Why Data Matters In Bank Lending - A Live Example

Banking and Data

You can ignore data when underwriting, but that would be a mistake as sometimes the differences are stark. Our best example comes from fresh data from the office supply sector. Normally, industry probabilities of defaults (“POD”) move by about 7% per year. In 2014 certain industries, like banks and retail office supply stores, moved with large rates of change and even multiple rates of change. While banks are moving in a positive direction and risk is being reduced, retail office supplies are moving in the opposite direction. By not taking into account the latest default data, you could be lending into a loss situation.


Cash flow from privately held office supply and stationary stores decreased from 2012 to 2013, according to data from Sageworks. In fact, of all retailers, office supply businesses already have razor thin and volatile margins compared to other retailers. Last year, higher wages, health care costs, cost of goods sold and lower sales caused cash flow (as calculated by earnings before taxes, depreciation and amortization (“EBITDA”)) go to near zero. Next year, our projections will be that not only will this trend will continue, but the average office supply business will lose money and PODs will rise dramatically.


Given that more and more offices are going paperless, not only are fewer reams of paper needed, but fewer notebooks, pens, staplers, in baskets, clips, shredders and hundreds of other products are also required. In addition, software and game sales, something that used to make up 10% of revenue at many of these stores, is now almost all downloadable and purchased online. Add to these trends competition from online retailers such as Amazon, Office Max, Staples, and the result is that smaller retailers are left to struggle.


If a bank were to make a $500k 5Y loan to a retail office supply business and not look at the data, they might price that loan around Prime + 1% and take a 1% loan loss reserve. If they ran the loan through a pricing model using their risk rating of “3” not only would they pat themselves on the back for printing a nice 3.75% margin, but they would be comforted in their expected 18.4% risk-adjusted return.


However, if the credit performs like the average of the industry, the above margins and expected return would be a fantasy in the worst order. If you adjust the model using 2013 data as a proxy, then your return drops to 0.4% as your expected loss almost triples. If you look at forward looking probability of default, well then the data predicts that your bank will most likely take a principal loss in the next 5 years and your return will likely be a negative 35.5%. That is a pretty big difference.


Data is going to play a larger and larger part in bank lending and with competitive margins, success is going to go to the bank that is more accurate in their predictions and underwritings. After all, banking is an empirical business.


Using Data In Banking: Office Supply Example 

Add new comment

Filtered HTML

  • Web page addresses and e-mail addresses turn into links automatically.
  • Allowed HTML tags: <a> <em> <strong> <cite> <blockquote> <code> <ul> <ol> <li> <dl> <dt> <dd>
  • Lines and paragraphs break automatically.

Plain text

  • No HTML tags allowed.
  • Web page addresses and e-mail addresses turn into links automatically.
  • Lines and paragraphs break automatically.