E9-7473. Federal Reserve Bank Services Private Sector Adjustment Factor

AGENCY:

Board of Governors of the Federal Reserve System.

ACTION:

Notice and request for public comment.

SUMMARY:

The Board requests comment on proposed modifications to its method for calculating the private-sector adjustment factor (PSAF). The PSAF is part of the Board's calculation, as required by the Monetary Control Act of 1980 (MCA), to establish the fees that Federal Reserve Banks (Reserve Banks) charge for certain financial services provided to depository institutions (DIs). Consideration of a new PSAF methodology was prompted by the reduction in clearing balances held by DIs at Reserve Banks following the Board's recent implementation of the payment of interest on required reserve balances and excess balances held at Reserve Banks, as well as by long-term changes in the structure of the market for providing payment services to DIs. The existing PSAF calculation model, which is built upon a correspondent bank framework, is driven primarily by the level of clearing balances held by DIs at Reserve Banks. The expected continued reduction in clearing balances will make the current PSAF calculation methodology less meaningful. Accordingly, the Board requests comment on the prospective need to change its methodology and its proposal to replace the current correspondent bank model for calculating the PSAF with a publicly traded firm model as described in this notice. If approved, use of this new model could be reflected in priced services fees as early as 2010.

DATES:

Comments must be submitted on or before May 29, 2009.

ADDRESSES:

You may submit comments, identified by Docket No. OP-1354, by any of the following methods:

• Agency Web Site: http://www.federalreserve.gov. Follow the instructions for submitting comments at http://www.federalreserve.gov/​generalinfo/​foia/​ProposedRegs.cfm.
• Federal eRulemaking Portal: http://www.regulations.gov. Follow the instructions for submitting comments.
• E-mail: [e-mail].
• FAX: [phone] or 202/452-3102.
• Mail: Jennifer J. Johnson, Secretary, Board of Governors of the Federal Reserve System, 20th Street and Constitution Avenue, NW., Washington, DC 20551.

All public comments are available on the Board's Web site at http://www.federalreserve.gov/​generalinfo/​foia/​ProposedRegs.cfm as submitted, except as necessary for technical reasons. Accordingly, your comments will not be edited to remove any identifying or contact information. Public comments may also be viewed electronically or on paper in Room MP-500 of the Board's Martin Building (20th and C Streets, NW.) between 9 a.m. and 5 p.m. on weekdays.

FOR FURTHER INFORMATION CONTACT:

Gregory L. Evans, Deputy Associate Director (202/452-3945), Brenda L. Richards, Manager (202/452-2753), Jonathan Mueller, Senior Financial Analyst (202/530-6253), or Rebekah Ellsworth, Financial Analyst (202/452-3480); Division of Reserve Bank Operations and Payment Systems. Telecommunications Device for the Deaf (TDD) users may contact 202/263-4869.

SUPPLEMENTARY INFORMATION:

I. Background

Under MCA, the Federal Reserve Banks must charge fees to DIs for certain financial services, known collectively as “priced services,” so as to recover, over the long run, all direct and indirect costs actually incurred in providing these services as well as the imputed costs that would have been incurred had the services been provided by a private-sector firm.^[1 2 ] MCA specifically identifies certain imputed costs that must be recovered via priced services fees, including taxes and return on equity (profit).

To set priced services fees in accordance with the requirements of MCA, the Board not only must estimate all actual direct and indirect costs incurred in providing priced services but also must impute costs that the Reserve Banks do not incur but would incur as private-sector entities. In determining a methodology for imputing these costs, the Board recognizes that there is no perfect private-sector proxy for the Reserve Bank priced services, but seeks a methodology that is theoretically sound and represents a reasonable approximation of the costs the Reserve Banks would incur if operating as private-sector providers. Because of the similarity between the services provided by Reserve Banks and many of the services offered by private-sector correspondent banks, the Board historically has derived these imputed costs, collectively known as the PSAF, and offsetting imputed revenue, known as net income on clearing balances (NICB), using a correspondent bank model. The PSAF and NICB are estimated annually, and the resulting net cost is incorporated each year when setting priced services fees and measuring cost recovery.^[3]

The Clearing Balance Program

The Reserve Bank clearing balance program was developed in connection with the implementation of MCA's requirement to establish fees for priced services. This program allows DIs to hold at Reserve Banks an agreed-upon level of clearing balances which serve several purposes, including facilitating settlement of transactions, protecting against overnight overdrafts, and paying for priced services through the generation of earnings credits. The Reserve Bank clearing balance program is largely modeled after similar programs offered by private-sector correspondent banks, wherein respondent banks maintain balances with their correspondents for some or all of the purposes listed above.

Under the Reserve Bank clearing balance program, a participating DI agrees to set and maintain a targeted minimum average clearing balance, known as the DI's contractual clearing balance, over a set period. A DI may hold balances in excess of its contractual clearing balance and is charged for deficiencies below the contracted minimum.

A DI accrues credits, known as earnings credits, on its contractual clearing balances (not on excess balances) held at a Reserve Bank at a rate currently equal to 80 percent of the 13-week moving average of the annualized coupon equivalent yield of the three-month Treasury bill. Earnings credits can only be applied toward priced services fees, and unused credits expire if not used within one year.

Calculating the PSAF

The Board's method for calculating the PSAF begins with developing a pro forma priced services balance sheet based on the projected average book value of Reserve Bank assets and liabilities to be used in providing priced services during the coming year.^[4] Additional elements on the priced services balance sheet are imputed as if the priced services were provided by a hypothetical private-sector correspondent bank. For example, a private-sector correspondent bank would be able to use the balances that its respondents deposit with it as a funding source for investments. Accordingly, the Board imputes investment income on clearing balances held at Reserve Banks based on an imputed portfolio of interest-bearing assets. Similarly, because private-sector correspondent banks are required to hold some portion of their deposit balances as vault cash or as balances at a Reserve Bank, the Board imputes a reserve requirement as a percentage of clearing balances. The imputed investment of clearing balances and the imputed reserve requirement both appear as assets on the priced services balance sheet.

The liability and equity components of the priced services balance sheet consist of clearing balances, short- and long-term liabilities related to providing priced services, imputed debt (if necessary), and imputed equity. The level of clearing balances on the priced services balance sheet increases or decreases at the discretion of the DIs maintaining those balances and provides a source of long-term financing for priced services assets.^[5] Using the correspondent bank model results in imputed debt only when core clearing balances, long-term liabilities, and equity on the priced services balance sheet are not sufficient to fund long-term assets; or when an interest rate sensitivity analysis indicates that a 200 basis point change in interest rates would change the percentage of priced services costs recovered (cost recovery) more than 2 percentage points. To satisfy the FDIC requirement for a “well-capitalized” institution, equity is imputed at 5 percent of total assets.^[6]

The imputed costs of the PSAF are derived from the priced services balance sheet. A target return on equity (ROE) rate is estimated and applied to the equity on the priced services balance sheet to determine the cost of equity. The ROE rate is estimated using the capital asset pricing model (CAPM), which calculates a firm's required ROE rate as the sum of a risk-free rate of return and a risk premium. In this model, the risk premium is the product of a firm-specific sensitivity factor, known as beta, which expresses the correlation of the firm's returns to the return of the market as a whole, and the expected return of the market in excess of the risk-free rate. In the PSAF calculation, the risk-free rate of return is based on the three-month Treasury bill rate, and the expected market risk premium is the average of the monthly returns of the market as a whole in excess of the risk-free rate over the most recent 40 years.^[7] The priced services beta of 1.0 assumes that, over time, priced services returns will be perfectly correlated with those of the overall market.

Given that Federal corporate income tax rates are graduated, State income tax rates vary, and various credits and deductions can apply, the correspondent bank model does not include an actual income tax expense. Instead, the Board targets a pretax ROE that would provide sufficient income for the priced services to fulfill their imputed income tax obligation. The imputed income tax rate used to calculate the pretax ROE is the median of the rates paid over the past five years by the top 50 bank holding companies (BHCs) ranked by deposit balances, adjusted to exclude any investment in tax-free municipal bonds. The PSAF also includes the estimated share of Board expenses that supports the priced services, imputed sales tax, and an imputed FDIC insurance assessment based on current FDIC rates and the level of clearing balances held at Reserve Banks.

Calculating NICB

The correspondent bank model includes imputed revenue, known as NICB, which is calculated each year along with the imputed costs of the PSAF. The NICB calculation assumes that, similar to a correspondent bank, the priced services would invest clearing balances, net of the imputed reserve requirement and balances used to finance priced services assets, in interest-bearing assets. To impute investment income, a rate of return equal to the yield on the three-month Treasury bill plus a constant spread is applied to the level of clearing balances available for investment on the priced services balance sheet. The constant spread is derived annually from a portfolio of investments comparable to the investment holdings of BHCs.^[8] The NICB calculation nets this imputed investment income against the actual cost of earnings credits, which represent the cost to the Reserve Banks of holding clearing balances.^[9]

Calculating Cost Recovery

The Board incorporates the PSAF and NICB into the projected and actual annual cost recovery calculations for Reserve Bank priced services. Cost recovery measures the percentage of priced services costs, including the PSAF, recovered through priced services fees and NICB. In the fall of each year, the Board projects the PSAF and NICB for the following year using the most recent clearing balance and rate data available (typically July data) during the process of establishing priced services fees. The Board also estimates cost recovery for the coming year using projected direct and indirect costs, revenue, and the net imputed cost generated from the estimated PSAF and NICB.

When calculating actual cost recovery for the priced services at the end of each year, the Board historically has used the estimated PSAF derived during the price-setting process with only minimal adjustments for actual rates or balance levels.^{[10]  [11]} The Board adopted this approach because the PSAF largely represents the fixed financing costs associated with the assets on the priced services balance sheet, which is updated annually. This method has proven to be reasonable and transparent without being unduly complex or burdensome. The Board updates NICB, however, to reflect actual interest rates and clearing balance levels throughout the year when calculating actual priced services cost recovery. Actual NICB, therefore, can vary from the projected amount used to determine priced services fees for a given year. For example, while the projected and actual PSAF for 2007 remained substantially unchanged at $132.5 million, actual 2007 NICB decreased from its $139.6 million projection to $133.8 million.

The Interdependence of Clearing Balances, the PSAF, and NICB

Changes in clearing balance levels directly affect the imputed costs and income that factor into priced services fees and cost recovery. Clearing balances not only represent the largest component of the priced services balance sheet but also drive the calculation of nearly all imputed elements included in priced services fees, including the financing costs, the cost of equity, and NICB. For example, clearing balances provide a major source of short- and long-term funding for the assets on the priced services balance sheet, representing 74 percent of total financing in 2007. Clearing balances thus reduce total imputed financing costs by eliminating the need to impute more costly forms of financing, such as debt.^[12] Clearing balances, in the form of imputed investments, also represent a significant portion of total priced services assets. Total assets, in turn, determine the level of imputed equity and the resultant imputed cost of that equity. In addition, the level of clearing balances influences the amount of funds available for investment in the imputed portfolio of investments and the cost of earnings credits, both of which are principal factors in the NICB calculation. These three elements—financing costs, the cost of equity, and NICB—are included in the net imputed cost that is recovered through priced services fees. Any change in the level of clearing balances, therefore, has a significant effect on the PSAF, NICB, and cost recovery.

Interest on Balances Held at Reserve Banks

Title II of the Financial Services Regulatory Relief Act of 2006 granted the Reserve Banks authority to pay earnings (interest) on balances maintained by or on behalf of DIs at Reserve Banks. Originally, this authority was to become effective in 2011. Section 128 of the Emergency Economic Stabilization Act of 2008, enacted on October 3, 2008, made the authority effective upon enactment. On October 6, 2008, the Board published an interim final rule amending Regulation D (Reserve Requirements of Depository Institutions). The interim rule directed the Reserve Banks to pay explicit interest on balances held at Reserve Banks to satisfy reserve requirements (required reserve balances) and on balances held in excess of both required reserve balances and contractual clearing balances (excess balances), effective October 9, 2008.^[13]

The Board has observed a significant decline in the level of clearing balances held at Reserve Banks following the implementation of interest on required reserve balances and excess balances and anticipates that this trend will continue. The daily average level of clearing balances over the two-week reserve maintenance period ending October 8, 2008 was $7.7 billion. As shown in figure 1, by the reserve maintenance period ending February 11, 2009, the daily average level of clearing balances had fallen to $4.6 billion. Over this period, the rate of interest paid on both required reserve balances and excess balances maintained at Reserve Banks was generally higher than the earnings credit rate paid on clearing balances.^[14] The interest rate on required reserve balances and excess balances as of March 2009 is 25 basis points, which is the top of the targeted range for the Federal funds rate and higher than the concurrent earnings credit rate for clearing balances. When the target Federal funds rate exceeds the earnings credit rate (the typical historical scenario), and absent a significant preference by DIs for implicit interest on clearing balances over explicit interest on excess balances held at Reserve Banks, DIs will likely continue to reduce clearing balances in favor of increasing excess balances to receive higher, explicit returns.

The expected continued decline in clearing balance levels could have significant implications for the imputed costs that factor into the Board's price-setting methodology. If clearing balance levels decline significantly, the priced services balance sheet will shrink dramatically, and the priced services will lose a major source of both funding and income. A continued reduction in clearing balance levels will decrease the similarities between the financial characteristics of the priced services and private-sector correspondent banks. Specifically, with low to zero clearing balance levels, it will be more difficult to draw the analogy between correspondent banks, whose balance sheets include large levels of deposit balances and related accounts, and the Reserve Bank priced services. Similarly, markedly reduced clearing balance levels will call into question the use of the FDIC's regulatory structure for well-capitalized depository institutions as a determinant of equity capital on the priced services balance sheet and will potentially nullify the calculation of an FDIC insurance assessment based on clearing balance levels. All of these factors challenge the continued applicability of a PSAF model based on a correspondent bank framework.

The potential for such circumstances, in conjunction with the ongoing changes in the nature of priced services competitors discussed below, has prompted the Board to consider changes to its approach to imputing the costs that MCA requires to be recovered through priced services fees. If approved, these changes could be effective as early as the 2010 pricing process. In determining the appropriate timing of such changes, the Board will consider trends in the level of clearing balances held at Reserve Banks and the extent to which the nature of the Reserve Banks' competitors, particularly in the check service, shifts away from correspondent banks.

The Board requests comment on the following:

If the explicit interest rate for required reserve balances and excess balances continues to be higher than the implicit rate paid on clearing balances in the form of earnings credits, is it reasonable to assume that DIs will continue to reduce or eliminate their level of contractual clearing balances in favor of holding additional excess balances? If not, why might DIs choose to maintain their clearing balances?

Will DIs raise and lower the level of clearing balances they hold at Reserve Banks depending on whether the earnings credit rate is above or below the rate on excess balances?

Are there any reasons why the Board should maintain its clearing balance program if demand for clearing balances continues to decline significantly?

Trends in the Banking and Payment Systems Industries

As noted above, when implementing the priced services provisions of MCA in the early 1980s, the Board identified private-sector correspondent banks as the most appropriate peer group for the priced services in adopting key elements of the policy. The Board considered correspondent banks to be a reasonable proxy for private-sector providers of priced services because they are the primary competitors of the Reserve Banks' check service, which historically has comprised more than 80 percent of the cost of Reserve Bank priced services activities. In doing so, the Board recognized that BHCs offer diverse services that extend well beyond the payment services that are provided by the Reserve Banks, and that these services largely drive BHC financial results; however, given that Reserve Banks and BHCs both hold customer balances that facilitate payment services, the Board considered it a reasonable comparison.

Recently, however, the analogy between private-sector correspondent banks and the priced services has become less applicable. The payment systems industry has sharply decreased its use of traditional check services and increased its use of electronic payment services. As a result, user-owned utilities, the Reserve Banks' typical competitors in electronic payment services, have increasingly replaced correspondent banks as the predominant competitors of the Reserve Banks in providing priced services. These user-owned utilities include such entities as the Clearing House Interbank Payment System (CHIPS), which is the primary competitor for Fedwire® funds transfer services, and the Electronic Payments Network (EPN), which is the only private-sector automated clearinghouse (ACH) operator. Both of these entities are part of a larger cooperative, The Clearing House Payments Company, LLC (TCH), which is owned entirely by its principal users. Unlike private-sector correspondent banks, user-owned utilities do not hold overnight balances for their participants. As paper check processing volumes continue to decline and the check service becomes more electronic, utilities will likely increasingly be key competitors of the Reserve Banks in providing priced services. These trends, in conjunction with the potential continued significant decline in clearing balances resulting from the ability of DIs to receive explicit interest on balances held at Reserve Banks, raise questions about the continued appropriateness of the correspondent bank model as the basis for the imputed costs that factor into the Board's pricing methodology.

II. The Proposed PSAF Model

The Publicly Traded Firm Model

The Board seeks to replace the current correspondent bank model with a model that is transparent, consistent with current financial theory and practice, and conceptually sound as a basis for efficient pricing in the market of payment services. To achieve these objectives, and given the difficulty in identifying and obtaining data for an applicable peer group, the Board proposes to replace the correspondent bank model with a “publicly traded firm model” for calculating the imputed costs that factor into priced services fees and cost recovery. This model recognizes the shift, in the priced services' financial characteristics and competitors, away from correspondent banks, as well as the difficulties inherent in a user-owned utility model as discussed below, and instead compares the priced services to the entire market of U.S. publicly traded firms.

Under the publicly traded firm model, the asset side of the priced services balance sheet would reflect only the projected portion of actual Reserve Bank assets used to provide priced services; no additional assets would be imputed. Any residual clearing balances maintained by DIs at Reserve Banks would not be included in the priced services balance sheet or in the calculation of the PSAF. Consequently, imputed investments and NICB would be zero by definition, and the priced services would impute additional equity and debt to meet the funding need on the priced services balance sheet. The publicly traded firm model would not include an imputed FDIC assessment, because the priced services' peer group would no longer be limited to private-sector correspondent banks and because, as noted above, any residual clearing balances would not be included in the priced services balance sheet or in the PSAF calculation. The imputed capital structure, debt and equity financing rates, and effective income tax rate would be based on data for the U.S. market as a whole and would be calculated using the various market data sources and time frames discussed below. The time frame selected for each of these imputed elements was chosen to minimize volatility in the PSAF from year to year. A one-year time frame was selected for elements that historically have been more stable; a five-year average was selected when data were more volatile historically or when changes in that element would have a larger impact on the PSAF.^[15] When averaging data for individual U.S. firms, the model would use value-weighted rather than equal-weighted averages.^[16]

The priced services imputed capital structure would be based on the most recent full-year value-weighted average capital structure (that is, total long-term debt to total long-term debt plus equity) of all U.S. publicly traded firms included in a commercially available financial database. The Board initially proposes using Standard & Poor's Compustat® database as the source for the capital structure and effective income tax rate of all U.S. publicly traded firms. The Standard & Poor's Compustat® database contains information on more than 6,000 U.S. publicly traded firms, which approximate the entirety of the U.S. market. Because of the timing of the price-setting process and the availability of relevant data, there would be a two-year lag in the data used in the PSAF calculation: for example, 2010 priced services fees, set in late 2009, would be based upon full-year 2008 data.^[17] Table 1 shows the value-weighted average capital structures for all U.S. publicly traded firms in the Standard & Poor's Compustat® database from 2003 to 2007. In 2007, based on the foregoing, the value-weighted average capital structure was 54 percent.

Table 1—Capital Structure (Capitalization Ratio) of U.S. Publicly Traded Firms

2003	2004	2005	2006	2007	Five-year average	Standard deviation
55%	53%	53%	52%	54%	53%	1.0%
Source: Standard & Poor's Compustat® data.

Because the PSAF resulting from the publicly traded firm model is not highly sensitive to capital structure and because the value-weighted average capital structure does not vary significantly from year to year, the Board believes that a one-year time frame is appropriate when imputing the priced services capital structure. This conclusion is supported both by financial theory, which states that changes in capital structure should not significantly affect the value of a firm, and by sensitivity analysis as shown in attachment 1.^[18]

The imputed effective income tax rate would be the five-year mean of the value-weighted average ratios of current tax expense to total net income for all U.S. publicly traded firms in the financial database. Table 2 shows the annual value-weighted average effective tax rates for all U.S. publicly traded firms in the Standard & Poor's Compustat® database from 2003 to 2007. For that period, the five-year mean of these tax rates was 24 percent. A five-year mean would be used because of the volatility of the annual effective tax rate from year to year and the sensitivity of the PSAF to this input, as shown in attachment 1.

Table 2—Effective Tax Rate of U.S. Publicly Traded Firms

2003	2004	2005	2006	2007	Five-year average	Standard deviation
19%	23%	27%	24%	29%	24%	3.4%
Source: Standard & Poor's Compustat® data.

The imputed long-term debt financing rate under the publicly traded firm model would be the five-year mean of an estimated average annual bond yield for the market as a whole. The Board proposes to use a five-year mean when imputing a long-term debt financing rate to be consistent with the treatment of the tax rate (both of these inputs are cost-related) and to reduce year-to-year volatility in the PSAF.^[19]

The Board initially proposes calculating the imputed long-term debt rate as the five-year mean of the Aaa and Baa Moody's bond yields published on the Federal Reserve Board's H.15 Statistical Release.^[20] The inclusion of only investment-grade debt is based on analysis of data on approximately 1,400 publicly traded firms in the Compustat database for which bond rating data are available.^[21] Given that the majority of outstanding debt for this population was investment grade, the Board considered an average investment-grade bond yield to be a reasonable proxy for the imputed priced services long-term debt financing rate. The Board considered two averaging techniques to determine the average investment-grade bond yield, which provided nearly identical results. Of these two approaches, the five-year mean of the Aaa and Baa Moody's bond yields was more simple and transparent.^[22] Table 3 shows the annual average yield from 2003 to 2007 using this methodology. For this period, the five-year mean was 6.0 percent.

Table 3—Average of Annual Moody's Aaa and Baa Bond Yields

2003	2004	2005	2006	2007	Five-year average	Standard deviation
6.2%	6.0%	5.7%	6.0%	6.0%	6.0%	0.2%

Using an average investment-grade bond yield as the imputed priced services long-term debt financing rate, however, does not take into account the effect of non-investment-grade debt on the average bond yield for the market as a whole. Inclusion of non-investment-grade debt would result in a somewhat higher imputed long-term debt financing rate. Accordingly, the Board could also calculate an average bond yield for U.S. publicly traded firms using five-year average yields for each bond rating, weighted by the relative proportion of debt outstanding in the market at each bond rating.^[23]

If short-term assets exceed short-term liabilities on the priced services balance sheet, short-term debt would be imputed at the average of the three-month AA and A2/P2 nonfinancial commercial paper rates as published on the Federal Reserve Board's Commercial Paper Release.^[24] This methodology is simple, transparent, consistent with the proposed approach to calculating the long-term debt financing rate, and based on publicly available data.

The Board considered other data sources for each of the imputed elements discussed above. These sources include the Flow of Funds Federal Reserve Board Statistical Release for capital structure, general corporate income tax rates as found on Internal Revenue Service (IRS) Form 1120 for the effective tax rate, and the ratio of “interest and related expense” to total debt for all publicly traded U.S. firms in the Standard & Poor's Compustat® database for the long-term debt financing rate.^[25] In each case, the Board considered the source set forth in the current proposal to be the superior alternative. The Flow of Funds release does not include data on U.S. publicly traded financial firms and provides only approximate market-value equity data. Use of the general corporate income tax rate published by the IRS would inappropriately exclude the effect of State and local taxes. A long-term debt financing rate calculated from the Standard & Poor's Compustat® database would be artificially high because of the inclusion of “related expense,” which includes items such as interest on deposits held at DIs, in the interest expense measure used in the numerator.

Under the publicly traded firm model, the imputed ROE rate would continue to be calculated using the CAPM with a beta of 1.0 and a 40-year average historical market premium. Given the sensitivity of the PSAF to the risk-free rate used in the CAPM, and because short-term Treasury bill rates are generally more sensitive to interest rate changes than longer-term rates, the Board considered replacing the current short-term risk-free rate with a longer-term risk-free rate. As shown in attachment 1, changes in the risk-free rate used in the calculation of the target ROE rate affect the PSAF more than any other imputed element. In 2005, the Board decided to use a three-month Treasury bill rate as the risk-free rate to impute the target ROE because this rate was consistent with that used to calculate NICB and would help minimize volatility in the net imputed cost caused by changes in interest rates.^[26] With the elimination of NICB under the proposed publicly traded firm model, however, using a longer-term Treasury rate, such as the 10-year Treasury bond rate, may be an appropriate way to minimize volatility in the calculation of the target ROE rate. A longer-term rate more closely matches the duration of stock market indexes used to estimate a beta, the expected life of the assets on the priced services balance sheet, and the investment horizon of a long-term investor.

Table 4 compares certain components for 2009 derived under the publicly traded firm model with the same components as derived under the baseline case.^[27] Using the elements discussed above, the publicly traded firm model returns a PSAF of $55.4 million compared with a baseline PSAF of $62.2 million (NICB of $48.8 million, net imputed cost of $13.4 million).

The baseline net imputed cost reflects clearing balance levels and interest rates as of July 2008. The correspondent bank model is highly sensitive to both of these variables. For example, using the lower clearing balance levels and interest rates from February 2009, projected 2009 NICB is less than half the amount that was projected for pricing purposes, leading to an increase in the 2009 net imputed cost. If clearing balances continue to decline, the variance between the PSAF calculated using the proposed methodology and the net imputed cost using the correspondent bank model will likely be significantly smaller than noted above. In contrast, as interest rates rise, the income generated on each dollar of clearing balances in the NICB calculation of the correspondent bank model will increase. Rising interest rates, however, will widen the spread between the interest rate on excess balances and the earnings credits rate, giving DIs more incentive to shift from maintaining clearing balances to maintaining additional excess balances. This expected reduction in clearing balances will reduce NICB, counteracting the effect of higher per-dollar earnings and likely leading to a net decrease in NICB. Consequently, rising interest rates could cause an overall increase of the net imputed cost of the correspondent bank model throughout the year. This increase could substantially shrink the variance between the PSAF of the proposed model and the net imputed cost of the current model.

Table 4—Comparison of Current and Proposed Model

	Balance sheet assets (billions)	Financing composition	Financing cost	Tax rate (percent)	Debt rate (percent)	PSAF (millions)	NICB (millions)
Baseline case: correspondent bank model	$9.2	Equity per FDIC guidelines	ROE of $46.2 M	32.6	(1)	$62.2	$48.8
Publicly traded firm model	1.3	54% long-term debt, 46% equity	$40.3M (ROE of $22.3M; debt cost of $18.0M)	24	6.0	55.4	0
1 No debt.

The Board believes that the publicly traded firm model would be an appropriate replacement for the current PSAF model for a variety of reasons. The publicly traded firm model is relatively simple to calculate and understand, easily replicable by the public, and uses objective, publicly-available data for all imputed inputs. Unlike the correspondent bank model, the publicly traded firm model is not linked to the level of clearing balances held at Reserve Banks. This characteristic is important given the uncertainty surrounding future clearing balance levels. Substantially lower clearing balances would not only affect the funding and income of the priced services but also undermine the basis for the use of an FDIC-based regulatory structure for depository institutions as a determinate of the priced services capital structure. A model that is not dependent on clearing balance levels is also appropriate in an environment where clearing balances are not relevant to a growing proportion of the Reserve Banks' competitors in providing priced services. Another advantage of the publicly traded firm model is its independence from a narrowly defined peer group, such as private-sector correspondent banks, that may become less relevant to the priced services over time. Unlike other models considered, the publicly traded firm model does not incorporate data from a limited number of comparable firms but rather from the entire U.S. market of publicly traded firms. This independence decreases the risk of price volatility that could result from changes in the characteristics or financial results of a limited peer group. The publicly traded firm model also is consistent with financial theory regarding capital structure and financing costs and is conceptually sound. In addition, the publicly traded firm model is consistent with the current approach to calculating the ROE using CAPM with a beta of 1.0, which compares the priced services to the market as a whole.

The publicly traded firm model also has a few drawbacks. If some level of clearing balances persists at Reserve Banks over the long term, excluding these priced-services-related balances from the calculation of the PSAF would depart from the Board's past practice of including all actual priced services assets and liabilities in the calculation of the PSAF and would disregard potential imputed income from these balances. A publicly traded firm model also departs from a model based specifically on the banking industry. This change in direction may conflict with the fact that the priced services are provided by Reserve Banks, which are, by definition, banks.

The Board specifically requests comment on the following:

Is using the U.S. market as a whole as a basis for the imputed capital structure, tax rate, and debt financing rates of the priced services reasonable? Is discontinuing the use of a correspondent bank model reasonable?

Are the proposed approaches to imputing the capital structure, effective tax rate, and long- and short-term debt financing rates appropriate?

Is it reasonable to include only investment-grade bond yields in the calculation of the imputed long-term debt financing rate? If not, what approach should the Board take to include other yields or rates in the calculation? What publicly-available data sources are best suited for obtaining data on non-investment-grade debt?

Is it reasonable to limit the calculation of the short-term debt financing rate to include only rated commercial paper even if the long-term debt financing rate calculation were expanded to include non-investment-grade debt, given the expectation that the need for short-term funding on the priced services balance sheet will be relatively small? If not, what approach should the Board take to include other rates in the calculation?

What publicly-available data sources are best suited for determining the effective tax rate, capital structure, and short- and long-term debt financing rates of the U.S. market?

Should the Board consider using a longer-term risk-free rate to calculate the target ROE to decrease the ROE calculation's sensitivity to changes in interest rates?

III. Other PSAF Models Considered

The User-Owned Utility Model

The Reserve Banks' major competitors in the provision of priced services increasingly are user-owned utilities rather than traditional correspondent banks. Accordingly, one approach to revise the methodology for imputing costs might be to model the priced services balance sheet and imputed capital structure, financing rates, tax rate, and other applicable costs on a user-owned utility. Under this methodology, the priced services balance sheet and imputed costs would reflect either the financial characteristics of a peer group of user-owned utilities currently existing in the market or theoretical assumptions about the behavior and characteristics of this type of organization.

A user-owned utility model is conceptually appealing because the Reserve Banks' competitors in the Fedwire® Funds, FedACH®, and, to a lesser extent, check services are increasingly user-owned utilities. Such a model also recognizes that, as clearing balance levels decline, providing priced services to DIs that do not maintain clearing balances could more closely resemble the operation of a user-owned utility than that of a traditional correspondent bank.

Selecting an appropriate peer group for this approach, however, is challenging. User-owned utilities typically provide a diverse array of services using various operational approaches. Although choosing a narrowly defined peer group of user-owned utilities, specifically one consisting of peers that provide services more closely resembling the priced services, could provide a more-comparable peer group, this approach may also introduce greater volatility in the PSAF because of the dependence on data from a small number of firms.

A user-owned utility peer group could present other problems as well. Publicly available financial data on user-owned utilities are often not published. For example, CHIPS and EPN provide services that compete with the priced services provided by Reserve Banks. These two entities, however, are both components of TCH, which does not publicly report its financial statements either by product line or in aggregate. Although data are more readily available to the public from several other user-owned utilities (such as SWIFT and the Depository Trust & Clearing Corporation), the services provided by these firms are less comparable to those provided by the Reserve Banks.

Basing this model on theoretical characteristics of user-owned utilities rather than on the actual data of a specific peer group could also prove challenging. User-owned utilities, by definition, lack incentive for profit maximization because the owners of these utilities are also their primary customers. Consequently, user-owned utilities tend to seek to maximize the benefit afforded to their users by providing low-cost services while remaining financially viable. Although the assumption that this characteristic could result in a lower required rate of return on equity is reasonable, establishing a methodology to calculate that rate using the limited economic literature available on the subject could be difficult. Further, establishing the means to calculate the other requisite imputed elements—capital structure, debt financing rates, and income taxes—using theoretical assumptions or academic studies could be similarly challenging.

The user-owned utility model exhibits some of the same drawbacks of the publicly traded firm model that the Board is proposing. For example, a user-owned utility model represents the same significant departure from a model based specifically on the banking industry. A user-owned utility model also would not include residual clearing balances, which departs from the Board's past practice of basing the PSAF on actual priced services assets and liabilities.

The Board specifically requests comment on the following:

Given that user-owned utilities reflect a significant portion of the Reserve Banks' competitors in providing priced services, would a user-owned utility model be more appropriate? If yes, are there approaches the Board should consider that would address the identified obstacles?

The Cost-Plus Model

In 2005, while commenting on proposed changes to the PSAF methodology for calculating the ROE, two commenters suggested a cost-plus model as a simple, straightforward method for calculating the PSAF. Accordingly, the Board investigated the possibility of using a cost-plus PSAF model based on priced services operating expenses. A cost-plus PSAF model would add a markup to the priced services operating expenses for the year. The markup would be calculated by applying an internal benchmark or market rate of return to the level of budgeted priced services operating expenses. Regardless of the method used to calculate the markup, residual clearing balances held at Reserve Banks would not be included in the calculation of net imputed cost, and NICB would therefore be zero by definition.

Calculating the markup for a cost-plus model requires a data source from which to develop the internal benchmark or market rate of return to be applied to budgeted operating expenses. In the case of an internal benchmark, the Board considered using an average of historical PSAF values. Such values, however, would not take current data into account and would reflect a correspondent bank model that is increasingly inapplicable given recent trends in the payments industries and the expected continued decline in the level of clearing balances. In addition, a static internal benchmark based on historical PSAF values would fail to reflect ongoing changes in the marketplace.

Alternatively, the Board could base the markup ratio applied to the priced services operating expenses on an external benchmark, such as the average markup over operating expenses for the U.S. market as a whole.^[28] Specifically, the Board could calculate the markup as the ratio of pretax income and interest expense to operating expense for all U.S. publicly traded firms. This markup could then be applied to the projected level of priced services operating expense, including imputed operating expenses such as sales tax, to determine the value of the imputed profit, debt financing cost, and income taxes to be factored into priced services fees. Applying a markup over expenses ratio based on value-weighted average data for all publicly traded U.S. firms in the Standard & Poor's Compustat® database to the 2009 budgeted priced services operating expense yields a projected 2009 PSAF of $157.5 million.^[29]

Although a cost-plus model is simple, transparent, and replicable by the public, it also has several weaknesses. A cost-plus model based on historical PSAF values is static and assumes continued use of the current correspondent bank model, which is increasingly inapplicable. In addition, basing a cost-plus model on accounting-based values captures only book, not market, values of financing and other costs. Such a model is also not consistent with current finance theory. As with the models discussed previously, a cost-plus model represents a departure from a model based specifically on the banking industry.

The Board specifically requests comment on the following:

Should the Board consider implementing a cost-plus model?

Are there other sources of data that the Board should consider using to calculate an appropriate markup over operating expenses or over another financial characteristic of the priced services?

Are there other approaches that the Board should consider to address the identified obstacles?

Continuation of the Current Correspondent Bank Model

The Board also considered the continued use of the current correspondent bank model to impute costs, with minor modifications. Using this model while also paying interest on required reserve balances and excess balances would result in a significantly smaller priced services balance sheet because of the anticipated decline in clearing balances and the associated imputed investment assets. Equity, which would still be imputed at the FDIC regulatory minimum for a well-capitalized depository institution, would shrink because of the reduction in size of the overall priced services balance sheet.

Residual clearing balances would continue to serve as a funding source for the priced services. If residual balances were not sufficient to meet the funding need, net of equity, on the priced services balance sheet, debt would be imputed. The imputed short- and long-term debt financing rates would be calculated using the same methodologies outlined for the imputed debt financing rates of the publicly traded firm model. Using average market debt financing rates in the correspondent bank model recognizes that as clearing balances fall and debt rises as a percentage of total priced services assets, the priced services balance sheet would look increasingly like that of a publicly traded firm and less like that of a correspondent bank.^[30] An average debt financing rate would also use readily-available public data and could be calculated with greater administrative ease. If residual clearing balances exceeded the funding need on the priced services balance sheet, NICB would be imputed.

Table 5 compares certain components for 2009 as derived under a continuation of the current correspondent bank model, with assumed residual clearing balance levels ranging from $0 to $4 billion, to the same components as derived under the baseline case. Using the values listed below, a continuation of the current correspondent bank model would return a net imputed cost between $50.7 million (PSAF of $50.7 million, NICB of $0) and $19.5 million (PSAF of $40.6 million net of $21.1 million in NICB).^[31] The increase in net cost is largely the result of the reduction or elimination of NICB caused by the decline in clearing balances levels. This increase is partially offset by a reduction in the cost of equity as a result of the reduced level of total assets and, consequently, of imputed equity on the priced services balance sheet.^[32]

Table 5—Correspondent Bank Model Under Different Clearing Balance Assumptions

Assumed clearing balance level	Balance sheet assets (billions)	Financing composition	Financing cost	Tax rate percent	Debt rate percent	PSAF (millions)	NICB (millions)	Net imputed cost (millions)
Baseline case: $7.4 B ($4 B in core clearing balances)	$9.2	Equity per FDIC guidelines; remainder clearing balances	ROE of $46.2 M	32.6	(1)	$62.2	$48.8	$13.4
$4 B ($2 B in core clearing balances)	5.0	Equity per FDIC guidelines; remainder clearing balances	ROE of $25.0 M	32.6		40.6	21.1	19.5
No clearing balances	1.3	Equity per FDIC guidelines; remainder debt	$35.6 M (ROE of $6.4 M; debt cost of $29.2 M)	32.6	6.0	50.7	0	50.7
1 No debt.

Continued use of the correspondent bank model for imputing costs would provide several advantages. Among these is its ability to draw upon a well-defined FDIC regulatory structure and a peer group with readily available data when establishing key imputed elements such as capital structure and rates. This model also would afford a means by which possible residual clearing balances held at Reserve Banks could continue to provide a low-cost funding source and potential source of imputed income.

A principal disadvantage of this model is the decreasing similarity between the financial and operational characteristics of the Reserve Bank priced services and traditional correspondent banks if the level of clearing balances held at Reserve Banks continues to fall. Historically, the Board has recognized that the financial characteristics of BHCs are not driven primarily by the payment services that compete with those offered by Reserve Banks, but has considered BHCs an appropriate peer group because they are the primary competitors to the Reserve Banks' check services and because both entities hold customer balances for the purpose of facilitating payments services. If clearing balance levels approach zero and as the check service declines as a percentage of priced services revenue and expenses, comparing priced services to correspondent banks for the purpose of establishing a PSAF model will be increasingly difficult. Dramatically reduced clearing balance levels will also call into question the applicability of an FDIC-based regulatory structure designed for depository institutions as the determinant of the priced services capital structure. Specifically, in an environment of low to zero clearing balance levels, applying the FDIC's regulatory structure could result in a priced services capitalization ratio of more than 85 percent, which seems unreasonable when compared to correspondent banks that are primarily funded by balances rather than long-term debt.

The Board specifically requests comment on the following:

Would continued use of the correspondent bank model to calculate the PSAF be appropriate given the expected reduction in clearing balances and changes in priced services competitors? If so, is the proposed approach for calculating a debt financing rate in the correspondent bank model reasonable?

IV. Competitive Impact

In its March 1990 policy statement “The Federal Reserve in the Payments System,” the Board stated that all operational and legal changes considered by the Board that could have a substantial effect on payment system participants are subject to a competitive-impact analysis.^[33] Under this policy, the Board evaluates whether a proposed change would have a direct and material adverse effect on the ability of other service providers to compete effectively with the Reserve Banks in providing similar services. These effects could be caused by differences in legal authority or constraints between Reserve Banks and private-sector competitors or by a dominant market position that the Reserve Banks might derive from such legal differences. If the proposed change creates such an effect, the Board must further evaluate the changes to determine whether its benefits—such as contributions to payment system efficiency, payment system integrity, or other Board objectives—can be retained while reducing the hindrances to competition.

The intent of the PSAF, and of setting priced services fees in general to fully recover the costs (including imputed costs and profits) to provide them, is to facilitate competition between Reserve Banks and private-sector providers of payment services to foster a more efficient payment system. Identifying a meaningful private-sector peer group for the purpose of calculating the PSAF, however, has been difficult given the specific nature of the priced services provided by the Reserve Banks. The correspondent bank model historically has provided a reasonable proxy for Reserve Bank priced services, although the Board recognizes that correspondent bank balance sheets and ROE are typically driven largely by services that are not similar to those provided by the Reserve Banks. As the Reserve Banks' check service becomes a smaller proportion of total priced services revenues and costs, user-owned utilities are increasingly becoming the Reserve Banks' key priced services competitors. Because correspondent banks will no longer represent the primary competitors of Reserve Banks in providing priced services, and because no reliable comparative data are available for the user-owned utilities, the Board believes modeling the PSAF on a publicly traded firm model is appropriate. Accordingly, the Board believes that such a change in the PSAF model, if made, would not have a direct and material adverse effect on the ability of other service providers to compete effectively with Reserve Banks in providing similar services.

V. Paperwork Reduction Act

In accordance with the Paperwork Reduction Act of 1995 (44 U.S.C. ch. 3506; 5 CFR 1320 appendix A.1), the Board has reviewed the proposal under the authority delegated to the Board by the Office of Management and Budget. The proposal contains no provisions subject to the Paperwork Reduction Act.

By order of the Board of Governors of the Federal Reserve System, March 30, 2009.

Jennifer J. Johnson,

Secretary of the Board.

Footnotes

1.  These priced services include the check, automated clearinghouse, Fedwire® Funds, and Fedwire® Securities (for activity not related to Treasury securities) services.

2.  12 U.S.C. 248a(c)(3).

3.  In 2008, actual direct and indirect costs represented approximately 88 percent of total priced services costs and the PSAF represented the remaining 12 percent. The PSAF constituted an estimated $108.3 million of the overall costs recovered by priced services activities, and was offset by approximately $101.7 million of NICB.

4.  The 2007 priced services balance sheet can be found in the Federal Reserve Board's 2007 Annual Report at http://www.federalreserve.gov/​boarddocs/​rptcongress/​annual07/​sec2/​c3.htm#nl12.

5.  Using clearing balances as a financing source is consistent with private-sector correspondent banks' use of their respondent balances to fund short- and long-term assets. In the correspondent bank model only the portion of clearing balances that has remained stable over time (core clearing balances), historically set at $4 billion, is used to fund long-term assets on the priced services balance sheet.

6.  Equity is imputed based on the FDIC definition of a well-capitalized depository institution for insurance premium purposes. The FDIC requirements for a well-capitalized depository institution are (1) a ratio of total capital to risk-weighted assets of 10 percent or greater, (2) a ratio of Tier 1 capital to risk-weighted assets of 6 percent or greater, and (3) a leverage ratio of Tier 1 capital to total assets of 5 percent or greater. Because the total capital on the priced services balance sheet has no components of Tier 1 or total capital other than equity, requirements 1 and 2 are essentially the same measurement. In addition, because risk-weighted assets have historically been considerably below actual assets on the priced services balance sheet, typically only requirement 3 has been binding for the priced services.

7.  Data on market returns are based on the French data series, which is the standard data series used to estimate the market risk premium (http://mba.tuck.dartmouth.edu/​pages/​faculty/​ken.french/​data_​library.html).

8.  These investments include short-term Treasury securities, government agency securities, commercial paper, long-term corporate bonds, and money market funds. For additional details on the calculation of the constant spread, refer to the notice of approval of modifications to the method for calculating the PSAF, 68 FR 61413-61418 (Oct. 28, 2003).

9.  Because clearing balances are voluntary, set by priced services customers, and held for clearing transactions or offsetting priced services fees, they are directly related to the priced services. The cost associated with holding clearing balances, therefore, is appropriately attributed to the priced services.

10.  Although the largest portion of the PSAF, the target ROE, is fixed, two minor elements of the PSAF calculation are variable. The first adjusts the imputed income tax expense for the difference between the projected and actual priced services net income by applying the imputed effective income tax rate to any difference. The second recalculates the imputed FDIC assessment using actual clearing balance levels and assessment rates.

11.  In light of the uncertainty about the long-term effect that paying interest on required reserve and excess balances held at Reserve Banks will have on the level of clearing balances, the Board will adjust the PSAF used in the actual cost-recovery calculation for 2009 using the actual clearing balance levels maintained throughout 2009.

12.  Historically, debt financing rates have been higher than the earnings credit rate, making debt a more costly source of financing for the priced services balance sheet. For the week ended February 11, 2009, the earnings credit rate paid on clearing balances held by DIs at the Reserve Banks was 0.09 percent versus 5.21 percent for the bond rate on Moody's Aaa-rated corporate bonds for the week ended February 13, 2009 (see http://www.federalreserve.gov/​releases/​h15/​20090105/​).

13.  73 FR 59482-59486 (Oct. 9, 2008), as amended by 73 FR 65506-65507 (Nov. 4, 2008), 73 FR 67713-67714 (Nov. 17, 2008), and 73 FR 78616 (Dec. 23, 2008).

14.  The recent plateau in clearing balance levels may be due to the small difference (often of less than 15 basis points) between the rates earned on excess balances and clearing balances in the current low interest rate environment. In a more normal rate environment, the absolute value of this difference will increase, giving DIs more incentive to shift from maintaining contractual clearing balances to maintaining interest-earning excess balances.

15.  Although MCA's requirement for cost recovery over the long run allows the Board to set fees to over- or underrecover costs in a given year to minimize price volatility, volatility in imputed costs makes the pricing process more complex. As a result, the Board has typically preferred to adopt PSAF methodologies that provide for stable rather than volatile imputed costs.

16.  Value-weighted averages assign equal weight to each dollar, while equal-weighted averages assign equal weight to each firm. The Board opted to use value-weighted averages to reflect more accurately the financial characteristics of the market as a whole rather than those of the “average” firm in the market.

17.  The two-year lag in the data used to calculate certain imputed costs in the PSAF is characteristic of the current model as well and is due in large part to the timing of the price-setting process.

18.  F. Modigliani and M.H. Miller (1958), “The Cost of Capital, Corporation Finance, and the Theory of Investment,” American Economic Review, 48, pp. 261-97. The Modigliani-Miller Theorem states that under some conditions and in an efficient market the value of a firm is unaffected by how that firm is financed.

19.  Although attachment 1 shows low levels of volatility in the average Moody's bond rates from 2003 to 2007, this stability has not been the historic norm. Given the PSAF's sensitivity even to small changes in the debt financing rate, the Board plans to use a five-year average to minimize volatility in the PSAF.

20.  http://www.federalreserve.gov/​releases/​H15/​data.htm. Moody's Aaa and Baa bond ratings represent the upper and lower limits of the range of investment-grade bonds.

21.  While the firms in this sample included only approximately 20 percent of publicly traded firms in the database, they represented more than 85 percent of the assets and debt of the complete population of over 6,000 firms. Analysis of data for this sample from 2003 to 2007 showed that 82 percent of outstanding long-term debt (which represents over 70 percent of the outstanding long-term debt for all firms in the database during that period) was investment grade.

22.  Alternatively, the Board could calculate an average investment-grade bond yield using five-year average annual bond yields for each investment grade, weighted by the relative proportion of debt outstanding for each grade in the population of approximately 1,400 firms. For 2003 to 2007, the weighted average bond yield using this technique differed from the five-year mean of the Aaa and Baa Moody's bond yields by 2 basis points.

23.  The relative proportions of outstanding debt would be based on the most recent five years of Standard & Poor's Compustat® data for which bond rating data are available.

24.  http://www.federalreserve.gov/​releases/​cp/​. AA and A2/P2 ratings for commercial paper approximate the same credit ratings as Moody's Aaa and Baa ratings for bonds. Since 2002, the priced services short-term funding need has been met by clearing balances, eliminating the need to impute short-term debt.

25.  Current corporate income tax rates can be found in the 2008 instructions for IRS Form 1120 at http://www.irs.gov/​pub/​irs-pdf/​i1120.pdf.

26.  70 FR 60347 (Oct. 17, 2005). NICB is based on an average three-month Treasury bill rate, while the target ROE CAPM calculation uses a current three-month Treasury bill rate for the risk-free rate.

27.  The baseline PSAF of $62.2 million, projected NICB of $48.8 million, and net imputed cost of $13.4 million are the Board-approved projected 2009 values using the correspondent bank model. 73 FR 65329-65340 (Nov. 3, 2008).

28.  The Board discarded the idea of basing the markup ratio on data for more narrowly-defined peer groups because of the challenges of comparability and data availability discussed previously.

29.  The Board could calculate a markup over expenses ratio using two averaging techniques: equal weighting and value weighting. The Board believes value weighting is more appropriate because it would yield less-volatile results and would better capture the characteristics of the market as a whole.

30.  For example, if clearing balances fall to zero, applying the FDIC regulatory structure to determine the capital structure on the priced services balance sheet would result in a capitalization ratio of over 85 percent.

31.  The results presented in Table 5 are based on a risk-free rate as of July 2008 of 1.67 percent. As interest rates increase, both the ROE costs of the PSAF and the earnings of the NICB portfolio would increase. The net effect of this increase would depend on the size and character of the priced services balance sheet.

32.  The decrease in total financing costs is offset in part by the cost of financing priced services assets with higher-cost debt instead of low-cost clearing balances.

33.  FRRS 9-1558.