On the interest rate risk of housing government sponsored enterprises.

摘要

Over the last decade, FannieMae and FreddieMac have enlarged their retained mortgage portfolios, increasing exposure to interest rate and prepayment risk. It is widely believed that this growth is partly due to the absence of market discipline, as investors perceive that losses will be borne by the Federal government, and has led to concerns that the size of the portfolios exposes the financial system to systemic risk. In this work, we quantify the risk using a stochastic factor cash flow model. We estimate an agency debt term structure and a mortgage contract rate model, and use it in two applications. First, while risk-taking is mitigated in principle by a regulatory risk capital test, we find that interest rate shocks exceeding those embodied in the test occur in about one percent of draws, and that this rate depends sensitively on the initial term structure. Additionally, other shocks not counted in the analysis may be equally severe. This suggests that the regulatory requirements may not be effective, and may be particularly lenient when initial rates are low. Second, as an alternative capital test, we construct a cash flow model of the combined firms and use the risk factors to quantify expected insolvency rates, loss-given-default, and value-at-risk curves. Critical assumptions for firm operation are that the effective asset-liability maturity gap is kept at its initial value, and that prepayment risk is partially hedged. For base case parameters and initial rates as of year-end 1999, we find a 480bps 10-year probability of insolvency, and a 10-year expected loss of {dollar}8 billion. The 99.9% Value-at-Risk measure is {dollar}300 billion. For conditions at year-end 2003, we find smaller values of 15bps, {dollar}200 million, and {dollar}70 billion, respectively. Provided hedging goals are met, we conclude that the risk of insolvency is highly dependent on the interest rate environment, and less so on the initial portfolio composition. We investigate the sensitivity of our results to model assumptions, and various regulatory proposals to mitigate the interest rate risk of the retained portfolio. The latter include portfolio growth rates, the effects of privatization and required risk capital levels.