How to predict if a borrower will pay you back

Without risk there is no reward.

All investors accept some degree of risk, that’s the cost of entering the game. Understanding this truth, however, is only the beginning. Investors must ask themselves how much risk they’re willing to incur in their reach for a larger return. Marketplace lending has made this question more complex.

Marketplace lending operates off a basic framework of joining borrowers with lenders. The cost of engaging in this form of finance is low due to the ease afforded by technological solutions. While this digital exchange enables speed it simultaneously obscures the understanding of what factors influence borrower behavior, namely repayments.

Lenders can benefit from greater insight into borrower behavior as a means of mitigating risk. If we can more accurately predict what causes borrowers to default we can make more strategic decisions when investing. The problem: P2P lending is a relatively new market and data has been scarce, until now.

Evaluating future behaviour of borrowers

The researchers behind the article Determinants of Default in P2P Lending exhaustively reviewed nearly 25,000 marketplace loans. Their goal was to form a more concrete understanding of what factors signal future borrower behavior. In short, they want to know what borrower characteristics are red flags for lenders.

One of their first findings confirms that the traditional risk/reward relationship is alive and well with marketplace lending. “The higher the interest rate, the higher the default probability is” determined the authors.

However, their deeper dive into the data also offered some unexpected revelations. “Loan purpose is also a factor explaining default: wedding is the less risky loan purpose and small business is the riskiest.” This discovery illustrates the complex nature of risk assessment; even non numerical data can illuminate default risk. Not surprisingly, factors like annual income, current housing, credit history and indebtedness are all powerful variables in predicting default risk.

Meanwhile, loan amount and length of employment had little or no bearing on a borrower’s ability to repay their obligation in their data set. Many investors would be surprised to learn that loan purpose has more predictive capability than employment data. For this reason the researchers warn that many lenders may give imbalanced or disproportional weight to a long list of factors. The result of this “asymmetry” is flawed risk analysis.

To avoid this problem investors can start judging default risk based on a few simple metrics. For example, one study in the Journal of Central Banking Theory and Practice determined that “Longer term loans are more risky than shorter term ones.” This makes intuitive sense given that over a longer time horizon there is a greater probability of life events that could arise and derail a borrower’s plans to repay. Specifically, 60-month loans are statistically more risky than 30-month loans. This same study had its own unique conclusions, finding that in terms of loan purpose, the least risky loans “are those for credit card payoff.” This outcome adds dimension to the above data.

Of course only so much information is available to lenders. To effectively gauge risk a lender must consider all available metrics while focusing more heavily on those that have been proven to predict outcomes.

Comments are closed.

1 thought on “How to predict if a borrower will pay you back”