Commercial Bank Valuation

Question 1

How might you conclude that a bank is overvalued or undervalued from an analysis of Public Comps?

Answer 1

You focus on the relationship between P / BV and ROE and P / TBV and ROTCE. If the bank you’re valuing has Returns-based metrics on par with the medians of the set, then its P / BV and P / TBV multiples should also be close to the medians.

The bank might be undervalued if its multiples are below the medians, but its Returns-based metrics
meet or exceed the medians; vice versa if the bank’s multiples are above the medians.

You might also look at metrics such as the CET 1 Ratios and Efficiency Ratios of the banks; if the bank
you’re valuing has worse performance metrics than its peers, it should be valued at lower multiples.

Question 2

How do you adjust a bank’s metrics and multiples if it has Excess or Deficit Capital?

Answer 2

If a firm has Excess Capital, e.g. its CET 1 Ratio is 15%, but peer companies are targeting only 12%, and the company itself is also targeting 12%, then you reduce its Equity Value, Book Value, and Tangible Book Value, and reduce its Net Income by the after-tax amount it earns on that Excess Capital.

If the bank has “Deficit Capital,” you do the opposite and increase all those metrics.

As a result, the P / TBV multiples tend to increase when a bank has Excess Capital, but the P / E multiples tend to decrease (the opposite applies for Deficit Capital).

Question 3

A bank’s Current Equity Value is $10,000, and its Tangible Book Value is $6,000. Its CET 1 is $5,500, its Net Income to Common is $600, and its Risk-Weighted Assets are $55,000. It is targeting a CET 1 Ratio of 12%.

If you assume a 2% rate of return on Excess / (Deficit) Capital and a 40% tax rate, what are this
bank’s P / TBV and P / E multiples after you adjust for Excess / (Deficit) Capital?

Answer 3

The bank’s CET 1 Ratio is currently $5,500 / $55,000 = 10%. Since it is targeting 12%, it has Deficit Capital of $55,000 * 2% = $1,100.

You add this Deficit Capital to its Equity Value and Tangible Book Value, so they become $11,100 and
$7,100, respectively.

You also increase the bank’s Net Income to Common by $1,100 * 2% * (1 – 40%), so it becomes $613.2.

So, the bank’s P / TBV multiple will be 1.56x, and its P / E will be 18.1x after these adjustments.

Question 4

How can management manipulate both Earnings-based multiples and Book Value-based multiples for commercial banks?

Answer 4

Management has significant discretion with both these multiples because the Provision for Credit Losses reduces Net Income, and the Allowance for Loan Losses reduces Book Value.

So, the management team could over-provision for Loan losses, which would reduce its Net Income and Book Value and increase the P / E and P / BV multiples as a result.

Question 5

What might be different about a set of Precedent Transactions for a bank vs. a set of Public Comps?

Answer 5

The main difference is that you’re less likely to adjust for Excess / (Deficit) Capital in the Precedent Transactions because the acquired companies may be quite different from each other. A regional, pure-play bank and a diversified, multi-national bank probably don’t target similar CET 1 Ratios.

Also, you may focus on historical metrics and multiples due to the lack of data for the projected metrics in Precedent Transactions.

If you do not have information on the ROA, ROE, and ROTCE for each acquired bank, you might focus on metrics such as the share-price premium or % Cash vs. Stock in each deal as well.

Question 6

Your co-worker claims that you can value a bank with a DCF, but only if you use a Levered DCF
so that you capture the bank’s Net Interest Income. Is this co-worker correct?

Answer 6

No! Although Levered FCF will capture the bank’s Net Interest Income, it still won’t be accurate because CapEx and Working Capital do not represent “reinvestment in the business” for a bank in the same way they do for normal companies.

So, if you wanted to use a Levered DCF, you would still have to modify it to capture the bank’s
reinvestment in some way (e.g., by capitalizing employee hiring and training costs).

Given the effort required, it’s easier to use the standard Dividend Discount Model.

Question 7

Walk me through a basic Dividend Discount Model (DDM) for a commercial bank.

Answer 7

First, assume an Asset Growth Rate, and make Risk-Weighted Assets a percentage of Total Assets.

Then, project Assets and Risk-Weighted Assets based on these figures. Assume a Return on Assets (ROA) for the bank, and use that to project Net Income.

Then, assume a Targeted CET 1 Ratio (e.g., 10% or 12%), and calculate the bank’s CET 1 for the year
based on that percentage times its Risk-Weighted Assets.

Back into the Dividends Issued by taking the bank’s CET 1, adding Goodwill and Other Intangibles, subtracting the beginning Shareholders’ Equity, and subtracting Net Income and anything else that might affect CET 1.

Then, discount the Dividends based on the Cost of Equity, and sum up all the discounted values.

Calculate the Terminal Value using the Multiples Method (typically based on P / TBV) or the Gordon Growth method, and discount it to its Present Value, also using the Cost of Equity.

Add the Present Value of the Dividends to the Present Value of the Terminal Value to determine the
bank’s Implied Equity Value. Compare this to the bank’s Current Equity Value.

Question 8

How is a Dividend Discount Model for a normal company different from a DDM for a bank?

Answer 8

For a normal company, you don’t need to “back into” Dividends based on targeted Regulatory Capital
ratios such as a 10% or 12% CET 1 Ratio.

Instead, you project Revenue down through Net Income, assume a simple Dividend Payout Ratio, discount and add up the Dividends, calculate Terminal Value based on P / E, discount the Terminal Value to Present Value, and add up everything to get the company’s Implied Equity Value.

Question 9

How do you calculate Cost of Equity for a bank, and what should you do if the normal method produces odd results?

Answer 9

You can still use the normal method: Cost of Equity = Risk-Free Rate + Equity Risk Premium * Levered Beta.

But banks tend to have very similar capital structure percentages, so you don’t need to un-lever the Betas from the comparable companies, find the median, and then re-lever it based on the company you’re valuing.

If you get odd results, you might need to expand the set of comparable banks or use Dividend Yield + Dividend Growth Rate to calculate the Cost of Equity.

Question 10

What are the flaws with using a DDM to value a bank?

Answer 10

Just like a normal DCF, it’s sensitive to far-in-the-future assumptions such as Terminal Value, and you might lack enough information to forecast Dividends properly.

Also, a DDM may not work well if the bank does not issue Dividends, and the results could be distorted if the bank issues/repurchases significant Stock or uses significant Stock-Based Compensation.

Question 11

What makes the biggest difference in a DDM: The Payout Ratio, the Discount Rate, the Net Income Growth Rate, or the Terminal Value?

Answer 11

As in a DCF, the Terminal Value tends to make the biggest difference because the PV of the Terminal Value often accounts for over 50% of the company’s Implied Equity Value.

After that, the Discount Rate also makes a big impact since it affects everything in the analysis. The Net Income Growth Rate makes a smaller impact, and the Payout Ratio makes an even smaller impact because a bank’s capital ratios constrain its Dividends.

Question 12

Could you use a Dividend Discount Model to value a bank that does not currently pay Dividends?

Answer 12

Yes, but you have to assume that it starts paying Dividends in the future, or the bank’s entire Implied
Equity Value will come from the PV of its Terminal Value, which makes the analysis silly.

If the bank is so new that it has no plans to start issuing Dividends anytime soon, a multiples-based valuation or Residual Income Model (see below) may be better.

Question 13

How do Stock Issuances/Repurchases and Stock-Based Compensation affect a DDM?

Answer 13

Ideally, you should exclude them from the DDM altogether so that only Net Income to Common and Dividends affect the bank’s Common Shareholders’ Equity.

If you do include them, you need to adjust the bank’s current share count based on their future impact.

For example, you might calculate the Present Value of the net amount the bank spends on these items, and then add up everything and divide by the bank’s current share price to estimate the number of new shares that will be created in the future.

Then, you could add those new shares to the bank’s current share count and use that new figure to
calculate the bank’s Implied Share Price.

Question 14

If you’re building a multi-stage Dividend Discount Model for a fast-growing bank that will eventually reach maturity, what changes in the analysis?

Answer 14

The ROA, ROE, Payout Ratio, and Targeted CET 1 Ratio will likely change in each period until the bank reaches maturity.

Also, the Cost of Equity might change in each year because a mature bank should have lower risk and lower potential returns than a younger, high-growth bank.

If that happens, then you need to calculate the Cumulative Discount Factor for each period and use
that to calculate the PV of Dividends and the PV of Terminal Value. You can’t just divide by ((1 + Cost of Equity) ^ Year #) anymore because the Cost of Equity changes in each period.

Question 15

How do you factor Excess / (Deficit) Capital into a DDM?

Answer 15

You shouldn’t need to factor it in if you’ve assumed that the bank issues Dividends such that it meets its Targeted CET 1 Ratio; any Excess Capital will be distributed in the form of Dividends.

And if the bank has Deficit Capital, it won’t be able to issue Dividends until it achieves its Targeted CET
1 Ratio.

If you have not set up the model like that, then you’ll have to add Excess Capital or subtract Deficit
Capital when you calculate the bank’s Implied Equity Value at the end.

Question 16

When is a Residual Income Model most useful, and how does it differ from a DDM?

Answer 16

The Residual Income Model is most useful for banks that are not currently issuing Dividends, but which have easy-to-establish figures for ROE and Cost of Equity.

It may also be useful if you believe that much of the bank’s Implied Equity Value comes from its current Balance Sheet, and you’re uncertain of the bank’s future Dividends.

The model setup is very similar to the one for a DDM, and you still “back into” Dividends based on the
bank’s Targeted CET 1 Ratio.

But instead of summing up the Present Values of the Dividends, you sum up the Present Values of the
Residual Income (also known as Excess Returns).

Residual Income equals ROE * Shareholders’ Equity – Cost of Equity * Shareholders’ Equity, and it
represents how much actual Net Income exceeds expected Net Income.

Then, you add the Present Values of these Residual Income figures to the current Book Value (or Common Book Value) of the bank to get its Implied Equity Value.

If the bank’s long-term ROE differs from its long-term Cost of Equity, you’ll also have to calculate the
Residual Income Terminal Value, discount that to its Present Value, and add it to everything above.

The intuition is that a bank’s Equity is worth about its current Book Value, but it might be worth more than that if its ROE exceeds its Cost of Equity in the future. If the opposite happens, the bank’s Equity might be worth less than its current Book Value.

Question 17

What are the advantages and disadvantages of a Residual Income Model?

Answer 17

The advantage is that the Residual Income Model is grounded in the bank’s current Balance Sheet, so
most of the bank’s Implied Equity Value comes from that rather than future Dividends.

The disadvantage is that it sometimes doesn’t tell you much beyond the obvious – that Book Value and Equity Value are closely related for banks.

Also, many bankers view this methodology as academic or overly theoretical, so it’s not common in
real life.

Question 18

Why does the Residual Income Model often assume no Terminal Value?

Answer 18

The Residual Income Model includes Terminal Value only if the bank’s Long-Term ROE differs from its Long-Term Cost of Equity.

You often assume that ROE differs from Cost of Equity in the explicit forecast period, but that the two converge in the Terminal Period, which means there is no Terminal Value.

If you did not make that assumption, you could calculate Terminal Value with:

Residual Income in Year 1 of Terminal Period / (Cost of Equity – Terminal Net Income Growth Rate)

Then, discount it to its Present Value with the Cost of Equity, and add it to the Sum of the PVs of Residual Income and the bank’s Book Value.

Question 19

How might you use a Regression Analysis to value a bank?

Answer 19

Since P / BV and ROE and P / TBV and ROTCE are related, you could gather data on all those metrics for a wide set of publicly traded banks in a certain region and then create a Regression that lets you
“predict” your bank’s multiples based on its Returns metrics.

For example, you might plot this data and get an equation such as:

P / TBV = 12.9 * ROTCE – 0.13

If your bank’s ROTCE is 12%, then its P / TBV should be 1.4x according to that formula. You might then
compare that Implied P / TBV Multiple to the bank’s Current Multiple to see if it’s valued appropriately. If there’s strong correlation in this data, you might take these results more seriously.
The Regression Analysis is useful as a “sanity check,” but not as useful as a strict valuation methodology.