Key Drivers of Default

According to the mortgage literature, the key determinants of default initially include observable socio-demographic characteristics, terms of the mortgage con­tract, mortgage characteristics, and macroeconomic conditions.

Terms of a credit contract are practically used as proxy variables to estimate the risk of a particular borrower. For example, mortgages with low loan-to value ratio (LTV) are attractive for non-liquid borrowers. The probability that they could face a serious problem of repayment of a loan is much higher. Moreover, borrowers with LTVs higher than 90 % think as holders, because they do not invest a lot of their own capital and are less motivated to overcome obstacles with repayment of a loan. For this reason, mortgages with high LTV are riskier, and lenders offer higher interest rates for these mortgage products. Loans that default tend to be adjustable – rate mortgages, are associated with higher initial LTV, and tend to be issued to borrowers with lower credit scores (Bajari et al. 2008). Campbell and Cocco (2011) empirically supported this idea by using simulated data. Their findings confirm that high LTV (loan-to-value) increases the probability of default.

Typically, mortgages have two types of interest rates—adjustable (ARM) and fixed (FRM). Fixed-rate mortgages are riskier, and the level of their interest rates depends on the stock index. According to Bajari et al. (2008) borrowers tend to have higher interest rates than the market rate.

Socio-demographic characteristics such as income of a particular borrower play a significant role to predict default, because they directly influence the ability to repay a mortgage. However, the debt-to-income ratio has larger effect on borrowers with low credit quality. The level of education could be regarded as a proxy for the level of financial literacy of a particular borrower, which could influence the probability of default as well.

A new stream of mortgage studies originated from the financial turmoil of 2007­2009 that began in the U. S. Several recent empirical findings by Dell’Ariccia et al. (2012), An and Qi (2012), Demyanyk and Van Hemert (2011), Ashcraft et al. (2011), and Mian and Sufi (2009) confirm the highly statistical significance of macroeconomic conditions in explaining mortgage default. Obtained results are consistent with the notion that a relaxation of lending standards, triggered by an increased demand for loans, contributed to the boom and the ensuing crisis, together with other supply-side explanations. Specifically, such supply factors include house price appreciation and mortgage securitization (Keys et al. 2009; Dell’Ariccia et al. 2012).

In addition, Dell’Ariccia et al. (2012) concluded that the development of the mortgage market led to the classical boom-bust scenario. This concept implies fast growth with subsequent relaxation of credit underwriting standards, debt service deterioration, and drop of market premiums. The warning signal of the onset crisis of 2007-2009 was an explosion of house prices in the real estate market during 2003-2005. A fall in house prices contributes also to an increase in defaults (Gerardi et al. 2009).

The contribution of local economic conditions and change of credit underwriting standards to default are also significant (Cutts and Merrill 2008). However, changes in mortgage default rates are most sensitive to changes in the structural component rather than the level of local unemployment rate (Querica et al. 2011). In addition, Querica et al. (2012) find that mortgage default and prepayment are more sensitive to structural unemployment than cyclical unemployment.

image242

Conclusion

The issues of credit risk modeling and analyzing the key default determinants are now at the center of the mortgage literature. Traditional models to predict the PD suffer from a sample selection bias. For this reason, advanced econometric techniques are applied, e. g. BVP model with selection. Empirical mortgage literature supports the idea that modeling the credit underwriting process and the borrower’s default decision simultaneously, with control for a sample selection bias, provides consistent results.

The initial research question is to understand key drivers behind the borrowers’ decision to default in the mortgage market, based on commonly observed characteristics. For this reason, a structural model on mortgage lending could be developed. It would take into account not only the proba­bility of approval and the probability of default, but also the probability of selecting a particular credit organization and terms of mortgage. The second research question is to assess the existence and the impact of sample bias in an empirical setting such as the Russian mortgage market. Available data sets include recent observations, allowing us to focus on the drivers behind the recent wave of mortgage defaults. The level of detail in the data allows us to control for various loan terms and borrower risk factors, and thus to control for a more comprehensive list of potential drivers of default.

The structural credit risk model could be incorporated into the decision making process of credit experts regarding the mortgage market and to contribute to the development of an effective risk management system. As a result, it leads to effective allocation of capital and will be beneficial for all credit market participants.

 

Acknowledgements The author would like to thank Anil K. Bera, Andreas A. Woudenberg and Alexander Karminsky for their helpful comments. This study was carried out with support from “The National Research University Higher School of Economics’ Academic Fund Program in 2013-2014, Research Grant No. 12-01-0130.” This survey was presented at the Perm Winter School-2013. The author is responsible for any errors that remain.

 

References

 

Ashcraft, A., Goldsmith-Pinkham, P., Hull, P., & Vickery, J. (2011). Credit ratings and security prices in the subprime MBS market. The American Economic Review, 101(3), 115-119.

An, M. Y., & Qi, Z. (2012). Competing risks models using mortgage duration data under the proportional hazards assumption. The Journal of Real Estate Research, 34(1), 1-26.

Bajari, P., Chu, C. S., & Park, M. (2008). An empirical model of subprime mortgage default from 2000 to 2007. NBER, Working Paper 14625.

Campbell, J. Y., & Cocco, J. (2011). A model of mortgage default. NBE, Working Paper No. 17516. Creel, M. (2008). Some possible pitfalls of parametric inference. Quantile, 4, 1-6.

 

Leave a reply

You may use these HTML tags and attributes: <a href="" title=""> <abbr title=""> <acronym title=""> <b> <blockquote cite=""> <cite> <code> <del datetime=""> <em> <i> <q cite=""> <s> <strike> <strong>