This dissertation examines the role that financial frictions play in the propagation of aggregate shocks, and the extent to which they are detrimental for firm value and welfare. The first chapter in this dissertation estimates a general equilibrium model of firm dynamics with endogenous leverage, bankruptcy, innovation, and entry decisions to quantify the private and public gains from resolving the debt overhang problem. The second chapter incorporates aggregate shocks to TFP, the level of idiosyncratic asset volatility, and the retained value of the firm upon bankruptcy in the model in chapter one, and analyzes the extent to which alleviating the debt overhang problem changes how aggregates and firm decisions respond to these different aggregate shocks. The third chapter develops a novel decomposition of changes in aggregate productivity that does not require the identification of firm-level TFP or production function coefficients, and implements this decomposition on U.S. public non-financial firms over the period of 1972-2012.
This dissertation develops three independent yet related frameworks to identify economic mechanisms through which financial frictions affect the aggregate economy over the business cycle and along the path of economic development. There are three chapters in this dissertation. In each chapter, a theoretical model is constructed based on motivating empirical facts, followed by quantitative analyses disciplined and evaluated by data at both the macro- and micro-level. Chapter 1, Financial Frictions and Agricultural Productivity Differences, explores the role of financial frictions in accounting for agricultural employment share and labor productivity differences across provinces in China. A two-sector general equilibrium model with a subsistence consumption requirement and financial frictions is constructed. Limited credit decreases the use of intermediate inputs and increases the use of labor input. As a consequence, workers are trapped in the agricultural sector and agricultural labor productivity is low. Since agricultural employment consists of a large percentage of total employment, aggregate labor productivity is also low. Quantitatively, financial frictions alone explain more than 25% of the observed employment share and productivity differences. Financial frictions amplify the effect of TFP differences on agricultural productivity differences by 30%. Cross-country sectoral value-added per worker differences are large. Value-added per worker is much higher in non-agriculture than in agriculture in the typical country, and particularly so in poor countries. Even though these agricultural productivity gaps (APG) are large, poor countries devote most of their employment to agriculture. Based on a novel data set of value-added at the sectoral level that is comparable across provinces, I find the same patterns across provinces in China. In the second chapter, Credit Constraints, Human Capital and the Agricultural Productivity Gaps, I explore and quantify the role of financial frictions in accounting for these puzzling patterns. A two-sector heterogeneous-agent model with human capital investment, occupational choices and financial frictions is developed. Financial frictions depress human capital accumulation and distort occupational choices of rural households. Quantitatively, our model could account for a substantial portion of the observed cross-province differences in sectoral productivities and the APGs. The financial friction alone could account for 80% of the across-province differences in AGPs. It also explains 1/3 of the sectoral productivity differences and 1/5 of the differences in the agricultural employment share and the aggregate productivity across provinces. In Chapter 3, A Search-Theoretic Model of Capital Reallocation, I investigate how search frictions in the capital market affects capital reallocation across firms and the price of used capital over the business cycles. A tractable dynamic general equilibrium model is developed to account for procyclicality of capital reallocation. Firms are heterogeneous in their productivities and they trade used capital in a market which is subject to search frictions. After idiosyncratic productivity shocks are realized, firms are able to adjust their capital stock to a more favorable level before production. In the booms, the demand of used capital increases and the market tightness of used capital market is small. Hence, capital reallocation is larger and the price of used capital is higher. During the recessions, buyers demand less used capital and the market tightness is large. Consequently, capital reallocation is smaller and the price of used capital is lower. Quantitatively, the model could generate a correlation coefficient between capital reallocation and output that is consistent with the data.
This thesis consists of three chapters on the macroeconomic implications of financial frictions. The first chapter investigates the implications of student loan debt on labor market outcomes. I begin by analytically demonstrating that individuals under debt tend to search less and end up with lower-paid jobs. I then develop and estimate a quantitative model with college entry, borrowing, and job search using NLSY97 data to evaluate the proposed mechanism under the fixed repayment plan and the income-based repayment plan (IBR). My simulation suggests that the distortion of debt on job search decisions is large under the fixed repayment plan. IBR alleviates this distortion and improves welfare. In general equilibrium, debt alleviation achieved through IBR effectively offers a tuition subsidy that increases college entry and encourages firms to post more jobs, further improving welfare. The second chapter, joint with Winston Dou, proposes a dynamic corporate model in which firms face imperfect capital markets and frictional product markets. We highlight the importance of the endogeneity of the marginal value of liquidity in determining the interactions between investment, financing and product price setting decisions. The model implies several testable predictions: (1) financially constrained firms are more inclined to increase their desired markups of products; (2) firms facing larger price stickiness tend to issue less external equity and conduct less big payouts; and (3) a large part of the cost from price stickiness is induced by financial frictions. Lastly, we provide stylized facts consistent with our model's predictions. The third chapter (joint with Era Dabla-Norris, Robert Townsend, and Filiz Unsal) develops a general equilibrium model with three dimensions of financial inclusion, depth, and intermediation efficiency. We find that the economic implications of financial inclusion policies vary with the source of frictions. In partial equilibrium, we show analytically that relaxing each of these constraints separately increases GDP. However, when constraints are relaxed jointly, the impacts on the intensive margin (increasing output per entrepreneur with access to credit) are amplified, while the impacts on the extensive margin (promoting credit access) are dampened. In general equilibrium, we discipline the model with firm-level data from six countries and quantitatively evaluate the policy impacts.
My dissertation is comprised of three papers on the causes and consequences of the U.S. Great Recession. The emphasis is on the role that financial frictions play in magnifying financial shocks, as well as in informing the effectiveness of potential policies. Chapter 1, "Financial Frictions, Investment Delay and Asset Market Interventions," co-authored with Shouyong Shi, studies the role of investment delay in propagating different types of financial shocks, and how this role impacts the effectiveness of asset market interventions. The topic is motivated by the observation that, during the Great Recession, governments conducted large-scale asset market interventions. The aim was to increase the level of liquidity in the asset market and make it easier for firms to obtain financing. However, firms were observed to have delayed investment by hoarding liquid funds, part of which were obtained through the interventions. We construct a dynamic macro model to incorporate financial frictions and investment delay. Investment is undertaken by entrepreneurs who face liquidity frictions in the equity market and a collateral constraint in the debt market. After calibrating the model to the U.S. data, we quantitatively examine how aggregate activity is affected by two types of financial shocks: (i) a shock to equity liquidity, and (ii) a shock to entrepreneurs' borrowing capacity. We then analyze the effectiveness of government interventions in the asset market after such financial shocks. In particular, we compare the effects of government purchases of private equity and of private debt in the open market. In addition, we examine how these effects of government interventions depend on the option to delay investment. In Chapter 2, "Housing Liquidity and Unemployment: The Role of Firm Financial Frictions," I build upon the role that firms' ability to obtain funding plays in the severity of the Great Recession. I focus specifically on how the housing crisis reduced the ability of firms to obtain funding, and the consequences for unemployment. An important feature I focus on is the role of housing liquidity, or how easy it is to sell or buy a house. I analyze how an initial fall in housing market liquidity, linked to rising foreclosure costs for banks, affects labor market outcomes, which can have further feedback effects. I focus on the role that firm financial frictions play in these feedback effects. To this end, I construct a dynamic macro model that incorporates frictional housing and labor markets, as well as firm financial frictions. Mortgages are obtained from banks that incur foreclosure costs in the event of default. Foreclosure costs also affect the ease with which firms can borrow, and this influences their hiring decisions. I calibrate the model to U.S. data, and find that a rise in foreclosure costs that generates a 10% fall in the firm loan-to-output ratio results in a 3 percentage point rise in the unemployment rate. The rise in unemployment makes it more difficult for indebted owners to avoid defaulting on their mortgage. This rise in default, on the order of 20 percent, creates further slack in the housing market by both increasing the number of houses on the market and reducing the amount of buyers. Consequently, there are large drops in housing prices and in the size of mortgage loans. Notably, when firm financial frictions are absent, I observe a counter-factual fall in the unemployment rate, which mitigates the effects on the housing market, and even results in a fall in the mortgage default rate. The results highlight the importance of the impact of the housing market crisis on a firm's willingness to hire, and how firms' limited access to credit magnifies the initial housing shock. In Chapter 3, "Housing Market Distress and Unemployment: A Dynamic Analysis," I add to the contributions of my second paper, and extend the analysis to determine the dynamic effects of the housing crisis on unemployment. In Chapter 2, I focused on comparing stationary equilibria when there is a rise in the foreclosure costs associated with mortgage default. However, a full analysis must also take into account the dynamic effects of the shock. In order to do the dynamic analysis, I modify the model in my job market paper to satisfy the conditions of block recursivity. I do this by incorporating Hedlund's (2016) technique of introducing real estate agents in the housing market that match separately with buyers and sellers. Doing this makes the model's endogenous variables independent of the distribution of households and firms. Rather, the impact of the distribution is summarized by the shadow value of housing. This greatly improves the tractability of the model, and allows me to compute the dynamic response to a fall in a bank's ability to sell a foreclosed house, thus raising the costs of mortgage default. I find that the results are largely dependent on the size and persistence of the shock, as well as the level of firm financial frictions that are present. When firm financial frictions are high, as represented by the presence of an interest rate premium charged to firms, and the initial shock is large, the shock is transferred to firms via an endogenous rise in the cost of renting capital. Firms scale back on production and reduce employment. The rise in unemployment increases the debt burden for households with large mortgages. They can try and sell, but find it difficult to do so because they must sell at a high price to be able to pay off their debt. If they fail, they are forced to default, thus further raising the mortgage costs of banks, further reducing resources to firms, and propagating the initial shock. However, the extent of the propagation is limited; once the shock wears off, the economy recovers to its pre-crisis levels within two quarters. I discuss the reasons why, and what elements would be needed for greater persistence.
