4
$\begingroup$

I am working on performing factor analysis on a loan portfolio. This is my understanding so far, and I was hoping that some of the smart folks here might be able to chime and guide me through this process.

Modern Portfolio Theory - Assets should be assembled such that the expected return is maximized for a given level of risk. Also, the return should not be assessed by itself, but by how it contributes to a portfolio's overall risk and return.

CAPM - Helps in determining the theoretically appropriate required rate of return of an asset to ensure a diversified portfolio.

Multiple Factor Models - Asset Pricing Models that can be used to estimate the discount rate for the valuation of financial assets. (extension of CAPM)

Sharpe Ratio - Measures excess return per unit of deviation in an investment asset or a trading strategy.

What does my data look like: short term loan information and how they were paid back.

My question for the smart people here is, how do I proceed? It seems like I understand the concept of what factor analysis is, but have difficulty tying it to the data.

$\endgroup$
3
  • $\begingroup$ Would you like to know quantitative Investment process? $\endgroup$ Commented Aug 31, 2019 at 0:44
  • $\begingroup$ OP, what did you end up doing here? It's an interesting question $\endgroup$ Commented May 27, 2020 at 17:55
  • $\begingroup$ No experience with loan portfolio analysis, but intuitively I would enumerate common attributes (factors) for each loan in the portfolio, convert them to numeric scale, on one hand, and compare against loan performance, on the other hand. Then run a regression model, e.g. OLSMultipleLinearRegression. $\endgroup$ Commented Jan 22, 2021 at 20:10

1 Answer 1

1
$\begingroup$

For the loan portfolios, the objective usually is the analysis of loss rate - expected loss, quantile etc, so one is normally interested in estimating the distribution of loss rates. This is slightly different settings compared to the standard (return vs variance) approach used in investment management, probably because loss is the key driver of the performance of loan portfolios.

But the factor analysis does not go away! It is used to simplify the modelling of correlation between different names/obligors. The simplest is the one factor model, and the far more common one factor model is the Vasicek portfolio loss model, based on on large homogenous portfolio assumption. This model forms the basis of the Basel capital framework for loan portfolios. But outside the regulatory requirements, usually a multi-factor model is used - which will have different factors, say for different sectors.

There are various alternative frameworks, the growth of CDOs before the crisis produced a lot of research in this area, so you can find as sophisticated models as you like! But I think the Vasicek loss model for homogeneous portfolio would be a good starting point, and you can then research copula, and multi-factor models, etc.

$\endgroup$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.