How to tell if a bank will fail?

brownstomach's picture
Rank: Monkey | 52

Hey monekeys!

In 2008, we saw several major banks, including Lehman Bros., collapse. Over the last decade, it seems like Deutsche bank has also been on a steady decline.

Perhaps someone who worked in either of these, or some other collapsed/collapsing bank could comment on this: Do you think there are some common traits or other signs characteristic of banks that meet this fate? In a recession, what differentiates a bank that fails from a bank that survives--or is it just up to fate at that point?

Comments (20)

Mar 14, 2019

House Lannister will probably fail within the next two months

Mar 14, 2019

Lehman was different than DB.

Lehman failed as it was holding Financial products and derivatives where Lehman lost money to put it simply.
Where as on the other end, DB is failing due to low profit margins due to increasing competition and a not so favorable market with additionally the underperformance of its Investment Bank.

    • 1
Mar 14, 2019

Agree with Pan European Monkey. DB isn't failing - it's just doing very poorly compared to what it used to. You can't really draw parallels to 2008 due to the heavy regulation imposed by both American and European authorities the last years.

    • 2
Mar 14, 2019

When you figure out the formula for predicting bank failures, the regulators would like to have a word with you.

    • 10
    • 1
Mar 14, 2019

The difficulty in doing downside analysis for financial institutions is the extreme sensitivity of earnings to small changes in asset quality. Banks are generally levered at 10x (i.e. 10% capital ratio) which magnifies these changes.

My first gig out of school was working in safety and soundness examination. A hot topic is always the adequacy of the provision for loan and lease losses. It works in the same way AR loss allowance and associated bad debt expense works for receivables.

So, in answer to your question, a bank (more specifically, a deposit holding financial institution), will start to have losses driven by higher and higher provisions for loan and lease loss expense which deteriorates the banks capital, which in turn, increases leverage beyond the regulatory thresholds. This happens relatively fast and is very hard to predict.

The more vulnerable (and more profitable in good times) banks will have lower T1 capital ratios and higher L/D ratios which magnifies the effect of losses on bank capital.

    • 3
Learn More

7,548 questions across 469 investment banks. The WSO Investment Banking Interview Prep Course has everything you'll ever need to start your career on Wall Street. Technical, Behavioral and Networking Courses + 2 Bonus Modules. Learn more.

Mar 14, 2019


Mar 14, 2019

im no bank regulator but probably when the liabilities are greater than the assets

    • 1
Mar 15, 2019

not exactly as its more to do with the spread and quality of the assets. Also everyone is ignoring the political element some banks are to big to fail as we dont have a free market banking system where they police themselves.

    • 1
Mar 14, 2019

For the sake of the argument, the points in your first sentence are consistent with mine. (1) contracting spread -> decreased profitability, which would only cause a bank failure if it depletes capital: depleted capital = liabilities are greater than assets (or in practicality, regulatory capital is below requirement). (2) asset quality issues -> provisioning expense -> decreased value of loans on balance sheet -> depleted capital (liabilities greater than assets).

Ovechkin and AmoryBlaines comments below are probably the best serious response here. Only thing AmoryBlaine's comment is wrong about is that NCOs do not flow through the income statement, it's the provisions that do. Ovechkin's comment adds the second important factor for bank failure which is liquidity (which I could argue is when short term liabilities are greater than short term assets). Note to OP - if you are serious on this topic you need to read into CECL, which is changing the way provisioning is going to be handled starting in 2020.

As AmoryBlaine pointed out, I would look into the Texas ratio as the most simple way to screen for potential bank failures. This screen gets you some scary results in southern Europe.

I'm not sure what your "we dont have a free market banking system where they police themselves" statement means.

Mar 15, 2019

All banks fail due to loan losses - so you need to understand the credit risks of the major genres of portfolio lending they are doing. Major risks for certain but not many banks now include cre and auto loans. Bank of the ozarks for example is certain to fail.

    • 2
Mar 16, 2019

With that level of confidence, you better already have an open short position...

Mar 15, 2019

There are generally two things to watch out for with a bank: solvency (regulatory capital) and liquidity (which for most banks is a combination of deposits and capital markets funding). As stated above, banks are highly geared with asset-to-equity ratios in the 20x range. Small changes in the value of assets like loans or securities held for trading can wipe out a bank's equity. For a commercial bank with no IB, loans to customers will likely be the main asset on the balance sheet. A downturn in the economy and an uptick in defaults coupled with worse than expected asset recoveries can lead to a big hit to equity. For IB's with large market making books, a market correction of those assets can also wipe out equity (see UBS and the 20 bn hit they took on sub-prime mortgages which required a state bail-out).

With regards to liquidity, if a bank looks like it's going to collapse you will likely see depositors rush to withdraw their money which causes a run on the bank. A great example of this in Europe was the Spanish bank Popular last year. Overnight the bank had to be rescued by Santander. Eqquity holders and tier 1 and 2 bond holders were wiped out in that transaction. Another example is in Ireland where the banks became too reliant on bond financing over deposit taking. When the credit markets dried up the banks couldn't roll over their debt and were forced to go cap in hand to the central bank for a bailout.

Mar 15, 2019

The metrics everyone has cited above make sense.

To the extent a bank's metrics look fine today but you want to know who might eventually fuck shit up, here's a few additional things.

Leadership: this is hard to identify but the thing that would generally make a bank less likely to fail is conservative leadership that's proactive rather than reactive. In other words, instead of trying to have a great year by chasing down the next hot product/market/trade, a CEO that's focused instead on his own agenda of slowly building the bank's competitive advantages to be better positioned 10 years from today.

Only obvious example is Jamie Dimon. GS has also been good at laying out its own path, but got sucked into the subrime thing pretty heavily before being saved by its prop desk hedging out the risk.

Revenue mix: getting very general now, but S&T/prop revenue is a lot more volatile/risky than IB revenue which in turn is somewhat riskier than PWM revenue. So if you looked up the revenue mix of all the banks and ranked them using that, it could be a measure of failure risk at the extremes.

    • 1
Most Helpful
Mar 15, 2019

Used to dabble in this area, so I'll offer a few thoughts:

  • As others mentioned, it's almost always driven by asset quality and credit losses. This can happen abruptly and is usually not a slow bleed of sub-par performance like DB
  • Bad assets usually show up in NPAs (non-performing assets, which includes NPLs and foreclosed real estate) and NPLs (generally loans 90+ days past due)
  • Once a charge-off is actually taken, it is labeled a NCO (net charge-off) and flows through the income statement
  • Banks set aside a loan-loss reserve on the balance sheet for eventual credit losses
  • Banks usually fail when the realized losses (NCOs) outstrip the reserve and cause earnings to go sharply negative
  • Since banks are usually levered 10:1 assets to equity, losses quickly eat away at the equity capital stack
  • The four best ways I've seen to monitor failure risk are: (1) closely watching the pipeline of bad loans - they usually track different categories like sub-standard assets->doubtful assets-> loss assets; if a high percentage of bad loans are making it all the way through to loss assets and you have a lot of new problem assets still showing up in substandard, there is likely a major credit problem; (2) look for banks that are trading well below tangible book value; during good times banks trade mostly on P/E basis, but during bad times investors look at liquidation/tangible book value; a very low P/TB means investors are expecting meaningful write-offs; (3) look at the Texas Ratio - this was first created during the savings & loan crisis that took place in the 80/90s mostly in Texas. The ratio takes the value of the bank's non-performing assets divided by the sum of the tangible equity and loan loss reserves. A ratio of more than 100% means that if the bank had to write off it's bad assets, it would eat entirely through the equity funding and existing reserves and be insolvent. Banks can bounce back from being over 100%, but this is a very useful ratio during downturns; lastly, (4) look for enforcement actions from regulators like the OCC. Bad banks will get slapped with cease & desist orders and other enforcement actions by their regulators. Usually these are for poor controls that are often a precursor to bad underwriting and potential credit losses.

Hope this helps.

    • 7
Mar 14, 2019

This is a good comment but it's important to note that NCOs DO NOT flow through the income statement and DO NOT affect earnings (under U.S. bank accounting standards).

Loan loss provisions appear on the income statement, and are an estimate of the amount a bank estimates will not be repaid. These provisions build a contra-asset on the balance sheet called Allowance for Loan Losses (ALL). Charge offs occur when a bank determines with ~certainty the value it will not get receive on the loan. A banks reduces the ALL by the amount of net charge offs during an accounting period. Loan loss provisioning expense is what decreases earnings. The provisioning estimate calculation is to change in 2020 with the implementation of CECL.

I would note I do not know too much about international bank accounting, but European rules are slightly different.

I would revise your sentence "Banks usually fail when the realized losses (NCOs) outstrip the reserve and cause earnings to go sharply negative" to "Banks usually fail when they are forced to take large provisions due to an increase in their estimate of future charge-offs, which causes the value of the bank's assets to fall below the market value of the bank's liabilities (i.e. capital is depleted)."

I think your Texas Ratio comment and explanation are spot on, as are your points on leverage and NPAs.

Mar 15, 2019

This is a fair distinction, but worth another follow-up. You are correct that the NCO itself is not directly on the income statement, but it does directly impact the earnings stream (and equity base). In my experience NCOs are also more closely followed than the provision itself. Below from Investopedia does a fair job explaining the relationship between the two and the income statement.

"A lender will reduce the loan loss provision by the amount of net charge-off during an accounting period and then refill the provision. The loan loss provision appears on the income statement as an expense and therefore will lower operating profits."

Higher NCOs drive a higher provision, which directly decreases earnings. From my experience, when performing bank ratio analysis for FIG IB, NCOs divided by loans is the most common measure of "true" credit costs. In my experience, NCOs were commonly viewed as "realized" credit losses and the better metric to look at.

    • 2
Mar 15, 2019

Besides structural and market factors, breaking the law is often what kills banks. A bank could be financially sound, but a large-scale scandal where the firm as a whole is not only looking at SEC violations, but actual felony charges would be more than enough to bring it down.

Drexel Burnham went down because it plead guilty to felony charges to avoid a RICO indictment. After the DOJ stripped most of the firm down what was left was an unprofitable heap that no one would loan to, and thus it filed Chapter 11.

If it comes out that Deutsche's been doing some really illegal stuff, this is a possible outcome.

"Work ethic, work ethic" - Vince Vaughn
Mar 15, 2019

Why don't you look up a death spiral during a liquidity crunch?

When people yell 'fire' in a crowded theater you tend to get exaggerated results that don't conform to models well, because the assumption that actions are independent are invalidated in those types of scenarios.

Mar 15, 2019
    • 2