Macro Roundup

Excellent review ofThe Industrial Revolution in Services

“…The reason for increased concentration appears to be technological, not anticompetitive. And, as might be expected from that cause, its effects are beneficial. Indeed, the story is both intuitive and positive....This means that competition is increasing, not decreasing, whether it is accompanied by an increase in concentration or not....The net effect is a decrease in the power of top firms relative to the economy as a whole, as the largest firms specialize more, and are dominant in fewer industries...Thus, summing up, technology has led to increased productivity as well as greater specialization by large firms, especially in relatively concentrated industries (exactly the opposite of the pessimistic stories):...”

Geoffrey Manne, "What if rising concentration were an indication of more competition, not less?,"Truth On The Market, December 14, 2019, https://truthonthemarket.com/2019/12/14/what-if-rising-concentration-were-an-indication-of-more-competition-not-less/

What if rising concentration were an indication of more competition, not less?

An oft-repeated claim of conferences, media, and left-wing think tanks is that lax antitrust enforcement has led to a substantial increase in concentration in the US economy of late, strangling the economy, harming workers, and saddling consumers with greater markups in the process. But what if rising concentration (and the current level of antitrust enforcement) were an indication of more competition, not less?

By now the concentration-as-antitrust-bogeyman story is virtually conventional wisdom, echoed, of course, by political candidates such as Elizabeth Warren trying to cash in on the need for a government response to such dire circumstances:

In industry after industry — airlines, banking, health care, agriculture, tech — a handful of corporate giants control more and more. The big guys are locking out smaller, newer competitors. They are crushing innovation. Even if you don’t see the gears turning, this massive concentration means prices go up and quality goes down for everything from air travel to internet service.

But the claim that lax antitrust enforcement has led to increased concentration in the US and that it has caused economic harm has been debunked several times (for some of our own debunking, see Eric Fruits’ posts here, here, and here). Or, more charitably to those who tirelessly repeat the claim as if it is “settled science,” it has been significantly called into question.

Most recently, several working papers looking at the data on concentration in detail and attempting to identify the likely cause for the observed data, show precisely the opposite relationship. The reason for increased concentration appears to be technological, not anticompetitive. And, as might be expected from that cause, its effects are beneficial. Indeed, the story is both intuitive and positive.

What’s more, while national concentration does appear to be increasing in some sectors of the economy, it’s not actually so clear that the same is true for local concentration — which is often the relevant antitrust market.

The most recent — and, I believe, most significant — corrective to the conventional story comes from economists Chang-Tai Hsieh of the University of Chicago and Esteban Rossi-Hansberg of Princeton University. As they write in a recent paper titled, “The Industrial Revolution in Services”:

We show that new technologies have enabled firms that adopt them to scale production over a large number of establishments dispersed across space. Firms that adopt this technology grow by increasing the number of local markets that they serve, but on average are smaller in the markets that they do serve. Unlike Henry Ford’s revolution in manufacturing more than a hundred years ago when manufacturing firms grew by concentrating production in a given location, the new industrial revolution in non-traded sectors takes the form of horizontal expansion across more locations. At the same time, multi-product firms are forced to exit industries where their productivity is low or where the new technology has had no effect. Empirically we see that top firms in the overall economy are more focused and have larger market shares in their chosen sectors, but their size as a share of employment in the overall economy has not changed. (pp. 42-43) (emphasis added).

This makes perfect sense. And it has the benefit of not second-guessing structural changes made in response to technological change. Rather, it points to technological change as doing what it regularly does: improving productivity.

The implementation of new technology seems to be conferring benefits — it’s just that these benefits are not evenly distributed across all firms and industries. But the assumption that larger firms are causing harm (or even that there is any harm in the first place, whatever the cause) is unmerited.

What the authors find is that the apparent rise in national concentration doesn’t tell the relevant story, and the data certainly aren’t consistent with assumptions that anticompetitive conduct is either a cause or a result of structural changes in the economy.

Hsieh and Rossi-Hansberg point out that increased concentration is not happening everywhere, but is being driven by just three industries:

First, we show that the phenomena of rising concentration... is only seen in three broad sectors - services, wholesale, and retail.... op firms have become more efficient over time, but our evidence indicates that this is only true for top firms in these three sectors. In manufacturing, for example, concentration has fallen.

Second, rising concentration in these sectors is entirely driven by an increase the number of local markets served by the top firms. (p. 4) (emphasis added).

These findings are a gloss on a (then) working paper — The Fall of the Labor Share and the Rise of Superstar Firms — by David Autor, David Dorn, Lawrence F. Katz, Christina Patterson, and John Van Reenan (now forthcoming in the QJE). Autor et al. (2019) finds that concentration is rising, and that it is the result of increased productivity:

If globalization or technological changes push sales towards the most productive firms in each industry, product market concentration will rise as industries become increasingly dominated by superstar firms, which have high markups and a low labor share of value-added.

We empirically assess seven predictions of this hypothesis: (i) industry sales will increasingly concentrate in a small number of firms; (ii) industries where concentration rises most will have the largest declines in the labor share; (iii) the fall in the labor share will be driven largely by reallocation rather than a fall in the unweighted mean labor share across all firms; (iv) the between-firm reallocation component of the fall in the labor share will be greatest in the sectors with the largest increases in market concentration; (v) the industries that are becoming more concentrated will exhibit faster growth of productivity; (vi) the aggregate markup will rise more than the typical firm’s markup; and (vii) these patterns should be observed not only in U.S. firms, but also internationally. We find support for all of these predictions. (emphasis added).

This is alone is quite important (and seemingly often overlooked). Autor et al. (2019) finds that rising concentration is a result of increased productivity that weeds out less-efficient producers. This is a good thing.

But Hsieh & Rossi-Hansberg drill down into the data to find something perhaps even more significant: the rise in concentration itself is limited to just a few sectors, and, where it is observed, it is predominantly a function of more efficient firms competing in more — and more localized — markets. This means that competition is increasing, not decreasing, whether it is accompanied by an increase in concentration or not.

No matter how may times and under how many monikers the antitrust populists try to revive it, the Structure-Conduct-Performance paradigm remains as moribund as ever. Indeed, on this point, as one of the new antitrust agonists’ own, Fiona Scott Morton, has written (along with co-authors Martin Gaynor and Steven Berry):

In short, there is no well-defined “causal effect of concentration on price,” but rather a set of hypotheses that can explain observed correlations of the joint outcomes of price, measured markups, market share, and concentration. As Bresnahan (1989) argued three decades ago, no clear interpretation of the impact of concentration is possible without a clear focus on equilibrium oligopoly demand and “supply,” where supply includes the list of the marginal cost functions of the firms and the nature of oligopoly competition.

Some of the recent literature on concentration, profits, and markups has simply reasserted the relevance of the old-style structure-conduct-performance correlations. For economists trained in subfields outside industrial organization, such correlations can be attractive.

Our own view, based on the well-established mainstream wisdom in the field of industrial organization for several decades, is that regressions of market outcomes on measures of industry structure like the Herfindahl-Hirschman Index should be given little weight in policy debates. Such correlations will not produce information about the causal estimates that policy demands. It is these causal relationships that will help us understand what, if anything, may be causing markups to rise. (emphasis added).

Indeed! And one reason for the enduring irrelevance of market concentration measures is well laid out in Hsieh and Rossi-Hansberg’s paper:

This evidence is consistent with our view that increasing concentration is driven by new ICT-enabled technologies that ultimately raise aggregate industry TFP. It is not consistent with the view that concentration is due to declining competition or entry barriers..., as these forces will result in a decline in industry employment. (pp. 4-5) (emphasis added)

The net effect is that there is essentially no change in concentration by the top firms in the economy as a whole. The “super-star” firms of today’s economy are larger in their chosen sectors and have unleashed productivity growth in these sectors, but they are not any larger as a share of the aggregate economy. (p. 5) (emphasis added)

Thus, to begin with, the claim that increased concentration leads to monopsony in labor markets (and thus unemployment) appears to be false. Hsieh and Rossi-Hansberg again:

e find that total employment rises substantially in industries with rising concentration. This is true even when we look at total employment of the smaller firms in these industries. (p. 4)

ectors with more top firm concentration are the ones where total industry employment (as a share of aggregate employment) has also grown. The employment share of industries with increased top firm concentration grew from 70% in 1977 to 85% in 2013. (p. 9)

Firms throughout the size distribution increase employment in sectors with increasing concentration, not only the top 10% firms in the industry, although by definition the increase is larger among the top firms. (p. 10) (emphasis added)

Again, what actually appears to be happening is that national-level growth in concentration is actually being driven by increased competition in certain industries at the local level:

93% of the growth in concentration comes from growth in the number of cities served by top firms, and only 7% comes from increased employment per city.... verage employment per county and per establishment of top firms falls. So necessarily more than 100% of concentration growth has to come from the increase in the number of counties and establishments served by the top firms. (p.13)

The net effect is a decrease in the power of top firms relative to the economy as a whole, as the largest firms specialize more, and are dominant in fewer industries:

Top firms produce in more industries than the average firm, but less so in 2013 compared to 1977. The number of industries of a top 0.001% firm (relative to the average firm) fell from 35 in 1977 to 17 in 2013. The corresponding number for a top 0.01% firm is 21 industries in 1977 and 9 industries in 2013. (p. 17)

Thus, summing up, technology has led to increased productivity as well as greater specialization by large firms, especially in relatively concentrated industries (exactly the opposite of the pessimistic stories):

op firms are now more specialized, are larger in the chosen industries, and these are precisely the industries that have experienced concentration growth. (p. 18)

Unsurprisingly (except to some…), the increase in concentration in certain industries does not translate into an increase in concentration in the economy as a whole. In other words, workers can shift jobs between industries, and there is enough geographic and firm mobility to prevent monopsony. (Despite rampant assumptions that increased concentration is constraining labor competition everywhere…).

Although the employment share of top firms in an average industry has increased substantially, the employment share of the top firms in the aggregate economy has not. (p. 15)

It is also simply not clearly the case that concentration is causing prices to rise or otherwise causing any harm. As Hsieh and Rossi-Hansberg note:

he magnitude of the overall trend in markups is still controversial... and... the geographic expansion of top firms leads to declines in local concentration... that could enhance competition. (p. 37)

Indeed, recent papers such as Traina (2018), Gutiérrez and Philippon (2017), and the IMF (2019) have found increasing markups over the last few decades but at much more moderate rates than the famous De Loecker and Eeckhout (2017) study. Other parts of the anticompetitive narrative have been challenged as well. Karabarbounis and Neiman (2018) finds that profits have increased, but are still within their historical range. Rinz (2018) shows decreased wages in concentrated markets but also points out that local concentration has been decreasing over the relevant time period.

None of this should be so surprising. Has antitrust enforcement gotten more lax, leading to greater concentration? According to Vita and Osinski (2018), not so much. And how about the stagnant rate of new firms? Are incumbent monopolists killing off new startups? The more likely — albeit mundane — explanation, according to Hopenhayn et al. (2018), is that increased average firm age is due to an aging labor force. Lastly, the paper from Hsieh and Rossi-Hansberg discussed above is only the latest in a series of papers, including Bessen (2017), Van Reenen (2018), and Autor et al. (2019), that shows a rise in fixed costs due to investments in proprietary information technology, which correlates with increased concentration.

So what is the upshot of all this?

First, as noted, employment has not decreased because of increased concentration; quite the opposite. Employment has increased in the industries that have experienced the most concentration at the national level.
Second, this result suggests that the rise in concentrated industries has not led to increased market power over labor.
Third, concentration itself needs to be understood more precisely. It is not explained by a simple narrative that the economy as a whole has experienced a great deal of concentration and this has been detrimental for consumers and workers. Specific industries have experienced national level concentration, but simultaneously those same industries have become more specialized and expanded competition into local markets.

Surprisingly (because their paper has been around for a while and yet this conclusion is rarely recited by advocates for more intervention — although they happily use the paper to support claims of rising concentration), Autor et al. (2019) finds the same thing:

Our formal model, detailed below, generates superstar effects from increases in the toughness of product market competition that raise the market share of the most productive firms in each sector at the expense of less productive competitors.... An alternative perspective on the rise of superstar firms is that they reflect a diminution of competition, due to a weakening of U.S. antitrust enforcement (Dottling, Gutierrez and Philippon, 2018). Our findings on the similarity of trends in the U.S. and Europe, where antitrust authorities have acted more aggressively on large firms (Gutierrez and Philippon, 2018), combined with the fact that the concentrating sectors appear to be growing more productive and innovative, suggests that this is unlikely to be the primary explanation, although it may important in some specific industries (see Cooper et al, 2019, on healthcare for example). (emphasis added).

The popular narrative among Neo-Brandeisian antitrust scholars that lax antitrust enforcement has led to concentration detrimental to society is at base an empirical one. The findings of these empirical papers severely undermine the persuasiveness of that story.

Matthew Smith, Owen Zidar, Princeton and Eric Zwick have a new update to Top Wealth in America: New Estimates and Implications for Taxing the Rich. Basically further highlighting that the primary driver behind rising incomes of ultra rich isn't a concentration of financial capital, but is a return to human capital

"...One of the most important takeaways from the research is that the primary driving force behind inequality between the richest Americans and everyone else isn’t, as many have argued, a concentration of financial wealth or capital. From 1980 to 2014, the rise in the top 1% share in national income due to capital income is only 2.4 out of 8.1 percentage points in total growth. Rather, the authors argue that much of the recent rise of top incomes and wealth represents a return to human capital, including the labor income of private business owners—for example consultants, lawyers, or doctors—often characterized as capital income of pass-through businesses for tax purposes...."
Matthew Smith, Owen Zidar, Princeton and Eric Zwick, "Wealth gaps are smaller than previous estimates suggest, and capital isn’t entirely to blame for rising inequality," Princeton University, April 29, 2020, https://economics.princeton.edu/2020/04/29/wealth-gaps-are-smaller-than-previous-estimates-suggest-and-capital-isnt-entirely-to-blame-for-rising-inequality/

Wealth gaps are smaller than previous estimates suggest, and capital isn’t entirely to blame for rising inequality

How rich are the richest Americans? A thorough answer to this question is necessary to address public concern over rising inequality, whether the distribution of resources is fair, and how policy ought to respond.

While over the last several years the U.S. has seen increased interest in addressing inequality through taxes on wealth or capital income (for example a capital gains tax), several questions remain. How much revenue could such taxes raise and, given the composition of wealth among the richest Americans, are these taxes the most effective way to reduce gaps in wealth and income? As the burden of coronavirus grows and policymakers start considering new tax policies, it’s important that the public have the most accurate estimates of top wealth and income available.

In a new update to their paper on measuring top wealth and inequality in the U.S. (PDF), Princeton’s Owen Zidar, Chicago Booth’s Eric Zwick, and Treasury’s Matt Smith find that the increase in wealth among the top 0.1% over the last 40 years is half as large as previous estimates would indicate.

Furthermore, they find that the financial assets and capital of the ultra-wealthy amount to a smaller share of national income than previously believed. Instead, they argue a bigger driver of top incomes is human capital and returns on labor often characterized in tax filings as private business profit.

Main findings

The authors find that the top 0.1% share of wealth increased from 7% to 14% from 1978 to 2016. While this rise is half as large as prior estimates, wealth is very concentrated: the top 1% holds nearly as much wealth as the bottom 90%. However, Americans in the 90th to 99th percentile, a group with more than $660,000 in wealth but less than $3.1 million, hold 36% of total wealth. Their wealth exceeds that of the bottom 90% or the top 1%. The chart below compares the authors’ new estimates to those produced in previous studies and using different approaches and data sources.

One of the most important takeaways from the research is that the primary driving force behind inequality between the richest Americans and everyone else isn’t, as many have argued, a concentration of financial wealth or capital. From 1980 to 2014, the rise in the top 1% share in national income due to capital income is only 2.4 out of 8.1 percentage points in total growth.

The paper also has implications for how much revenue a wealth tax might raise. These new estimates of top wealth suggest total revenues would be far less than many expect. A one percent tax on the top 0.1% in 2016 would generate $112 billion assuming taxpayers didn’t change their behavior in response. A graduated tax, which taxes wealth above $50 million at 2% and adds a surtax of 1% of wealth exceeding $1 billion, would raise $117 billion in 2016 (assuming no change in behavior). Prior estimates, calculated using an “equal-returns” approach (see next section for more detail), estimated $207 billion in revenue from such a graduated tax. The new estimates suggest a graduated tax would bring in only 57% of that estimated revenue before accounting for behavioral responses, which would lower revenue estimates further.

Why the new estimates differ from previous estimates

The approach the authors take to estimate wealth builds on an approach first employed by others (Giffen, 1913; Stewart, 1939; Saez and Zucman, 2016) to scale up or “capitalize” income observed on tax returns to estimate top wealth. That is, they take an individual’s income, reported in official fiscal statistics, and estimate how much wealth created it.

The strength in the authors’ new approach, however, is that they use new data to discipline how income maps to wealth for different groups of people. They start by looking at several types of income: fixed income (from liquid assets, deposits, bonds, etc.), public equity income, private business income, etc.

Then, in estimating the amount of wealth that creates these different types of income, they allow for the possibility that rates of returns for each income type can vary across people. Accounting for this empirically relevant variation produces different wealth estimates than earlier work that assumes equal rates of return across people.

As one example, they show that fixed income portfolios of the rich skew toward high-yield bonds and loans, whereas the fixed income portfolios of the non-wealthy are mostly bank deposits. Simply put, fixed income portfolios for the wealthy differ in nature, risk, and liquidity from those for the less wealthy. This difference in the composition of people’s portfolios means there are unequal effective returns on fixed income for the rich and everyone else.

While the authors are not the first to argue that the rich earn higher rates of return on their income (it’s a key point of Thomas Piketty’s Capital), an important contribution of the research is the extent to which they test both their assumptions and previous assumptions against administrative and survey data that does exist.

When the authors produce wealth estimates using their assumption that the rich earn higher rates of return, their results line up nicely with estate tax data and the wealth data reported by the Federal Reserve’s Survey of Consumer Finance (SCF). When they produce estimates using the assumption that everyone earns equal rates of return, their results overstate fixed income assets reported by the SCF, especially at the top. For more detail on how the authors’ new approach differs from the 2016 study by Emmanuel Saez and Gabriel Zucman, read section 10.1 of the full paper or the authors’ response to comments on their earlier draft.

Another contribution of the Research is showing how the authors’ updated wealth estimates change what we know about income inequality and how national income is distributed across income groups. For more detail on how wealth estimates relate to distributional national accounts and inequality statistics, read Eric Zwick’s recent testimony before the Joint Economic Committee of the U.S. Senate (PDF).

Conclusion

While there are many ways to continue improving estimates of wealth and income, this Research further improves our understanding of why inequality has increased and the most effective ways to address growing gaps in wealth and income.

To learn more about the findings and methodology, please download the full paper.

Matthew Smith, Owen Zidar, Eric Zwick, " Top Wealth in America: New Estimates and Implications for Taxing the Rich," National Bureau of Economic Research, April 24, 2020, https://scholar.princeton.edu/sites/default/files/zidar/files/szz-wealth.pdf

“….This paper uses administrative tax data to estimate top wealth in the United States. We build on the capitalization approach in Saez and Zucman (2016) while accounting for heterogeneity within asset classes when mapping income flows to wealth. Our approach reduces bias in wealth estimates because wealth and rates of return are correlated. We find that the top 0.1% share of wealth increased from 7% to 14% from 1978 to 2016. While this rise is half as large as prior estimates, wealth is very concentrated: the top 1% holds nearly as much wealth as the bottom 90%.However, the “P90-99” class holds more wealth than either group after accounting for heterogeneity. Private business and public equity wealth are the primary sources of wealth at the top, and pension and housing wealth account for almost all wealth of the bottom 90%. Our approach substantially reduces estimates of mechanical wealth tax revenue and top capital income in distributional national accounts, which depend on well-measured estimates of top wealth. From 1980 to 2014, capital income accounts for 2.4 out of 8.1 percentage points of the rise of the top 1% income share…. Figure 1A shows how our preferred approach alters estimated wealth for the top 0.1%. The top 0.1% wealth share in 2016 is 14% when accounting for return heterogeneity, and around 20% when assuming equal returns. Top 1% and 0.01% shares fall by 22 percent and 36 percent, respectively. The growth in top wealth shares is also less dramatic. Our approach reduces the growth in top shares since 1978 by half, leaving the recent wealth estimates above the estate tax series and closer to the SCF. Overall, wealth is very concentrated: the top 1% holds nearly as much wealth as the bottom 90%.However, the “P90-99” class holds more wealth than either group.Our approach also alters the composition of top wealth. We find a larger role for private business wealth and a smaller role for fixed income wealth, consistent with the composition of top wealth in the SCF and estate tax data. Private business and public equity wealth are the primary sources of wealth at the top, and pension and housing wealth account for almost all wealth of the bottom 90%...”

Just an FYI, Smith/ Zidar/Zwickupdated Top Wealth in Americathatwe looked at over the summer. Basically further highlighting that the primary driver behind rising incomes of ultra rich isn't a concentration of financial capital, but is a return to human capital