As long time readers of this publication know, when it comes to lending, we believe that cash flow is king. If there is a 2nd place metric, it is geography. Outside of those two factors, the data shows everything else pales in comparison. As banks, we spend way too much time worrying about items like collateral value and borrower guarantees. While both are important, lenders need to consider the order of magnitude impacting a default. To underscore our point, we recently looked at a variety of 26k CRE loans across the nation and focused on performance from 2004 to 2010. What we found may help banks better underwrite and price their lending risk, placing more resources in understanding the quality of cash flow.
To place the importance of debt service coverage (DSCR) in perspective, it helps to compare the magnitude of change in the contribution to a default. For example, given the current market, the difference between a probability of default (PD) on a 70% LTV loan and an 80% LTV loan is about 30% greater. The LTV is more of a driver for loss given default (recovery of principal) than the PD; however there is an impact on PD, as a borrower will come out of pocket to protect their equity up to about an 86% LTV. After that, the borrower will repay a loan only as global cash flow and reputation allow. By comparison, the PD of a loan with a 95% versus an 85% LTV is only greater by about 20%. In similar fashion, the PD increases by less than 10% for a loan with a 65% LTV compared to a loan with a 55% LTV.
In comparison to LTV, while we talk in percentages about the impact of different LTVs on the PD, we talk in terms of multiples when we discuss cash flow coverage. For example, a loan with a DSCR of 1x cash flow is 412%, or more than 4x more likely to go into default than a loan with 1.2x coverage. A loan with a 0.8x DSCR is 7x more likely to go into default as one with a DSCR of 1.3. These are big differences and the takeaway here is when it comes to underwriting, understanding the sanctity of cash flows should garner more analysis than any other single item.
As a rule of thumb, the most important level of debt service coverage to achieve is 1.3x. Cash coverage lower than that dramatically increases the probability of a loan going into default, making loan performance much more sensitive. Above that, the decrease in the PD (the rate of change) tends to moderate. Once you get above 1.9x cash flow, the PD becomes much less sensitive. In fact, when it comes to sensitivity, outside of the 1.3x, the important levels to keep in mind are: 0.5x, 0.7x, 1.7x and 1.9x. It is around these points that cash flow has the largest implied influence on the PD, as the rate of change decreases materially after each point.
Please keep in mind that the above data and conclusions are national averages for all types of CRE and are based primarily on data from CMBS transactions so they may not be the same for your area or underwriting. Hospitality, for example, tends to be more sensitive than any other loan type to cash flow, while office loans, the least. TX and MD are a whole lot less sensitive than Southern CA to cash flow changes. To better understand how the interplay of DSC, LTV, borrower guaranties and geography down to the zip code level impact defaults (and pricing); be sure to subscribe to our relationship profitability model, BIGProfit. If you don't have a loan analytical tool that helps underwriting, just remember that the lower the DSCR below the magic 1.3x, the more your bank better be sure those cash flows are going to be stable and or growing in the future to avoid defaults.