A Better Way to Assess the Economic Impact of the Basel Proposal

The Federal Reserve’s Vice Chair for Supervision, Michael Barr, has indicated that under the U.S. banking agencies’ Basel proposal, the average loan would see a cost increase of a mere 3 basis points. As a result, the Vice Chair contends that the proposal’s impact on borrowing costs will be negligible, and its benefits for financial stability will outweigh the costs.

This assessment is incomplete, because it relies solely on a simple average of the proposed changes in capital charges. Furthermore, the economic impact analysis in the proposal omits half of the $2 trillion increase in risk-weighted assets for operational risk. This omitted half, tied to fee income, is known as the services component of the operational risk capital charge.[1] A share of this services component is tied to lending-related business lines or functions; another share to trading activity; and the remainder to other sources of noninterest income that are part of financial intermediation from banks, and therefore cannot be excluded from any analysis of the proposal’s costs and benefits. A comprehensive evaluation of the capital proposal requires careful examination of each significant component of the proposal and its potential effects on each of a bank’s various business lines and intermediation functions, across lending, trading and other important financial intermediation activities.

In this note, we recommend ways the agencies can more comprehensively evaluate the Basel proposal’s impact on bank lending, trading and other financial intermediation activities, including by accounting for the omitted $1 trillion services component and by drilling down by business line. Not only banks’ lending but their trading and other financial intermediation activities are crucial to support economic growth and broader financial market liquidity. Moreover, the banking industry believes that the proposal will likely result in significant increases in costs for certain business lines, financial intermediation functions and borrower segments. These increases will particularly affect borrowers with limited ability to absorb them.

In other words, the current estimate of a 3-basis-point aggregate impact on lending rates has three major shortcomings, which a more comprehensive assessment must address. First, the estimate of lending impact needs to consider the lending related share of the missing $1 trillion RWA from the services component of the operational-risk capital-charge calculation (“op risk services component” for brevity). Second, the overall impact assessment needs to consider RWA effects on banks’ trading and other financial intermediation activities. Third, reporting only an aggregated impact masks key distinctions across business lines and intermediation functions, as well as across institutions.

BPI’s analysis, using publicly available data, indicates that about one-third of the RWA generated from the op risk services component is linked to lending-related business lines and functions. This includes revenues from credit cards, leases and loan commitment fees. The remaining two-thirds are related to trading activities and nonbanking services, such as asset management, securities underwriting and fees and commissions from securities brokerage and fiduciary activities. Although the large capital charge associated with the latter two-thirds of the services component may not directly affect lending, it could discourage banks from diversifying beyond pure lending services. Ultimately, the higher capital charge would still affect consumers because they would bear the increased costs of financial intermediation services.

Within the services component, the proportion of banking and non-banking services driven by differences in banks’ business models varies widely. For example, the lending-related share of the op risk services component will be close to 100 percent at banks specializing in credit card and auto lending. As a result, the proposed rule’s capital impact on lending at these institutions will be relatively large. With other banks that specialize in asset management, payments and custody services, customers will also feel the impact of the new rules driven by the increase in each bank’s cost of providing those services.

The economic impact analysis needs to follow the business practice of banks, where capital is allocated specifically to different business lines. If a bank’s return on equity for a particular business line fails to exceed its cost of capital (considering all changes in capital requirements), the bank may either pass the higher funding cost on to borrowers or reduce those exposures. This will result in the bank redirecting capital to other business lines or returning it to its shareholders.

Publicly available data offer only a basic method for categorizing fee income across business lines. This limits the ability of BPI or others relying on public data to accurately allocate the op risk services component among lending-related activity, trading and other nonbanking services. However, the agencies can leverage information from the Federal Reserve’s quarterly FR Y-14 regulatory reports to assess the impact of the RWA generated from the op risk services component, and we recommend they do so.

The FR Y-14 data collection gathers data on noninterest income by line of business. These data would allow the agencies to allocate the $1 trillion of RWA in the services component to the different business lines—mortgages, credit cards, small business lending, commercial lending, asset management, investment banking, custody services and sales and trading, and so on. The FR Y-14 data would therefore enable more accurate allocation of the op risk services component among lending-related activity, trading and other nonbanking services, allowing for more accurate assessment of the overall impact of the proposal on those activities.

Finally, it is important to emphasize that the agencies should consider reducing the risk weightings for certain business lines. Currently, some of these lines may be subject to excessively stringent capital requirements, especially when compared with their historical loss experiences. A comprehensive assessment of the proposal’s benefits and costs should acknowledge that a capital-neutral approach is inappropriate for business lines like mortgages, retail loans, small business lending and trading activities. These business lines often face disproportionately high capital charges once the impacts of operational risk and stress testing are considered, as detailed in previous BPI posts.[2]

The Proposal’s Economic Analysis Excludes a Key Component

The agencies have estimated the proposal’s effect on the lending and trading activities of covered banks. They did so by allocating the share of risk-weighted assets across lending and trading activities. The proposal includes changes to the calculation of capital requirements for four risk stripes: credit risk, market risk, operational risk and credit valuation adjustment (CVA) risk. Credit risk and a portion of operational risk were allocated to lending activities; market risk, CVA risk and a portion of operational risk were assigned to trading activities. However, as we explain in this section, the agencies did not allocate the $1 trillion increase in RWA from the services component to either lending or trading activities. This $1 trillion accounts for nearly 50 percent of the increase in aggregate RWA for all banks subject to the proposal, so it is obviously a highly material exclusion.

Although the agencies’ analysis is not fully transparent, we estimated how much of the proposed operational risk charge is tied to lending activities and how much to trading activities. The proposal states:

The agencies estimate risk-weighted assets associated with banking organizations’ lending activities would increase by $380 billion.[3] (page 64,169)

Since the agencies have estimated a $400 billion decline in risk-weighted assets for credit risk, we can implicitly estimate that the lending portion of operational risk is $780 billion. The $380 billion increase in risk-weighted assets represents a 3.5-percent rise, or a 0.3-percent increase in required capital. If we assume that the cost of equity is 10 percentage points higher than the cost of debt, this leads to a 3-basis-point increase in lending costs.

With respect to the impact of the proposal on trading activity, the proposal states:

The agencies estimate that the increase in RWA associated with trading activity (market risk RWA, CVA risk RWA, and attributable operational risk RWA) would be around $880 billion for large holding companies.[4] (page 64,170)

Applying a similar approach to trading as to lending, we can implicitly estimate that the trading-activity portion of operational risk is $172 billion. The results are summarized in Figure 1.

fig 1 - allocation of op risk rwa

However, despite the agencies estimating a $1,950 billion increase in risk-weighted assets due to the operational risk charge, they have omitted almost $1 trillion of this increase in their economic impact analysis. This unallocated $1 trillion is excluded from the agencies’ estimation of the effects on lending and trading due to the proposed rule. The omitted amount, which the agencies seem to associate implicitly with fee income linked to other banking services, represents nearly half of the estimated increase in aggregate risk-weighted assets. As a result, its significance is substantial.

Business Lines: A Better Way to Assess the Economic Impact of the Basel Proposal

Banks often compare the expected returns on various lines of business with the required capital to determine whether to expand or scale back a particular line. A common metric used is the risk-adjusted return on capital:[5]

risk adjusted return on capital

For example, the greater the required capital for a specific business unit, the lower its risk-adjusted return. Business lines with lower risk-adjusted returns tend to contract, while those with higher returns are likely to expand.

Typical business lines at banks include commercial lending, mortgage lending, credit card lending, small business lending, investment banking, private equity, trading, asset management and treasury services, among others. Ideally, the agencies’ impact analysis would assess the effects of the proposed changes on required capital, breaking down these effects in more detail, such as by individual business lines.

Analyzing the proposal’s effects at the business-line level is crucial for two reasons. First, focusing solely on the average capital effect across all business lines obscures important details about how the proposal will affect various lending categories. Previous BPI posts have raised significant concerns about specific lending segments such as mortgage loans and retail loans. Second, for the purpose of more fully and accurately accounting for RWA impacts from the services component of operational risk, we must analyze the sources of fee income that comprise the $1 trillion unallocated portion of operational risk. The rest of our discussion focuses on the latter issue.

Accounting for the Omitted Portion of the Op Risk Services Component

Public data on bank income are published in the quarterly FR-Y9C regulatory reports. Although these data present only a rudimentary breakdown of noninterest income across business lines, they are adequate for developing a rough allocation across lending-related, trading and other fee-generating activities. Toward this goal, we start by analyzing the components of noninterest income reported in Schedule HI of the FR Y-9C, which correspond to the currently unallocated (services) component that has been excluded from the agencies’ analysis. This helps to identify which categories of fee income are associated with lending-related business lines and which with trading and other intermediation-related activities. The operational-risk charge related to these activities, properly allocated, should be incorporated into the RWA economic impact assessment.

Seven items of noninterest income are included in the services component of operational risk, ranked here by order of importance, along with the item’s percent contribution to the total services component on average over the past three years:[6]

  • Other Noninterest Income (34 percent). This category includes all other operating income such as bank card and credit card interchange fees, lease income and loan commitment fees.
  • Investment Banking, Advisory Fees (26 percent). This category includes fees and commissions from underwriting securities, private placements of securities, investment advisory and management services, merger and acquisition services and other related consulting fees.
  • Fiduciary Activities (15 percent). This category includes gross income from services offered by the trust departments of banks.
  • Fees and Commissions from Securities Brokerage (15 percent). This category includes fees and commissions from securities brokerage activities, including from the sale and servicing of mutual funds, from the purchase and sale of securities and money market instruments where the bank is acting as agent for other banking institutions or customers, and from the lending of securities owned by the bank.
  • Service Charges on Deposits Accounts (7 percent). This category includes amounts charged to depositors in domestic offices.
  • Servicing Fees (1 percent). This category includes income from servicing real estate mortgages, credit cards and other financial assets held by others, or any other insurance activities net of expenses.
  • Insurance Fees (1 percent). This category includes income associated with premiums earned by holding company subsidiaries engaged in insurance underwriting or reinsurance activities, and income from insurance product sales.

Clearly, the bulk of fee income related to lending-related business lines and functions would be contained in the other non-interest income category. An additional substantive share of lending-related fee income would derive from the investment banking category. This is because banks help firms of all sizes in raising equity and debt from market investors, categorizing these services as lending-related activities. In addition, these fees are connected to trading activities, since banks often make markets and hold these securities in their portfolios for this purpose.

figure 2 - share of fee income across business lines by bank

Figure 2 illustrates the distribution of fee income across these seven categories, as reported in Schedule HI by various banks. Note the wide variation in bank shares across different components of noninterest income. For instance, credit card banks and those focusing on auto lending report a share of other noninterest income in the op risk services component ranging from 82 to 99 percent. In contrast, universal banks and domestic regional banks, which typically have substantial loan portfolios, report a share of other noninterest income in the op risk services component from 35 to 60 percent, with one universal bank being an exception. On the other hand, trading banks often report a share for investment banks and advisory fees ranging from 60 to 80 percent. And custodian banks report a share in fiduciary revenues that varies from 80 to 95 percent.

How Much of Other Noninterest Income Is Related to Lending?

The “other noninterest income” category accounts for about one-third of all fee income in the op risk services component of the operational-risk charge. To gain a better understanding of this broad category, we analyze the memoranda items in Schedule HI of the FR Y-9C. This item includes seven standardized subcategories, including credit card interchange revenues.[7] Banks are required to report any item of other noninterest income that exceeds $100,000 and represents more than 7 percent of their noninterest revenues. They must also report any other revenue sources not included in the seven standardized subcategories, provided these revenues exceed the set regulatory thresholds. In such instances, the language used in the write-in revenue fields is not standardized, and each bank words it individually.

Among the standardized items, “Bank card and credit card interchange fees” account for about 42 percent of other noninterest income. The remaining six items, which total 21 percent of other noninterest income, are not lending-related, rarely exceed the regulatory reporting thresholds, and therefore are not itemized. The write-in items make up about 37 percent of other noninterest income.

Capital charges tied to bank card and credit card interchange fees will directly affect the pricing and profitability of both consumer and small business credit alongside capital charges for credit risk. With credit cards, the combined capital charge for credit and operational risks are likely to be passed on to consumers through higher annual fees, increased interest rates, reduced card rewards or other diminished benefits. Alternatively, a portion of the increased capital charges might be passed on to small businesses in the form of higher fees for payments and other services. This could disrupt existing banking relationships by driving small-business customers away, thereby also affecting access to credit for small businesses. Furthermore, some issuers may find that certain consumer segments or small businesses can no longer be served profitably due to reduced card usage prompted by these changes.

As for the write-in items, to assess their ties to lending-related business lines and functions, we first classify them into more standardized categories amenable to such an analysis:

  • Credit Card Fees (12 percent of other noninterest income). This category includes net revenues associated with card fees, annual credit card fees, transaction processing revenues and other credit-card-related fees.
  • Operating Lease Fees (8 percent of other noninterest income). This category includes operating lease revenues, lease income, rent on operating leases to others and other lease income.
  • Loan Commitment Fees (7 percent of other noninterest income). This category includes syndication underwriting fee income, mortgage loan fees, net of broker premiums, fees on loans and lending commitments and other noninterest fees on loans.
  • Income from Affiliates (4 percent of other noninterest income). This category includes affiliate service charges, income from dividends, fees and reimbursements from affiliates and revenue transfers.
  • Change in Valuations (2 percent of other noninterest income). This category includes a broad range of items, namely FX losses on hedged items, write-downs due to credit spreads, net change in fair value of derivatives and changes in the fair value of mortgage loans held for sale.
  • Miscellaneous (4 percent of other noninterest income). This category includes all other write-in items that are not classified above.

In this classification, about 25 percent of the other noninterest income category (from the first three groups listed) is directly linked to lending. Since credit cards are generally considered a business line with a lending focus, the 42 percent associated with credit card interchange fees are to be added on, so that in total 67 percent of the other noninterest income category is connected to lending-related business lines and functions.

Furthermore, servicing activities are generally related to lending. However, since servicing fee revenues are reported net of fees, whereas the op risk services component requires gross amounts for calculation, estimating the size of gross servicing fees in the services component is challenging. Recharge income should also be excluded from the services component because it applies only to subsidiaries, specifically U.S. subsidiaries of foreign banks.

And as noted, among the other components of noninterest income, investment banking and advisory fees are also closely related to lending in addition to being tied to trading activities; in the absence of more granular data, we shall presume an even split. In sum, the lending-related share of the op risk services component is two-thirds of the other-noninterest-income category, or about 23 percent, plus half the investment banking share, or 12 percent. Based on this analysis, about one-third of the op risk services component could reasonably be attributed to lending-related business lines and functions. The other two-thirds are tied to trading activity and other sources of fee income. However, the level of disaggregation provided in public reports poses challenges for achieving a more detailed level of disaggregation beyond what is accomplished in our analysis.

The Fed’s Stress-Testing Data: The Answer to a Better Economic Impact Analysis

Fortunately, the agencies do not have to rely on FR Y-9C data to assess the changes in capital requirements across business lines. The Federal Reserve collects more detailed information on noninterest revenues through its confidential data on the stress tests. As shown in Figure 3, the quarterly FR Y-14 regulatory data collection gathers data on noninterest income by line of business. For instance, these data only allow the agencies to allocate the $1 trillion of RWA in the op risk services component to the various business lines, such as mortgages, credit cards, small business lending and commercial lending.

Figure 3: FR Y-14Q Regulatory Report—Noninterest Income Schedule

fig 3 - fr y - 14q reg report
fig 3 - fr y - 14q reg report continued

Previous BPI research examined the effects of the proposal on residential mortgages and card credit lines, demonstrating substantial increases in risk weights in some risk segments, driven largely by the new operational risk charge. The increased capital charges will adversely affect the price and availability of credit within these segments, with a disproportionate impact on access to credit among low- and moderate-income households.

A final important point is that the U.S. proposal imposes risk-weight surcharges, relative to the international Basel III agreement, for which there is no underlying risk justification (20 percentage points in each mortgage LTV category, and 10 percentage points for credit card and other retail). These surcharges apparently reflect the goal that the overall change in capital requirement for these products be neutral. As a general proposition, that notion is misguided. The agencies should consider reducing the risk weights from current levels for some business lines.

Moreover, in the United States, the annual stress tests bring more risk sensitivity to banks’ capital requirements. They offer added assurance that banks hold sufficient capital to absorb extreme downturn losses. 

Conclusion

In this note, we recommended ways the agencies can more comprehensively assess the impact of the Basel proposal on banks’ lending activities. The focus of these recommendations is on accounting for the lending activity share of the $1 trillion services component of the operational-risk capital-charge calculation, currently missing from the agencies’ published estimate of overall capital impact. This is best addressed through a granular analysis of the op risk services component by business line. We have used the FR Y-9C reports to show how such an analysis can be conducted. Our illustrative calculations also document the materiality of the missing $1 trillion component, since we estimate that almost half of the op risk services component is lending-related. Fortunately, the agencies have access to the more detailed FR Y-14Q data, which present revenue information broken down by line of business.

In actual practice, banks allocate capital at the business-line level, so it’s essential to have a business-line focus when appropriately assessing the effects of increased capital charges on lending activity. If a bank’s return on equity for a particular business line fails to exceed its cost of capital (after factoring in all changes to capital requirements), the bank will either pass the added funding cost on to borrowers or reduce those exposures, redirecting capital to other business lines or its shareholders.

Performing a revised economic analysis that addresses the deficiencies in the proposal’s current analysis will lead to a significant increase in its estimated costs and show that the original analysis greatly understated these costs. Not only would this overcapitalization fail to achieve a clear benefit that would outweigh the associated costs, it would actually discourage banks from diversifying from pure lending services.


[1] See Francisco Covas, “The Trillion Dollar Omission in Vice Chair Barr’s Cost Analysis,” Bank Policy Institute (October 12, 2023), available at https://bpi.com/the-trillion-dollar-omission-in-vice-chair-barrs-cost-analysis/

[2] See Paul Calem and Francisco Covas, The Basel Proposal: What It Means for Retail Lending, Bank Policy Institute (Nov. 8, 2023), available at https://bpi.com/the-basel-proposal-what-it-means-for-retail-lending/; and Paul Calem and Francisco Covas, The Basel Proposal: What it Means for Mortgage Lending, Bank Policy Institute (Sep. 30, 2023), available at https://bpi.com/the-basel-proposal-what-it-means-for-mortgage-lending/.

[3] See Federal Register, Regulatory Capital Rule: Large Banking Organizations and Banking Organizations With Significant Trading Activity, Vol. 88, No. 179 (September 18, 2023), available at https://www.govinfo.gov/content/pkg/FR-2023-09-18/pdf/2023-19200.pdf.

[4] See note 3.

[5] See John C. Hull, Risk Management and Financial Institutions, 6th ed., pp. 599–616. New York: Wiley, 2023.

[6] Servicing fees are reported by deducting expenses from revenues. The proposal mandates that banks capitalize based on gross servicing fees. Therefore, servicing fees will account for more than 1 percent of the op risk services component. Moreover, these data are available in the quarterly FR Y-14 regulatory reports, but this information is confidential.

[7] The seven itemized items in other noninterest income are: “Income and fees from the printing and sale of checks,” “Earnings on/increase in value of cash surrender value of life insurance,” “Income and fees from automated teller machines (ATMs),” “Rent and other income from other real estate owned,” “Safe deposit box rent,” “Bank card and credit card interchange fees,” and “Income and fees from wire transfers.”