Using Your Credit Data As An Early Warning Signal Against Defaults

Predictive Data Analytics

The volatility that is kicking up in the market is causing a resurgence for banks to develop early warning signals on credit. Banks sit on a treasure trove of data that is only partially utilized. While many banks monitor loan reviews and borrower financial information, there is a rich amount of data available in which to draw actionable credit intelligence. In fact, we have identified over 35 factors that play a material role in monitoring credit. These factors have been found to be statistically relevant enough to predict future defaults without giving an abundance of false alerts on performing loans.


Three Easy Metrics To Monitor


Some of these factors, such as debt service coverage, are well known. However many secondary credit factors, such as credit spreads, area effective rents, and staff reductions are rarely utilized. Of these, three of the most effective for commercial and industrial loans, all involve monitoring credit and are all data that the bank controls. These three credit factors include monitoring line of credit usage, installment/line loan repayment performance and checking account overdrafts.


This effort doesn’t take a data scientist, just the will and small amount of extra resources to add key items to your current credit risk monitoring scorecard. For example, monitoring past due loan repayments is an obvious marker that many banks currently utilize. However, this only gives about 60 days of warning. By adding an alert any time a non-historically utilized line of credit is drawn on, banks can gain at least an extra seven months of warning on average for those companies that also have an outstanding term loan.


Similarly, we have found that a non-seasonal 40% increase for three months in general working capital (non-acquisition or expansion) usage of the line that is not correlated to an increase in revenue is also a precursor of loan problems for many borrowers. This is to say, if a line of credit is unexpectedly drawn, either two things are happening – either business is booming, or business is tanking. Either way, a conversation is needed.



Other Common and Overlooked Early Warning Signs


When lines of credit draws do not get repaid as they have done in the past, this is also a precursor to possible future credit problems. Another accurate credit measure is the monitoring of extension of overdraft credit. The nice part about this factor is that it is statistically significant for almost all types of credit, including commercial real estate operating accounts, commercial loans, agriculture, residential mortgages and other consumer lines of credit. Here, for example, usually a 30% increase in checking account overdrafts for a quarterly period can also foretell future credit problems.


Four Credit Metrics That Predict Default



Past Due Loans: Many banks monitor the frequency that a loan goes past due and the average tenor of the past due period. Good thing as 60% of non-performing loans have a history of frequent periods where the loan is past due. Comparatively, only 5% of performing loans exhibit frequent past due payment problems. As a result, the frequency and tenor of past due periods are accurate predictors of future defaults. Average time between last credit trigger and default: Seven months.


Overdue Installment Credit: Less frequently used, but still common is the monitoring of overdue payments on installment or lines of credit. 32% of troubled lines of credit showed a history of delayed repayment of scheduled outstanding balance reductions. This compares to only 6% of performing loans that had a history of delayed repayment. Average time between last credit trigger and default: Five months.


Credit Line Utilization and Checking Overdrafts: Perhaps the two most overlooked early warning statistics are non-seasonal/non-revenue correlated credit line utilization and overdrafts in checking accounts. These two occur 28% and 27%, respectively in problem loans, but only about 2.5% of the time in performing loans. Average time between last credit trigger and default: Four months and two month, respectively.


Making A Difference


Setting up a monitoring system to alert you can go far in providing banks an early warning sign as to credit problems. Increasing monitoring effectiveness can reduce the probability of default in the range of 10% to 60%, reduce the loss given default around 20% and reduce the exposure at default by about 40%.


Having an effective early warning system allows bankers to negotiate for more collateral, restructure existing terms to prevent future problems, reduce future exposures and get the borrower some additional expertise on their potential problems.


Looking at two bank examples as highlighted by the red and green lines below, there is about a 40% point difference in net exposure at default (on the x-axis) between a bank that leverages early warning signals and ones that are reactive to credit problems. 


That difference in performance could be saving $3mm or more in losses for an average $500 sized bank. Another way to look at this is having an early warning system could either reduce loan loss provisions by an average of 15% or lower risk-weighted capital levels by 10%.


Net Exposure At Default


Putting This Into Action


For those customers that have a term loan and a line of credit or checking account, try making sure you can tie the information together to give you the best chance of identifying problems early. Doing so will give you a big jump on other banks that have extended credit but do not have the effective credit monitoring system that you do.