Firm Investment During Large Crises: The role of Credit Conditions

Abstract Business fixed investment in Greece collapsed during the Great Depression. In 2014 it was about half the level it attained in 2009. The large crisis was characterized by both a drop in domestic demand and an increase in the shadow cost of capital. In a standand model, firm investment depends on the marginal product of capital which is driven by profitability and on the shadow cost of capital. What was the relative importance of these two factors for the observed investment collapse? Regarding the shadow cost of capital, was its variation mostly due to real investment adjustment costs or frictions in the credit market? To answer these questions, we use a novel, firm-level dataset of Greek manufacturing firms. We find that profitability is not enough to explain observed drop in firm investment. In addition, financing constraints on firm access to capital were the predominant friction from the capital cost side. Our strategy is to estimate a dynamic model of firm investment separately for the periods before and during the crisis. Capital adjustment cost entails both convex and non-convex adjustment costs. We also allow for debt finance subject to collateral constraints, with aggregate financial shocks affecting maximum debt capacity, as in Khan and Thomas (2013). As a first step, we estimate firm profitability and observe extensive heterogeneity both across firms and across sectors before and during the crisis. Intriguingly, while we observe a substantial drop of investment rates in all sectors, in some firm profitability does not drop during the crisis. This is an indication that variation in the cost of capital is important. Our structural estimates show a pronounced increase in the importance of capital adjustment costs during the crisis. A firm model with only real adjustment frictions adequately captures certain data moments like investment rate dispersion and inaction. However, when simulated for the crisis period, conditional on the firm distribution of capital stock and firm profitability, the model can account for less than half the observed drop in the investment rate. In a second step, we augment the model with financial frictions that can generate further variation in the cost of capital across firms. An aggregate financial shock that increases the collateral requirement for borrowing, induces some firms with a good profitability shock realization to reduce investment. Our results highlight the important role of the near collapse of the Greek banking system and extreme tightening of firm financing conditions in generating the collapse in investment during this large crisis.