Charting Banking XXIV: pre-tax pre-provision earnings
If tangible common equity multiples have their issues, what is a good valuation alternative?
If you trust the bank to survive its credit headwinds while at the same time its total revenues are growing, or at least stabilizing, I think it is fair to use some sort of multiple of normalized earnings. And if that is the case, pre-tax pre-provision earnings are a good start.
- Why pre-tax? Two reasons. First, with all the recent losses many banks are not going to be paying taxes for a couple of years. And second, it makes easier the comparisons among banks and across time when taxes paid are fluctuating.
- Why pre-provision? Provisions are the main way credit issues get reflected in the income statement, building up reserves for consequent charge-offs. Therefore, if we want to isolate the earnings power from the credit issues this is the main line to subtract. The other way credit issues impact earnings is loses related to sales of real estate owned REO.
And the idea is that after getting a standardized PTPP, you can use your personal estimate on steady state provisions and taxes to get to an estimate normalized earnings. We will get to that later.
There is one little problem though: that is all the consensus on PTPP there is. Each analyst and company do their own adjustments:
- One-offs like goodwill write-downs and gains or loses on securities sold
- CVAs and other mark to market distortions
- REO gains and loses
- Depreciation and amortization
- Credit related operating expenses like mortgage servicing
And there is no consensus even on how to name pre-tax pre-provision earnings:
- pre-tax pre-credit earnings
- pre-tax pre-provision income
- pre-provision operating income
- core earnings
And of course each one of those come with its own abbreviation.
That makes life hard because you are not sure what each company or analyst is including and how to compare across companies and analysts. My opinion is that there is no way around it: you have to get your own estimates … and I will share one fast, but not very precise, recipe:
- Get cash from operations from the cash flow statement, and reverse the following adjustments
- Subtract back all working capital adjustments that are usually the lines that start with “increase” or “decrease”
- Subtract back stock compensation
- Subtract back tax provisions from the cash flow statement
- Subtract the equipment purchases (capex) from cash from investing
- Add back taxes (taxes provisions) from the income statement
Or the same, start from net income in the cash flow statement and add depreciation, amortization, provisions, one-time loses/gains, taxes, and subtract equipment purchases. Each company has different line disclosures in the cash flow statement so you still have to season to taste. But at least is much more standardized than using analysts and company estimates.
At the end, you get something very similar to what Buffett refers as owner’s earnings only that it does not include taxes. Among its benefits, it excludes non-cash charges and is somewhat conservative because it does not adjust for high administrative expenses related to foreclosures and REO administration (ie: Bank of America has 30,000 FTEs entirely dedicated to solving mortgage issues).
These are the numbers I am using as estimates for the Big 4 and the big challenger. Use them only as a reference. As I said, there is no way around it: you have to get your own estimates.
Do not try to be overprecise, I usually round up numbers trying to sin on the conservative side. Better to be roughly right than precisely wrong.
One way to double check the estimates is to compare profitability ratios (do it with tangible figures if you prefer) across companies and across time to have a better idea of the earnings power of the bank. These are some high level numbers across companies,
Now how to get from PTPP to a normalized valuation? Well, the long way is to do a discounted cash flow with normalized provision, taxes and growth. I will recognize that I prefer simplicity, with some common sense, and generally use a 10x multiple of PTPP:
- Very simple to calculate.
- 10x pre-tax is close to 15x after tax.
- 15x historic earnings multiples imply growing at the same rate as rest of the economy and returning just cost of capital.
- Big banks cannot grow a bigger share of the economy forever so a conservative multiple looks good.
- Substantial non-earning assets.
- Some of them still have high cost financing (prefs, trups, tarp), despite high liquidity, that will be reduced over time.
- The banks that I am interested in, the ones under some distressed valuations, will not pay taxes for years.
- Non crisis provisions of around 0.3% of revenues are less than what many banks pay in mortgage mess related expenses.
There is at least one big exception to these oversimplified assumptions, banks heavy in credit cards or similar lines based on a model of high normalized provisions in line with high net interest margins. A clear example of this is Capital One. My recommendation, go to previous 10Ks and double check the normalized provisions.
Having estimated a normalized value, it is good practice to go through the credit/legal issues and estimate how many years would take to solve them, apply your preferred margin of safety discount based on the company’s specific risks and growth prospects, and voilà.
That was today’s look inside the sausage factory.
Long BAC, C
PS: criticism of these numbers based on real figures is very much welcomed!
PS 2: I adjusted USB PTPP to reflect its 2011 growth after comments at the Corner