Essays on U.S. Federal Housing Credit Policy
Fieldhouse, Andrew Jack
Congress has heavily intervened in U.S. mortgage markets ever since the Great Depression, when federal housing credit policies were first deployed to resuscitate housing and mortgage markets. Congress chartered federal agencies and government-sponsored enterprises (GSEs) to promote access to mortgage credit by purchasing or guaranteeing mortgages. Since the onset of the Great Recession the Federal Reserve has conducted its own form of housing credit policy aimed at reducing the cost and increasing the availability of mortgage credit: large-scale purchases of government agency mortgage-backed securities (MBS). Whereas monetary policy targets a term structure of risk-free interest rates, credit policies aim to alter the allocation of credit by absorbing or subsidizing lending risks. Government purchases of mortgage debt might simply displace private mortgage lending, or housing credit policies could channel resources toward housing by subsidizing a reduction in mortgage risk premia. Despite the federal government's expansive use of housing credit policies, evidence on the macroeconomic effects of government mortgage purchases has been constrained by an identification problem. Regressing housing or mortgage market activity on government agency mortgage purchases would capture reverse causality bias arising from policy endogeneity and profit motives; these sources of reverse causality bias would similarly undermine common macroeconometric identification strategies. Studying the Fed's MBS purchases during the Great Recession faces related intrinsic challenges. My dissertation develops a novel identification strategy to circumvent these challenges to inference regarding the macroeconomic effects of government agencies purchasing or selling mortgage debt. I construct instrumental variables from observable policy interventions over 1967--2006 and use them to tease apart the intended versus unintended causal effects of government purchases of mortgage debt. The novel identification strategy underpinning my dissertation is a narrative analysis of regulatory policy changes affecting government agency purchases of mortgage debt. The narrative approach to time series identification exploits the historical record for exogenous policy shocks, as opposed to backing out shocks from latent variables with modeling assumptions or statistical techniques. I contribute the first narrative analysis of U.S. housing credit policies, using primary sources to identify and quantify regulatory shocks affecting the mortgage holdings of Fannie Mae, Freddie Mac, Ginnie Mae, the Federal Reserve, and the U.S. Treasury Department. Regulatory policy changes that I classify as not cyclically motivated are intended as instrumental variables for government agency mortgage purchases, circumventing concerns about reverse causality bias. The first chapter of my dissertation exploits identifying variation from my narrative analysis to document whether government mortgage purchases advance stated housing policy objectives and to study interactions with monetary policy. We find that agency purchases boost total mortgage lending and lower mortgage rates, indicating that policymakers are capable of directing credit toward housing, as intended, as opposed to simply crowding out private mortgage lending. Agency purchases also advance periodic policy objectives of increasing housing starts and homeownership rates. We identify similarities in the transmission of monetary policy shocks and agency mortgage purchases, and document significant interactions between monetary and housing credit policies; Congress frequently used cyclically motivated credit policies to cushion housing and mortgage markets from contractionary monetary shocks. If housing credit policies expand targeted lending volumes by subsidizing or absorbing private credit risks, they may inadvertently reduce other lending. My second dissertation chapter empirically tests whether the mortgage purchases of Fannie Mae and Freddie Mac unintentionally displace commercial lending and related real activity by subsidizing an expansion in mortgage lending. I use my narrative analysis for exogenous variation in the mortgage purchases of Fannie and Freddie. Regulatory shocks to GSE mortgage purchases boost private home mortgage lending yet unintentionally reduce commercial real estate and business lending. U.S. housing credit policies similarly reallocate construction activity toward housing and away from commercial real estate, negating any intended stimulus to aggregate construction spending or employment. The third chapter of my dissertation is the narrative analysis underpinning the identification strategies behind the first two chapters. My findings are relevant to efforts in Congress to resolve the fate of Fannie and Freddie, as well as to the ongoing reduction in the Fed's MBS holdings. Government agency mortgage purchases are capable of increasing mortgage borrowing, reducing mortgage rates, boosting housing investment, and raising homeownership rates, as intended, through a subsidy channel. Subsidizing mortgage borrowing, however, involves an unintended tradeoff with respect to commercial lending and commercial real estate investment. Government agency mortgage purchases appear ineffectual as a tool of stabilization policy---as they have been employed on and off since the Great Depression.
Public Finance; macroeconomics; Economics; Credit Policy; Government-Sponsored Enterprises; Mortgages; Quantitative Easing
Barseghyan, Levon; Karolyi, George Andrew; Nimark, Preben Kristoffer
Doctor of Philosophy
dissertation or thesis