Firms borrow against earnings more than they do against their assets. How does this affect the aggregate response to financial crises? I take a dynamic stochastic general equilibrium model of heterogeneous firms that choose their capital and debt subject to a borrowing constraint and examine the recovery from a financial crisis when firms can use different types of collateral. I compare between two collateral types: assets, and earnings. I find that when firms borrow against earnings recessions are deeper, but recoveries are quicker compared to when firms borrow against assets. I also find that neither type of collateral can, buy itself, completely explain the recovery from the Great Recession. Instead, the path of investment after the 2007-2008 Financial crisis is better captured by firms borrowing against earnings than by firms borrowing against assets, but this is reversed when looking at the path of output. This suggests that a combination of collateral types is required to fully capture the recovery from the Great Recession.

Governments spend substantial sums subsidising firm investment in Research and Development (R&D) with the aim of increasing output. How effective are these subsidies? Could alternative policies boost output more? I develop a quantitative model of the US economy in which heterogeneous firms choose R&D, physical capital, and debt subject to a collateral constraint distinguishing between physical capital and R&D. I find that negative spillovers from higher aggregate R&D render R&D subsidies largely ineffective at increasing output. Conversely, subsidising entry instead of R&D achieves ten times the increase in output because entry subsidies do not generate large negative R&D spillovers. My results suggest that a policymaker seeking to increase output should focus on incentivising entry rather than R&D.

The entry of new firms and the introduction of new products are both pro-cyclical. Which of the two contributes most to business cycle fluctuations? I develop a dynamic stochastic general equilibrium model with endogenous entry and exit that allows heterogeneous multi-product incumbent firms to invest in R&D to affect their probability of introducing a new product in the future. Firms are subject to aggregate productivity shocks that drive business cycles. I simulate the economy’s recovery from a recession and find that potential entrants’ responses accounts for more than 90% of the path of output during the recovery. Conversely, incumbent firms’ responses account for roughly 20% of this recovery path. Hence, a policymaker that wants to speed up the recovery from a recession should focus on encouraging new entry.

The US firm-size distribution has changed substantially over time. How does this affect the aggregate response to a financial crisis? I take a dynamic stochastic general equilibrium model of heterogeneous firms that choose their capital and debt subject to a borrowing constraint and examine the recovery from a financial crisis when firms have a choice of whether or not to borrow against capital or earnings. I compare the economy’s response between the firm-size distribution that prevailed at the start of the 2007 Financial Crisis to the one that prevailed in 1980. I find that the trough of the recession is shallower under the 1980 firm-size distribution due to the higher proportion of smaller firms. These small firms increase their capital and output to ameliorate the effect of the tighter collateral constraints. I also find that the majority of the recession is driven by the tightening of the capital collateral constraint, indicating that a capital constraint alone might be sufficient to model the Great Recession. 

Advertising has become more effective over time. How has this affected the long-run economy and business cycles? I develop a dynamic stochastic general equilibrium model in which heterogeneous multi-product firms invest in advertising and in Research and Development (R&D). Advertising increases demand for existing products while R&D creates new products. I find that historical improvements in advertising effectiveness have increased output and consumption by between 1.4% and 5.8% due to a reallocation of resources from low productivity to high productivity firms. I also find that increasing advertising effectiveness speeds up the economy’s recovery from a recession and reduces the volatility of the business cycle. Moreover, advertising and R&D are substitutes or complements depending on the part of the business cycle in which the economy finds itself. These results highlight the importance of advertising on the aggregate economy.. 

The size-distribution of firms in the US has shifted rightward over the previous 40 years, such that BDS data indicate the average size of a firm has increased by almost 20% between 1978 (21 employees, on average) and 2016 (25 employees). At the same time, there is evidence (such as that in De Loecker, Eeckhout & Unger (2020)) that firm mark-ups have increased by as much as 27% over the period 1986 - 2016. We investigate the extent to which the rightward shift in the firm-size distribution can explain the increase in mark-ups using a model of heterogeneous firms with endogenous entry and exit. Firms in our model charge a mark-up that varies according to their size via the Kimball aggregator, which allows us to examine the effect of different distributions of firms on the average mark-up. We calibrate our model to match the firm-size distribution in each of 1986 and 2016 and find that the changes in firm size over that timeframe imply an increase in mark-up of 17%. In other words, changes in the firm-size distribution can explain roughly 63% of the observed change in mark-ups over the same period. 

We extend the standard SEIR model to include consumption and labour decisions of households to capture endogenous variations in the transmission rates of a viral infection in the presence of aggregate uncertainty about policy intervention. We explore and contrast the economic and epidemiological effects of various policy interventions: a baseline laissez-faire decentralised equilibrium with no policy intervention, severe restrictions, moderate restrictions, and a conditional lockdown based on the number of hospital admissions. We find that accounting for agents’ uncertainty regarding the timing and size of any restrictions being imposed has substantial effects on the outcomes of those policies. For example, even in the laissez-faire baseline case, accounting for this uncertainty means that agents’ endogenous responses to increasing infections are smaller than they would be in the absence of this uncertainty. This effect is driven by agents having some non-zero expectation regarding the imposition of a future lockdown and associated reduction in utility, such that they compensate by not reducing their consumption and labour responses by as much as they would without this uncertainty.