When asked how to value financial institutions, I’ve always taken the cop out plea: They’re outside my sphere of competence. What makes that statement extremely interesting is that I own and operate one and I still can’t value them!

Robert posted one of the finest, most eloquent, and thoroughly researched answers to the question of why it is so difficult to value financial services companies.

Marc asked:

Hi All,
Can someone please advise of an appropriate method for calculating the IV for financial services companies? From what i have read, i believe that it needs to be approached slightly differently.
In particular we have some strong performing banks (both traditional and investment) in Australia and i would like do the numbers to see how they stack up.
Cheers
(sic)

I am going to reprint Robert’s response here, or you can view the comment string for yourself – starting here.

Robert’s Response

Dissatisfied with my earlier response to Marc concerning evaluation of financial firms, I performed some research. While dated, there was nothing of value in any of the finance, accounting, and investment texts from graduate business school – beyond noting the necessity of banks to secure a positive spread between investment returns and payouts to customers, the need for accurately predicting the economy’s movement and degree in order to choose between short-term and longer-term investments, and the necessity of factoring withdrawals.

Next, I consulted Philip Fisher’s Common Stocks and Uncommon Profits, since Buffett was a friend and admirer of Fisher – claiming the he was 25 percent Fisher and 75 percent Graham. There was nothing on point that I found in Fisher’s book.

As for Graham, Security Analysis (first published in 1934) offers support for my previous message. In my copy (ISBN: 0-07-024496-0, McGraw-Hill), Graham writes:

It may be logically contended that since dealing in securities is an integral part of the investment-trust business, the results from sales and even the changes in portfolio values should be regarded as ordinary rather than extraordinary elements in the year’s report. Certainly a study confined to the interest and dividend receipts less expenses would prove of negligible value. If any useful results can be expected from an analysis of investment-trust exhibits, such analysis must clearly be based on the three items: investment income, profits or losses on the sale of securities, and changes in market values. But the reader must bear in mind the basic fact that security analysis, however intelligently conducted, cannot be more dependable than the conditions with which it deals. No study of investment-trust reports during the pyrotechnic years of 1928 to 1933 could possibly shed light upon the future results to be expected from these enterprises. The value of such a study would have to be confined to certain comparisons between one company and another. Even here it would be difficult to distinguish confidently between superior management and luckier guesses on the market. (Page 358, emphasis in the original.)

Graham next extends this uncertainty to other financial firms, with the statement,

The same problem is involved in analyzing the results shown by insurance companies and by banks. (Ibid.)

On the next page, Graham extends the negative assessment to the issue of increased volatility attributable to this lack of clarity / granularity inherent in financial firms.

The fact that the operations of financial institutions generally – such as investment trusts, banks, and insurance companies – must necessarily reflect changes in security values, makes their shares a dangerous medium for widespread public dealings. Since in these enterprises an increase in security values may be held to be part of the year’s profits, there is an inevitable tendency to regard the gains made in good times as part of the “earning power,” and to value the shares accordingly. This results of course in an absurd overvaluation, to be followed by collapse and a correspondingly excessive depreciation. Such violent fluctuations are particularly harmful in the case of financial institutions because they may affect public confidence. It is true also that rampant speculation (called “investment”) in bank and insurance-company stocks leads to the ill-advised launching of new enterprises, to the unwise expansion of old ones, and to a general relaxation of established standards of conservatism and even to probity.” (Ibid, pg. 359.)

[Note: The relaxation of regulations concerning investments and investment reporting (under the current US administration) strikes me as interesting, in light of this last line. Of course, sub-prime mortgage loans was not a banking or Glass-Steigall problem but, rather, largely occurred outside of regulated banking.]

In Security Analysis Graham ultimately draws the following conclusion on the subject:

The securities analyst, in discharging his function of investment counselor, should do his best to discourage the purchase of stocks of banking and insurance institutions by the ordinary small investor.” (Ibid.)

Security Analysis was the first of two oft-quoted texts by Graham, and it was the first published. The second, The Intelligent Investor was first published in 1950 and revised four times – with the last revision in 1973. It was recently re-published by Collins (ISBN-10: 0-06-055566-1 & ISBN-13: 978-0-06-055533-5), with preface by Buffett and commentary by Jason Zweig. Where Security Analysis expresses evident bias informed by the then-recent market crash in the late 1920s, The Intelligent Investor is marginally more accommodating, due to the added safety of increased regulation, but he offers no specifics on how to value such institutions. On page 360, Graham writes:

A considerable variety of concerns may be ranged under the rubric of “financial companies.” These would include banks, insurance companies, savings and loan associations, credit and small-loan companies, mortgage companies, and “investment companies” (e.g., mutual funds). It is characteristic of all these enterprises that they have a relatively small part of their assets in inventories – but on the other hand most categories have short-term obligations well in excess of their stock capital. The question of financial soundness is, therefore, more relevant here than in the case of the typical manufacturing or commercial enterprise. This, in turn, has given rise to various forms of regulation and supervision, with the design and general result of assuring against unsound financial practices.

The next paragraph compares equity appreciation results for financial companies with other firms between 1940 and 1970, and he concludes that financials have achieved comparative results. This is the middle paragraph of just three devoted to this topic, and he concludes by writing:

We have no very helpful remarks to offer in this broad area of investment – other than to counsel that the same arithmetical standards for price in relation to earnings and book value be applied to the choice of companies in these groups as we have suggested for industrial and public-utility investments.

This last statement is important because, in his earlier writings, he notes the difficulties in relying on reported investment earnings by financial firms and, separately, he notes the difficulties associated with applying book value to investment holdings. Should you, for example, value them at your initial purchase price, the price in the market when publishing the annual statement, the aggregation of book values for each internal investment by the firm, or some other method (perhaps including depreciation, inflation, estimated breakup value, etc). Each has its problems, as Buffett has repeatedly noted in his letters to BH investors – recognizing that a significant component of this holding company is an investment house, and the largest categorical component is insurance and re-insurance.

Surely, however, there have been advances since Graham described these difficulties in the early 1970’s and when I graduated business school in the early 1990’s. Well, consulting firm McKinsey & Company and then-current and -former partners Tom Copeland, Tim Koller, and Jack Murrin authored Valuation: Measuring and Managing the Value of Companies, Third Edition (ISBN: 0-471-36190-9 cloth, 0-471-36191-7 paper, 0-471-39748-2 cloth with CD.) This excellent text devotes just two chapters to valuing financial firms (banks are covered in chapter 22 and insurance firms are addressed in chapter 23, with no chapter devoted to investment houses).

Copeland, et al, undertake bank valuation by converting the financials to free cash flow to bank shareholders with the following adjustments:

Income Statement:
Interest Income
+ Fee Income
- Interest Expense
- Provision for Credit Losses
+ Non-Interest Revenue
- Non-Interest Expense
+ FX Income
– Taxes
= Net Income
+ Extraordinary Items
+ Depreciation
= Cash from Operations

Balance Sheet Sources
Gross Loans Due
– Provisions and Unearned Income
= Net Loans Paid
+ Increase in Deposits
+ Increase in External Debt
+ Increase in Other Liabilities
+ Increase in Accounts Payable
= Sources

Balance Sheet Uses
New Loans
+ Increase in Securities Held
+ Increase in Accounts Receivable
+ Increase in Net Tangible Assets
+ Increase in Other Assets
- Decrease in Deposits
– Decrease in External Debt
= Uses

Free Cash Flow to Bank Shareholders = Cash From Operations + Sources – Uses

This level of granularity is often unavailable to non-insiders or through secondary sources such as Yahoo and Morningstar. Consequently, this free cash flow conversion is the initial step commend by the authors for internal analysis, and, to the extent available, for outsider analysis. It would, therefore, be necessary to do this based on the SEC filings for the past 5 years to determine trends and targeted portfolio levels by loan and investment class.

Insiders are advised to calculate and apply the spreads earned on balances, since income is derived from interest returns on investing-member funds minus the interest paid back. This, of course, varies depending on current interest rates (borrowing longer-term and paying shorter-term rates when favorable, for example). There is, however, a gains and losses mismatch due to the necessity of rolling over shorter-term instruments more frequently than their longer-term counterparts. Consequently, the bank is constantly wagering on the direction and degree of interest rate changes in the out years. Imagine, for example, their difficulty predicting the latest 75-basis point decrease in US rates as the Federal Reserve sought to combat the liquidity crisis – a reversal to expected increases designed to address projected growth in inflation. In any event, the insider must calculate and apply corrections to the mismatch in order to project cash flows in future years. The authors advise against using perpetuity calculations to by-pass this step. In their example (using prior data for Citi), a 23 percent ROE under a perpetuity model becomes a 0.5 ROE by year 3 if adjusting for the mismatch.

Here is how they summarize the “mismatch” challenges:

The key to handling the problem of mismatch gains or losses is to build a good forecast that takes into account:

1. The way spreads are forecasted to change with the changing interest rate environments.
2. The inflow of funds from loans being paid off and the outflow of funds at new rates as new loans are made.
3. The substitution between interest-bearing and non-interest-bearing deposits as interest-rate environments change.
4. The portion of mismatch profits that is sustainable because forward rates tend to be higher than their corresponding realized spot rates.

It is not easy to build all of these variables into a forecast. Even if you decide not to do so, it helps to understand the illusion of mismatch profits. (Page 434).

As for outsider valuations, the authors assert:

Determining the quality of loans is the most difficult problem for an outsider’s valuation, and little information is available to help solve it. Consider loans to emerging market countries or commercial real estate. Althought they are sometimes sold in secondary markets for 50 cents on the dollar, this kind of markdown must be viewed with healthy cynicism. The loans that banks keep are probably worth more than those they choose to sell in the secondary market.

The market value of the loan portfolio evolves with changes in interest rates and in the creditworthiness of debt in the bank’s loans portfolio. It is possible to find out what percentage of the portfolio is represented by emerging market, leveraged buyout, or commercial real estate lending. These can then be marked to market (at least approximately) as market conditions change. (Ibid.)

Next they turn to the question of estimating the difference between possible, potential, and likely dividends paid to the various classes of investors and finally recommend creation of 10-year pro forma’s that adjust each line-item based on interest rate and inflation expectations, leading to the discounted cash flow analysis. As with the spread analysis, it is necessary to perform this for each class of loan undertaken by the institution.

The authors go into other significant consideration, such as:

• The stability of loan values due to FDIC coverage for loan loss,
• The differences between wholesale and retail banking,
• Inter-bank and extra-bank loans,
• Capital structure,
• The cost of equity,
• Treasury borrowings,
• Risk management, and

What they describe as “the shared-cost problem,” which is the normal issue of how to allocate expenses by branch, service line, and headquarters (which is more of an internal issue, that is largely ignored when valuing non-financial firms).

I will not provide the same treatment for the valuing of insurance companies, but the authors indicate that this is similarly complex:

In Insurance companies, operations and financing are intertwined, as they are in banks. As a result, the equity, rather than enterprise, discounted cash flow approach must be employed to value insurance companies. In addition, insurance companies have unique operating characteristics that warrant further discussion. (Page 449.)

Investment houses, as previously mentioned, are not addressed in this text, but the same issues apply. Investment houses have similar liquidity flow and redemption concerns, spead issues, portfolio allocation challenges (especially for margin accounts and in cases where they make the market – some do, others don’t). Marc’s question, blessedly, was limited to banks, and I’ve marginally extended this to insurance since banks are a common source of home and car insurance.

My apologies for the long-winded response, but Marc asked an intelligent and difficult question – one that has been raised by others and addressed with marginally less detail by Joe. Given his greater experience, I hope he will correct any inadvertent errors I may have made – with the same request extended to the CPAs who frequent this blog. While I am a management professor, finance is not an area of professional focus. You are, therefore, invited to consider the source and apply the appropriate credibility discount to all that I’ve written.

Robert

My Response

Thanks Robert. That was awesome! In short, they’re still outside my sphere of competence!