Most investors are familiar with the valuation techniques used to value non-financial companies. A number of different models are used from relative valuation, to discounted cash flows, multiples comparison, dividend yield, and others. To a beginning investor these techniques seem foreign and complicated, but after some use they become accepted and familiar.
For a reason I don't understand investors are comfortable with industrials, but not financials. Often banks are lumped into the "too hard" pile. An investor might feel comfortable investing in an industrial who's product they can't explain or understand, yet will avoid investing in a bank whose product is used daily and whose business model is both simple and straight forward. If I were to summarize banking in a few sentence it would read as follows:
"A bank takes money from depositors and lends it to borrowers. The bank makes a spread between the rate borrowers pay the bank and what depositors are paid. Out of their spread they pay their operating expenses, taxes and are left with net income."
Investors are willing to put their money in IBM, an extremely complex business, but avoid banks, a very simple business. I would love to see a survey of IBM's investors answering the following question: "Why would a customer choose WebSphere over Weblogic?" I would put a question like this on the same level as "What's the difference between a McDonald's hamburger and a Burger King hamburger?" Imagine trying to summarize IBM, or Medtronic in a few sentences like above, I'm not sure it's possible.
I've made a case in the past that banks are a portion of the market that investors can't ignore. Of the 14,077 stocks in the US 5,628 are classified as financials. Of those classified as financials 1,200 are banks. To disregard financials completely is throwing away 1/3 of the stocks in the US.
Let me share some stats I shared with my newsletter readers regarding community banks. There are 6,739 banks in the US, of which 5,734 have less than $1b in assets and are profitable. There are only 275 traded banks with more than $1b in assets. This means of the 1,200 or so traded banks almost 1,000 are below $1b in assets, the magic threshold where most of Wall Street tunes out. I spoke with an analyst recently who said that anything below $10b in assets is considered too small to consider. His fund has limited themselves to looking at about 70 banks, I'm sure they're not alone in this view.
Maybe I've convinced you that it's worth looking at community banks as an investment. The next question is how do you value a community bank? I recently read a PDF by David Moore where he detailed his approach on valuing community banks. I agree with his methodology and have summarized it below. There isn't one way to value a community bank, but multiple ways. Each way could yield a different value, but ultimate all methods of valuation should somewhat agree on a potential value.
This is the most common valuation technique and the one most accepted by the market. The technique is simple, a bank is compared to a set of their peers across a number of financial metrics. If their peers are trading at 1.5x book and they're at 90% of book it's reasonable to assume they are undervalued. Moore contents that the market is driven by relative valuations in the short term. That is if you're looking at a 2-3 year window for investing in a bank relative valuations are paramount. The key to determining if a bank is relatively undervalued is getting the correct peer group. What constitutes a correct peer group could be up for endless debate. The FDIC sets arbitrary peer group classification by asset size.
In my view a good peer group is one that's composed of banks competing in a similar market that are similar sizes. The First Bank of Tennessee shouldn't be compared against Regions Financial, even if they have branches in the same area. Just because Wells Fargo and Bank of America have blanketed the country with branches doesn't mean they are automatically peers to every other bank in the US.
Moore posits that in the short term relative valuation rules the market. Over a 2-3 time period a bank should trade in line with their peers.
Discount Dividend Valuation
Much of the market disregards dividend discount models as too simplistic but Moore makes the case that over the long term the dividend discount model (DDM) value of a bank will approximate its long term shareholder return.
The dividend discount model is the present value of future dividends the bank is expected to pay over a specified period of time. If one can accurately predict earnings growth, as well as the future dividend payout ratio and discount rate this method of valuation can be accurate. In my view the problem with a DDM model is there are too many assumptions taken into account. A difference between a 9% and a 12% discount rate can result in a dramatically different terminal value. The same could be said about earnings growth or a bank's payout ratio.
Along with the DDM value another potential forecasted value is the deposit premium of the bank. This can be calculated by taking the spread between the bank's borrowing cost and their deposit cost discounted at the 10-year Treasury rate. This future value of the bank's deposit premium could be added to either book value, or checked against peers to evaluate whether the bank is trading at a premium to their deposits or a discount to their deposits.
A bank's take-out value is my preferred way to value a bank. Moore is somewhat dismissive of this model. He states that he's heard rumors of bank sales that have never come to fruition, and betting on a bank merger can be foolish. While I agree that it's foolish to speculate on potential mergers I disagree that this value has no merit. I believe that a bank's take-out value is ultimately what a well informed private buyer might pay for the bank. I also believe that stocks are mean reverting with the mean being the private market value of businesses. This means that the anchor of value for banks should float around the take-out values of similar banks.
To estimate the take-out valuation of a bank one needs to look at a bank and estimate the expenses that could be eliminated in an acquisition as well as the potential improvement to earnings for a potential acquirer. Many barely profitable one branch banks can suddenly become attractive acquisition candidates if executive pay were eliminated. Consider a bank barely scraping by with earnings of $30k. If acquired the acquiring bank might eliminate the CEO/CFO/COO and other redundant personal resulting in a $500k or more savings a year. Earnings could jump 10x in an acquisition by retiring the executive team alone. Other potential cost savings could be realized by economies of scale too.
The last valuation model discussed in the paper is one I'm very familiar with, but also one that's extremely rare in the banking world, liquidation value. Banking is a regulated industry and if a bank finds themselves in trouble the FDIC will step in and force an orderly liquidation or acquisition. It's very rare that a bank enters liquidation proceedings willingly and without FDIC meddling.
The liquidation value for a bank is calculated in a similar way to how one might calculate it for a non-financial company. A bank's liabilities are viewed as being worth 100% of their stated value whereas their assets are discounted based on an investor estimate of quality. Cash and certain securities are given full value, loans are discounted based on the type and quality. Other assets are discounted as well.
A bank's liquidation value should be viewed as the lower bounds of potential value unless the bank is living under the shadow of a potential FDIC takeover.
There is no single correct way to value any company. A company's valuation can change depending on market circumstances, management circumstances, or owner circumstances. A company that is attractive when long term rates are 4% might not be attractive when rates are at 8%.
A bank investor should use as many valuation tools as they see reasonable to determine a range of values for a community bank. If a bank's price is at enough of a discount to the range of values it should be considered for a portfolio. The idea of using multiple valuation techniques is because each model uses different assumptions, and the combination of varying assumptions should flush out any faulty assumptions an investor might be making about a company.