AEJ: Macroeconomics, 18(3), Vol. 18, No. 3, July 2026, pp. 265-303
AEA: Papers & Proceedings, Vol. 113, May 2023, pp. 56-60
with Lena Anayi, Nick Bloom, Phil Bunn, Paul Mizen, Greg Thwaites, Ivan Yotzov
We study how firms' financing of innovation shapes the transmission of monetary policy to productivity. Using a new panel of U.S. firms' balance sheets matched to loan contracts, we compare firms with and without access to cash flow-based borrowing, credit extended against earnings rather than physical collateral. Although contractionary monetary policy shocks reduce cash flow similarly for both groups, they lower R&D much more among firms without access to this form of financing. To interpret these findings, we develop a New Keynesian model with endogenous growth in which firms with access smooth R&D through debt issuance, while firms without access cut innovation investment more sharply. A 25-basis-point monetary tightening pushes output about 1 percent below its pre-shock trend on impact and leaves a persistent loss of about 0.12 percent. Extending cash flow-based borrowing to all firms reduces this persistent output loss by about one-third. The persistent decline falls disproportionately on firms without access, which are younger and produce more and higher-quality patents. The medium-run cost of monetary policy is therefore not only lower innovation, but also the misallocation of innovation away from successful innovators.
How do firms adjust their investment in response to sales shocks and what determines the response? Using a unique firm-level survey, we propose a novel approach to estimate UK firms’ marginal propensity to invest (MPI) out of additional income: the forecast error of their sales growth expectations. Investment responds significantly to these sales surprises, with a 1 percentage point unexpected growth in sales translating into a 0.31 percentage point increase in capital expenditure. We find attentive firms to be more responsive, consistent with sales growth surprises providing firms with information about their demand. Sales growth surprises also cause firms to increase their prices, supporting this interpretation. We do not find evidence that these results are driven by financial frictions, uncertainty, or productivity shocks.
This paper studies the effect of asset-based versus cash flow-based debt contracts on the transmission of monetary policy to firm-level investment and borrowing. Using information from detailed loan-level data matched with balance sheet data and stock return data, I document that in response to a contractionary monetary shock, asset-based borrowers experience sharper contractions in borrowing and investment than cash flow-based borrowers. Despite the fact that asset-based borrowers contribute only 15% to aggregate investment, they are responsible for 64% of the total investment response. To understand the channels and provide a microfoundation for the endogenous choice of these debt contracts, I set up a heterogeneous firm New Keynesian model with limited enforceability. The quantitative model shows that the traditional collateral channel explains this heterogeneous sensitivity as cash flow-based borrowers are less susceptible to collateral damage from changes in asset prices. This result indicates that the prevalence of asset-based debt contracts increases the strength of the financial accelerator channel and thereby shapes monetary policy transmission.
How does the dispersion of firm-level shocks affect the investment channel of monetary policy? Using firm-level panel data, we construct several measures of dispersion of productivity shocks, time-pooled and time-varying, and interact high-frequency identified monetary policy shocks with these measures of idiosyncratic shock volatility. We document a novel fact: monetary policy has dampened real effects via the investment channel when firm-level TFP shock volatility is high. Our estimates for dampening effects of volatility are statistically and economically significant - moving from the tenth to the ninetieth percentile of the volatility distribution approximately halves point estimates of impulse response functions to contractionary monetary policy shocks. Given that dispersion rises in recessions, these findings offer further evidence as to why monetary policy is weaker in recessions, and emphasize the importance of firm heterogeneity in monetary policy transmission.
Financial Constraints Revisited (with Nick Bloom, Phil Bunn, Paul Mizen)
Financial constraints are central to firm investment and to the transmission of monetary policy, yet they are never observed: for decades the literature has inferred them from balance-sheet proxies. We measure them directly instead. In the Bank of England's Decision Maker Panel, senior decision makers report whether, and on which margin, finance constrains their planned investment, and we match their answers to each firm's archived balance sheet. A financial constraint turns out not to be one latent state but a set of distinct margins that split internal from external finance: the balance sheet identifies them through different characteristics, profitability and dividends marking the internal margin and leverage the external, a separation we test formally and that no single index reproduces. Standard proxies recover constraint status barely above chance out of sample, and the internal margin is the hardest of all to predict. A model with two primitives, internal net worth and one external-finance supply curve, organises the margins into two transmission channels: the internal margin moves with cash-flow surprises, the cost margin with the interest rate, and adverse shocks transmit more strongly than favourable ones, a firm-level counterpart of the aggregate 'pushing on a string' asymmetry.
NonLinearities in the Effects of Monetary Policy
with Russell Cooper
Market Concentration and Monetary Policy Transmission in a Monetary Union
with Livia Chitu and Federica Romei
High Frequency Firm Responsiveness
with Nick Bloom, Phil Bunn, Paul Mizen, Greg Thwaites, Ivan Yotzov