Economic Capital and

the Assessment of Capital Adequacy

5Supervisory Insights Winter 2004

The assessment of capital adequacyis one of the most critical aspectsof bank supervision. In completing

this assessment, examiners focus on acomparison of a bank’s available capitalprotection with its capital needs basedon the bank’s overall risk profile.

Bank management must likewisecontinuously evaluate capital adequacyin relation to risk. In recent years, manybanks have adopted advanced modelingtechniques intended to improve theirability to quantify and manage risks.These modeling techniques frequentlyincorporate the internal allocation of“economic capital” considered neces-sary to support risks associated withindividual lines of business, portfolios,or transactions within the bank. As aresult, economic capital models canprovide valuable additional informationthat bankers and examiners can use intheir overall assessment of a bank’scapital adequacy.

As will be discussed later, economiccapital models or similar risk and capi-tal adequacy assessment processes areimportant to banks adopting therevised Basel framework. But revisionsto capital regulations have not beenthe driving force behind the develop-ment of these models as such method-ologies have been in use for more thanten years at some of the nation’slargest banks. Economic capital hasalso become a useful and sometimesnecessary tool for other insured insti-tutions. Several regional banks andsome community banks have devel-oped or are exploring implementationof economic capital models with morebanks likely to do so in the future.This article provides an introductionto the concept of economic capital,describes the relationship betweeneconomic capital and the revised Baselframework, and discusses examinerreview of economic capital models as

a part of the supervisory assessment ofcapital adequacy.

Economic Capital

Economic capital is a measure of risk,not of capital held. As such, it is distinctfrom familiar accounting and regula-tory capital measures. The output ofeconomic capital models also differsfrom many other measures of capitaladequacy. Model results are expressedas a dollar level of capital necessaryto adequately support specific risksassumed. Whereas most traditional meas-ures of capital adequacy relate existingcapital levels to assets or some form ofadjusted assets, economic capital relatescapital to risks, regardless of the exis-tence of assets. Economic capital is basedon a probabilistic assessment of potentialfuture losses and is therefore a poten-tially more forward-looking measureof capital adequacy than traditionalaccounting measures. The developmentand implementation of a well-function-ing economic capital model can makebank management better equipped toanticipate potential problems.

Conceptually, economic capital canbe expressed as protection againstunexpected future losses at a selectedconfidence level. This relationship ispresented graphically in Chart 1 (seenext page).

Expected loss is the anticipated aver-age loss over a defined period of time.Expected losses represent a cost ofdoing business and are generallyexpected to be absorbed by operatingincome. In the case of loan losses, forexample, the expected loss should bepriced into the yield and an appropriatecharge included in the allowance forloan and lease losses.

Unexpected loss is the potential foractual loss to exceed the expected loss

6Supervisory Insights Winter 2004

and is a measure of the uncertaintyinherent in the loss estimate.1 It is thispossibility for unexpected losses to occurthat necessitates the holding of capitalprotection.

Economic capital is typically definedas the difference between some givenpercentile of a loss distribution and theexpected loss. It is sometimes referredto as “unexpected loss at the confi-dence level.”

The confidence level is established bybank management and can be viewed asthe risk of insolvency during a definedtime period at which management haschosen to operate. The higher the confi-dence level selected, the lower the proba-bility of insolvency. For example, ifmanagement establishes a 99.97 percentconfidence level, that means they areaccepting a 3 in 10,000 probability ofthe bank becoming insolvent during thenext twelve months. Many banks usingeconomic capital models have selecteda confidence level between 99.96 and

99.98 percent, equivalent to the insol-vency rate expected for an AA or Aacredit rating.

The primary value of economic capitaland the reason that banks have alreadyadopted such methodologies is its appli-cation to decision making and riskmanagement. Specifically, the use ofsuch models can:

■ contribute to a more comprehensive

pricing system that covers expectedlosses,

■ assist in the evaluation of theadequacy of capital in relation to thebank’s overall risk profile,

■ develop risk-adjusted performance

measures that provide for better eval-uation of returns and the volatility ofreturns,2 and

■ enhance risk management efforts byproviding a common currency for risk.

The following example illustrates howeach of these potential uses could beapplied at a bank.3 This example

1Unexpected loss is often described as the volatility of loss around the average over time. 2Risk-adjusted performance is typically measured at the business unit level, but can also be used to evaluatehow individual business unit returns contribute to a bank’s overall profitability and risk profile. 3Specific methodologies, such as the use of a default-only measure of credit risk discussed in the example,should be viewed as potential approaches rather than as the only or best alternative.

Freq

uenc

y of

Los

s

Expected Loss Confidence Level

Economic Capital

Amount of Loss (increasing to the right)

Chart 1. Economic Capital

Economic Capitalcontinued from pg. 5

7Supervisory Insights Winter 2004

describes only credit risk quantificationand its translation to economic capitalfor commercial lending activities. Obvi-ously risks are evident in activities otherthan commercial lending, and commer-cial lending itself involves numerousrisks in addition to credit risk.4 Banksthat use economic capital models gener-ally identify and quantify all types of riskacross all lines of business throughoutthe bank.

Example: Economic Capital Allocation for Commercial Credit Risk

At its most fundamental level, creditrisk is associated with loan losses result-ing from the occurrence of default andthe subsequent failure to collect in fullthe balances owed at the time of default.5

Expected credit losses associated withdefault can therefore be determinedfrom parameters associated with the like-lihood of a loan defaulting, or an esti-mate of the probability of default (PD)during a defined time period, and theseverity of loss expected to be experi-enced in the event of a default, or anestimate of loss given default (LGD).Naturally, this ratio would be applied to ameasure of estimated exposure at default(EAD) to convert loss expectations todollar amounts. The resulting formula:

Expected losses ($) = PD(%) * LGD(%) * EAD($).

PD and LGD parameter estimatesare drawn from the bank’s historicalperformance or from a mapping ofinternal portfolio risk assessments to

external information sources for PDand LGD parameters. This requiresthat banks have in place processesthat enable them to periodically assesscredit risk exposures to individualborrowers and counterparties withrobust internal credit rating systemsthat reflect implicit, if not explicit,assessments of loss probability. Defini-tions of credit grades should be suffi-ciently detailed and descriptive toclearly delineate risk level betweengrades and should be applied consis-tently across all business lines.

For example, a bank could have a tengrade credit rating system with associ-ated one-year probabilities of defaultdrawn from their historical default expe-rience within each grade as shown inTable 1 (next page). In this example, thehistorical default rate experienced forloans internally graded as a “6” hasbeen one percent, which is approxi-mately equivalent to the long-termdefault frequency associated with anS&P credit rating of BB.

Estimates for loss severity in the eventof default could likewise be constructed.LGD grades assigned to loans are oftenassociated with factors such as loan type,collateral type, collateral values, guaran-tees, or credit protection such as creditdefault swaps.6

Pricing Implications: A credit facilitywhich is the same in all other respectsmay be priced differently based on itsexpected loss.7 Table 2 shows expectedlosses for three different borrowers withthe same loan structure and collateral

4Such as interest rate risk and operational risk associated with underwriting and servicing of loans. 5The example describes a default-only perspective to derive a loss distribution; i.e., loan defaults create creditlosses. Some banks have adopted a more robust perspective for credit loss which considers the probabilitydistribution of obligor grade migration and resulting changes in the economic value of the loan; i.e., a decline inthe credit quality of a loan regardless of any default creates credit losses. 6Some banks consider guarantees and credit protection as substitutes for the borrower and therefore use guar-antor or counterparty PDs in place of borrower PDs, while other banks retain the borrower PD and considerguarantees and credit protection in determining LGD. 7Pricing models are considerably more complex than the simplistic approach shown in this example. This discus-sion is merely intended to show that expected losses are often built into the pricing of loans.

8Supervisory Insights Winter 2004

support resulting in a 40 percent lossseverity in the event of default. Thehigher risk credit grade has five timesthe expected loss of the lower riskcredit grade.

If a bank made middle market loanswhich fell into the three grade bandsshown in Table 2, but priced most ofthese loans with an implicit loss expec-tation of 50 basis points, the bank isovercharging stronger borrowers andundercharging weaker borrowers. One

potential result is that the bank couldend up with stronger borrowers exitingthe bank and find its loan pool progres-sively weaker and portfolio returns inade-quate for losses experienced.

Although such a highly quantitativeprocess may appear somewhat foreign tomany bankers, a form of probability ofdefault estimates is considered in the use

of consumer FICO scores or banks’ owninternal loan scorecards. Furthermore,many banks, including many communitybanks, are already relating this type ofanalysis to their allowance for loan andlease loss determination.

Capital Adequacy: The allocation ofeconomic capital to support credit riskbegins with similar inputs to deriveexpected losses but considers otherfactors to determine unexpected losses,such as credit concentrations anddefault correlations among borrowers.Because borrower defaults are notperfectly correlated, the default riskof a credit portfolio is less than the sumof the risks contained in the underlyingloans. Economic capital credit riskmodeling therefore measures the incre-mental risk that a transaction adds to aportfolio rather than the absolute levelof risk associated with an individualtransaction. Complex models arerequired to derive this measure of port-folio loss volatility and translate that intoan associated economic capital charge.

Table 3 shows an example of credit riskeconomic capital allocations (credit riskonly) determined using the PD and LGDparameters previously discussed and amodel translation of those parametersinto a credit risk capital charge.8 The

8The credit economic capital allocations shown in the table were derived using the regulatory capital calculationfor corporate credit exposures under the revised Basel framework. Refer to International Convergence of CapitalMeasurement and Capital Standards, June 2004 text, Basel Committee on Banking Supervision. As discussedlater in this article, the regulatory capital calculation under the revised framework differs in important ways fromeconomic capital methodologies, but is used for illustrative purposes in this example as a proxy for an economiccapital methodology to avoid disclosing information about proprietary models used by any bank. The tableincludes nine obligor grades and nine facility grades; the tenth borrower grade previously discussed was fordefaulted loans and is not shown as the methodology for estimating risk in defaulted exposures varies consider-ably among institutions.

Table 1

Internal Loan Grade 1 2 3 4 5 6 7 8 9 10Average Probability of Default 0.03% 0.06% 0.10% 0.25% 0.50% 1.00% 2.50% 8.00% 22.00% 100.00%Mapping to External Ratings AA A BBB+ BBB BB+ BB B+ B CCC D

Note: The mapping to equivalent external credit ratings, in this case S&P ratings, is an approximation and provided here only for reference.

Example Obligor Grades and Associated Default Probabilities

Table 2

Loan Grade PD * LGD = Loss5 0.50% 40% 20 basis points6 1.00% 40% 40 basis points7 2.50% 40% 100 basis points

Expected Loss

Economic Capitalcontinued from pg. 7

9Supervisory Insights Winter 2004

bank’s obligor grades and associatedPDs are shown at the top of this table.The bank’s facility grades and associ-ated loss severity estimates are shownon the left-hand side of the table. Theassociated capital charges representthe dollar amount of capital neededto support a $100 one-year maturitycommercial loan based on parameterinputs (such as the PD estimate) andmodel assumptions (such as defaultcorrelations).

Credit economic capital allocations fora non-defaulted $100 one-year maturitycommercial loan using this modelwould range from as low as 13 cents toas high as $36.43. Everyone intuitivelyexpects increased risk to be associatedwith lower-quality graded loans or loanswith higher loss severity, but the alloca-tion of economic capital estimates thelevel of risk associated with a particulargrade band and differentiates riskamong bands.

For example, commercial loans gradedas a 5 or a 6 with an LGD of 40 percentin the table above would not likely besubject to regulatory classification orcriticism; i.e., both credits would be“pass” credits. However, the economic

capital allocations show a considerabledifference in the inherent risk betweenthese loans. A $100 one-year maturitycommercial loan that is graded a 6would receive a $5.21 credit economiccapital allocation compared with a$3.71 allocation for a similar loangraded 5, an approximately 40 percentincrease in estimated risk.

Risk-Adjusted Performance Measures:Economic capital is also used to evalu-ate risk-adjusted performance; withoutsome quantification of risk associatedwith an activity, it is not possible tomeasure performance on a risk-adjustedbasis. Several techniques have beendeveloped with two such approachesthat incorporate economic capitalallocations demonstrated below:

■ Risk Adjusted Return On Capital(RAROC), a percentage measureof performance = Economic NetIncome / Economic CapitalAllocation

■ Economic Profit, or ShareholderValue Added (SVA), a dollar measureof performance = Economic NetIncome – (Economic CapitalAllocations * Hurdle Rate)9

9The hurdle rate can be viewed as the firm-wide cost of capital. Returns above the hurdle rate add to share-holder value and those below, while perhaps profitable, detract from shareholder value.

Table 3

Facility Grades Obligor Grades and Associated Default Probabilitiesand Associated 1 2 3 4 5 6 7 8 9Loss Given Default 0.03% 0.06% 0.10% 0.25% 0.50% 1.00% 2.50% 8.00% 22.00%

A 10% $0.13 0.23 0.33 0.62 0.93 1.30 1.84 2.84 4.05B 20% 0.27 0.46 0.66 1.23 1.85 2.61 3.67 5.69 8.10C 30% 0.40 0.69 1.00 1.85 2.78 3.91 5.51 8.53 12.14D 40% 0.54 0.91 1.33 2.46 3.71 5.21 7.35 11.38 16.19E 50% 0.67 1.14 1.66 3.08 4.64 6.51 9.18 14.22 20.24F 60% 0.81 1.37 1.99 3.70 5.56 7.82 11.02 17.06 24.29G 70% 0.94 1.60 2.32 4.31 6.49 9.12 12.86 19.91 28.33H 80% 1.08 1.83 2.66 4.93 7.42 10.42 14.69 22.75 32.38I 90% 1.21 2.06 2.99 5.55 8.35 11.72 16.53 25.60 36.43

Example Economic Capital Allocations ($) for $100 1-Year Maturity Commercial Loan

10Supervisory Insights Winter 2004

Assume that a bank is considering theperformance of two loan portfolios:Portfolios X and Y, with Portfolio Xassumed to be higher risk and produc-ing a higher return relative to Portfolio Y(see Table 4). Using internal gradingparameters and economic capitalmodeling for credit risk, managementcan strengthen its evaluation of the riskreturn trade-off of the two portfolios.10

Please note, this example considersonly credit risk. Bank managementwould incorporate assessments of otherrisks in determining risk-adjustedperformance.

Initially, bank management may havebeen inclined to select Portfolio X,based on simple return characteristics,as shown below. On a risk-adjustedbasis, however, Portfolio Y is thepreferred alternative. Although Portfo-lio X produces higher expected bookand economic net income, the volatil-ity of Portfolio X’s return (i.e., risk)is not adequately compensated for incomparison to Portfolio Y. Portfolio Ygenerates a higher RAROC and resultsin a greater economic profit on asignificantly lower economic capitalallocation.11

10In many banks, risk-adjusted performance measures are built into the determination of compensation for lineof business managers and staff, directly influencing behavior at the business line level. Often, both a dollarlevel of risk-adjusted performance, such as SVA, and a percentage measure, such as RAROC, are used.

• Percentage measures of performance are often used because dollar measures may not providesufficient information to distinguish between alternative acceptable investments. For example, twoportfolios could produce the same dollar measure of risk-adjusted performance, but one could requiresubstantially larger capital allocations.

• Dollar measures of performance may be used because managers might be inclined to reject an invest-ment that would generate positive SVA if that investment generated a RAROC that was lower than theirexisting business line RAROC. For example a manager might choose to reject an otherwise desirableinvestment with a 20 percent RAROC if his line of business had an average RAROC of 25 percent.

11Note that Portfolio X and Portfolio Y, when considering credit risk only, would be acceptable to managementas both generate positive economic profit assuming a hurdle rate of 10 percent.

Table 4

Portfolio X Portfolio YPortfolio Balances $100,000,000 $100,000,000Net Income before Losses* $1,400,000 $1,100,000Loan Parameters:- PD 0.50% 0.25%- LGD 50% 40%- EL (in bps) 25 10

Expected Losses $250,000 $100,000Income after Expected Losses $1,150,000 $1,000,000Economic Capital (credit only)** $4,640,000 $2,460,000RAROC 24.8% 40.7%Economic Profit (10% hurdle rate) $686,000 $754,000

* Net income before losses = loan interest + fees + soft dollars - funding costs - operating costs.

** Determined from the economic capital charges shown in Table 3.

Example Risk-Adjusted Performance Measures

Income after Expected Losses $1,150,000 $1,000,000Flat Capital Charge (e.g., 8%) $8,000,000 $8,000,000Return on Equity 14.4% 12.5%

Economic Capitalcontinued from pg. 9

11Supervisory Insights Winter 2004

Although the decision reached in thisexample resulted in lower overall creditrisk, economic capital models are notdesigned to always favor strategies thatproduce lower risk. Economic capitalshould be viewed as a tool to enhancerisk identification and selection. Deci-sions resulting in the acceptance ofhigher credit risk can be expected tooccur when supported by transactionlevel returns that compensate for higherrisk or increased portfolio diversifica-tion benefits.

Risk Management: The implications forloan pricing, capital analysis, and risk-adjusted performance measures relatedirectly to risk management, buteconomic capital, as a common currencyof risk, can provide additional potentialapplications to the risk managementprocess. For example, some banks usecredit economic capital allocations inplace of or in addition to more tradi-tional credit hold limits based onnotional exposures which may not fullycapture factors such as potential lossseverity, default correlations with therest of the credit portfolio, or maturityeffects on default probability.12

The preceding example focused oncredit risk. But similar assessment andquantification efforts can help banksidentify, monitor, and manage otherrisks in other lines of business as well.

The effectiveness of a bank’s riskmanagement practices is an importantconsideration in the supervisory evalu-ation of an institution and directlyinfluences the regulatory assessmentof capital adequacy. Strong riskmanagement practices can compensatein part for higher levels of inherent

risk in a bank’s business activities.Recognizing this relationship, therevised Basel capital frameworkpromotes the adoption of stronger riskmanagement practices throughout thebanking industry by incorporatingindustry advances in risk modeling andmanagement into regulatory capitalrequirements.

Economic Capital and Basel II

The revised Basel framework seeks tocreate more risk-sensitive regulatorycapital requirements in order to addressconcerns that the regulatory capitalmeasures established by the 1988 BaselAccord do not adequately differentiaterisk, and to reduce regulatory capitalarbitrage activities which have erodedthe relevance of current risk-based capi-tal measures at some institutions. Manyindustry participants and observers haveassociated economic capital with thecalculation of minimum regulatory capi-tal requirements under the first pillar ofthe revised framework and the supervi-sory review process under the secondpillar. As discussed below, however,economic capital and regulatory capitalunder the revised framework are notsynonymous.

The First Pillar—MinimumCapital Requirements

The calculation of minimum regulatorycapital under the revised frameworkrelies heavily on certain inputs from thebank’s assessment of its individual riskprofile. For example, the calculation ofthe capital charge for credit riskconsiders the distribution of a bank’sspecific credit exposures among inter-

12Intuitively, longer maturity loans to the same borrower entail greater credit risk; i.e., the default risk of a fiveyear loan to a borrower, even a borrower of strong credit quality, is significantly greater than a six-month loan tothe same borrower. However, traditional credit hold limits, such as notional exposures by loan grade, rarelycapture this maturity effect. Credit economic capital allocations frequently adjust the one-year PD estimates foran obligor to reflect the differences in credit risk resulting from facility maturity.

12Supervisory Insights Winter 2004

nally assigned PD and LGD grades. Thetranslation of that risk profile into acapital charge, however, is consistentfor all institutions. The required inputsfollow specific slotting criteria and areapplied against regulatory risk weightcurves which are the same for all institu-tions.13 This need to ensure consistencynecessarily creates differences betweena bank’s internal capital allocations andthe minimum regulatory capital charge.

Potential differences also exist inthe inputs used. For example, in itseconomic capital model, a bank may usea long-term estimate of LGD that coversall economic cycles, but for regulatorycapital purposes, the LGD estimateshould reflect economic downturn condi-tions for exposures where loss severitiesare expected to vary substantially witheconomic conditions.14

More fundamentally, the risks capturedunder regulatory and economic capitaldiffer. The regulatory capital chargecaptures only credit, market, and opera-tional risk. Furthermore, the regulatorycapital calculation does not fully addresscertain aspects of these risks, such ascredit concentration risk. As previouslydiscussed, economic capital modelsgenerally address all risks arising fromthe bank’s business activities.

Economic capital also typically incorpo-rates a diversification benefit which isnot considered in the regulatory capitalcalculation. This diversification benefit isa top-line measure of how changes in therisk associated with each business activ-ity occur in relation to changes in risk inall other activities.

Chart 2 (next page) provides a graphicexample of some of the potential differ-ences between regulatory capital underthe revised Basel framework andeconomic capital at a hypothetical bank.In this example, total economic capitalallocations are higher than the regula-tory minimum capital charge. While thistypically may be expected to be thecase, in some instances a bank couldreasonably have lower economic capitalallocations than regulatory capitalrequirements depending on the specificrisk characteristics of the bank and thesignificance of the diversification benefit.

As demonstrated by the above discus-sion, a bank is not required to have a fullyfunctional economic capital model todevelop the necessary inputs for thecalculation of the minimum regulatorycapital charge. These inputs generally canbe determined independent of anycomprehensive risk measurement andmanagement process. However, thesecond pillar of the revised frameworkcreates a more direct link to a bank’s ownrisk and capital adequacy assessments.

The Second Pillar—Supervisory Review Process

The second pillar establishes a regula-tory expectation for the evaluation ofhow well banks assess their own capitalneeds. The second pillar does not explic-itly require banks to adopt economiccapital models. It does, however, estab-lish an expectation for banks to performa comprehensive assessment of the risksthey face and to relate capital adequacyto these risks.15

13The regulatory risk-weight curves serve as a proxy for default correlations, with the expected default experi-ence among weaker commercial borrowers (credits with higher PDs) assumed to be less correlated withsystemic risk (overall economic conditions).14Paragraph 468 of the revised Basel framework. International Convergence of Capital Measurement and CapitalStandards, June 2004, Basel Committee on Banking Supervision. 15Paragraph 732 of the revised Basel framework: “All material risks faced by the bank should be addressed in thecapital assessment process. While the Committee recognizes that not all risks can be measured precisely, aprocess should be developed to estimate risk.” International Convergence of Capital Measurement and CapitalStandards, June 2004, Basel Committee on Banking Supervision.

Economic Capitalcontinued from pg. 11

13Supervisory Insights Winter 2004

Furthermore, the bank’s own capitalanalysis is expected to encompass allrisks, not only those risks captured bythe minimum regulatory capital calcula-tion. The revised Basel frameworkdescribes three areas not addressed inthe minimum capital calculation thatshould be specifically considered underthe second pillar:

■ Risks that are not fully capturedunder the first pillar, such as creditconcentration risk

■ Risks that are not considered underthe first pillar, such as interest raterisk, and

■ Factors external to the bank, such aseconomic conditions.16

The supervisory qualification and on-going validation of a bank’s compliancewith regulations implementing therevised framework will necessarily incor-porate review of a bank’s risk quantifica-tion efforts and capital analysis. Whilethere is no supervisory requirement foreconomic capital methodologies to beemployed in this process, many largeinstitutions appear likely to use theireconomic capital models to demonstratecapital adequacy in relation to risk underPillar 2.

16Paragraph 724, International Convergence of Capital Measurement and Capital Standards, June 2004, BaselCommittee on Banking Supervision.

Example Comparison of Minimum Regulatory Capital with Economic Capital

Minimum Regulatory Capital

Capital ($ billions)

Regulatory Model

Model Drivers

Other Model Differences:- Different confidence levels used- Variations in input data

-10

-5

0

5

10

15

20

25

30

35

Funding sources anduses stress scenarioanalysisMeasure ofpotential earningsvolatilityEconomic Valueof Equity (EVE)resultsValue-at-Risk overa liquidationperiod plus stressscenario analysisFrequency andseverity lossdistributions andother factorsPD, LGD, EAD, andmaturity as imputs;observed correlationsused; creditconsiderationsconsidered; credit lossesrelated to changes ineconomic value

Capital $21

Market Risk4

Operational Risk5

Credit Risk12

PD and LGD bands, EAD,and some maturity dataas inputs; regulatoryrisk curves used tocapture correlations;credit losses relatedto default

Frequency and severityloss distributions andother factors

10-day Value-at-Riskplus specific riskcharges

Model Drivers

Economic Capital

Bank Model

Capital $25

Liquidity Risk4

Business Risk4

Inerest Rate Risk4

Market Risk5

Operational Risk5

Credit Risk10

DiversificationBenefit

-7

Vector analysis ofrisk correlations

Chart 2

14Supervisory Insights Winter 2004

Supervisory Review ofEconomic Capital

Regulators expect certain large orcomplex banks to perform appropriaterisk quantification and capital analysisregardless of whether the bank is subjectto the revised Basel framework. This isparticularly important at banks wheremore traditional capital adequacy meas-ures may not adequately capture theinherent risk of their business activities,such as at banks heavily engaged insecuritization activities. An economiccapital model is one tool available forsuch analysis.

At banks where economic capitalmodels are used, considerable supervi-sory effort is focused on the process.Examiners consider both the adequacyof economic capital processes and theresults of such processes in their supervi-sory evaluation of the bank. Further-more, as discussed later in this article,examiners may find it beneficial tomodify certain traditional examinationprocedures to more fully evaluate riskmanagement practices associated withthe economic capital process and otherrisk modeling techniques.

Process Review

When properly used, economic capitalmodels can improve risk managementand the evaluation of capital adequacy.However, these models can suffer fromdata limitations, erroneous assumptions,inability to sufficiently quantify risks,and potential misuse or misunderstand-ing of model outputs. Examiner assess-ment of the appropriateness of a bank’scapital adequacy analysis, potentiallyincluding economic capital methodolo-gies, can be a consideration in thesupervisory evaluation and rating ofbank management. Institutions foundto have material weaknesses in theirmethodologies may be directed tostrengthen risk measurement andmanagement capabilities.

The supervisory approach used to eval-uate a bank’s economic capital processwill necessarily vary based on thecomplexity of the institution and theextent of use of the economic capitalprocess by bank management. Examina-tion guidance on economic capitalmodels is limited. Federal Reserve BoardSupervisory Letter SR 99-18 and thesecond pillar of the revised Basel frame-work do not specifically addresseconomic capital methodologies, butboth documents describe the supervisoryreview of a bank’s capital analysisprocess. Many of the principles discussedin these documents are included in thegeneral review concepts examiners maywant to consider that are discussedbelow.

Evaluate the adequacy of board andmanagement oversight concerningeconomic capital. Management isresponsible for understanding the natureand level of risks undertaken in thebank’s activities and how these risks fitwithin the overall business strategy ofthe bank. To evaluate this oversight,examiners could review:

■ specific board approval of risk toler-ances and associated capital levels

■ periodic economic capital reportsprovided to the board and seniormanagement. Such reports shouldbe sufficient to allow the board andmanagement to evaluate risk expo-sures, determine that the bank holdssufficient capital relative to identifiedrisk, and incorporate capital needsinto the strategic planning process.

Determine that economic capitalmethodologies appropriately incorpo-rate all material risks. At a minimum,this should include assessments of credit,market, operational, liquidity, and busi-ness risks. To make this determination,examiners could review:

■ a mapping of data inputs to materialexposures, ensuring accuracy andcompleteness

Economic Capitalcontinued from pg. 13

15Supervisory Insights Winter 2004

■ documentation supporting the appro-priateness of specific risk quantifica-tion techniques

■ analysis supporting the reasonable-ness and validity of stress tests andscenarios used

■ analysis and testing of model sensi-tivity to key assumptions and datainputs used

■ model validation work, including,where appropriate, the evaluation ofdevelopmental evidence, processverification, benchmarking, andback-testing.

Evaluate the control environment.Controls should be in place to ensure theintegrity of data inputs and the overallmanagement process. In evaluating suchcontrols, examiners could consider:

■ the quality of management informa-tion systems, including the timelinessof incorporation of changes in thebank’s risk profile

■ internal or external audit programreview of economic capitalmethodologies

■ the corporate governance structureas it relates to risk management andeconomic capital.

Determine the extent to which theeconomic capital process is used indecision making, such as in settingrisk limits or evaluating performance.Economic capital processes that are inplace but not integrated with the institu-tion’s risk management proceduresgenerally are ineffective.

Results Review

The results of economic capitalmodels can provide examiners anothertool in the supervisory evaluation ofcapital adequacy, enabling examinersto compare tangible capital levels (capi-

tal available to support risk) witheconomic capital levels (the bank’sown measure of its risk). As has alwaysbeen the case, an institution found tohold inadequate capital in relation torisk, regardless of the institution’scompliance with minimum regulatorycapital requirements, is expected totake appropriate actions to reduce riskor increase capital.

Banks generally operate with a capitalcushion above the level of risk measuredby the economic capital model, recogniz-ing the imprecision inherent in such esti-mation and the need for the bank to beresponsive to potential changes in condi-tions. Several factors can be consideredin determining the appropriate cushion,including:

■ the robustness of the bank’seconomic capital methodologies,

■ the quality of data inputs, assump-tions, and parameters,

■ volatility of the business model,

■ the composition of capital,17 and

■ external factors, such as businesscycle effects and the macroeconomicenvironment.

Incorporation into the OverallSupervisory Process

The development and implementationof risk models such as economic capi-tal often represents a significantchange in a bank’s overall risk manage-ment philosophy and practices. Like-wise, the overall supervisory processfor banks adopting economic capitalmodels can be affected as examinationfocus may shift more to process evalua-tion. Transactional testing wouldcontinue to figure prominently in theexamination function, but the purposeof transactional testing may be redi-rected to validation.

17This is particularly critical when considering the capacity of various elements of capital to absorb losses understress scenarios.

16Supervisory Insights Winter 2004

For example, the earlier discussion ofcommercial lending credit risk highlightsthe need for examiners to focus on vali-dating the accuracy of the loan gradingprocess at all grade bands rather thanconcentrating their attention primarilyon large or criticized facilities. The classi-fication of individual loans becomes inte-grated with the evaluation of the bank’sinternal loan grading system.

Furthermore, economic capital resultscan provide useful information for risk-scoping. Examiners can incorporate thebank’s risk quantification efforts andtrends in economic capital allocationsas another tool to better focus supervi-sory efforts on areas of high or increas-ing risk.

The use of economic capital and otherrisk modeling techniques is expected tocontinue to evolve and expand to moreindustry participants. Supervisory evalua-tions of banks are also changing toappropriately incorporate such advancesby the industry.

Robert L. Burns, CFA, CPA

Senior Examiner, LargeFinancial Institutions

The author thanks numerouscolleagues within the FDIC and at otherregulatory agencies who provided invalu-able edits, comments, and suggestionsfor this article.

Economic Capitalcontinued from pg. 15