Securitization versus Traditional Banks: An Agnostic View of the Future of Fannie Mae, Freddie Mac, and Banks

Introduction and Background

The current financial crisis was precipitated by a bubble in house prices and its subsequent burst, which led to a wave of foreclosures, the seizure by the Federal government of the main vehicle for securitization (the Government Sponsored Enterprises [GSEs] Fannie Mae and Freddie Mac), the obliteration of the “private label” securitization market, the failure of 92 banks so far this year, and bailout costs for the remainder of the banking system. No one has come out smelling like a rose. The question we address is what should happen to the historically most important players in the mortgage market: Fannie Mae, Freddie Mac, and the banks.

Broadly speaking there are two models for funding mortgages (and other loans): the portfolio lender model, which entails financial institutions (e.g., banks) originating and holding loans in their portfolio and funding them with debt (e.g., deposits), and thesecuritization model, which entails buying loans and putting them into pools and selling (perhaps structured [1]) shares in the pools to capital market investors. Many of the current financial arrangements are combinations of the two. The easiest way of looking at the two models is to think of them as applying to institutions called “banks” and “securitizers” and to view the rules and benefits that apply to them as their “charters.”

There is a trade-off between the advantages and disadvantages of securitization: securitization provides a low cost and elastic source of funds from the capital market, but it has “agency costs” because of information asymmetry between investors (or whoever takes on the credit risk) and loan originators. Both structures have transactions costs; for instance banks have costs of running branches and the things necessary to attract depositors, and securitization has costs of setting up and marketing deals as well as costs of monitoring agents’ behavior. A competitive balance among the charters depends on the balance of the above as well as implicit and explicit subsidies, particularly in the form of guarantees, but also taxes, received by them.

Guarantees have been a central aspect of both charters. This began with the Great Depression and federal deposit insurance. Fannie Mae and Freddie Mac (hereafter FF) can be understood within the deposit insurance framework: Their GSE status gives them an “implicit” guarantee that is similar in function to deposit insurance, and they use this guarantee to compete with banks. This competition has been referred to as “dueling charters.” [2]

Differences between banks and GSEs can be overrated. For instance, the GSE structure has been criticized as being unworkable because of its dual role as a private company but with a public purpose. However, that is not especially unusual, especially relative to banks, which via the Community Reinvestment Act (CRA) are in much the same position. Public purpose regulation is not that unusual in banking. [3]

During the current financial crisis with the ubiquitous exercise of the Too-Big-To-Fail policy, it has become apparent that virtually all major financial institutions (banks included) and systems of institutions are GSEs. Public policy choices need to realize this and manage it. We suggest that instead of picking a model, we diversify and allow the two charters to compete by making the playing fields between them as level as possible. We do not see the elimination of guarantees as likely or especially desirable. It is probable that the model chosen will be less important than the regulatory structure that accompanies it. Our best guess is that the bank model will focus on Adjustable Rate Mortgages funded with deposits and the securitization model on fixed rate mortgages funded with long term securities. But there are lots of ways to skin a cat, and both charters have ways of doing more or less the same thing.

A Framework

Our point of departure is the “Modigliani-Miller Irrelevance Theorem” (henceforth “MM”).[4] Briefly, the theorem is that under a set of assumptions, which mainly involve competitive markets, low transaction costs, and widely agreed on information, the liability structure of the firm is irrelevant in the sense that changing the way the firm finances its assets will not affect its “all-in” cost of funds. This is because different liability structures are simply different ways of rearranging the same cash flows from the firm’s assets, [5] and competition will assure that all structures will be priced so that the sum of the parts will equal the whole, no matter how the parts are chosen.

Taken literally, the theorem suggests that the different charters should not affect mortgage rates. A softer version is that advantages of different structures are likely to be small, and because of very elastic long run supply curves, small advantages of one source of funding (e.g., from a subsidy or lower transaction costs) can lead to big effects on market share but with perhaps small effects on interest rates and resource allocation.

Institutions[6]

The structure of the U. S. mortgage market has changed dramatically in the last quarter century because of the rise of securitization, which was largely brought on by the three secondary market Agencies: Fannie Mae, Freddie Mac, and Ginnie Mae (a Government Owned Enterprise). [7] In addition, the Federal Home Loan Bank (FHLB) system provides liquidity to the banking institutions by discounting their mortgage-related assets. Their growth has been accompanied by a decline in the market share of the traditional lenders, banks, and thrift institutions (e.g., savings and loans). That growth was reversed by the growth of the “non prime” (subprime and “Alt-A”) [8] market after 2003, and renewed with the collapse of that market.

In 2008, FF were put into conservatorship. The government has become the major shareholder in both companies. The conservatorship has made the guarantee much closer to explicit, particularly because the Treasury has agreed to inject up to $400 billion in capital as needed.

Banks, the main alternative to securitization, also have a low-cost source of funds in the form of insured deposits, but this source of funding has not been as elastic as the one coming from capital markets in general, which can be tapped quickly by the secondary market. As a result, banks sometimes have trouble raising money quickly, especially relative to the Agencies.

Banks, like all corporations, are essentially structured deals with subordination: the common stock holders are the most subordinated, followed by preferred stock, subordinated debt, etc. The MM Theorem of course applies to this structure. A key difference, which applies to any ongoing business, is the option of management-to-change portfolio composition, something that contractually is not usually allowed in structured securitization deals. Details differ, but structured securitization deals and banks have a lot in common, and it is not clear which is less transparent.

Recent Securitization

The beginning of modern securitization in the 1970s and early 1980s relied primarily on the standard pass-through security, which bundled mortgages together and sold shares in them, usually with Agency guarantees of credit risk. While Agency MBSs have little or no credit risk, they have two types of interest rate risk: the usual risk of any long-term security that its value will fall when rates rise, and a second risk that is similar to that of callable bonds, because borrowers have the option to refinance (i.e., call the mortgage) which they tend to do when rates fall. Guaranteeing credit risk allowed investors to focus on managing interest rate risk.

Beginning in 1983, the first collateralized mortgage obligations (CMOs) established tranches that received principal payments in sequence. In this way a complicated pool of 30-year callable securities was broken into a sequence of bonds, which could be sold to different types of investors. A popular form of structuring segments prepayment risk by taking parts of the pools and designating one piece that takes the first part of prepayments with the other piece taking the risk only above some level. This is a form of subordination, in this case of prepayment risk.

Both Fannie and Freddie also fund mortgages with debt.[9] Debt allows them to issue very homogenous and easy to understand securities, which can trade very efficiently, almost like Treasury debt, but it requires them to manage the interest rate risk. The share of debt financing is now around one-third.

The Non-Agency (Private Label) Market

This market operates mostly in areas not eligible for the agencies. A big part of this market has been loans with balances too large for Agency purchase. More recently, the subprime market, which consists largely of loans to borrowers with poor credit histories, has grown very rapidly. A related market, for “Alt-A” loans, which are loans that are prime except for low documentation of income and wealth, has also grown rapidly. The two combined have grown from around 5% in the 1990s to 10% in the early 2000s and then to roughly 33% after 2003. Commercial mortgages are also funded with securities (Commercial Mortgage-Backed Securities or CMBSs).

Non-Agency securities are different from Agency MBSs because they do not have the benefit of Agency guarantees on credit risk. As a result, credit risk is typically managed by subordination, as with the CMOs, but in this case to rearrange credit risk. Typically there is a series of subordinated tranches that take the default losses up to some amount and senior tranches, which take the rest. Functionally these are mini corporations, like special purpose banks. An alternate credit enhancement tool has been insurance on the pool, e.g., credit default swaps.

Unbundling, the Securitization Process, and Agency Problems

The traditional portfolio lender performs all aspects of the mortgage bundle: it originates the mortgages, it services them, [10] it takes the risk of default (perhaps along with a private or government insurer), and it raises money in the deposit market to fund it. The secondary market evolved by unbundling this package.

Unbundling takes advantage of division of labor and promotes competition among the suppliers of the various bundles, but it has a cost. In particular, there is a “principal/agent” problem: the principals (e.g., ultimate investors) depend on agents (e.g., the institutions originating and servicing the loans) to perform as promised. For those who end up taking the risk, the major principal-agent problem has come from the reliance on originators and servicers to originate good loans and service them properly.

This is all in contrast with the traditional, bundled bank, which had all the elements of the bundle under its control and was less worried that the part of the firm that originates mortgages would take advantage of the part of the firm that evaluates, monitors credit risk, and enforces the loan contracts.[11] Hence, banks have an advantage at controlling agency costs.

Modigliani and Miller

So where does Miller-Modigliani fit into this? In the pre-secondary market world in the U.S. where banks (savings and loans) did the lending, transaction costs and asymmetries were more or less the same for everyone, as were regulations. In that world MM was violated because holding less capital lowered “all-in” costs to the banks because it allowed better exploitation of the deposit insurance guarantee. A price paid for this was that funding was forced through the deposit market, which has been shown not to be the most efficient vehicle for funding fixed rate mortgages (recall Reg. Q ceilings on deposit interest rates).

The advent of GSEs didn’t add or subtract much in terms of the existence of guarantees, but it changed the types of guarantees and the possible ways of operating and exploiting the guarantees, by allowing institutions to get access to the bond market through MBSs. These markets have lower transaction costs, a more elastic supply of funds, and a better way of managing interest rate risk. But the GSEs were forced, because they were secondary markets, to take on some asymmetric information problems that banks did not have to take on. So MM was still violated, but it was violated in different ways.

The subprime market was largely unburdened by regulation and guarantees. It violated MM because of asymmetric information on steroids.

Policy Issues: Structures and Guarantees

The two central policy issues are:

1. Resource allocation in normal times.
2. Systemic risk when times are bad.

The first refers to incentives for pricing and risk taking as they affect investment in housing versus other investment and allocation of housing among households. Guarantees invite moral hazard if, as is inevitable, they are imperfectly managed. The second issue refers to the center of much of the recent concern because of spillovers from housing finance markets to other markets, which sent the financial system overall into collapse and contributed to a major recession.

Guarantees have been a way of controlling systemic risk, but at the expense of better resource allocation. For instance, savings and loans in the 1980s had similar incentives to take risk and misallocate resources and provoke bailouts (of the government insurance fund that supported the S&Ls). Their collapse was the consequence of taking interest rate risk followed by excessive investment in commercial real estate after the passage of the Garn-St Germain Act (1982) and the collapse of the oil price bubble in 1986, but because of deposit insurance it did not produce bank runs to speak of or have broad effects on the economy.

The story until recently was that the main source of moral hazard was government guarantees, but we have discovered that the least guaranteed sector, private label securitization, indulged in a very large amount of moral hazard. The private market, for reasons related to policy decisions (e.g., delegating too much regulatory authority to rating agencies), did not monitor risk very well.

Moral hazard is complicated. Financial institutions do not always exploit it fully, to the extent that they have reputations or franchises to preserve. Investment banks don’t always put really bad loans into pools because they want to establish a reputation for good service and survive to collect high fees again. But institutions will ramp up risk-taking when they are in trouble or when the benefits are bigger than the expected loss in reputation or franchise value.

There are lots of trade-offs. The most obvious is that guarantees tend to produce poor resource allocation but less panic and a more stable source of funds.

Going Forward

Right now, Fannie Mae, Freddie Mac, and Ginnie Mae dominate the market, and there is little prospect for private capital moving into FF, and little prospect for banks moving heavily into mortgages in the near future. For now they are the only game in town. But what about the future?

The trade-off between agency costs and fundraising costs (and supply elasticity) would appear to be the most important issue for securitization vs. banks, but how much is not obvious. For instance, FF were clearly subject to agency costs and spent resources on managing it, but their General and Administrative costs, which include these costs, were low, much lower (under .10% of asset balance) than those of banks, and their loan default performance has been better than industry performance. On the other hand, large banks have found non-deposit ways of funding that approximate the efficiency of the Agency access to capital markets.

Recommendations

Our proposed solution is to be agnostic and set regulations that convey the same subsidy (hopefully zero) to each charter and let them duel. This will require addressing weaknesses of the banks and bank regulation along lines parallel to those of the GSEs. The dueling charter model is imperfect, but alternatives, like relying on banks alone or pretending we have a stable, unguaranteed private sector, are worse bets. Here are the GSE, bank, and market proposals that we suggest need to be followed lest risk simply be transferred to one or the other along with their guarantee:

Capital

New capital rules will have to be stronger, but not just by increasing minimum capital levels. FF had two capital rules applied to them: one was a series of stress tests and the other a minimum if they passed the stress tests. Clearly the minimum was too low, but simply raising it is not enough. Both companies managed their risks so as to keep stress test capital below the minimum because the minimum is less volatile and easier to manage. Raising the minimum alone will probably not change risk because the incentive will be to increase stress test risk to match the higher minimum. A suggestion is to make the two tests additive rather than substitutes (with some allowance for time to adjust to more volatile capital requirements). Capital rules must be risk based, and stress tests are likely to be the best single measure.

Additionally, banks and thrifts have been subject to a risk-based capital standard since FDICIA passed in 1991, as well as Prompt Corrective Action standards if capital falls below certain specified amounts. Also, since 1991, banks and thrifts have been subject to risk-based deposit insurance premiums regulated by the FDIC. As the current banking and thrift crisis demonstrates, the regulators did not apply these with vigor and let the risk-based standards lapse and, for large banks, have refused to apply the standards of Prompt Corrective Action. Rather, the Basel II capital accords that were going into effect in 2006 lowered the capital requirements for the largest 20 banks just at the time their capital should have been increasing and a year before their stock prices began to dip.

Further, less costly (e.g., for tax reasons) forms of capital, such as subordinated debt that is credibly subordinated by automatically converting it into preferred or common stock if common stock price falls below some preset level, should be considered. This form of debt could have been sold relatively easily a few years ago and could have avoided to a considerable degree the capital meltdown later.

Accounting capital rules are likely to be too slow to catch some risks, but in some cases marking to market will overstate losses to long term investors when the market evaporates and trading is based on the “lemonness” of assets rather than long-run value. [12] An alternative is to base capital regulation on a mark-to-market basis of net cash flows from all operations such that it meets a certain risk-based rule in terms of debt coverage and operational needs. [13]

Running a Portfolio

What is on or off balance sheet is an accounting rather than economic distinction, which is not obviously related to risk. [14] For instance, credit risk, which has been the source of recent problems, is accepted by FF in largely the same way whether the loans are “sold” (with a put back to FF) or debt-funded.[15] The portfolio issue is mostly interest rate risk, which is serious and was a major source of problems for Fannie Mae in the early 1980s as well as the genesis of the problem for the S&Ls. However, the size of the portfolio is not a good measure of the risk for two reasons: a very large part of the risk can be hedged by selling long-term callable debt or using other forms of debt and forward and option contracts, and second, on the other side, a large amount of risk can be hidden in a small portfolio by holding nasty CMO tranches.

An important element of risk is liquidity risk. In the current crisis, the uncertainty about the guarantee produced a liquidity problem; FF borrowing spreads went up sharply in August 2008. This is a problem that cannot be solved by interest rate swaps; it can be solved by real matched funding, that is, funding long term mortgages with long term callable bonds.

Bank and GSE portfolios are not much different; they both can take on lots of interest rate risk. However, this is an area where stress tests are relatively easy to use as a risk control, e.g., by requiring enough capital to withstand a big shock to interest rates.

Carrots, Sticks, and Market Power

A part of the balancing act for any structure is market concentration. From the standpoint of resource allocation, there is a second-best case to be made for market power as an offset to excessively low mortgage rates directing too much money to housing and as an incentive to take on less risk (to preserve the franchise); this, however, is probably not the most effective way of equating private and social costs. Market power and concentration do increase political power and limit or eliminate useful results of competition, such as innovation in alterative mortgage products and risk management. We don’t know what the right size is for financial institutions or how to set up rules to get there.

Finally, the guarantees should be explicit, and FF should be charged a risk-based user fee, which should substitute for their mission regulations (same for banks and CRA). The division of labor should be to use the certainty of the guarantee to control systemic risk and to continue to provide an elastic supply of funds, and use pricing and capital to control resource misallocation and allow the dueling charters to compete.

† Gerald A. Hanweck is professor of finance and chair of the finance area in the School of Management at George Mason University.

†† Anthony B. Sanders is Professor of Finance in the School of Management at George Mason University where he is the Distinguished Professor of Real Estate Finance. He has previously taught at University of Chicago (Graduate School of Business), University of Texas at Austin (McCombs School of Business) and The Ohio State University (Fisher College of Business). In addition, he served as Director and Head of Asset-backed and Mortgage-backed Securities Research at Deutsche Bank in New York City.

††† Robert Van Order is Professor of Economics, George Washington University, Washington, DC, and former Chief International Economist, Freddie Mac, McLean, VA.

[1] “Structured” means taking the cash flows from the pool of mortgages and selling them in non–pro rata ways, for instance, selling the interest payments to one group of investors and the principal payments to another or prioritizing the impact of default losses by having one group take the first loss. Some of these structures, particularly the latter, are discussed below.

[2] See Robert Van Order, "The U.S. Mortgage Market: A Model of Dueling Charters," Journal of Housing Research, 11 (2000): 233-255.

[3] Bernard Shull and Gerald A. Hanweck, Bank Mergers in a Deregulated Environment: Promise and Peril, (Westport, CT: Quorum Books, 2001). ch. 2.

[4] Franco Modigliani and Merton Miller, "The Cost of Capital, Corporation Finance and the Theory of Investment," American Economic Review 48 (3) (1958): 261–297

[5] This assumes limited liability by the issuer, so that the structure is forced to depend on just the asset cash flows. But if, for instance, there is an outside guarantee on debt funding, then debt funding might be preferable (see below).

[6] For a discussion of some of the history of the secondary market, see John C. Weicher, “The Development of the Housing GSEs,” in Fannie Mae and Freddie Mac: Public Purposes and Private Interests, Volume 1. ed. Peter J. Wallison (Washington D.C.: American Enterprise Institute, 1999).. For the decline of the thrift industry see Federal Deposit Insurance Corporation 1997, “History of the Eighties,” available at http://www.fdic.gov/databank/hist80

[7] See Davidson, Andrew, Anthony B. Sanders, Lan-Ling Wolff, Securitization: Structuring and Investment Analysis. Wiley, 2003 for an in-depth discussion of securitization.

[8] Alt-A loans are loans that have characteristics of prime loans (e.g., good credit history) but have less than full documentation of income and wealth.

[9] For example, Freddie Mac’s debt can be found on their website at http://www.freddiemac.com/debt/

[10] Servicing means collecting the payments and sending them to wherever they are supposed to go and managing delinquency and foreclosure.

[11] That is not to say that there is no risk. Compensation schemes could induce conflicts of interest inside the firm. The point is that conflicts inside the firm are easier to resolve.

[12] See Akerlof, George A. (1970). "The Market for 'Lemons': Quality Uncertainty and the Market Mechanism". Quarterly Journal of Economics 84 (3): 488–500.

[13] See Hyman Minsky, Discount Mechanism (Washington, D.C : Board of Governors of the Federal Reserve System, 1973).

[14] An advantage of stress tests is that they simply look at cash flows whether on or off balance sheet.

[15] For instance, subprime securities have been held in portfolio, but they could have been (and to some extent were) as easily re-securitized in some sort of CMO format with the same amount of credit risk.