The role of financial frictions in amplifying and propagating economic shocks has received significant attention in explaining fluctuations over the business cycle. Financial frictions introduce a wedge between the cost of external finance and the opportunity cost of internal funds. This implies that the strength of firms’ balance sheets will affect the manner in which their investment activity reacts to economic shocks. Current firm investment affects future balance sheet strength, creating a dynamic feedback loop that propagates economic shocks over time. Theoretical models of this so-called “financial accelerator” have played an important role in the literature (e.g. Bernanke and Gertler (1989); Kiyotaki and Moore (1997); Shleifer and Vishny (1992), Bernanke (2007)).
In spite of their importance, empirically testing financial accelerator models has proven to be difficult. While a vast literature exists examining the presence of financial frictions, these frictions serve only as a necessary ingredient for financial accelerator models. See, for example, Lamont (1997), Rauh (2006), Hennessy and Whited (2007) and Kaplan and Zingales (1997), Hubbard and Kashyap (1992), and Rajan and Ramcharan (2012). The essence of financial accelerator models—namely, the role that financial frictions play in propagating economic shocks—remains understudied empirically. There are at least three reasons why this is the case. First, it is difficult to measure exogenous shocks that affect firm productivity. Second, measuring firm productivity, in and of itself, is quite challenging. Indeed, standard productivity measures, such as TFP, are often residuals of regressions relating (mismeasured) outputs and inputs. Finally, it is difficult to obtain clean measures of collateral values, which often play an important role in financial accelerator models.
In a recent project we test the central predictions of financial accelerator models by focusing on a novel setting – the agricultural sector in Iowa. This sector provides a natural environment, rich in data, to examine how shocks to productivity are propagated, both during normal times as well as during crises. As a source of identification, we use exogenous shocks to productivity arising from variation in weather. To analyze productivity, and relate it to productivity shocks as well as measures of financial constraints, we exploit the rich data available on farm crop yields. Finally, focusing on the agricultural sector provides us with a measure of collateral values: land is a main source of collateral for farms and data on land prices are readily available.
We find that the effect of weather shocks is indeed persistent: past weather-driven shocks to productivity affect both current farm yields as well as current land values, up to two years following the shock. We also find these effects are weaker for counties with higher per-capita income. Consistent with the importance of financial frictions in accelerator models, our results show that the sensitivity of farm yields and land values to past weather shocks increases during the 1980s farm debt crisis. The effect is economically substantial, with the sensitivity of yields to past shocks increasing during the debt crisis by a factor of more than three. The result highlight how temporary shocks to productivity can have long lasting effects.
Rajkamal Iyer is an Associate Professor of Finance at MIT Sloan School of Management.
Contributors to this post include Nittai Bergman, Associate Professor of Finance at MIT Sloan and Richard Thakor, a doctoral student at MIT Sloan.