100 Years of Rising Corporate Concentration

The dominance of large firms in the US economy has drawn growing attention in recent years. From Walmart to Apple, many salient examples point to the prominence of large companies in day to day life. Standard datasets corroborate this impression. For instance, comprehensive economic census data available since the 1980s show that the largest firms account for an increasing share of output in many industries.

A common presumption is that rising concentration of economic activities in the largest American companies is a new and unusual phenomenon. Its occurrence is often attributed to recent events, from information technology and globalization to population aging and shifts in antitrust enforcement. The focus on the recent decades prevails as standard datasets provide much more information about this period.

Is this time really different? Is rising concentration an abnormal phenomenon, driven by special features of today’s world? In this paper, we collect long-run data on the concentration of production in the U.S. from 1918 to 2018, and document the extent to which a small set of top businesses account for a large share of production assets or output. We do so by digitizing the size distribution of U.S. business from the Statistics of Income report, published annually by the Internal Revenue Services (IRS). The historical IRS publications provide comprehensive information about the population of corporations. For some years we also have data on the size distribution of noncorporations (partnerships and proprietorships), and our findings are similar when noncorporations are included.

In the long-run data, we observe that concentration (measured using the share of assets, sales, or net income of top businesses) has increased persistently over the past century.  For example, since the early 1930s, the asset share of the top 1 percent corporations has increased from 70 percent to 97 percent, and the share of the top 0.1 percent has increased from 47 percent to 88 percent. As the largest firms expand, the share of the medium sized firms shrink, while the smallest firms (e.g., the bottom half) always account for a minimal fraction of aggregate economic activities.

We also study top business shares at the broad industry group level. Again we observe a secular increase in most of the main industry groups. Interestingly, the timing differs across industries. In manufacturing and mining, rising concentration took hold before the 1970s; in services, retail, and wholesale, rising concentration occurred primarily after the 1970s. We perform a number of checks to make sure the data are reliable and the trends are consistent. Importantly, our data cover the population of U.S. businesses, not just publicly listed firms, which miss large private firms and have very limited coverage at the industry level.

Why does concentration increase over the long run? Establishing causal mechanisms is inevitably challenging. Nonetheless, our evidence suggests that recent events are not the full story, and persistent economic forces appear important. It might be possible that technological development has ameliorated the knowledge problem of society, so that production through large organizations has become increasingly feasible. It might also be possible that winner-take-all effects are becoming stronger.

Indeed, the inevitable dominance of large firms was a popular topic of discussion a century ago, including in the legal community. The Modern Corporation and Private Property by Adolf Berle and Gardiner Means is one example. Curiously, although George Stigler remarked in 1980 that “There are not many books fifty years old whose central argument is identified for modern economists simply by naming the work, but Berle and Means’s The Modern Corporation and Private Property is surely one,” this book is not well remembered among economists by now. And neither is their analysis of the dominance of large companies. In Chapter 3 titled “Concentration of Economic Power,” the book tabulated the largest 200 non-banking corporations by assets in 1930, and estimated that they accounted for 49.2% of total non-banking corporate assets. The prominence of large companies sets the stage for their subsequent argument that the separation of ownership and control among these firms have great social significance.

Other contemporaries of Berle and Means saw even greater social significance in the dominance of large firms. As one extreme example, Vladimir Lenin pieced together government statistics in the U.S. and Europe, and proclaimed that “The enormous growth of industry and the remarkably rapid concentration of production in ever-larger enterprises are one of the most characteristic features of capitalism. Modern production censuses give most complete and most exact data on this process.” To Lenin, the rise of large firms supports the feasibility and necessity of central planning: it takes advantage of the economies of scale from modern industrial technology, and avoids having economic power controlled by the elite capitalists.

Over the past century, the limitations of government-directed planning have become better understood, yet the scope of planning through large-scale business enterprises has continued to expand. The latest technological development raises interesting open questions going forward. Some postulate that generative AI could further ameliorate the knowledge problem of society (so that large enterprises are easier to manage) or strengthen winner-take-all effects. Meanwhile, maybe decentralized coordination will eventually become efficient with future technological advancement. If history is any guide, the dominance of large firms will continue to be an important topic in society, and time will tell whether the long-run trends towards greater concentration will revert one day.