We have all heard the adage - "Don't put all your eggs in one basket". We presume this means that if your eggs were in multiple baskets, not all would break if you stumbled and fell with your basket in hand. This of course assumes that you are not carrying all your baskets at the same time when you fall.
Many analysts and consultants advise banks to diversify their loan portfolio. Diversification has been touted to help banks decrease risk and at the same time increase overall return. The idea is that different asset classes (let's stick to loans for now) offer a different balance of risk and return. Diversifying a portfolio of loans in our example allows
banks to create a better risk/reward profile overall.
Diversification works primarily because of a principle known as "correlation" (where risk and reward of various loans move in different patterns and so are less correlated to one another). The higher the correlation between two loans, the more their returns tend to move in tandem. A well-diversified portfolio of loans is not perfectly positively correlated, reducing the bank's risk for a given level of expected return.
That is interesting, but will diversification of the loan portfolio
help a community bank decrease their risk? Further, can it increase return and what practical diversification tool can help? Since before the credit crisis, the issue of choosing between a lending focus (or specialization) and diversification has been the center of debate for many bank management teams.
The empirical evidence on the benefits of loan diversification vs. focus for banking is mixed. In the seminal study by Acharya, Hasan & Saunders in 2001, foreign banks' risk/return performance was measured over many years. The authors considered the benefits of diversification on three variables: 1) geographic, 2) industrial (for example lending to doctors versus manufacturing companies), and 3) sectoral (for example lending to consumers, commercial, municipalities, etc.). While the conclusions are complex (and depend on the riskiness of the bank, size and other factors), the results show there are significant benefits to bank performance from geographic diversification; sectoral diversification leads to ambiguous risk/return results (meaning it depends on the bank) and industrial diversification leads to negative risk/return results.
For community bankers, the implications of the study are tenuous. Most community banks do not have the ability to deploy significant geographic diversification, so that is out. Furthermore, because of small size and limited resources, community banks focus on a limited number of sectors. That leads to a problem, because as community banks try to diversify, they quickly run out of the ability to allocate resources to focus on key aspects of the underwriting business (defeating the benefits of diversification due to loss of focus and reduced expertise in new sectors).
While community banks have historically been specialized and focused lenders in their community (largely to small businesses), we believe there is benefit from diversification to the extent the following strategies exist: it allows the sharing of underwriting skills and competencies between business lines; the bank can leverage its brand from one line to another; diversified lines can share marketing skills and economies of scale at some level at least are realized.
Banks that add contiguous footprint can gain some level of geographic diversification and those that leverage knowledge between similar but uncorrelated industries can realize the benefit of industrial diversification.
Lastly, loan structure diversification can be simple. Consider that many banks will offer the same 10Y loan with a 5Y reset and a 25Y amortization to most of their customers regardless of industry. Doing so stacks refinance risk, interest rate risk and marketing risk into the same basket. To change this, offer various repricing, maturity and amortization options.
The right diversification strategy can help bankers enjoy any basket of eggs as an omelet vs. getting it on one's face.