A number of developments in the past few years have dramatically changed the framework for evaluating capital structure alternatives for U.S. insured depository institutions of all sizes. First, the implementation of Financial Accounting Standards Board Statement No. 141, effective in 2001, precluded the use of pooling of interest accounting, thereby creating goodwill in premium merger transactions. Second, the adoption of FASB Interpretation No. 46 in 2003 disallowed General Accepted Accounting Principles recognition of minority interest treatment for trust preferred securities in favor of recognition as long-term debt. Third, the Federal Reserve's final capital rule in April confirmed the Tier 1 capital treatment of trust preferred, but mandated the deduction of goodwill from core capital elements beginning in 2009. Finally, and most recently, Basel II will permit large, internationally active banks to allocate capital based on a sophisticated assessment of their credit, operating and interest-rate risk rather than assets outstanding.
Among these developments, the proposed implementation of Basel II may have the greatest impact as it will allow large, internationally active banks to significantly reduce the risk-weighting of their assets and, correspondingly, the capital they must hold to support such assets. Basel II could prompt large banks to gain a significant pricing advantage over small and mid-sized banks since they will be able to hold materially less Tier 1 capital, relative to risk-weighted assets, than their smaller competitors. And while roughly 15 to 20 U.S. bank holding companies are expected to adopt Basel II, the vast majority of U.S. financial institutions will not, and will instead continue to adhere to Basel I capital and risk-management guidelines.
U.S. regulatory agencies are very concerned about the potential competitive implications of this dichotomy. Their concern was highlighted when the Federal Reserve Board, Office of Comptroller of the Currency, Federal Deposit Insurance Corporation and Office of Thrift Supervision announced in a joint press release in late April that they were slowing the U.S. adoption of Basel II. They took this position because the preliminary analysis from participating institutions indicated evidence of "material reductions in the aggregate minimum required equity and significant dispersion of results across institutions and portfolio types," according to a Fed statement. Representatives from America's Community Bankers and the Independent Community Bankers of America reiterated this concern in their testimony before Congress on May 11, 2005.
Amidst all of this activity, executives and boards of directors at non-Basel II-adopting banks should focus on designing an optimal capital structure that provides a cost-effective way to structure Tier 1 capital to lower the weighted average cost of such capital, and help offset some of the competitive advantage Basel II-adopting banks will have. To do this effectively, non-Basel II-adopting banks need a new framework for developing an optimal capital structure, as traditional corporate finance theories, including the Modigliani Miller theorem, are not particularly helpful for evaluating capital structures for insured depository institutions.
This new framework addresses the four central questions surrounding optimal capital structure for bank holding firms: What form of capital should be issued? What is the optimal amount of each type of capital to be issued? How should the capital be issued? And how can non-Basel II banks use capital structure to level the competitive playing field with Basel II adopters?
The type of capital to be issued is very much a function of matching capital needs with the type of capital that most efficiently and cost effectively meets those needs. In addition to meeting key objectives by matching needs with types of capital, another key determinant is the after-tax coupon cost of the capital. This is reasonably straightforward for capital instruments that have a stated coupon payment, since this payment is a function of the size and financial strength of the paying institution. For ease of reference, it is assumed that all issuance would be done by non-rated issuers in a pooled transaction. This assumption is reasonable, given that about 90 percent of the more than 8,000 banks in the U.S. have less than $1 billion in assets. Moreover, only about 1.7 percent, or 133 banks, have an investment-grade long-term senior debt rating of at least Baa3 or better from Moody's Investors Service. Based on the market conditions in April, the approximate pre-tax coupons for trust preferred securities and non-cumulative perpetual preferred were three-month LIBOR plus 2.00 percent and 3.60 percent, respectively. Assuming a 40 percent tax rate, the after-tax coupon cost would be about 3.12 percent and 7 percent, respectively.
The determination of the after-tax cost of equity, however, is a bit more challenging. While there are many ways to estimate this figure, three of the most frequently used are the capital asset pricing model, normalized return on equity, and dividend growth methods. In fact, the Federal Reserve uses versions of these three methods to determine the cost of equity when estimating reimbursement of cost of capital amounts for services provided to the Fed.
The most traditional, and widely accepted, academic approach of the three is the CAPM, which states that the cost of equity is equal to the risk-free rate plus a market-risk premium adjusted for the volatility of a particular company relative to the market. Typically, issuers have used the 10-year U.S. Treasury rate as a proxy for the risk-free rate, 600 basis points as a proxy for the market risk premium and the beta of the issuer's common stock as a proxy for the volatility of the particular company relative to the market. However, the majority of depository financial institutions in the U.S. are small- and mid-sized banks that may have limited trading activity in their common stock. As such, the beta of the common stock may not be a particularly accurate gauge of the volatility of the common stock of these smaller institutions. Thus, while this method has some limitations, it is still one of the most widely used models for calculating the cost of capital in the financial industry. As of April, the CAPM cost of equity was about 10.26 percent.
The normalized return on equity method, otherwise known as the comparable-accounting earnings model, is a more practical way to evaluate the implied cost of equity. This approach simply uses the bank's targeted level of ROE as a proxy for the bank's after-tax cost of equity, with the reasoning that every dollar of equity invested has to yield that threshold amount, lest the targeted ROE decline. As of April, and based on the median return on average equity for more than 800 banks with assets of between $100 million and $10 billion, the cost of equity using this method approximated 11.25 percent.
The dividend growth method, otherwise known as the Gordon growth model, is yet another way of estimating the implied cost of equity based on a future series of dividends that grow at a constant rate. This method uses the current dividend yield on the common stock and the expected growth rate in earnings per share. For those banks not paying a dividend, this approach is obviously less helpful in determining an implied cost of equity. As of April, and based on the median dividend yield and growth rate for over 800 banks with assets of between $100 million and $10 billion, the cost of equity using this method approximated 13.12 percent.
Based on the three methods described above, the after-tax cost of equity ranges from 10.26 percent to more than 13 percent. In this analysis, an assumption of 12 percent was used as a proxy for the after-tax cost of equity for depository financial institutions.
The two most important steps in answering the question of the optimal amount of each type of capital to be issued are to examine the key components of current regulatory capital guidelines and then minimize the weighted average cost of Tier 1 capital. Since trust preferred securities represent the lowest after-tax cost of Tier 1 capital, the optimal amount of such capital would seem to be the maximum amount that an issuer would be permitted to count as Tier 1 capital. However, the analysis is more complicated because for Tier 1 capital calculation purposes, the Fed's final capital rule of April noted that bank holding companies were limited to a combination of trust preferred securities and non-cumulative perpetual preferred not exceeding voting common equity.
Conservatively, the combination of trust preferred and non-cumulative perpetual preferred may not exceed 40 percent of Tier 1 capital. Therefore, the determination of the optimal amount requires an iterative model that accounts for trust preferred being limited to 25 percent of core capital elements (less goodwill net of deferred tax liability, beginning in April 2009); non-cumulative perpetual preferred increasing core capital elements and, as a result, the permitted amount of trust preferred; and the 40 percent limitation mentioned above.
Since trust preferred represents the lowest after-tax cost form of Tier 1 capital and the permitted issuance amount is limited to 25 percent of core capital elements, the optimal amount to issue may be simplified as (1/3) x core capital elements. Similarly, since non-cumulative perpetual preferred represents the next lowest cost form of Tier 1 capital, the calculation for the optimal amount may be simplified as non-cumulative perpetual preferred = (1/4) x (common equity + goodwill - deferred tax liability). By combining this framework for amounts of each type of capital with a calculation of the after-tax coupon cost and after-tax cost of equity, small- and mid-sized banks can develop an optimal capital mix that results in a lower weighted average cost of Tier 1 capital. By lowering the weighted average while complying with the Fed's capital rules to remain well-capitalized, a banking company can increase its competitiveness and the value of cash flow available to shareholders.
In issuing subordinate debt, trust preferred or non-cumulative perpetual preferred stock, there are three primary forms of market access for financial institutions to raise capital: public market, pooled market and private market.
Public-market access is generally only available to issuers with asset sizes greater than $5 billion, securities ratings of at least investment grade BBB- and at least five years or more of satisfactory operating experience. Consequently, the pooled market has rapidly emerged as the preferred option for middle-market banks-those with between $100 million and $5 billion in assets. Indeed, since 2000 there have been more than 47 pooled transactions aggregating $21 billion in trust preferred and subordinated debt completed for middle-market financial institutions, according to our research. This market is generally available to issuers with investment-grade financial strength ratings and at least four to five years of satisfactory operating history. For those issuers that require a rapid turnaround, have less than $100 million in assets, or insufficient operating history or financial results, the private market can be an attractive alternative.
For the issuance of common stock, the primary determinant of how the capital should be raised is the offering size. Transactions in which the offering sums are less than $20 million are usually transacted privately. A pooled market for common equity issuance hasn't yet developed, due to the lack of homogeneity in valuing equity securities. However, a pooled market has recently developed for non-cumulative perpetual preferred securities of middle-market banks, as evidenced by the launch by Nuveen Asset Management and Spectrum Asset Management of the first ever closed-end funds focused on purchasing these types of securities.
Large, internationally active bank holding companies that adopt Basel II, and lower their risk-weighting of assets accordingly, can potentially gain a pricing advantage over non-adopting banks. For example, if such a bank can lower its risk-weighting of assets from 75 percent to 50 percent, and, assuming a cost of Tier 1 capital of 12 percent (100 percent equity funded), also lower the required return from 90 BP to 60 BP, it would gain 30 BP in after-tax pricing advantage, or 50 BP, in pre-tax pricing advantage (assuming a 40 percent tax rate) over a non-adopting bank. As every community banker knows, strong customer service may allow his or her bank to charge some premium, but 50 BP, tests that limit.
So, the key question is: How can the non-Basel II adopters use nimble capital management to help offset some of this pricing disadvantage? The answer boils down to reducing the weighted average cost of Tier 1 capital.
For example, if a community bank lowered its weighted average cost of Tier 1 capital from 12 percent to 9 percent while keeping the risk-weighting of its assets constant at 75 percent, then the required return would drop from 90 BP to 67.5 BP. The Basel II-adopting bank would still have a pricing advantage of 7.5 BP over the non-adopting bank, but the price difference would be dramatically reduced. As a result, the bank would be better able to compete based on its higher service level and other competitive advantages over a larger bank.
Also, keep in mind that starting in April 2009, large, internationally active banks will be limited to 15 percent of restricted core capital elements, such as trust preferred securities, to the sum of core capital elements, including restricted core capital elements (net of goodwill). All other bank holding companies will be permitted to have 25 percent of restricted core cap- ital elements. Since coupon payments on trust preferred securities are tax-deductible, such capital generally represents the lowest after-tax cost of Tier 1 capital. Therefore, the banks not subject to this 15 percent limitation will have a potential advantage over large, internationally active banks.
With the adoption of Basel II on the horizon, the time for small to mid-sized bank executives and boards of directors to develop optimal capital structures is now. Clearly, the ACB and the ICBA are very concerned about the competitive landscape, given their testimony on the potential anti-competitive impact of Basel II. It is now critically important that all bankers understand the linkage between weighted average cost of capital and risk-based capital ratio, and also realize the opportunity to enhance competitiveness with optimal capital management.
Large banks historically have had more options at their disposal than small and mid-sized banks in developing optimal capital structures, but thankfully this is changing. Mid-sized banks, for example, now can use the pooled market for trust preferred securities or the recently developed pooled market for non-cumulative perpetual preferred securities. By carefully selecting the appropriate mix of trust preferred, non-cumulative perpetual preferred and common stock, a financial institution can lower its weighted average cost of Tier 1 capital and, as a result, increase its franchise valuation. Surely, this is something all constituents of bank holding companies-from executives and shareholders to regulators-should and will welcome.
Thomas W. Killian is a principal of the Investment Banking Group of Sandler O'Neill & Partners.
