Housing Finance — What Should the New System Be Able To Do?


Chairman Frank, Ranking Member Bachus, and distinguished membersof the Committee, I thank you for the invitation to appear attoday's important hearing. I am Mark Calabria, Director ofFinancial Regulation Studies at the Cato Institute, a nonprofit,non-partisan public policy research institute located here inWashington, DC. Before I begin my testimony, I would like to makeclear that my comments are solely my own and do not represent anyofficial policy positions of the Cato Institute. In addition,outside of my interest as a citizen, homeowner and taxpayer, I haveno direct financial interest in the subject matter before theCommittee today, nor do I represent any entities that do.

Housing and Mortgage Market Principles

Any set of proposals to restructure our system of mortgagefinance should begin, and be consistent, with a well defined set ofprinciples. The principles which should guide the shape of ourmortgage finance system are as follows: · Private, at risk,capital should serve as the foundation of our mortgage financesystem.

  • To the extent that government provides insurance, guarantees orsubsidies, those should be structured to act in a counter-cyclicalmanner. Too much of the current structure magnifies the booms andbusts in our housing markets. Policy should dampen cycles, ratherthan exaggerate them.
  • While policy should dampen housing cycles, we are unlikely tocompletely avoid property cycles - they remain a recurringphenomenon in our history. Accordingly, policy should explicitlyrecognize that housing booms and busts are likely to occur. Anypolicies based upon faulty assumptions, such as ever rising homeprices and "it's always a good time to buy" - should berejected.
  • In planning for housing booms and busts, policies should alsoexplicitly plan for the failure of institutions engaged in mortgagefinance. While efforts should, of course, be made to reducefailures, the system should be robust to the failure of any one ortwo companies.
  • Policies should avoid concentrating credit and interest raterisk into a small number of entities. As long as the risk isclearly understood, spreading that risk among many parties willreduce the impact of the failure of any one entity.
  • The costs and benefits should be transparent and credible.Subsidies should be on-budget and easily understood. The Americantaxpayer has a right to know what they are obligated for;accordingly, subsidies and contingent liabilities must be properlyaccounted for.
  • Housing policy should be tenure-neutral. The vast majority ofbenefits to homeownership accrue to the homeowners themselves andtheir immediate communities. The benefits to society at large havebeen grossly exaggerated and renting should be treated as arespectable alternative. Accordingly, policy should abandon anyfocus on a particular homeownership rate. Tenure-neutral, however,does not imply a "subsidies for everyone" approach.
  • To the extent that policies encourage homeownership, thathomeownership should be sustainable. Encouraging families to becomeowners with little or no equity ultimately harms the very familieswe wish to help.
  • Housing policy should also focus on housing as shelter, not asa speculative investment.
  • To the extent that subsidies are provided, they should becarefully targeted only to those who would not otherwise be able tohave a home, or achieve homeownership. The vast majority of currentsubsidies go to households that would have owned without thesubsidy. Subsidies should also be tied to incomes, not home pricesor rents. A disproportionate share of subsidies currently goes toupper income households. There is no compelling policy rationale toprovide housing subsidies of any kind to wealthy households.
  • To the extent that subsidies are provided via the mortgagefinance system, great care should be taken to insure that thosesubsidies end up with homeowners, and not simply passed along tothe housing or mortgage industry.
  • The current levels of leverage, both on the part of householdsand financial institutions, in our mortgage finance system shouldbe reduced.
  • The level of maturity mismatch in our mortgage finance systemshould be reduced.
  • Elements of our mortgage finance system that are little morethan disguised transfers of wealth should be rejected, includingattempts to cross-subsidize highrisk borrowers.
  • Policies whose impact is largely to run up housing pricesshould be rejected.
  • Mortgage finance should be insulated from politics. During aboom, political pressures will generally favor further inflatingthe boom.

A mortgage finance recovery

Aside from addressing the future of mortgage finance is theimmediate question of what to do about the current state ofmortgage finance. While a variety of problems face the mortgageindustry, the most important is the future direction of houseprices, and the expectation of such. As long as there is asubstantial chance of further declines in home prices, investorswill have difficulty projecting losses on mortgage relatedsecurities. Accordingly, first Congress and the Administrationshould end all efforts to prop up house prices. The harm from aquick reduction in home prices, or even an over-shooting on the waydown, is far less than the harm that results from holding pricesabove marketclearing levels. Once housing markets have reached thepoint where buyers and investors believe prices can fall nofurther, than both homebuyers and capital will return to themortgage market in strength.

To encourage private capital to return to the mortgage market,Congress should strongly affirm the importance of respectingprivate contracts. Repeated calls for mortgage "cramdowns" andother threats of expropriation increase the difficulty of pricingmortgage investments and encourage investors to place their wealthelsewhere. As long as investors believe Congress may ex postere-write the terms of their investments, they will hesitate toinvest at other than punitive rates. This is illustrated in therecent comments of a senior MetLife executive, who stated that"MetLife will not buy new securities until it knows what willhappen to the current ones - and whether investors will have toabsorb the resulting losses."1

As the financial crisis has receded, investors' flight toquality has also receded. The marginal investor is now againlooking for higher yields. Once investors are certain that higheryields can be found in the mortgage market, and that such returnswill not be subject to expropriation, then private money willreturn to the mortgage market in force. We are already witnessingthe early stages of several private sector mortgagesecuritizations. Just as important is what we did not see: a shockto the mortgage market from the winding up of Federal Reservepurchases of agency MBS. For the right price, investors are willingto supply credit to the mortgage market. Current marketdifficulties are compounded when uncertainty as to credit risk isexasperated by political risk.

On the 30 year fixed-rate mortgage

Any discussion of reforming our mortgage finance system has toaddress the central role of the 30 year fixed rate mortgage. Firstwe must begin with the very simple, yet critical, observation thatsomeone, the homebuyer, the financial sector or the taxpayer, mustbear the interest rate risk inherent in the 30 year fixed. A fixedrate mortgage does not eliminate interest rate risk, it simplytransfers it. In the case of the savings and loan crisis, and therecent bailouts of Fannie and Freddie, much of that risk wasinvoluntarily transferred to the taxpayer. It is worth rememberingthat most homeowners are taxpayers, so simply moving interest raterisk from homeowners to taxpayers does not make homeowners, as agroup, better off. We may end up making homeowners, as taxpayers,worse off if they were not fully informed as to this transfer exante.

As the taxpayer bears some of the interest rate risk of the 30year fixed, the price facing homebuyers is artificially low,relative to adjustable rate mortgages. Were the taxpayer no longerbearing this risk, I believe financial institutions would stilloffer 30 year fixed rate mortgages, however the spread of thosemortgages would increase relative to adjustable rate mortgages.Historically the difference between the 30 year fixed and the 1year ARM has been about 100 basis points. Without governmentsupport, my educated guess is that spread would increase to around130 basis points. While this may seem like a major increase, it is1) not large enough to adversely impact homeownership rates and 2)below some of the highs of earlier this decade - for instance in2003, certainly a "strong" housing market, these spreads approached240 basis points.

Proposals for reforming Fannie Mae and FreddieMac

While there are many important and critical issues to be decidedin restructuring our mortgage finance system, no issue is morecentral than the future of Fannie Mae and Freddie Mac. Ultimategoal of any GSE reform should be to create a system, where in atime of mortgage market stress, a GSE can fail, without cost to thetaxpayer or significant disruption to the financial and mortgagemarkets. To some extent, this will require making such markets lessreliant on the GSEs and reducing the extent to which theirsecurities permeate the financial system.

Consistent with the principles above, I recommend the followingsteps in reforming Fannie Mae and Freddie Mac:

  • The more, the merrier. Whether purely public or purely private,having only two Fannie/Freddie like institutions guarantees thatthese entities will be bailed out if they become insolvent. Theonly way to make failure a credible option is to have several. Iwould suggest breaking up Fannie/Freddie into about a dozen, equalsized entities.
  • Reduce ambiguity around debt status. Subject all GSE securitiesissues to requirements of 1933 and 1934 Securities Acts. Alsoremove all statutory treatment of GSE securities as "government"debt.
  • Allow only issuance of MBS - no unsecured debt, no portfolio.Also eliminates risk of GSE default on money market mutualfunds.
  • Get GSEs out of guarantee business. MBS should represent a"true" securization, not a retaining of credit risk on balancesheet.
  • Eliminate loan limits, set loan sizes based upon income, say 3times median state income, also allows elimination of housinggoals.
  • Require bank regulators to treat bank holdings of GSE debt asnon-governmental, corporate debt. Also limit any insured depositoryfrom holding more than a small percentage, say 5%, of its assets inGSE debt.
  • Charters should be issued/removed by regulator, not Congress.Consistent with having more GSEs, allow regulator to issue newcharters and conversion of other financial institutions into newGSE charter.
  • Limit or bar holdings of GSE debt by foreign central banks.Fannie/Freddie bailout was as much a foreign policy decision as aneconomic one.
  • Require all mortgages purchased by GSE to have a minimum cashdownpayment of 10 percent - no piggybacks. To avoid disruptions tothe mortgage market, this requirement could be phased in over a fewyears, starting with a cash requirement of 5 percent.
  • Subject GSEs to bankruptcy code. Conservator/receiver modelincreases chance of bailouts, and reduces market discipline on thepart of debtholders.
  • New Fannie/Freddie privatization model could be based uponco-op model of the FHLBs. Require lenders selling loans to purchaseequity, similar to FHLB advance model. This would better alignincentives of lenders with the risks taken by Fannie/Freddie.

Toward a Countercyclical Mortgage FinanceSystem

U.S. Housing Markets have tended toward a regular pattern ofboom and bust. While some degree of cyclicality is likelyunavoidable, federal mortgage policy has often contributed to thesewide swings in housing activity. Mechanisms can be created thatdampen the incentives for households and financial institutions toengage in bubble behavior. These mechanisms should, of course, bedirectly related to a national interest. Entities that do not posea systemic risk to the financial system or receive backing from thetaxpayer, implied or otherwise, should be free to innovate andsucceed or fail.

Housing bubbles are driven foremost by the speculative behaviorof households. Current federal and state policies encourage suchspeculation. For instance several states, such as California,require that residential mortgages be non-recourse. That is, in thecase of a default, the lender can only pursue the house and not anyof the borrower's other assets or income. Recent research from theFederal Reserve Bank of Richmond indicates that "recourse decreasesthe probability of default when there is a substantial likelihoodthat a borrower has negative home equity."2 Not only does a lack of recourse increase defaultsduring the bust phase of the cycle, but such also likely increasesthe incentive of buyers to enter the market with greaterspeculative intent. Where there is a federal interest, allmortgages should contain recourse provisions and such provisionsshould be exercised.

The scholarly literature on speculative bubbles concludes thatsuch bubbles are more likely to develop the lower are transactioncosts and the lower is the required holding period of the asset inquestion3. It is for thisreason that many countries, such as Canada, whose mortgage marketscontain substantial pre-payment penalties, did not witness the samelevel of boom and bust as the U.S. housing market. We shouldreverse the policy trend toward eliminating pre-payment penaltiesand instead encourage significantly broader use of such. The easeof repeated re-financings, coupled with equity-extractions, greatlyadded to the severity of the boom and bust.

The most important predictor of mortgage default is equity, orlack thereof4. Owners thatare underwater are significantly more likely to default thanhomeowners with equity. Requiring reasonable downpayments when themortgage has a federal interest would significantly reduce theseverity of housing cycles. Ultimately federal policy should worktoward a cash downpayment of 10 percent. During booms this can beraised. For instance requiring a downpayment that is the higher of10 percent or last year's national house price appreciation wouldgreatly reduce housing cycles. Similarly the capital whichfinancial institutions, including GSEs, are required to holdagainst residential mortgages should be linked to house priceappreciation.

Systemic Risk and Mortgage Finance

While Fannie and Freddie were rescued for a variety of reasons,prominent among those is that fact that their securities, bothequity and debt, permeate our financial system. For instance, morethan 40% of money market mutual fund holdings were in the form ofGSE securities. Were a receiver to impose substantial losses onshort-term unsecured GSE debt, hundreds, if not thousands, of moneymarket mutual funds would have "broken the buck." Same with insuredcommercial depositories. According to the FDIC, before the burstingof the housing bubble, holdings of government-sponsored enterprise(GSE) securities, bonds and mortgage-backed securities as well aspreferred stock, constituted more than 150% of Tier 1 capital forinsured depositories. If we thought bank losses from the reducedvalue of Fannie and Freddie preferred shares was a problem, theselosses would have been rounding errors compared to bank losses fromFannie and Freddie debt. Sadly Wall Street was also infected. Forinstance, the Federal Reserve has reported that more than 50% ofMaiden Lane One assets, the toxic assets that the Federal Reserveguaranteed in order to persuade JPMorgan to buy Bear Stearns, wereGSE securities.

Our country has witnessed housing booms and busts before,although not one of this magnitude. The fallout from such a largebubble bursting was guaranteed to be painful and prolonged.However, the resulting financial crisis did not have to result. Thefinancial crisis resulted from the fact that so much of thesoundness of our financial system is build upon the sand of houseprices.

Innovation, Standardization and the UnknowableFuture

Given the clear role that many facets of our current mortgagefinance system played in creating the housing boom and bust, it istempting to proscribe a set of standards for the mortgage marketand require all participants to meet those standards. Such would bea tragic mistake. The better path would be to allow essentially twosystems: one for institutions that place the taxpayer and thefinancial system at risk, and one for nondepositories and non-banksthat do not place the taxpayer and system at risk. Entities shouldbe able to choose under which system they operate, ultimatelyallowing the free choice of individuals to determine the bettersystem. Such a system would also allow innovations that improveconsumer welfare without putting the financial system atrisk.5 We have already seenthe result of concentrating mortgage risk into a small handful ofentities; we must avoid repeating that mistake. In addition toavoiding the concentration of risk into a few entities, we shouldalso avoid the concentration into a few business models. According,we should closely examine the possibility of utilizing variousforms of mortgage finance, including, but not limited to coveredbonds, portfolio lending, and mortgage backed securities.

1 Quote from Nancy MuellerHandal in Aline van Duyn "A Business Decision," FinancialTimes February 23, 2010 p.7.

2 Ghent and KudlyakRecourse and Residential Mortgage Default: Theory and Evidence fromU.S. States. Federal Reserve Bank of Richmond. Working PaperWP09-10.

3 Barlevy. "Economictheory and asset bubbles." Economic Perspectives. 2007.Federal Reserve Bank of Chicago.

4 Kristopher Gerardi, AdamHale Shapiro, and Paul S. Willen. Decomposing the ForeclosureCrisis: House Price Depreciation versus Bad Underwriting. FederalReserve Bank of Atlanta Working Paper 2009-25

5 See generally, "TheImpact of Deregulation and Financial Innovation on Consumers: TheCase of the Mortgage Market." with Paul Willen, Kristopher S.Gerardi and Harvey Rosen. Journal of Finance,forthcoming.