Business cycles reshape product market dynamics, but these responses differ markedly across markets. This heterogeneity arises from differences in consumer demand structures and their interaction with producers' entry and exit decisions–an aspect largely overlooked in the literature. This paper examines how product entry and exit over the business cycle depend on the structure of consumer demand. I develop a model of strategic firm dynamics under imperfect competition with heterogeneous producers and consumers, allowing demand composition to shape the evolution of market structure. Using detailed scanner data from the consumer goods sector, I estimate the model separately for each differentiated product market to analyze producer responses during the Great Recession. I find that negative aggregate demand shocks lead firms to reduce entry and increase exit, with the magnitude of these responses varying systematically with the price elasticity of demand. Markets with less elastic demand experience larger contractions in product variety, and interactions between elasticity and product characteristics further amplify recessionary effects. I also show that the welfare costs of the business cycle are substantially larger when accounting for demand heterogeneity, with most losses stemming from declines in consumer surplus. In the absence of consumer heterogeneity, the welfare loss is significantly understated. Overall, the results demonstrate that consumer heterogeneity is a key driver of how recessions affect market structure and welfare.

Presented at: (2025) SEA Annual Meeting, Fall Midwest Macroeconomics Meeting, Texas Macro Job Candidate Conference, BC PhD Conference in Economics, Texas A&M Macro Mini-Conference (poster) (2024) SEA Annual Meeting, Federal Reserve Board, North American Summer Meeting of the Econometric Society, ASSA Annual Meeting (2023) SEA Annual Meeting

This paper studies the critical but underexplored role of subcontracting in shaping the spatial and firm-level effects of federal government spending. Leveraging newly available data on defense subcontract awards since 2011, linked with NETS establishment-level data, we examine prime–subcontractor relationships across counties, industries and time. We document three stylized facts: (1) subcontracting leads to widespread geographic relocation of federal dollars; (2) it reallocates spending across sectors, notably from service-sector primes to manufacturing subcontractors; and (3) large firms dominate subcontracting networks, even receiving subawards from smaller primes. Accounting for this geographic relocation shows that conventional estimates understate local multiplier effects by approximately 20%. While subcontracting broadens the spatial reach of federal spending, its average local impact is smaller than that of prime contracts. Establishment-level evidence shows that, subcontractors—especially large ones and those in goods sectors—exhibit weaker and less persistent employment and revenue gains than prime contractors, reflecting the shorter and less stable nature of subcontracts. These weaker multipliers also stem from the skewed distribution of subcontracts toward large manufacturers. Overall, our findings reveal substantial heterogeneity in how procurement opportunities diffuse through the private sector and shape the effects of federal spending.

We decompose the employment effect of the minimum wage into changes in employment within continued establishments and changes due to the exits of establishments. For small-sized establishments, we distinguish between other labor adjustment margins, such as hours worked, hiring, and separation at the individual level. Using employer–employee matched panel data, we show that the magnitude and channels of the employment effect differ by establishment size and industry. We find a negative effect of the minimum wage on employment growth, which is highly concentrated among small establishments. In particular, in the food and lodging and manufacturing industries, a substantial part of the effect is driven by business closing.