The Regional Economist • October 2002

www.stlouisfed.org

B y M a r k D . V a u g h a n a n d D a v i d C . W h e e l o c k

Deposit insurance reform is on Congress'agenda.The House of Representatives passed a bill in May,

and the Senate is likely to debate reform this fall.Although the pending reform bill would change federaldeposit insurance in many ways, none of its features hassparked more controversy than a proposal to raise thecoverage ceiling, currently set at $100,000 per account,to $130,000 per account. On the one hand, the case forraising the ceiling seems simple: Inflation has erodedthe real value of coverage considerably since the lastincrease, in 1980. On the other hand, bank failures dra-matically rose and the thrift industry nearly collapsed inthe wake of that increase. Was the spike in failures acoincidence, or did the higher deposit insurance ceilinghave something to do with it? The 20th century U.S.experience suggests that boosting the coverage ceilingmay not be such a good idea.1

Why Insure Deposits?

The economic argument for deposit insurance stemsfrom the macroeconomic fallout from banking panics.In the past, depositors often had difficulty distinguish-ing financially sound from financially shaky banks.News about a regional economic shock would makedepositors nervous because they could not determinethe actual condition of their bank. Depositor nervous-ness grew out of the potential loss of uninsured fundsshould their bank fail. Before federal insurance, badeconomic news would sometimes spook depositors intowithdrawing funds from all banks. The U.S. bankingsystem was uniquely prone to mass withdrawals or"runs" because the typical bank was small and its loanportfolio was undiversified. Severe episodes of runs—banking panics—intensified economic downturns bydisrupting the payments system and cutting the flow ofcredit to business firms. Indeed, some economists haveattributed the depth and length of the Great Depressionto the severe banking panics of the early 1930s.2

Deposit insurance is an antidote to banking panics.When their funds are insured, depositors will not viewbad economic news with alarm. And the absence ofalarm means the absence of panics.

Following the Panic of 1907, eight states establishedinsurance systems. Then, in response to the panics ofthe early 1930s, when 9,000 banks—some 30 percent ofthe nation's total—went under, Congress establishedthe Federal Deposit Insurance Corp. (FDIC). The FDICoffered coverage to all U.S. commercial banks. In 1933,a temporary ceiling on coverage was set at $2,500; in1935, a permanent ceiling was fixed at $5,000. In 1950,

the ceiling was raised to $10,000; in 1966 to $15,000; in1969 to $20,000; in 1974 to $40,000; and finally in 1980to $100,000. Federal deposit insurance succeeded instabilizing the banking system; since the 1930s theUnited States has experienced no banking panics and,until the 1980s, almost no failures of insured institutions.3

So, What's the Problem with Deposit Insurance?

Although deposit insurance eliminated bankingpanics, it sometimes encouraged imprudent risk-taking.Deposit insurance encouraged bank risk-taking becausethe price of coverage—premiums and deductibles—was not set by the principles that guide private insur-ance companies.

Private insurance companies reduce their risk expo-sure with premiums and deductibles. In an automobilecollision policy, for example, the insurer sets premiumsbased on expected payouts, which in turn reflect thechance an insured driver will crash and the cost ofrepairing his car if he crashes. To deter insured partiesfrom driving recklessly because they are covered—aphenomenon that economists call moral hazard—privatecompanies charge higher rates to accident-prone driversand insist on deductibles from all drivers. Deductiblesencourage safe driving—that is, they combat moral haz-ard—because insured parties must bear some of thecost of accidents. Insurers also control risk exposures bypre-screening applicants. Pre-screening prevents reck-less drivers from disproportionately obtaining coverage—a phenomenon that economists term adverse selection.Deductibles reduce adverse selection as well as moralhazard because reckless drivers will steer clear of poli-cies that force them to share the cost of crashes.

Unlike private insurance, deposit insurance planstypically have not linked premiums to expected payouts.Instead, public plans have used flat premiums—that is,rates set as a fixed percentage of deposits—because theyare simple to administer. Marginal analysis—a staple inthe economist's tool kit—can demonstrate the resultingincentive problems. Bankers take on risk up to the pointwhere the extra, or marginal, benefit of risk-taking equalsthe marginal cost. The marginal benefit of risk-taking toa banker is the greater prospect of profits. The marginalcost of risk-taking is the increase in interest demandedby uninsured depositors, the increase in premiumsdemanded by the deposit insurer and the increase inlosses from risks that do not pan out. Because covereddepositors are shielded from losses, insurance eliminatesthe incentive to demand higher interest rates from riskybanks. So, with flat-rate deposit insurance premiums,

[5]

The Wrongirection?

Even without the proposed increase

to $130,000, the inflation-adjusted

ceiling on federal deposit insurance

is high by historical standards. The

chart tracks the real value of cover-

age from 1933 to the present. The

upward jags correspond to statutory

increases in the coverage ceiling;

other movements reflect changes in

the price level. The thrift debacle

followed the 1980 statutory increase^

which provided real coverage

exceeding three times the level of

the 1930s. Inflation since 1980 has

brought the real value of coverage

closer to the 1933-79 average. i | H §

the only check on risk-taking is a bank'snet worth. Net worth is the differencebetween the value of assets and thevalue of liabilities; it represents the stakethe owners have in the bank and oper-ates much like a deductible for insur-ance coverage. When net worth is high,the owners have much to lose from risksthat do not pay off. If net worth falls tozero, however, the owners have nothingto lose, and the marginal cost of risk-taking is essentially zero. Under suchcircumstances, bankers may yield to thetemptation to take imprudent risks—that is, succumb to the moral hazard indeposit insurance.

Government supervision of the bank-ing industry can combat moral hazard indeposit insurance. Bank supervisors canmonitor risk-taking with regular on-siteexaminations and continuous off-sitesurveillance. Supervisors can also insistthat bank net worth remains at highlevels. Finally, they can impose sanctionson risky institutions by, for example, pro-hibiting dividend payments, removingbank officers or denying merger applica-tions. Still, if supervisors fail to spot ris-ing risks or lack the resources to disciplinerisky banks, then imprudent risk-takingcan lead to waves of failures and the col-lapse of the deposit insurance system.

How the State Systems Fared

The fate of the state deposit-insur-ance systems in the early 20th centuryillustrates the consequences of poordesign and lax regulation. None of thesesystems survived more than 20 years.Oklahoma established its deposit-insur-ance system in 1907; within two years,Kansas, Nebraska, South Dakota andTexas offered coverage. By 1917,Mississippi, North Dakota andWashington were running such pro-

grams. Although details differed, thestate systems shared features that con-tributed to their demise. In particular,premiums did not rise with bank risk;they equaled a fixed percentage ofdeposits. Banks could be assessed addi-tional premiums should the state'sreserve fund run low, but these sur-charges were often limited by statute.4

Inadequate diversification of the loanportfolios contributed to the problems ofthe state-run systems. The states offer-ing coverage were mostly rural and agri-cultural, and the insured banks weremostly small concerns that lent locally.When commodity prices and farm prof-its soared during World War I, the num-ber of banks and the size of their loanportfolios mushroomed—especially instates with deposit insurance. In 1920-21,however, commodity prices collapsed,farm income plummeted and loandefaults skyrocketed. As bank failuresmounted, the reserves of the state depositinsurance funds evaporated. Premiumswere raised, but of the eight state insur-ance systems, only that of Texas had suf-ficient reserves to cover insured depositsin all failed banks. By 1929, each state haddismantled its deposit insurance system.

The failure of the state deposit insur-ance systems went deeper than thepost-World War I collapse of agriculturalprices; adverse selection played a keyrole. Risky banks were eager to join thestate systems because coverage madeattracting deposits easier. Easier accessto deposits meant fewer barriers to risk-taking. Well-managed, conservativebanks had little interest in joining sys-tems that benefited the depositors oftheir risky competitors. When given thechance, they opted out, leaving thereserve funds to be supported by banksat higher risk of failure. In the end,states with deposit-insurance systemssuffered disproportionately high bankfailure rates, with insured banks postingthe highest failure rates of all. Withoutcontributions from low-risk banks to paydepositor claims from high-risk banks,the reserve funds dried up.

Moral hazard compounded the prob-lems of the state deposit-insurance pro-grams. Because all covered banks paidthe same fixed-percentage premiums,the deposit insurance programs did notdeter risk-taking. Because insureddepositors, confident in their state's pro-gram, did not demand higher interestrates from risky institutions, depositordiscipline did not deter risk-taking.Because state bank examiners lacked theresources to force banks to act conserva-tively, government supervision did notdeter risk-taking. Insured banks couldmake risky loans with relative impunity—the marginal cost of risk-taking was

[6]

The Regional Economist • October 2002

www.stlouisfed.org

low. Imprudent lending helped producehigh default rates, widespread bank fail-ures and bankrupt deposit-insurancefunds. As a contemporary commentatornoted about one state's program, "It gavethe banker with little experience and care-less methods an equality with the managerof a strong and conservative institution.Serene in the confidence that they couldnot lose, depositors trusted in the guaran-teed bank. With increased deposits, thebank extended its loans freely."5

Did the Feds Do Any Better?

When designing a federal deposit-insurance program, Congress sought toavoid the problems that had broughtdown the state systems. To combatadverse selection, Congress insisted thatall national banks and members of theFederal Reserve System accept coverage—thereby preventing larger, and typicallystronger, banks from opting out. Thenationwide scope of the program alsoreduced the likelihood that a geographicor industry shock—like the collapse ofagricultural prices in the 1920s—wouldbankrupt the insurance fund. To combatmoral hazard, the new program requiredthat insured banks undergo regular fed-eral safety and soundness examinations.As a further check, Congress limitedentry into banking markets. Limits onentry shielded existing banks from com-petition, allowing them to reap highprofits and build up net worth. High networth, in turn, made bankers think twiceabout undertaking risky activities. Putanother way, close government scrutinyand stiff entry barriers deterred risk-taking by increasing the marginal cost.Federal insurance of thrift deposits alsobegan in the 1930s with the creation ofthe FSLIC, the Federal Savings and LoanInsurance Corp. The savings and loandeposit-insurance program was similar tothe program administered by the FDIC.

Moral Hazard Redux: The S&L Crisis

Although the federal deposit insur-ance system improved on the state-runprograms, it did not eliminate moralhazard. Premiums were, once again, setat a fixed percentage of an institution'sdeposits. Because premiums were nottied to failure risk, the deposit insurancesystem imposed no marginal cost onrisk-taking. Only the net worth ofinsured institutions and the watchfuleye of bank and thrift supervisors heldexcess risk-taking in check.

A dramatic rise in interest rates inthe late 1970s and early 1980s triggeredmassive moral hazard in the thrift indus-try and paved the way for the collapse ofthe thrift deposit insurance program.

Savings and loans concentrate on takingshort-term household deposits andmaking long-term mortgage loans.Rising rates increased the cost of servic-ing deposits relative to the revenue fromoutstanding mortgage loans. Collectivelosses from the interest rate squeezewiped out the industry's net worth. Themagnitude of the problem—thousandsof savings and loans were technicallyinsolvent—prevented supervisors frompolicing each institution's appetite forrisk. Thrift supervisors also came underintense political pressure to keep insol-vent institutions open. Inadequatesupervision, coupled with the low mar-ginal cost of risk-taking, led thrifts tomake highly speculative business andreal estate loans with insured deposits.Many of these gambles did not pay off,compounding the losses from the inter-est-rate squeeze. In 1989, Congressdissolved the FSLIC. But unlike the dis-solutions of the state deposit insurancesystems, which imposed no cost on statetaxpayers, the dissolution of FSLIC costU.S. taxpayers $150 billion.

The Federal Deposit Insurance Corp.Improvement Act of 1991 (FDICIA) wasdesigned to prevent another thrift-typedebacle. The act beefed up supervisionby mandating four things: annual safety-and-soundness exams, prompt correc-tive action, risk-based deposit insuranceand least-cost failure resolution. Fre-quent exams improve the flow ofinformation between bankers andsupervisors so that emerging problemscan be addressed quickly and decisively.Prompt corrective action, which man-dates specific supervisory responses todeteriorating bank net worth, guaran-tees that emerging problems will beaddressed quickly and decisively. Risk-based premiums, which currently rangefrom zero to 27 cents annually per$100 of deposits, increase the cost ofcoverage as bank risk rises, therebymaking deposit insurance more likeprivate insurance. Least-cost resolution,which forces the FDIC to clean up fail-ures in the least costly way for thedeposit-insurance fund, shifts more ofthe losses to uninsured depositors.And greater loss exposure increases theincentive to demand higher interestrates from risky institutions. The con-sensus is that FDICIA has reduced thechances of another deposit insurancemeltdown, though the act has not beenput to the test by a banking crisis.6

If It Ain't Broke, Why Fix It?

The booming economy of the 1990s,with some help from FDICIA, producedthe strongest banking conditions inrecent memory. Indeed, the decade saw

[7]

Advances fromFederal HomeLoan BanksCould Set BackInsurance Fund

One issue not on the table inthe current debate that per-

haps should be is the impact ofcollateralized bank liabilities onthe FDIC. Collateralized liabilitiesare funds that must be securedby bank assets. An example isdeposits from municipal govern-ments; state and local law oftenrequires that banks set asideU.S. government securities asbacking for municipal deposits.A much more important sourceof collateralized funding for mostU.S. banks is advances from theFederal Home Loan Bank (FHLB)system. These advances increasepotential losses to the deposit-insurance fund.

The FHLB system is a government-chartered but member-owned enterprisethat works to increase the liquidity of mort-gage markets. The FHLB increases liquidityby advancing funds to institutions that origi-nate mortgages; mortgage loans, in turn,collateralize the advances. Congress estab-lished the system in 1932 tolend to thrift institu-tions; membershipwas opened in 1989to others in the mort-gage-originationbusiness. Now theFHLB offers thrifts,commercial banksand credit unions awide range of prod-ucts and services tofund mortgage loans,to manage interest-raterisk and to meet the otherchallenges of a competitive banking environment. Between 1992 and 1999, total FHLBassets grew by 262 percent. During thatsame period, community bank membershipin the system increased four-fold, and out-standing advances increased 16-fold.

The growing volume of FHLB advancesthreatens the deposit-insurance fund throughtwo channels. Advances encourage a bor-rowing bank to take more risk. Also, becauseadvances are collateralized, the FHLB has

first crack at the bank's assets should fail-ure occur. The first channel increases

the likelihood that a borrowingbank will fail and that theFDIC will have to dip intothe reserve fund. Thesecond increases the

cost of a failure to theFDIC—that is, the size ofany necessary dip intothe reserve fund.1

Advances encour-age risk-taking because

Home Loan banks havelittle incentive to demand moreinterest or to withdraw funding

when the credit risk of a borrow-ing bank increases. Advances are heavily col-lateralized—the market value of mortgagecollateral typically covers 125 to 170 percentof the advance; the FHLB can also lay priorityclaim to other assets of a borrowing bankshould it fail and should mortgage collateral

net worth ratios soar and failure ratesrumble. At the same time, the FDICreserves swelled, allowing a premium cutfor healthy banks. By 2000, fewer than10 percent of U.S. banks paid any premi-ums at all. With all this good bankingnews, why would anyone want to fiddlewith the federal deposit-insurance system?

Much of the pressure for raising thecoverage ceiling comes from communitybankers. Community banks are relativelysmall institutions; the Financial Moderni-zation Act of 1999 set the asset limit forregulatory purposes at $500 million.They specialize in making loans to andtaking deposits from distinct regions,such as small towns or city suburbs.

In the 1990s, large banks merged ata record pace; these mergers producedsizable cost savings and put intense pres-sure on community banks to cut expenses.At the same time, community banks lostconsumer loans and retail deposits totax-exempt credit unions.

Community bankers argue that ahigher coverage ceiling would give thema better shot at luring large householddeposits, retirement accounts andmunicipal deposits away from largebanks. Raising the ceiling is only fair,these bankers believe, because largebanks enjoy "too big to fail" status,which effectively extends coverage toall deposits.

Finally, they note that rising priceshave considerably eroded the real valueof coverage since 1980; just compensat-

ing for 22 years of inflation means raisingthe ceiling to $218,0007

Reasons Not to Raise the Ceiling

Raising the deposit-insurance ceilingdoes have a downside—it could exacer-bate the moral hazard problem that hasplagued 20th century deposit-insurancesystems. A higher ceiling would reducethe marginal cost of risk-taking forinsured banks because a larger portion ofdeposits would be shielded from losses.These incentive effects are not just the idledaydreams of theorists; the thrift debaclefollowed on the heels of the last increasein the coverage ceiling—a hike to $100,000per account from $40,000 in 1980.

There are other reasons why raisingthe deposit-insurance ceiling may be asolution in search of a problem. For onething, the current ceiling can alreadyprovide much more than $100,000 incoverage. By one economist's calcula-tion, a family of four could insure upto $3.2 million in a single institution,thanks to joint and multiple individualaccounts.8 The case for inflation adjust-ment is also weaker than it first appears.If the starting point for indexing is the1935 ceiling of $5,000, rather than the1980 ceiling of $100,000, the current capshould be reduced to $65,954. Finally, ifimplicit or explicit subsidies to largebanks and credit unions put communitybanks at a competitive disadvantage, itwould be better to level the playing field

[8]

prove insufficient. This protection explainsthe system's stellar record in avoiding loanlosses: No Home Loan bank has ever lost apenny on an advance.

Because advances are essentially freeof credit risk, the individual Home Loanbanks can set terms that are largely inde-pendent of the failure risk of the borrower.Put another way, borrowing from the FHLBenables a bank to sidestep any market-imposed penalties for failure risk. As aconsequence, banks with a greater tastefor risk will be more interested in joining theFHLB and funding growth with advances—an example of adverse selection. Oncein the system, member banks can makerisky new loans with advances and see littlerise in the marginal cost of risk-taking—anexample of moral hazard. In short, accessto advances could end up increasing thelikelihood that a borrowing bank will fail.

Advances could also threaten thedeposit-insurance fund by weakening theFDIC's position in failure resolutions. UnderU.S. bankruptcy law, collateralized claimslike advances are settled first during failureresolution. So when a member bank fails,the FHLB stands first in line for repayment

—even before the FDIC. Other thingsequal, fewer losses to the FHLB implygreater losses for the FDIC. In short,even if FHLB funding does not encour-age risk-taking, losses to the insurancefund will be higher when a failure occursbecause the FHLB gets first shot at theassets of the failed bank.

One solution might be for the FDIC tofactor FHLB advances into the premiumscharged for deposit insurance. Each timean insured bank would borrow from theFHLB, the FDIC could re-price coveragebased on the new likelihood of failureand the new likely cost of any failure thatdid occur. Under current law, however,risk-based premiums reflect only thefinancial condition of the insured bank,not any additional losses to the FDICfrom FHLB advances. Also, mostobservers believe that the current capon premiums—27 cents a year per$100 of deposits—is too low to deterrisk-taking, much less to cover any addi-tional losses from collateralized funding.

1 Stojanovic, vaughan and Yeager (2000) providemore details about the FHLB system.

by eliminating those subsidies ratherthan introducing more distortions.

A look back at the economic justifica-tions for deposit insurance strengthensthe case against raising the ceiling. Evenif there were no insurance, the U.S. bank-ing system today would be less vulnera-ble to panics than it was before thecreation of the FDIC. The banking sys-tem is less vulnerable because the typicalU.S. bank is larger and, as a result ofextensive branching, better diversifiedthan earlier in the century. In 1934, forexample, the United States had 14,146banks and the average bank held$43.2 million in assets (expressed in 2001dollars). Collectively, these banks oper-ated 17,237 branches. By 2001, the totalnumber of banks had dropped to 8,080,the size of the average bank had jumpedto $813 million and the sum total ofbranches had multiplied to 73,644.Larger, more diversified banks meanthat economic shocks are less likely toundermine depositor confidence in thebanking system. Even if shocks didunnerve depositors, the Federal Reservehas learned from the experiences of the1930s and will intervene when necessaryto prevent banking problems fromthreatening the macroeconomy.

Other Changes

Other proposed reforms in depositinsurance make a great deal of sense.The House bill would consolidate the

reserve funds for the bank and thriftdeposit insurance programs, a move thatwisely reflects the narrowing differencesamong depository institutions. Also, pro-posed changes in the method of replen-ishing the reserve fund would allow theFDIC to better prepare for a rainy day.Under current law, the fund must be keptat 1.25 percent of total insured deposits.The House bill would give the FDICsome flexibility to move the percentagewhen reserves run low.

Still, economic theory and historicalexperience suggest that boosting thecoverage ceiling is a bad idea. It is possi-ble that the final bill will include theFDIC's proposal to make risky institu-tions pay much higher premiums forinsurance coverage.9 It is also possiblethat other FDICIA safeguards, togetherwith memories of the thrift debacle, willprompt bank supervisors to counter anyimprudent risk-taking. The history of theeight state deposit-insurance systemsand the thrift deposit-insurance systempoints to a more likely outcome—anincrease in moral hazard. As Yogi Berrahas phrased George Santayana's warn-ings about the lessons of history—it maybe like deja vu all over again.

Mark D. Vaughan is the supervisory policy officer inthe Banking Supervision and Regulation Departmentof the Federal Reserve Bank of St. Louis. David C.Wheelock is an assistant vice president in the ResearchDivision of the Federal Reserve Bank of St. Louis. Theauthors would like to thank Tim Bosch, Gary Comer, AltonGilbert, Dan Nuxoll and Tim Yeager for helpful commentsand Tom King for excellent research assistance.

Vic Regional Economist • October 2002

ENDNOTES

1 See Furlong and Kwan (2002) formore on other aspects of reform.

2 See Calomiris and White (2000) formore on the economic rationale fordeposit insurance.

3 See Bradley (2000) for more on thehistory of federal deposit insurance.

4 Calomiris and White (2000) providemore details about the various state-run systems. See Wheelock andKumbhakar (1995) for an in-depthlook at incentive problems in theKansas system.

5 Harger (1926, p. 278).

ft See Benston and Kaufman (1998) formore on the thrift debacle and FDICIA.

7 See ICBA (2000).

s See Thomson (2001).

9 See FDIC (2001) for more details.

REFERENCES

Benston, George J. and Kaufman,George G. "Deposit Insurance Reformin the FDIC Improvement Act: TheExperience to Date." Federal ReserveBank of Chicago Economic Perspectives,Second Quarter 1998, pp. 2-20.

Bradley, Christine M. "A HistoricalPerspective on Deposit InsuranceCoverage." FDIC Banking Review,Vol. 13, No. 2, pp. 1-25.

Calomiris, Charles W. and White,Eugene N. "The Origins of FederalDeposit Insurance," in U.S. BankDeregulation in Historical Perspective.Cambridge, Mass.: CambridgeUniversity Press, 2000, pp. 164-211.

Federal Deposit Insurance Corp."Keeping the Promise:Recommendations for DepositInsurance Reform," April 2001.

Furlong, Fred and Kwan, Simon."Deposit Insurance Reform—WhenHalf a Loaf is Better." Federal ReserveBank of San Francisco Economic Letter,No. 2002-14, May 10, 2002.

Harger, Charles Moreau. "An Experi-ment That Failed: What KansasLearned by Trying to Guarantee BankDeposits," Outlook,Vo\. 143 (June 23,1926), pp. 278-79.

Independent Community Bankers ofAmerica. "Federal Deposit Insurance:Time for Reform and IncreasedCoverage," Fall 2000.

Stojanovic, Dusan; Vaughan, Mark D.;and Yeager, Timothy J. "FHLB Funding:How Secure is the FDIC's Perch?"Federal Reserve Bank of St. LouisRegional Economist, October 2000,pp. 4-9.

Thomson, James B. "Who Benefitsfrom Increasing the Federal DepositInsurance Limit?" Federal ReserveBank of Cleveland EconomicCommentary, Sept. 15, 2001.

Wheelock, David C. and Kumbhakar,Subal C. "Which Banks ChooseDeposit Insurance? Evidence ofAdverse Selection and Moral Hazardin a Voluntary Insurance System,"Journal of Money, Credit and Banking,Vol. 27, February 1995, pp. 186-201.

[9]