend toward concentration in many industries largely explains both the diminishing opportunities for upward mobility through entrepreneurship, and the reduced competition among employers for wage employees.36 Meanwhile, thanks largely to changes in tax law since the early 1980s, major U.S. corporations have used almost all their profits in recent decades to reward their shareholders with dividends and stock buyback schemes. That leaves little for investment in productive enterprise, or for raising the wages of rank-and-file workers.37

Potential certainly exists for our children to inherit a far more productive and broadly prosperous society than exists today. Yet for this to occur, it is not enough to focus primarily on individuals or even on the problem of “the 1 percent” growing richer. To turn the negative generational trends around requires that we reverse the deep changes in our political economy that have led to mass inequality across generations.

About the Author

Phillip Longman is the author of numerous books and articles about public policy in realms ranging from demographics to health care and competition policy. He currently is a policy director with New America’s Open Markets program, a senior editor at Washington Monthly, and a lecturer at Johns Hopkins University.


  1. Ball, Robert M. 1965. "Is Poverty Necessary?" Social Security Bulletin, p. 18. 
  2. Greenstone, Michael, and Adam Looney. 2011. “Trends: Men in Trouble,” Milken Institute Review, pp. 8-16. 
  3. Madrick, Jeff, and Nikolaos Papanikolaou. 2008. “The Stagnation of Male Wages,” Policy Note (New York: Schwarz Center for Economic Policy Analysis, The New School), p. 3, Tables 1.1, 1.2.
  4. Kearney, Melissa S., Brad Hershbein, and Elisa Jácome. 2015. “Profiles of Change: Employment, Earnings, and Occupations from 1990-2013” (Washington, DC: The Hamilton Project, Brookings Institution), Figure 1.
  5. Federal Reserve Board. 2014. “2013 Survey of Consumer Finances” (Washington, DC: Board of Governors of the Federal Reserve System), Table 1 89-98 01-13.
  6. Shapiro, Robert J. 2015. “Income Growth and Decline under Recent U.S. Presidents and the New Challenge to Restore Broad Economic Prosperity” (Washington, DC: Center for Effective Public Management, Brookings Institution), Table 1, p. 2. See also worksheet.
  7. Ibid. See also Table 2.
  8. Acs, Gregory. 2011. “Downward Mobility from the Middle Class: Waking Up from the American Dream” (Washington, DC: Pew Charitable Trusts, Economic Mobility Project), Figure 6, p. 14.
  9. Chetty, Raj, et al. 2014. “Is the United States Still a Land of Opportunity? Recent Trends in Intergenerational Mobility.” Working paper 19844 (Cambridge, MA: National Bureau of Economic Research).
  10. Steuerle, C. Eugene, et al. 2013. “Lost Generations? Wealth Building among Young Americans” (Washington, DC: Urban Institute), Figure 3.
  11. Survey of Consumer Finances 2012. 2014., cited by Neil Howe, “Are you Born to be Better Off Than Your Parents?”, Figure 4, July.
  12. Pew Research. 2013. “Retirement Security Across Generations,” (Washington, DC: Pew), Figure 11, May. 
  13. The share of men of prime working age—those 25 to 54 years old—who are in the workforce declined by 5.2 percent between 1992 and 2012. Bureau of Labor Statistics. 2013. "Labor Force Projections to 2022: The Labor Force Participation Rate Continues To Fall," Monthly Labor Review, December, table: “Civilian Labor Force Participation Rate by Age, Sex, Race, and Ethnicity.”
  14. Appelbaum, Binyamin. 2014. “The Vanishing Male Worker: How America Fell Behind,” New York Times, December.
  15. Schieber, Sylvester J. 2012. The Predictable Surprise (New York: Oxford University Press).
  16. Social Security Website.
  17. Bureau of Labor Statistics data. 2014. cited by Annie Lowrey, “Changed Life of the Poor: Better Off, but Far Behind,” New York Times, April.
  18. The Pell Institute and PennAHEAD. “Indicators of Higher Education Equity in the United States, 45 Year Trend Report,” (Washington, DC: Pell Institute, 2015), p. 28. 
  19. Dettling, Lisa, and Joanne Hsu. 2014. “The State of Young Adult’s Balance Sheets: Evidence from the Survey of Consumer Finances,” (St. Louis: Federal Reserve Bank), p. 13, May.
  20. Student Debt and the Class of 2013, Project on Student Debt, November 2014, 
  21. For health-care expenditures, see Centers for Medicare and Medicaid Services, “National Health Expenditures; Aggregate and per Capita Amounts, Annual Percent Change and Percent Distribution, by Type of Expenditure: Selected Calendar Years 1960-2012” (Washington, DC: CMMS, n.d.), Table 1. For hourly earnings, see Bureau of Labor Statistics Employment, “Hours and Earnings from the Current Employment Statistics Survey” (national) (Washington, DC: BLS, 2015.
  22. Girod, Christopher, et al. 2014. “2014 Milliman Medical Index” (Seattle: Milliman, inc.).
  23. See, for example, Gawande, Atul. 2015. “Overkill”, The New Yorker, May. Brownlee, Shannon. 2010. Overtreated: Why Too Much Medicine is Making Us Sicker and Poorer (New York, Bloomsbury).
  24. Pew Research. 2013. “Retirement Security across Generations” (Washington, DC: Pew), Table 1.
  25. Current Population Survey/Housing Vacancy Survey, Series H-111, U.S. Census Bureau, Homeownership Rates by Age of Householder: 1994 to Present, Table 19.
  26. Dettling, Lisa J., and Joanne W. Hsu. 2014. “The State of Young Adults’ Balance Sheets: Evidence from the Survey of Consumer Finances,” Federal Reserve Bank of St. Louis REVIEW, p. 316 EN1426368479.
  27. Lynn, Barry C., and Lina Khan. 2012. “The Slow-Motion Collapse of American Entrepreneurship” Washington Monthly, July/August.; Edsal, Thomas. 2015. “Has American Business Lost its Mojo?” New York Times, April.
  28. Yellen, Janet. 2014. “Perspectives on Inequality and Opportunity from the Survey of Consumer Finances” a speech to the Federal Reserve Board, October.
  29. Kauffman Foundation. 2015. “The Future of Entrepreneurship: Millennials and Boomers Chart the Course for 2020” (Kansas City: Kauffman Foundation).
  30. The National Marriage Project. 2012. State of Our Unions, 2012 (Charlottesville, VA: University of Virginia). As recently as the 1980s, only 13 percent of children born to mothers with only a high school degree were born outside of marriage. By the late 2000s, that figure had risen to 44 percent.
  31. Smith, Mark, et al., editors. 2012. Best Care at Lower Cost: The Path to Continuously Learning Health Care in America (Washington, DC: Institute of Medicine, National Academies Press), pp. 3-11.
  32. “The Real Reason College Tuition Costs So Much,” New York Times, April 4 2015.
  33. Brookings, Tax Subsidies for Asset Development: An Overview and Distributional Analysis, EN1428352351.
  34. Lynn, Barry C., and Phillip Longman. 2010. “Who Broke America’s Jobs Machine?” Washington Monthly, March/April.; Lynn, Barry C. 2010. Cornered: The New Monopoly Capitalism and the Economics of Destruction (New York: Wiley).
  35. Lazonick, William. 2014. “Profits without Prosperity,” Harvard Business Review, September; Carpenter, Dan. 2015. “What Piketty Missed: The Banks,” Washington Monthly, March/April.

Firms as Drivers of Growth and (In-)Equality

By Christian Moser Ph.D. candidate, Princeton University
Christian Moser
Christian Moser


1. A trade-off between growth and equality?

The consequences for the U.S. economy of the 2007-2008 financial crisis have proven to be prominent and persistent. Employment dropped precipitously to a thirty-year low, while aggregate output took a hit in excess of $600 billion. The two panels of Figure 1 illustrate that, at present, these measures show little sign of recovery, with both employment and GDP lagging behind pre-crisis trends. Yet, these aggregate trends mask significant heterogeneity.

Figure 1. Left: Civilian Employment-Population Ratio; right: Real GDP vs. Pre-Recession Trend

Source: Author’s own calculations using U.S. Bureau of Labor Statistics (2015) and U.S. Bureau of Economic Analysis (2015).
Shaded areas mark U.S. recessions.

Take as an example the nation’s median household income, which fell by 7.4 percent between the onset of the Great Recession in 2007 and 2012. Real earnings of households at the twentieth percentile of the income distribution in 2012 are lower than they were in 1996. Meanwhile, earnings of families in the highest quintile have experienced robust growth over the past decades, with little turbulence during the most recent recession. The left panel of Figure 2 illustrates the evolution of family income percentiles for three groups: twentieth percentile, median, and eightieth percentile (with 1996 income levels normalized to 1.0 for ease of comparison). Real incomes of these groups grew in roughly equal proportions during the late 1990s but started to diverge with the onset of the 2001 recession, with lower-income groups lagging behind throughout later years. The right panel summarizes the rise in income inequality by a common measure, the Gini coefficient over family income. By this measure, the United States ranks among the most unequal of all developed countries in the world.1

Figure 2. Left: Normalized Percentiles of Real Family Income Distribution; right: Income Gini Ratio of Families

Source: Author’s own calculations using U.S. Census Bureau (2015), U.S. Bureau of the Census (2015), and Federal Reserve Bank of St. Louis (2015)
Shaded areas mark U.S. recessions.

The trend of increasing income inequality presents a predicament for policymakers: whether to pursue policies that are pro-growth or pro-equality? The implicit assumption is that there exists a trade-off between growth and equality such that improvements in one will necessarily exacerbate the other. While the U.S. experience over recent decades supports this claim, effective policy design requires an understanding of the microeconomic drivers behind both inequality and growth.

Recent research has underscored the relevance of firms for determining both aggregate productivity and pay. Work in this area has found that firms played a notable role in driving inequality in the United States over the past decades (Barth et al. 2014; Song et al. 2015). These studies show that conditional on worker characteristics such as educational attainment and work experience, the place of employment is a significant predictor of workers’ pay. Studies of other countries (see Card et al. 2013 for Germany and Mueller et al. 2015 for the UK) confirm that increasing disparities between firms are associated with rising income inequality. Yet, three important questions remain unanswered: first, to what extent are income differences causally attributable to differences across firms? Second, what attributes lead firms to pay more or less generously than the average? Third, what feasible policies could lead to a reversal of the inequality trajectory followed by the United States while also promoting productivity growth?

The goal of this short paper is to address the third question and propose an answer by drawing on the experience of a large upper-middle-income country—namely, Brazil. By comparing and contrasting the link between firms and inequality in Brazil with what we know about the United States, some general lessons on the micro-foundations of inequality and growth emerge. While a holistic treatment of the first two questions is beyond the scope of this paper, references to recent research findings shed some light on these related issues.

The rest of the paper is structured as follows: Section 2 summarizes stylized facts about income inequality in the United States and then discusses sources behind its rapid increase. Section 3 draws on the Brazilian case to explore parallels and differences with the U.S. experience. Finally, Section 4 concludes by suggesting three ways in which policies targeted at firms can aim to reduce income inequality, while also enhancing aggregate productivity.

2. Firms as drivers of rising income inequality in the United States

Using administrative tax records, Kopczuk et al. (2010) document that income inequality in the United States has risen precipitously since the early 1970s.2 Dissecting the rise in inequality, the authors show that a large part is driven by rising inequality in the top half of the distribution. Research on the sources behind the rapid increase in inequality has explored many possible explanations, including skill-biased technological change (Autor et al. 2008) and the related rise in the college premium (Pew Research Center 2014), the decline of labor unions (Card 2001), and a continued decrease in the real minimum wage (Lee 1999). Yet, these works have limited ability to speak to the relative importance of individual- vs. employer- vs. institution-specific wage components. While imperative for diagnosing the root causes of rising inequality, the distinction between these three levels of wage determination remains an under-researched topic.

Addressing this gap, recent work has stressed the importance of firm-level wage determinants and their divergence over the past decades. Linking a panel of workers to their employers using the U.S. Longitudinal Employer-Household Dynamics (LEHD) data, Barth et al. (2014) find that the increase in income inequality between 1970 and 2010 is largely attributable to changes in the supply side of the labor market.3 To demonstrate this, the authors proceed in two steps: first, they show that, in a Mincer regression framework, almost two-thirds of the overall increase was due to rising between-employer disparities in pay, and most of the between-employer changes are due to changes in estimated establishment fixed effects after controlling for worker characteristics. Second, they confirm that average real wage growth by income percentile between 1992 and 2007 closely tracks the shape of the growth in estimated effects attributed to establishments employing workers in the respective percentiles. The latter finding suggests that, absent selection issues, changes in establishment-specific pay components explain almost all of the increase in overall income inequality over this period.

In related research, Song et al. (2015) construct a matched employer-employee dataset using administrative data from the U.S. Social Security Administration to quantify how much of the overall income inequality increase is due to between-employer changes.4 In accordance with the results from the LEHD data, they find that almost all of the increase in wage dispersion between 1978 and 2012 is accounted for by the evolution of average wages across firms. Interestingly, the individual-specific income growth rates are matched closely by firm-level pay growth for workers up to and including the top income percentile, whereas within-firm inequality appears to account for a small fraction of the overall income inequality increase for this group. This finding suggests that firms have significant explanatory power for both overall inequality trends and also for inequality among the group of top earners. Similar findings obtain across regions, industries, sex, and age, thus corroborating the proposition that firm-level changes explain the observed rise in income inequality.

3. Firms as drivers of falling income inequality in Brazil

What general lessons does the U.S. experience teach us about the relation between firms and income inequality? One may hypothesize that something inherent about the evolution of firms over the past decades has led to greater income dispersion among workers at different types of employers. Instances of such firm characteristics may include their choice of production technologies, organizational structure, or remuneration policies. Yet, it remains to be examined whether firm-level changes are causally related to changes in inequality. Taking a step in that direction, it would be informative to identify an example in which convergence of employer average pay resulted in decreased overall inequality.

The history of Brazil presents such a case that can be used to shed light on the inherent link between firms and inequality. Over the past two decades, Brazil has experienced rapid growth and falling income inequality, spurred by a sequence of political and economic reforms such as opening to international trade and gradually increasing the minimum wage. The country’s concurrent growth in aggregate output with declining inequality is summarized in Figure 3. The left panel plots real GDP in trillions of real Brazilian Reais along with a linear trend from 1996–2008 for comparison to the U.S. growth experience. The right panel shows the evolution of various percentiles of the real income distribution for working-age males employed in Brazil’s formal sector.5

Two striking facts emerge: First, Brazil grew faster on average than the United States did during those years.6 Second, labor income growth has been particularly pronounced at the bottom of the distribution, with the twentieth percentile of the income distribution growing by 140 percent between 1996–2012, while growth at the eightieth percentile was significant, but distinctly lower at approximately 75 percent. These differential growth rates imply a rapid decline in dispersion of the income distribution throughout this period. These trends are noteworthy also because the Brazilian experience of growth and falling inequality over the past two decades is representative of the Latin American region more broadly.

Figure 3. Left: Real GDP vs. Pre-Recession Trend; right: Normalized Percentiles of Individual Real Labor Income Distribution

Source: Author’s own calculations using OECD (2015) and RAIS data
Shaded areas mark Brazilian recessions.

Motivated by the Brazilian experience, which stands in contrast to that of the United States, Alvarez et al. (2015) investigate the sources of the decline in inequality. Using employment records from the Relação Anual de Informações Sociais (RAIS), an administrative matched employer-employee dataset covering the universe of formal workers in Brazil, the authors conclude that firms were an important determinant of the country’s inequality evolution from the late 1980s onward, particularly during the period of declining inequality from 1996–2012. To the extent that firms can be attributed the inequality decline in Brazil, this further strengthens the view that firms can drive inequality movements in either direction.

Figure 4 summarizes the correlation between worker income growth and firm average income growth across the income spectrum, following the method used by Barth et al. (2014) and Song et al. (2015). The graph is based on data on working-age male employees in Brazil’s formal sector7 and is constructed as follows:

First, to focus on relative movements in the income distribution, we normalize the earnings distribution in both 1996 and 2012 by the median earnings in that year. We then rank all formal sector workers by their normalized labor income in 1996 and compute average income values for each percentile group. We repeat this procedure for 2012 and compute the growth rate of average incomes for each percentile group. The resulting relative growth rates by income percentile are plotted as a solid blue line with circles. It is worth noting that this line is downward sloping almost throughout the entire income distribution. This downward slope indicates that lower-income groups grew faster than higher-income groups, which implies that there was compression throughout the income distribution.

Second, we go back to the percentile group classification of workers in 1996 and compute for every individual the average income their current employer is paying all of its employees in that year. We then take the average by income percentile over these employer average incomes. Repeating the exercise for 2012, we can again compute growth rates of firm average incomes by percentile groups of the income distribution and plot the results as a dashed red line with squares.

Figure 4. Individual and Firm Average Log Normalized Real Income Ratio by Income Percentile, 1996–2012.

The most striking insight from Figure 4 is that worker income growth rates and their employer average income growth rates align closely throughout the income distribution. For example, average income of the twentieth percentile of the income distribution grew approximately 30 percent faster than the median growth rate did, while average firm income of workers in that percentile grew by the almost exactly the same amount.8 This is consistent with the view that changes between firms can account for a large share of the overall inequality decline in Brazil during this period.

To put these results into perspective, it is helpful to compare the U.S. evolution of growth and inequality with Brazil. Contrary to widely held views, the Brazilian experience suggests that there need not be a trade-off between growth and inequality. Further research on the microeconomic mechanisms behind observed changes in the United States and in Brazil is warranted in order to understand the determinants of firm-driven growth and inequality in both countries.

4. Implications for policies promoting growth and equality

Recent research findings emphasize the role of firms as determinants of aggregate productivity in the economy (Buera et al. 2015). The fact that changes in firm-level pay are an important driver of changes in inequality—growing in the United States and declining in Brazil—suggests that firms also have important distributional functions. An immediate question is: how can the link between firms, productivity, and the distribution of economic rents guide the design of policies to promote both growth and equality?

To deduce policy recommendations from the previous section’s findings, it is crucial to determine to what extent income differences are causally attributable to differences across firms. To this end, it is worth noting that the explanatory power of firms remains high when controlling for differences in the composition of workers across firms along observable characteristics, including education, work experience, race, and gender, as shown to be the case in the United States by Barth et al. (2014). Further controlling for workers, sorting across firms along unobservable but time-invariant attributes, as Card et al. (2013) do for Germany and Alvarez et al. (2015) do for Brazil, suggests that firms are important drivers of inequality dynamics.

What firm attributes can policy feasibly influence to spur both growth and equality? To answer this question, our previous analysis suggests that firms should be viewed as dual drivers of both aggregate output and the distribution of economic rents. Three aspects of firm-level wage setting are central to both these ends:

First, firm characteristics such as productivity (Faggio et al. 2010) and export status (Schank et al. 2007) have been shown to be closely related to both total factor productivity and also to firm pay differences. Policies that help firms to adopt more productive technologies and access larger product markets are promising candidates to increase firm profits and, at the same time, boost workers’ pay.

Second, conditional on firm characteristics, the allocation of workers across heterogeneous employers has important aggregate as well as distributional consequences (Davis and Haltiwanger 2014; Blanchard and Katz 1992). Policies aimed to reduce obstacles to worker mobility, which include stringent job protection laws and red tape costs from firm size-dependent policies, can help successful businesses grow, while at the same time offering better job opportunities to workers.

Third, wage-setting policies determine how economic rents are split between a firm’s stakeholders and its workers. In the presence of labor-market frictions, which prevent the efficient allocation of workers across employers, firms can appropriate a larger share of the benefits from the employment relationship. In such an environment, institutional wage policies, such as a minimum wage, can act as a transfer of rents from firms to workers, while also increasing overall productive efficiency (Card and Krueger 1994).

Contrary to conventional wisdom, this short article argued that firms have the potential to spur both growth and equality. Carefully designed and implemented policies of the three classes discussed above have the potential to enhance aggregate productivity, while also inducing a more equal distribution of economic rents.

About the Author
Christian Moser is a Ph.D. Candidate in Economics at Princeton University. His research interests include Macroeconomics, Public Finance, and Development Economics, and his work in these fields focuses on the causes and consequences of economic inequality. Prior to Princeton, he completed his B.Sc. in Economics and Mathematics with First Class Honours from the University of St. Andrews in 2010.


  1. See for example the OECD Income Distribution Database (IDD).
  2. Comparable results are obtained when using alternative inequality measures, including the Gini coefficient, income percentile ratios, and the variance of log incomes. The income concept used in their analysis is labor income, and the income unit is the individual.
  3. Their analysis focuses on establishments, not firms, as the unit of analysis.
  4. In contrast to Barth et al. (2014), these authors conduct their analysis at the firm level.
  5. Details of the microdata, sample selection, and robustness to alternative income definitions are described in Alvarez et al. (2015).
  6. From the right panel of Figure 1, it can be inferred that the United States grew by approximately (15.2/10.5)1/16 − 1 = 2.3 percent per year between 1996 and 2012. In comparison, Figure 3 shows that Brazil grew by approximately (2.2/1.3)1/16 − 1 = 3.3 percent over the same period.
  7. Qualitatively similar trends are observed for different age classes, as well as for female workers. The data do not comprise workers in Brazil’s informal economy, but similar inequality trends are observed in Brazilian household survey data, such as the Pesquisa Mensal de Emprego or the Pesquisa Nacional por Amostra de Domicílios, which are representative of Brazil’s entire economy. Unfortunately, these survey data contain no detailed information on workers’ place of employment.
  8. Taking the exponential of the log normalized real income ratio, we get exp(0.26) = 0.30. 


Alvarez, Jorge, Niklas Engbom, and Christian Moser. 2015. “Firms and the Decline of Earnings Inequality in Brazil.”

Author, David H., Lawrence F. Katz, and Melissa S. Kearney. “Trends in U.S. Wage Inequality:

Revising the Revisionists.” Review of Economics and Statistics, May 2008, 90 (2), pp. 300–323.

Barth, Erling, James C. Davis, and Alex Bryson. 2014. “It’s Where You Work: Increases in Earnings

Dispersion across Establishments and Individuals in the U.S.” NBER Working Paper 20447.

Blanchard, Olivier Jean, and Lawrence F. Katz. 1992. “Regional Evolutions.” Brookings Papers on Economic Activity, 23 (1), pp. 1–76.

Buera, Francisco J., Joseph P. Kaboski, and Yongseok Shin. 2015. “Entrepreneurship and Financial
Frictions: A Macro-Development Perspective.” NBER Working Papers.

Card, David. 2001. “The Effect of Unions on Wage Inequality in the U.S. Labor Market.” ILR Review, 54 (2), pp. 296–315.

Card, David, and Alan B. Krueger. “Minimum Wages and Employment: A Case Study of the Fast-Food Industry in New Jersey and Pennsylvania.” American Economic Review, 1994, 84 (4), pp. 772–793.

Card, David, Joerg Heining, and Patrick Kline. “Workplace heterogeneity and the rise of West German wage inequality.” Quarterly Journal of Economics, August 2013, 128, pp. 967–1015.

Davis, Steven J., and John Haltiwanger. 2014. “Labor Market Fluidity and Economic Performance.”

Faggio, Giulia, Kjell G. Salvanes, and John Van Reenen. 2010. “The evolution of inequality in productivity and wages: panel data evidence.” Industrial and Corporate Change, 19 (6), pp. 1919–1951.

Federal Reserve Bank of St. Louis. 2015. “US. Bureau of the Census, Income Gini Ratio of Families by Race of Householder, All Races GINIALLRF], retrieved from FRED.”

Kopczuk, Wojciech, Emmanuel Saez, and Jae Song. “Earnings Inequality and Mobility in the United States: Evidence from Social Security Data since 1937.” Quarterly Journal of Economics, 2010, 125 (1), pp. 91–128.

Lee, David S. “Wage Inequality in the United States During the 1980s: Rising Dispersion or Falling Minimum Wage?” Quarterly Journal of Economics, August 1999, 114 (3), pp. 977–1023.

Mueller, Holger M., Paige P. Ouimet, and Elena Simintzi. 2015. “Wage Inequality and Firm Growth.” Working Paper.

OECD. 2015. “Gross Domestic Product by Expenditure in Constant Prices: Total Gross Domestic Product for Brazil [NAEXKP01BRQ652S], retrieved from FRED, Federal Reserve Bank of St. Louis.”

Pew Research Center. 2014. “The Rising Cost of Not Going to College.” Technical Report, Pew Research Center.

Schank, Thorsten, Claus Schnabel, and Joachim Wagner. “Do exporters really pay higher wages? First evidence from German linked employer-employee data.” Journal of International Economics, May 2007, 72 (1), pp. 52–74.

Song, Jae, David J. Price, Fatih Guvenen, and Nicholas Bloom. 2015. “Firming Up Inequality.”

U.S. Bureau of Economic Analysis. 2015. “Real Gross Domestic Product [GDPC1], retrieved from FRED, Federal Reserve Bank of St. Louis.”

U.S. Bureau of Labor Statistics. 2015. “Civilian Employment-Population Ratio [EMRATIO], retrieved from FRED, Federal Reserve Bank of St. Louis.”

U.S. Bureau of the Census. 2015. “Real Median Family Income in the United States [MEFAINUSA672N], retrieved from FRED, Federal Reserve Bank of St. Louis.”

U.S. Census Bureau. 2015. “Current Population Survey, Annual Social and Economic Supplements.”

Session Summary:


New Entrepreneurial Growth Conference: Inequality

Income inequality has become a topic of intense public debate in the United States over the past few years. This partly has been stoked by Thomas Piketty’s Capital in the Twenty-First Century, as well as data showing a widening income gap in America. There can be different interpretations of such inequality, pointing toward different paths of action. Accordingly, at Kauffman’s New Entrepreneurial Growth conference, we asked participants to debate the proposition: “Something should be done about inequality.” This type of discussion takes on particular complexity in the context of entrepreneurship because, almost by definition, entrepreneurship can generate skewed, or unequal, outcomes. Yet, successful entrepreneurs—even if their personal enrichment outpaces others by orders of magnitude—can lessen inequality by challenging incumbents, sharing rewards, and broadly enhancing productivity. At the same time, it remains an open question as to whether growing income inequality narrows the pipeline of potential entrepreneurs.

Contributors observed that we have seen dramatic growth in wage inequality in the last three decades, and a whole generation has seen no real wage gains. Furthermore, they said, inequality is continuing to rise. Between 1999 and 2000, the United States saw enormous growth in productivity and jobs, but some people were left behind. While growth has now slowed, we’re still leaving many people behind. The employment-to-population ratio has fallen and has been slow to recover since the recession. Participants noted that young, less-educated black males are doing especially poorly, and the digital age isn’t going to help them. While some inequality can be an incentive to work hard and pursue success, it can also be a problem for democracy when a generation hasn’t shared in the gains.

Other contributors suggested that there isn’t a single relationship between inequality and growth, that inequality is not always associated with growth, that inequality may not hurt growth, and that there are other causes of inequality to be considered. Participants were interested in greater research concerning why inequality affects growth and a closer look at data on the relationship between inequality and growth.

Causes of inequality. Participants noted that high unemployment contributes to low wage growth, but there are many other factors at play, as well. Rent-seeking behavior, for example, causes inequality, wastes physical and human capital, adds no social value, and allows people to become rich without promoting growth. A contributor mentioned the example of expenditures to shave time off high-speed trades in the financial industry, which simply shifted the profit rather than increasing the size of the pie.

Another participant suggested that the benefits of many recent innovations have been captured entirely by a specific sector and aren’t creating any welfare gains. Citing the example of the cost of financial transactions or services, he explained that the costs of the inputs into those transactions (computers, etc.) have decreased, but prices haven’t gone down over the last thirty years. In contrast, Google and its innovations enhance social welfare.

Changes in family structure and welfare also were discussed as causes and consequences of inequality and declining social mobility and as an important factor in discussions of human capital. Massive incarceration, the increase in families with single parents, more children born out of wedlock, and the lack of economic security in households headed by women play a significant part in declining social mobility, and barriers to mobility start with young children before elementary school.

Participants contended that our inefficient U.S. social welfare system fails to support these families, contributing to greater inequality and constraining new business formation. Policy, participants said, could be shaped to create a more robust safety net and invest in and support children and lower- and middle-class families, including policies related to family leave, paid maternity leave, and childcare. Participants explained that the payroll tax currently squeezes the safety net, and there is no sign of it abating. In many firms, we have seen enormous growth in human resources to administer retirement accounts and health care benefits because the country doesn’t have a sufficient safety net. Firms, in effect, have to run the welfare state, in addition to their primary roles as innovators. Scandinavian countries, by contrast, have deregulated social welfare over the last twenty years, creating an efficient safety net and making it easier to start businesses and hire and fire employees.

A participant pointed out that health care and education are the two areas of our economy that have most impinged on equality, noting that these two areas of family budgets have grown enormously in price, have stripped people of wage gains, are heavily government funded, and are terribly regulated. In both fields, we have significant regulation, but it’s not the regulation and performance monitoring we need to get better-quality services at lower costs. Participants emphasized that we need the right regulation, rather than more or less regulation.

Social mobility. One participant suggested that inequality is actually two distinct problems that aren’t truly linked: a problem of inflated income at the top and a separate decline in social mobility in the bottom. Each of these problems, he said, needs to be addressed independently. Social mobility, another contributor suggested, is actually more important than inequality. We need, he said, to distinguish between inequality of outcomes and inequality of opportunity.

Consumption inequality. Another participant suggested that a focus on consumption inequality and standards of living rather than simply income inequality yields a more complete and positive picture. Consumption inequality, he said, hasn’t increased at the same pace as income inequality. Consumers in the United States can now purchase a bigger bundle with the same wage, and even average households are now accessing much more technology that improves their lives.

Along these lines, there was some disagreement regarding the impact of the liberalization of economies and gains from trade on lower-income families. While Walmart offers low-cost products that benefit lower-income people, it also has promoted the movement of industrial manufacturing and lower-skilled jobs overseas as it seeks out low-cost suppliers in China and elsewhere. One participant suggested that lower-income people in the United States are hurt by low-skilled jobs going abroad, but they benefit more as Walmart consumers of the low-cost products that are created in China. Another participant disagreed, responding that low-wage workers lose much more than they gain through Walmart’s cheaper goods. Cheap goods would benefit them if wages were constant, but wages have declined, leaving families less able to afford the goods they need. Participants suggested that more data on this issue would be useful, but emphasized that we cannot look at U.S. inequality in isolation from the rest of the world. Global inequality has actually fallen over the past few decades.

A participant pointed out that just as the developing world has many people who can do low-skilled jobs at lower wages, it also has highly skilled people who could replace doctors and lawyers in the United States at lower wages. We don’t outsource those jobs, however, because we have policies to protect them. He would like to see highly paid professionals in direct competition with their counterparts in the developing world.

Another contributor responded that we already allow some of that competition through immigration, as many highly skilled immigrants come to the United States to become doctors. Many of these immigrants, however, return home to practice, meaning that we are effectively educating our competition. Another participant suggested that we could educate people to be doctors and lawyers at lower costs or we could design our trade agreements so that these skilled immigrants could practice in the United States and compete.

Political inequality. Finally, the connection between economic inequality and political inequality was also raised. Money is a political resource, and those without wealth are unable to influence the government and have a strong voice in politics. If we could make government less dependent on lobbyists for information, participants said, we could reduce the inequality of information flowing into government.

Solutions. Participants proposed that there are policy changes that could decrease inequality and boost dynamism in the economy. One participant emphasized that we need both predistributive strategies that increase economic growth, as well as distributive strategies that allow for greater equality. Redistribution alone, she said, creates disincentives for growth. Another participant emphasized that solutions must bring people on the bottom up, rather than bringing those on the top down.

Political obstacles to potential solutions also were discussed. One participant emphasized that policy entrepreneurs who can generate political support for new policy ideas will be at a premium in coming years. Another contributor argued that liberal anti-inequality measures, such as more unionism in private-sector firms and better family leave policies, will not pass given the current government polarization and gridlock. However, the libertarian inequality agenda, which includes addressing the massive levels of incarceration and policies that allow for rent-seeking and junk entrepreneurship, may be achievable. Contributors also suggested the following policy changes and goals to reduce inequality and increase dynamism:

  • Fund training programs more strategically, with less duplication and better regional allocation. Technology can be used to make them more demand-driven.
  • Reduce incarceration rates.
  • Limit the high salaries received by nonprofit executives, especially those in higher education.
  • Tax the broken financial sector.
  • Remove incentives for tax shelters and the tax avoidance industry.
  • Change accreditation and professional certification processes to reduce barriers to entry and liberate innovators. Low-income young people could get professional certifications through their high schools that would allow them to work. Furthermore, relaxing licensing for physicians would allow other health care professionals to do some of physicians’ tasks.
  • Consider a negative income tax that would guarantee a decent income to those at the bottom of the income scale.

Other policy suggestions focused more specifically on increasing dynamism. These included:

  • Reduce the amount of regulation on businesses, especially those that adversely affect young businesses or that exclude entrance to the market.
  • Increase regulatory certainty.
  • Reform the patent system to unlock ideas and reduce artificial rents generated by too much protection.

Projections for the future. Panelists were asked to predict what would happen in twenty years if the status quo doesn’t change. One panelist saw reasons for optimism, suggesting that more balanced trade will lead to more U.S. jobs and that health care costs will necessarily come into line. He indicated that it isn’t possible to sustain the factors that have led to the growth in inequality, and it will, therefore, eventually decrease.

Another participant also suggested that there are promising changes occurring, even if the prospect for national change is unlikely. Online education, for example, will increase and may promote reform in the system. And, we are seeing experimentation at the state level with more support for families, including sick days and family leave policies that may spread to more states. One contributor noted that the tax reform in 1986 that cleaned out an inefficient tax code and created a better foundation for entrepreneurial growth was achieved in a similarly difficult political environment, indicating that there may be a chance for similar changes in the near future.

Other panelists, however, were decidedly pessimistic, suggesting that machines will continue to do more, and our inefficient social welfare system and the lack of support we provide for young children and families will mean that they will not see gains in the next twenty years. We will see a class of people at the upper end of the income spectrum who live as if they are in a golden age, especially the highly educated, dual-income families with broad social networks. There will be no gains for the middle class, they said, and people at the bottom will continue to struggle and face even more friction while trying to climb the ladder. Reminding the audience of the discussion of policy and agency from the debate in the second session, another panelist explained that these are not automatic structural mechanisms; rather, they are a function of agency and the decisions we make. While innovation can take the form of real solutions to human problems, he anticipates that we primarily will see growth in innovation that either solves problems that are created by the state itself or innovation that exploits government vulnerability or arbitrage. These types of innovation draw down the legitimization of capitalism and perpetuate rent-driven inequality.