Like Degas obsessed with dancers, bankers have been brought up to obsess over the composition of their loan portfolio. Egged on by our examiners, auditors and random pundits, we slice and dice our loan portfolios as if it means something and pat ourselves on the back when we can show a nice pie chart with lots of even looking slices. The reality is many of our sectors move together and offer little in the way of diversification. When the economy turns, tenants aren’t looking for new office space any more than they are looking for new industrial space.
Take a typical bank commercial real estate portfolio. The below matrix is composed of 10Ys worth of loan valuation data and shows that most bank-related real estate sectors move in lockstep with each other. Banks like to add office loans because they think the move offers diversification to retail loans. The reality is that office loans and retail loans are 95% correlated. Bankers are careful to add multifamily exposure to their commercial real estate holdings, but the reality is that loans on apartment buildings are about 88% correlated to the rest of a bank’s commercial loan portfolio.
Prior to the 1980’s, it was a standard concept to hold 25% of all your bank assets in Treasuries. While the yield was relatively pitiful, when the economy went into a recession, falling rates caused these quality fixed rate assets to gain in value offsetting the drop in value of the floating rate loans that suffered in credit quality. These days, banks have moved into securities with greater risk often taking on credit, optionality or liquidity risk in exchange for yield. While this helps earnings, it is done at the altar of diversification as investments with those additional risks are more correlated to a bank’s loan portfolio. While there is nothing wrong with any of these asset classes, our point here today is the need to design your balance sheet instead of letting your balance sheet design the bank. As a rule of thumb, the more correlative risk you have in your loan portfolio, the larger your low-credit risk, low-optionality and high-liquidity investment portfolio your bank should have.
For commercial real estate, one trick that some banks do is to divide their portfolio into metro and sub-markets. Instead of correlation in the 90%-range, you end up getting correlations in the 70%-range depending on the market. Presently, this concept is in full effect as real estate in major markets is about 12% above their pre-recession highs, while many of the suburbs are still 10% below their peaks. Suburban markets often lag their metro cousins between two and seven years. Now, for example, the difference in capitalization rates is near a record-wide 1.89%, which means the sub-markets are set for a rebound over the next three years, while most metro markets are projected to remain stable. Thus helps reduce the correlation. Managing your portfolio by geography/market is often more advantageous to banks than managing by sector. In other words, by changing pricing, sales/marketing resources and underwriting standards based on geography instead of sector, banks can achieve a better risk/reward profile than banks that just manage by sector.
C&I is another area that merits a review. If your C&I portfolio is full of businesses like insurance companies, title companies, real estate developers, entertainment, transportation and companies that are heavily dependent on commodities (steel, rubber, etc.), then you are not as diversified as you might think. However, if you can track your C&I portfolio by NAICS code or industrial sector, then you may be on to something. Businesses like consumer products, technology, chemicals and healthcare have relatively low and sometimes negative correlations to real estate. These businesses may need more aggressive pricing because they offer better value to the bank to offset some of the correlations found in the rest of the balance sheet.
In a similar vein, paying attention to owner-occupied buildings may also move banks to a more optimized risk/reward trade-off. While banks often break out their investment property loans with their owner occupied property loans, correlations of these two segments are 99% according to FDIC data. At 99% why the different treatment? Why even go through the risk management exercise of trying to break the categories out?
Of course, it doesn’t have to be like that. There is only a 99% correlation because banks largely price different owners the same. Smart banks will treat their owner-occupied real estate loans closer to a C&I transaction with collateral than a real estate loan. You may have 75% LTV and 1.25x debt service coverage on both loans A and B (with identical terms and performance), but if loan A is to an independent pet store owner (increasing probabilities of default) and loan B is to a chemical company (decreasing probabilities of default), Loan A has more than 5 times the risk. By pricing these loans differently and managing resources to construct your owner occupied real estate by design, you get a decidedly different outcome. Instead of a 99% correlation, you get an 88% correlation resulting in greater diversification.
There are hundreds of ways to achieve loan portfolio diversification and we have just presented some today. The future of banking is more granular risk management, dynamic pricing and targeted resource allocation. By being more proactive in your balance sheet construction, banks can achieve greater profitability and more stable earnings by putting loans into categories that matter.
Submitted by Chris Nichols on December 02, 2014