The administration has told us that American financial markets are in a perilous state because of recent and prospective losses of financial institutions that threaten the adequacy of their capital1 and because of the illiquidity and presumed market underpricing of Mortgage Backed Securities. Both of these have increased banks' demand for liquidity and tightened bank lending standards. A credit crunch, which would worsen an economic slow down, could result if markets for normally liquid assets such as MBSs and bank capital are not restored.
The proximate source of these problems is a higher than expected default rate on mortgages (mainly adjustable rate mortgages to subprime borrowers) and uncertainty about how high the default rates will go. The increase in defaults and foreclosures (the takeover of the house collateral that backs the defaulted loan) has followed closely the decline of grossly inflated housing prices, but they are not directly caused by falling housing prices. They are caused by the inability of mortgage borrowers to meet their monthly payments or their decision that it is not economically advantageous for them to do so. Falling housing prices may affect the second of these, especially for "flippers" (people who buy property on speculation that housing prices will rise).
A secondary source of these problems is the complexity of modern financial instruments, which has increased the interdependence of financial institutions on each other through credit default swaps and other forms of insurance and warrants, and made it difficult to know who actually bears the losses from mortgage defaults.
A contributing factor in the excessive extension of mortgage loans to those now unable or unwilling to repay them (in addition to the government's policy of promoting greater home ownership among low-income households) was weak oversight of compliance with lending standards, which may also have been too low. The enforcement of these standards was made more difficult by the widespread securitization of mortgage loans (the sale of loans by their originators to other investors, which tends to reduce the originator's interest in the soundness of the loan). These weaknesses will be corrected by loan originators and mortgage pool consolidators who want to protect their businesses, and hopefully by a carefully focused strengthening of the regulatory restrictions on lending. The market's self-corrections will be more complete the more heavily the costs of the recent past weaknesses and mistakes are borne by those who made them.
The administration faces the need to prevent an immediate credit crisis and to adjust the regulatory regime in order to reduce the prospects of similar problems in the future. Short-run fixes, especially under the pressure of a crisis, always run the risk of undermining or weakening what is needed for the long run. Any short-run help should be designed to minimize the "moral hazard" of bailing out the mistakes of those who took risks that have gone bad. Such accountability for our actions (accepting the consequences whether good or bad) is a cornerstone of the market's regulation of economic activity.
The urgent, immediate need is to provide banks and the financial market with the liquidity and capital they need in order to resume or maintain normal lending. The mystery of Secretary Paulson's $700 billion proposal is that the Federal Reserve has been quick and imaginative in addressing the market's liquidity problems. "Commercial banks and bond dealers borrowed $217.7 billion from the Federal Reserve as of yesterday [September 25, 2008], more than double the prior week..."2 In addition, AIG has drawn "down $44.6 billion on its $85 billion credit line from the Fed, up from $28 billion as of Sept. 17 ..."3 The use of MBSs as collateral for loans from the Fed has large advantages over their proposed purchase by the Treasury. The almost impossible task of pricing the very diverse MBSs correctly (near-best estimates of their "true" value) for purchases by the Treasury is avoided by their use as collateral. The Fed's lending facilities are in place and operating and do not require new legislation from Congress. While overpricing them would allow banks and other holders of comparable MBSs to reflect these higher values and thus improve reported capital in their financial statements, overvaluation might discourage the revival of trading them in the market and hence, market-provided liquidity. Overpricing would be paid for by taxpayers and underpricing would be paid for by bank capital.
As noted above, the proximate cause of the crisis is the increase in defaults on mortgages. Can — and should — these be prevented? The downward adjustment in housing prices from their severely inflated levels of two years ago should not be prevented. These are needed to make housing affordable again and to revive the housing market (both construction and resale). Though the pace of price decline has slowed4 and sales of existing homes are recovering slightly from year-end lows5, prices will need to fall another 10 percent or so to fully return to the average affordability level of the 2000 to 2003 period.6 Everything possible should be done to facilitate the restructuring of defaulting mortgages where borrowers are willing and able to pay reduced monthly charges if these adjustments preserve greater value to the lenders than what they would receive for foreclosing on deflated homes. Lenders have every incentive to help borrowers avoid foreclosures in this way, and hundreds of thousands of such loan restructurings have been voluntarily undertaken on a one-by-one basis. There is no conceivable way government bureaucrats could organize to perform this task better. Going beyond this to bail out borrowers will create a dangerous moral hazard that will encourage reckless borrowing in the future in the hopes of another bail out.
There is a growing consensus among economists outside of government and Wall Street that a better approach is to directly help banks to increase their capital.7 Such help might take the form of mandating the suspension of dividend payments (adding the retained earnings to capital) for any banks whose capital is below, say, 6 percent of risk-weighted assets, and requiring recapitalization through new share issues if needed. But these tools are already in place and in use — though bank shareholders would naturally rather be bailed out than have to pay up themselves.8 There is also a good case for lowering capital adequacy requirements cyclically during business downturns as long as estimates of loan and other losses are properly made and provisioned (reserves set aside to cover them, should they be realized).
In fact, there is little evidence of widespread serious capital deficiencies among banks. Only a small number of banks have failed and the United States has reasonably good and efficient procedures for resolving insolvent or critically undercapitalized banks. These procedures should be used (as they have been) one case at a time. Bear Stearns and AIG are not banks, and the legal resolution tools for banks are not available to the authorities for dealing with them. Under the circumstances, the Federal Reserve did about the best it could to remove them from play as smoothly and efficiently as possible. Consideration should be given to enacting similar tools for any class of financial institution that might pose systemic risks if it should fail.
The general view in the market is that banks have already written off all or most of their expected losses. Are depositors and other creditors really panicking to the extent that a massive government intervention is needed to restore confidence in the system? The run on money market mutual funds last week clearly unnerved the authorities, but Fed lending to banks has enabled them to replace the withdrawn funds (by buying fund commercial paper), and fund depositors will need to put their withdrawn money somewhere.
But what if depositors lose confidence in banks more generally, creating a truly systemic crisis? The onset of recession will surely add addition losses to banks. While weak and poorly managed banks should be allowed to fail for their mistakes, it would not serve the public interest to let a large proportion of the banking system go under for reasons largely outside of their control. In this unlikely situation, the government should intervene more forcefully. The Swedish approach for dealing with its banking crises in the early 1990s has been raised as a model that the United States should consider. Stefan Ingves, who has been the governor of Sweden's central bank (Riksbank) since January 1, 2006, oversaw the formulation and implementation of Sweden's banking crisis resolution, first as undersecretary for financial markets and institutions at the Swedish Ministry of Finance and then as director general of the Swedish Bank Support Authority in 1993-94. Stefan was also my direct boss at the International Monetary Fund from 1999 until I retired in 2003, and the lessons of the Swedish experience played an important role in its advice to the many countries that experienced a banking crisis over the last fifteen or sixteen years.
The end of Sweden's property boom in 1991 and the collapse of real estate prices in 1991–92 put a large number of its banks underwater (negative capital). In September 1992, Prime Minister Carl Bildt announced that the government was guaranteeing all bank deposits and other creditor claims on all banks in Sweden, thus immediately ending bank runs. It also ended any market discipline of bank soundness on the depositor side. The government focused instead on maximizing the discipline and oversight of shareholders (bank owners). Banks were required to immediately and fully write off all losses in order to have an honest picture of their condition. Those needing government funds to recapitalize (if they chose to stay in business) were required to surrender shares first, thus putting the regulators temporarily in charge of such banks. Sweden earned high marks for its skillful and principled approach to its banking crisis.9
The situations in Sweden in the early 1990s and in the United States now have similarities and important differences. Sweden did not have the legal bank resolution tools that the United States now has available, and thus, had to improvise. Because of these tools, the United States can take over and resolve a failing bank without fear of systemic consequences. We need to continue applying these tools, including extending them to cover investment banks and insurance companies, and the Fed needs to continue supplying all of the liquidity needed. If public confidence in banks erodes, deposit insurance limits might be raised. I see little purpose in injecting government capital directly into undercapitalized banks outside of the existing bank resolution mechanisms, but it would seem preferable to Secretary Paulson's $700 billion purchase of MBSs.
Another proposal worthy of further consideration is to insure nonagency (privately issued) MBSs in the same way those issued by Fannie and Freddie are insured.10 This would nationalize such insurance along the lines of the Federal Deposit Insurance Corporation. The contracted mortgage payments would be assured, but the return to those buying and holding them would be reduced by an insurance premium that would reflect the expected loss from defaults by the ultimate borrowers. If the market is currently underpricing these MBSs, as is assumed by many,11 such insurance would increase their market price (and the capital of banks holding them). Of equal importance, it would remove further uncertainty over the value of such MBSs, which should restore their liquidity in the market. If MBSs could be correctly categorized by default risks from the underlying mortgages (a challenging task) and insurance premiums assigned accordingly, there would be no cost to taxpayers. The new MBS insurance agency would be responsible for monitoring mortgage originators for compliance of the underwriting standards applicable to each risk (insurance premium) group. It is also quite a challenge (but not necessarily impossible) to put such a scheme in place in a matter of weeks. Until the premiums for each MBSs are determined they are not likely to be traded.
If Congress eventually adopts a version of the administration's $700 billion MBS bailout proposal, the most important decision will be how to price MBSs purchased by the Treasury (or some entity created by the legislation for this purpose) and the insurance premiums paid to guarantee them. If each MBS is correctly priced, that is, if it is purchased at a price that matches its ultimate yield to the holder (which is unknowable in advance), the operation will cost the taxpayers nothing. Such a price will be above what these MBSs can be sold for in the market today because it will remove uncertainty over the price and thus the risk premium attached to current trading. It will provide a better price for valuing comparable MBSs in financial institutions' portfolios. Banks that are still seriously undercapitalized after these repricings should be dealt with in accordance to existing bank insolvency legislation. Failing institutions that are considered too big to fail should be recapitalized by the Treasury (under the systemic emergency provisions of the Federal Deposit Insurance Corporation Improvement Act of 1991), thus placing them under temporary government control. The draft legislation also authorizes "the Secretary to encourage the servicers of the underlying mortgages, considering net present value to the taxpayer, to take advantage of the HOPE for Home owners Program under section 257 of the National Housing Act or other available programs to minimize foreclosures."12
The use of reverse auctions to establish the Treasury's purchase prices should break MBSs into the same risk classes as would be needed for setting the premiums for an insurance program as discussed above. This will be a challenging task, as MBSs are very diverse. Banks will obviously offer their worst MBSs for sale in such auctions, but if each auction pertains to a relatively narrow risk class, the range of actual expected values within each class will also be narrow. The auction price that clears supply and demand for that auction should then be relatively close to the best possible assessment of value given the information available at the time. Auctions should proceed quickly. Financial institutions that participate in sales to the Treasury will be subject to restrictions on executive compensation. "Troubled assets" may be guaranteed. These guarantees would be financed by premiums paid by issuers based on their risk class.
For best results, all MBSs of each risk class purchased by the Treasury in reverse auctions should be insured using an insurance premium that matches the discount from face value of the auction price paid by the Treasury. To illustrate, if MBSs of risk class X with a face value of $100 are sold to the Treasury for $60, all such MBSs held by financial institutions subject to mark-to-market accounting would be valued at $60 per unit. All such MBSs would be guaranteed — that is, the trusts (or SPVs) holding the mortgage pools backing this class of MBSs would be paid by the Treasury insurance fund for any shortfalls or defaults in payments by the mortgagees. Thus, the trusts would be guaranteed by the U.S. government to receive full value for their mortgage pools. The insurance premium they should pay for this guarantee should be exactly the $40 discount established by the reverse auction to the Treasury. Owners of this class of MBSs would receive the full value of the underlying mortgages, less the insurance premium (or $60 [$100 – $40]). These combined measures will have accomplished several important things. MBSs that might have traded in a risk-averse market at $50 or $45 per unit are now valued at $60 by all who hold them, thus improving their capital. Furthermore, that value is now guaranteed by the U.S. government and these MBSs can trade in the market at the price of $60 with the same risk as U.S government securities (i.e., no risk). This should ensure the marketability (liquidity) of these assets. If performance of the underlying mortgages improves or worsens over time, the gain or cost accrues to the taxpayers underwriting the guarantees. If mortgage performance deteriorates to $50 per $100 face value, for example, holders of the MBSs will still receive $60 and the insurance fund will incur costs of $50, which is $10 more than the insurance premium of $40. Combining Treasury purchases of MBSs along with guarantees for those MBSs remaining in the market should provide a substantial strengthening of market liquidity. The existing tools for providing liquidity to banks and for resolving problem banks will still be needed, and should be sufficient.
On Monday, September 30, 2008, a majority of the House of Representatives rejected the Administration's proposals. No convincing evidence has been presented that existing liquidity and bank resolution tools are not up to the job of seeing us through the adjustments to earlier excesses which are now underway. As long as that remains the case, the Federal Reserve and FDIC should continue to rely on them. Should fresh evidence indicate otherwise, the authorities should turn first to increasing deposit insurance limits, establishing a mortgage insurance agency (to replace Fannie Mae and Freddie Mac), and — only as a last resort — should the injection of taxpayer-financed capital be put into (and thus, government takeover of) financial institutions that are judged too big to fail. The Senate passed a version of the same proposal Wednesday that is expected to gain more support in the House, which is now expected to pass the bill Friday. If so, the Treasury should consider using the authority given them in the bill to inject capital (ala Sweden) in under capitalized banks rather the buying their MBS.
By far the best idea I have heard comes from Leon W. Louw, President of The Free Market Foundation in South Africa and fellow member of the Mont Pelerin Society, who proposes that the government guarantee to underwrite a portion, say 50 percent, of all actual losses from defaults on residential mortgages. This plan could be fully implemented instantly (the bailout legislation before Congress is so broad that it would surely cover this possibility within its authority). The actual disbursement of taxpayer money would be spread over time as defaulted mortgages were foreclosed and the exact losses determined. It poses no requirement to estimate probable losses (or determine reasonable prices for existing MBSs). Whatever owners of mortgages (directly or indirectly) currently think those losses are likely to be, and thus the estimated value of their mortgage related holdings, Leon's proposed guarantee would immediately cut that estimate in half (or whatever the government's share would be).
In this example, the difference between the current market value of mortgages and their full face value would be immediately cut in half to the immediate benefit to the capital of the banks and others who own them. The primary cause of this and last week's financial panic and sizing up of credit flows (fear that credit losses will exceed bank capital or reduce it to unacceptably low levels)will be directly addressed with an infusion of taxpayer financed capital without complicated administrative procedures and without eliminating the financial incentive for mortgage lenders to lend prudently. Those who made mistakes would continue to pay a price for those mistakes but not the full price. After the current crisis passes, the expected financing cost of these guarantees could be shifted from taxpayers to mortgage holders in the form of the risk-based insurance premiums I discussed above.
1 According to Scott Lanman, "The subprime-mortgage collapse has led to $522 billion of writedowns and losses at major financial institutions since the start of 2007." "Federal Reserve Doubles Lending as Crisis Worsens," September 26, 2008, www.bloomburg.com. See also Coats, Warren. "The U.S. Mortgage Market: the Good, the Bad and the Ugly," Association of Banks in Jordan, June 22, 2008; "Fannie and Freddie: More Good, Bad and Ugly." July 31, 2008; "The D E Fs of the Financial Markets Crisis," Cato Institute, September 26, 2008; and Coats, Warren, and Aarno Liuksila. "Institutional and Legal Impediments to Efficient Insolvent Bank Resolution and Ways to Overcome Them." USAID Conference, Banking Supervision in a Global Environment "Meeting International Standards," Sofia, Bulgaria, July 7-9, 1999.
4 Laurie Goodman, "MBS: Update on the Crises," UBS, September 23, 2008, and S&P/Case Shiller Composite-20 Home Price Index.
6 Deutsche Bank, "MBS Special Reports; Update: Projecting Mortgage Losses," September 22, 2008.
8 Good bank supervisory practice and the Federal Deposit Insurance Corporation Improvement Act of 1991 (designed to overcome supervisors' reluctance to act as capital becomes weak) mandate that, as capital falls below prudent levels, supervisors must take increasingly severe actions before it has fallen to critically low levels (2%), at which point the FDIC must take over the bank. This puts great weight on having adequate capital (shareholder stake in the successful management of a bank) as the basis for leaving most of the rest to market regulation.
10 Over 60 percent of all outstanding MBSs were issued by Fannie Mae and Freddie Mac and are thus already guaranteed (now with the full backing of the government). The serious weaknesses of the government-sponsored enterprises and their contribution to the existing subprime crisis are discussed in "Fannie and Freddie: More Good, Bad and Ugly," July 31, 2008. Any insurance program contains potential moral hazards and requires careful monitoring of compliance as well as appropriately priced underwriting standards. A properly charted insurance fund is much more likely to be successful at this than was the mixed-mandated, private government-sponsored enterprises. The FDIC provides a good example.
12 Discussion draft of Emergency Economic Stabilization Act of 2008.