This conference is supported by Grant #2018-1897 from Smith Richardson Foundation
How much power do employers have to suppress wages below marginal productivity? It depends on the firm-level labor supply elasticity. Leveraging data on job applications from the large job board CareerBuilder.com, Azar, Berry, and Marinescu estimate the wage impact on workers' choice among differentiated jobs in the largest occupations. They use a nested logit model of worker's utility for applying to jobs with varying wages and characteristics, including distance from the potential worker's home. The researchers account for the endogeneity of wages by using several different instrumental variable strategies. They find that failing to instrument results in implausibly low elasticities, whereas plausible instruments result in more elastic estimates. Still, the implied market-level labor supply elasticity is about 0.6, while the firm-level labor supply elasticity is about 5.8. This implies that workers produce about 17% more than their wage level, consistent with employers having significant market power even for the largest occupational labor markets.
From 1977 to 2012, the share of US manufacturing employment fell from 26 to 11%, with half of the change occurring among continuing firms. Ding, Fort, Redding, and Schott develop a model of structural transformation within and between firms and examine the contribution of trade towards this reallocation. Their model highlights the role of intangible knowledge capital that can be redeployed across disparate sectors within the firm and complementary design and production activities. As reductions in trade barriers lower the price of intermediate inputs and reduce production costs, this complementarity increases the share of design in overall firm costs. Using separate data on firm input and output shocks, Ding, Fort, Redding, and Schott provide empirical evidence in support of this prediction.
Jarosch, Nimczik, and Sorkin develop an approach to measuring labor market power that builds on the structure of a canonical search model. They relax the common assumption of a continuum of firms and assume that large employers exert market power in the wage bargain by effectively eliminating their own future job openings from a worker's threat point. This granular extension yields a micro-founded concentration index similar to the Herfindahl and a structural mapping between the index and worker compensation. The intuition for the index is that it captures the discounted probability that an unemployed worker will run into the same employer two or more times in a row. The researchers extend the model to allow for productivity differences across firms. They then use the model as an accounting device to measure the contribution of labor market concentration to the level and evolution of mean wages in Austrian labor markets. To do so, Jarosch, Nimczik, and Sorkin define labor markets by clustering firms based on worker flows.
In addition to the conference paper, the research was distributed as NBER Working Paper w26239, which may be a more recent version.
When competing firms possess overlapping sets of investors, maximizing shareholder value may provide incentives that distort competitive behavior, affecting pricing, entry, contracting, and virtually all strategic interactions among firms. Backus, Conlon, and Sinkinson propose a structurally consistent and scaleable approach to the measurement of this phenomenon for the universe of S&P 500 firms between 1980 and 2017. Over this period, the incentives implied by the common ownership hypothesis have grown dramatically. Contrary to popular intuition, this is not primarily associated with the rise of BlackRock and Van- guard: instead, the trend in the time series is driven by a broader rise in diversified investment strategies, of which these firms are only the most recent incarnation. In the cross-section, there is substantial variation that can be traced, both in the theory and the data, to observable firm characteristics - particularly the share of the firm held by retail investors. Finally, the authors show how common ownership can theoretically give rise to incentives for expropriation of undiversified shareholders via tunneling, even in the Berle and Means (1932) world of the "widely held firm."
Neiman and Vavra show that over the last 15 years, the typical household has increasingly concentrated its spending on a few preferred products. However, this is not driven by "superstar" products capturing larger market shares. Instead, households increasingly focus spending on different products from each other. As a result, aggregate spending concentration has in fact decreased over this same period. The researchers use a novel heterogeneous agent model to conclude that increasing product variety is a key driver of these divergent trends. When more products are available, households can select a subset better matched to their particular tastes, and this generates welfare gains not reflected in government statistics. Neiman and Vavra's model features heterogeneous markups because producers of popular products care more about maximizing profits from existing customers, while producers of less popular niche products care more about expanding their customer base. Surprisingly, however, their model can match the observed trends in household and aggregate concentration without any resulting change in aggregate market power.
In addition to the conference paper, the research was distributed as NBER Working Paper w26134, which may be a more recent version.
Many online markets are dominated by a handful of platforms, raising concerns about the exercise of market power in the digital age. Spotify has emerged as the leading interactive music streaming platform, and Aguiar and Waldfogel assess its power by measuring the impact of its promotion decisions - via platform-operated playlists - on the success of songs and artists. The researchers employ discontinuity and instrumental variables approaches to identification and find large and significant effects of playlist inclusion on success. Aguiar and Waldfogel's results provide direct evidence of a prominent platform's power and suggest a need for continued scrutiny of how platforms exercise their power.
The rise in national industry concentration in the US between 1977 and 2013 is driven by a new industrial revolution in three broad non-traded sectors: services, retail, and wholesale. Sectors where national concentration is rising have increased their share of employment, and the expansion is entirely driven by the number of local markets served by firms. Firm employment per market has either increased slightly at the MSA level, or decreased substantially at the county or establishment levels. In industries with increasing concentration, the expansion into more markets is more pronounced for the top 10% firms, but is present for the bottom 90% as well. These trends have not been accompanied by economy-wide concentration. Top U.S. firms are increasingly specialized in sectors with rising industry concentration, but their aggregate employment share has remained roughly stable. Hsieh and Rossi-Hansberg argue that these facts are consistent with the availability of a new set of fixed-cost technologies that enable adopters to produce at lower marginal costs in all markets. They present a simple model of firm size and market entry to describe the menu of new technologies and trace its implications.
In addition to the conference paper, the research was distributed as NBER Working Paper w25968, which may be a more recent version.
In the past several decades, the U.S. economy has witnessed a number of striking trends that indicate a rising market concentration and a slowdown in business dynamism. Akcigit and Ates make an attempt to understand potential common forces behind these empirical regularities through the lens of a micro-founded general equilibrium model of endogenous firm dynamics. Importantly, the theoretical model captures the strategic behavior between competing firms, its effect on their innovation decisions, and the resulting "best versus the rest" dynamics. The authors focus on multiple potential mechanisms that can potentially drive the observed changes and use the calibrated model to assess the relative importance of these channels with particular attention to the implied transitional dynamics. Their results highlight the dominant role of a decline in the intensity of knowledge diffusion from the frontier firms to the laggard ones in explaining the observed shifts. Akcigit and Ates conclude by presenting new evidence that corroborates a declining knowledge diffusion in the economy. The researchers document a higher concentration of patenting in the hands of firms with the largest stock and a changing nature of patents, especially in the post-2000 period, which suggests a heavy use of intellectual property protection by market leaders to limit the diffusion of knowledge. These findings present a potential avenue for future research on the drivers of declining knowledge diffusion.
In addition to the conference paper, the research was distributed as NBER Working Paper w25756, which may be a more recent version.