“It’s the deposits, stupid!” How to prevent your bank from starring in a sequel of 2008

In 2008 we saw many banks that heretofore appeared to be investor darlings, implode in a tsunami of defaults and skyrocketing LLPs that wiped out profitability.  What happened then?  Could it happen again now?

When analyzing bank financial statements or listening to quarterly calls, we often hear management discuss deposits in the context of “cost of funding”.  This treatment, similar to “cost of goods sold” on a manufacturing income statement, is not only short-sighted, it also leads to shareholder value destruction, and, in the – not rare – extreme, banks going to bank heaven.  How can deposits, and a bank’s lack of strength in them, lead to troubles on the credit side of the balance sheet?


First, some analytics:

There is a strong correlation between a bank’s shareholder value metrics, and its deposit portfolio strength.

We define deposit portfolio strength as characterized by the following two conditions:

  1. Lower loan-to-deposit (LTD) ratios

  2. Higher % of DDA balances in the deposit mix

The two conditions above are highly correlated with Total Shareholder Returns (TSR) and P/B ratio.  The table below shows the degree of correlation between these (using univariate regressions):


Analyses include U.S. publicly traded banks, above $5B in assets, whose primary income statement driving activity is mainstream core banking (e.g., no specialty, money center, etc.).


When analyzing these metrics jointly, correlation increases.  In a multivariate regression, LTD ratio together with %DDAs have an R2 of 49% with TSR, and 90% with P/B.  This means that banks with anemic deposit portfolios, i.e., those who have both a high LTD ratio and a low % of DDAs in their deposit mix, tend to have low TSR and even lower P/B.

These weaker deposit portfolios also tend to be the hardest hit when rates go up, since LTD also correlates with deposit repricing betas: R2 = 58%.

Furthermore, banks with high LTD ratios seem to also have a low % of DDAs and NOW balances: R2 = 81%.  Lower % of DDA balances typically reveals a lower number of core deposit relationships anchored by active transaction accounts.

Why is this happening?

A high R2 does not necessarily mean causality.  The challenge to “causality” is certainly valid in the case of deposits (i.e., interest expense) seemingly explaining half the variability of profitability.  How can a line item that is smaller in absolute terms relative to e.g., interest income, be that important?  Is there business logic behind this result?

Yes, there is!

Banks that have (a) high LTD ratios, (b) low % of DDAs in their deposit mix, and (c) high interest expense are, by definition, banks that are weak in their abilities to gather and keep low-cost, relationship-based, retail (i.e., consumer and small business) deposits.  We call these “Deposit-Weak Banks” or DWBs.

When rates rise, especially from low starting points, DWBs are highly exposed on the liability side of the balance sheet.  Their deposit customers, most of whom do not have the primary DDA account (or any DDA for that matter) with the bank, tend to be less loyal.  Bombarded with marketing messages from the competition offering attractive deposit rates, these customers flee.  To maintain a funding source for their loans, DWBs resort to increasing deposit rates on their MMDA and CD products.

What further accelerates these deposit rate hikes is that shareholders expect growth no matter what the rate environment.  At the same time, DWBs tend to be particularly strong on loan origination.  For most DWBs, loans are the growth engine while deposits lag behind (hence the high Loan-to-Deposit ratios).

Because of the above, this is the sequence of events that will, with near mathematical certainty, lead to very bad outcomes:

  1. Rates rise in the market.

  2. Because deposit relationships are not anchored around a primary DDA account, deposits flee in search of higher rates (which now become more prevalent and more heavily advertised).  Since a DWB’s retail network is weak in acquiring DDAs and core deposit relationships, it also tends to be weak in retaining them, thus being largely ineffective in any defensive plays to stave off the deposit flight.

  3. Because a DWB has a high Loan-to-Deposit ratio, it cannot afford to strategically simply let some of these non-loyal deposits run off.  Instead, the bank resorts to raising deposit interest rates on MMDA and CD products.  Branch incentives are raised in hopes they will be the silver the bullet for opening new core retail deposit accounts.  Higher goals accompany the new higher incentives.  Alas, a silver bullet they are not.  The only deposit “strategy” that seems to work for a DWB is constantly hiking interest rates (note: this explains the high deposit betas in our analysis above).

  4. Bank resorts to increased market funding, which is even more sensitive to rate increases.

  5. NIM shrinks.  Management explains this for a few quarters as “the repricing of our loans will take longer than the repricing of our deposits; this is normal”.  In other words, “time will heal”.  Unfortunately, in this case it will not.

  6. Management realizes that with the ever-increasing interest expense, the only way to deliver the NIMs that shareholders and analysts expect is to pursue higher rate loans (which naturally carry more credit risk).  The pursuit of these loans is not simply for incremental additions to the existing loan portfolio.  In some cases, they replace lower rate loans that are, in one way or another, taken off the balance sheet.  In other words, the entire loan portfolio is aggressively re-structured to deliver a higher blended interest rate.

  7. The loan portfolio now carries more credit risk.  Not only are these higher rate loans inherently riskier, the bank resorts to seek them outside its historical hunting grounds and beyond existing well-understood client relationships.  Out-of-footprint lending and new-to-bank clients increase.  Lending grows disproportionally fast to customer segments and in products that were historically smaller (and whose risk the bank does not understand as well).  The result?  A bank that is very exposed towards not well understood credit risk

  8. It is at the tail end of the growth cycle, that the Fed raises rates to try to tame the end-of-cycle growth burst.  This burst is fueled by misplaced consumer and business confidence after years of rising tides lifting all boats.  Given historical patterns, and starting the clock from when the Fed started raising rates, we should not be more than 2 years away from the next recession.  While it is a losing game to try to forecast the exact timing of the next downturn, the next recession is probably closer rather than farther from arriving on our doorsteps.   When it hits, the bank is sitting on a loan book that has a higher proportion of riskier loans which (and this is a critical part) has grown disproportionally in the bank’s balance sheet.  The high deposit rates have pushed the bank to lend outside of its comfort zone; outside of its area of expertise (i.e., new geographies, new customer segments, new-to-bank customers, new credit products, new collateral classes).

    (Optional; might not happen this cycle due to new regs).  In the last cycle, just before the credit winds changed direction, many banks were blindsided by the lack of credit losses and reduced Loan Loss provisioning.  In a few cases, banks even posted negative quarterly LLPs, thinking they were over-reserved.

  9. Recession arrives.  Defaults and losses in the “transformed” loan portfolio rise at alarming rates, Loan Loss Reserves are depleted, LLPs skyrocket.  Profitability falls below hurdle rates and short of shareholder expectations; quarterly losses may appear.

  10. Bank management gets re-shuffled, but failure is increasingly inevitable.

This notion of a failure as a result of the 10 steps above was proven beyond any doubt in the prior cycle.  By combining LTD and %DDA into a single metric that we termed Deposit Strength Index (DSI) we observed a very strong correlation (R2 = 98%) to the likelihood of a bank failing.


Includes all 600 US banks above $1B in assets existing as of Dec 2004. Failures were observed to Dec 2008.
Source: Delos Advisors analytics, FDIC data.


So why are deposits ignored despite their proven importance to bank sustainability through business cycles?

 Many reasons.  A common observation is the background of the management teams.  It is fair to argue that there is nothing more complicated in banking than credit.  After all, understanding credit risk and managing credit portfolios is the raison d’être of banks.  The brightest, most educated, minds in banking often rise through the ranks of lending.  Despite these impressive credentials, most lenders have an obvious blind spot: retail in general and retail deposits in particular.

Another reason, related to the above, is a bank’s underinvestment in profitability analytics at the customer household and product levels.  As an example, banks that use correctly calculated and applied Funds Transfer Price curves (note: indeterminate maturity deposits in particular should be given appropriate expected durations far beyond their contractual “on demand” nature) and Capital Charge calculations at the account level, have helped management teams, credit backgrounds or not, to shed light on the blind spot described above.

As a rule – appreciating that all rules have exceptions – banks whose management teams are heavily dominated by career lenders tend to under-appreciate and under-value the retail business and the value of core retail deposits.  Also as a rule, banks that do not have account level profitability analytics are also driving blind.

If my bank fits the above “at-risk” profile today, how can it escape this path?

We are still early enough in the cycle that there is time to prevent this unfortunate sequence of events.  Some options to consider:

  1. Prepare the market and the analysts on the notion that you may choose to slow growth or shrink profitability while you build your retail deposit engine.  Explain that you plan to build a Retail business that will not depend on high deposit rates but on another source of differentiation.  Present this as what it is: a balance sheet restructuring strategy on the liability side.

  2. Avoid sudden changes to your credit strategy in pursuit of higher interest income to protect you NIM.  It’s better for NIM to suffer in the short term than to add ticking credit time-bombs in your portfolio, ready to go off once the next recession hits.

  3. Focus on the Retail business.  Invest in it.  Don’t simply ask your front line what to deliver.  In addition to goals, develop a real strategy (we will discuss this in a future white paper) that guides the front line on how to achieve success.

  4. If you have a strong capital position, and if the right target is available, consider acquiring the missing Retail component (strategy, management team, retail processes, and all…) by seeking a bank with a strong retail team and powerful retail strategy.

History has a nasty habit of repeating itself.  Smart bankers, regardless of background, are the ones who study the past to learn from it and guide their banks to create shareholder value no matter which way the market winds blow.

While the proverbial high tide will rise all boats, it is the seasoned, prudent, and well-studied captain whose ship will safely navigate the harshest of storms.

We are entering storm season again.  It’s not a question of “if” but “when”…