Robert Van Order on Fannie and Freddie

When I left Madison for Washington six years ago, it was to follow my wife, who was offered a terrific job practicing geriatric medicine for under-served communities at the Washington Hospital Center. I decided to go to Freddie Mac at the time, in large part because I admired many people there, including Ed Golding, who was in charge of financial research, and Bob Van Order, who was the chief economist for many years.

Bob wrote the following on the raison d'etre for Fannie and Freddie, and with his permission, I am passing it along:

UNDERSTANDING FANNIE AND FREDDIE
By Robert Van Order
University of Aberdeen and University of Michigan
July 2008
Financial markets are different from other markets. They deal intensively in information and misinformation. Most of the time information is good enough and financial markets work fine, but when there are serious doubts about the quality of information entire blocks of investors, e.g., institutional investors who are not confident about information, exit and markets break down. In the case of lending markets borrowing rates go up abruptly and some borrowers have trouble getting loans at any price. Something like this has been happening in mortgage markets, especially subprime markets, and it seems to be spilling over into other markets.
Since the Great Depression when financial markets really went crazy we have developed institutions to try to control financial panic. A big part of this development has been deposit insurance, which provides bank depositors with assurance that they can get their money no matter what their bank does. Most of the time deposit insurance has served us very well. We don’t see bank runs to any great extent; in 1987 when the stock market crashed in a way that was not that different from 1929 we saw not a whiff of a bank panic. Of course the security has come at a cost. Banks can use deposit insurance as a basis for risk-taking and cause large costs to tax payers and distort resource allocation, as was the case with the Savings and Loans in the 1980s.
Fannie Mae and Freddie Mac (FF) are a part of this apparatus. They are usually referred to as Government Sponsored Enterprises or “GSEs.” That is, they are enterprises (they are privately owned), but they have special charters and benefits primarily in the form of implicit guarantees, for which they do not pay, and regulation (for instance, FF are limited to the mortgage markets and have regulations on their capital and on lending to targeted groups) which constrain their operation. They buy mortgages from lenders and they fund the purchases by issuing their own debt and (most often) mortgage backed securities (securities backed by particular pools of mortgages). They take credit risk because they are responsible for paying off investors in their securities in the event of default. They also take interest rate risk to the extent their debt funding is out of sync with their assets’ cash flows. Interest rate risk has not been the issue recently, but credit risk certainly has.
The GSEs have several purposes. Much of the recent focus has been on subsidizing homeownership, but the really important function has been to provide “liquidity” to the mortgage market, which basically means that they keep the market open even when times are tough, like now. Along with this they provide an element of standardization, which helps lenders and investors better understand what they are originating and buying. They can do this in part because their implicit guarantee allows them to raise money in bad times in the same way that deposit insurance allows banks to raise money even when they are in trouble.
Guarantees involve a subsidy, if they are not paid for. In FF’s case their debt has been rated AAA or AAA+ because of the government connection; whereas on its own it has been rated in the low AA range. The difference in borrowing rates between the two is around a quarter of a percent to .40%, which is a rough measure of their subsidy. Banks get a similar subsidy, which varies from bank to bank. The subsidy is probably larger now.
The analogy with deposit insurance is deliberate. Banks are de facto GSEs. Indeed, so are most major financial institutions around the world, which is in part why we have had relatively stable financial markets for decades. That does not mean that all GSEs are good things or that regulation can’t be improved on, but we’re not in uncharted water here. Nor are we looking at unprecedented support for mortgage markets. Since at least the 1950s almost all U.S. mortgages have benefited from guarantees: initially direct insurance like FHA and deposit insurance, especially via the Savings and Loans, and more recently via GSEs. The only part of the market without substantial government presence has been the subprime market.
What we have is a trade off. The GSEs provide stability. A measure of some of the current benefit from having FF in the market can be seen from a natural experiment that comes from the way the GSEs are regulated. There is a limit on the size of loan that FF can buy (It has been changed recently; at the beginning of this year it was $417,000. The limit is indexed to house prices over time). For years the loans above this cutoff (so called “jumbo” loans) paid rates about a quarter of a per cent higher than on loans below the limit (on 30 year fixed rate mortgages). The loans above the limit are not much different from those below; they are prime loans (the borrowers have good credit histories) with similar characteristics. However, at the end of last summer, as the subprime news hit the fan, the spread rose by over 1% and has stayed in that neighborhood. Again there is nothing in the data to suggest anything like that big a difference in credit risk on these loans. The difference is due to the existence of FF as liquidity providers in their part of the market. There is every reason to believe that without FF interest rates would have gone up in a comparable manner in most of the market.
The trade off is that the GSE structure, particularly the implicit guarantee, invites risk-taking in the same way that deposit insurance invites it. This distorts resource allocation (in this case diverting resources toward housing and homeownership). Perhaps more important, politically, it risks bail-outs (like with the S and Ls in the 1980s). Both the benefits and costs of GSEs are real and need to be balanced.
Right now the two GSEs are in trouble, and it is important to understand why. We can think of their business as having two parts: the “core” business of buying and funding prime mortgages, mostly 30 year fixed rate mortgages, and a portfolio of other “non Agency” mortgage-backed securities, which are not backed by prime loans, but rather by riskier loans like subprime and “Alt-A” mortgages. FF bought pieces of subprime and Alt-A deals that had insurance and subordination (meaning there were other investors in the loan pool that took risks first (e.g., the first 15% of the pool losses), which were supposed to protect them, so their parts of the deal (tranches) were rated AAA. These have performed worse than AAA and have fallen sharply in value. The non Agency portfolio is less than 10% of the combined $5 trillion in mortgage related assets held by the two, but it has been most of the recent controversy.
The core business of FF has been experiencing large defaults, almost certainly larger than either has ever experienced. This does not appear to have been due to changes in risk-taking or growth in the companies: the prime mortgages they bought have not changed much over time, and their performance has actually been better (lower delinquency rates) that those for prime mortgages in general. Nor was excessive growth a problem; the FF market share and level of purchases dropped sharply after 2003, as the subprime share rose. As far as one can tell so far the answer to “why?” is mostly that house prices have declined faster than at any time since these two have been around (It may have been worse in the 1930s). Both companies have taken write downs from this and more is likely to come. However, while this will certainly cut sharply into profits for several years it is likely that the two will weather the storm.
How do we know this? Well, we don’t know for sure, but we can get some ideas from history. Comparable declines in prices in some recent bad regional declines suggest that there is room for quite high default rates without collapse. An underappreciated tool that the FF regulator uses to assess FF capital is a series of “stress tests, which simulate the companies’ performance under stressful (in this case large credit loss situations). While the tests are a bit complicated the principle is simple: take the worst regional experience in recent history (for which we have data; this is the “oil patch” states in the 1980s) and project it nationwide and ask if the institution has enough capital reserve to survive ten years (leaving positive capital behind). The tool has not gotten much publicity because the institutions have generally (this includes the first quarter this year) had no trouble passing . It is important in part because the stress test appears to be actually happening. It looks like the two still pass (they had considerable excess in the first quarter of this year) but not without considerable pain. They will need added capital cushions both to make sure they get through the stress and so that they can grow. This part of their problem is a little controversial, but not a lot.
It is the non agency portfolios that are the big issue. A problem in assessing the non Agency portfolio, and the economic position of the two GSEs in general, is that the two standard measures of their position—two measures of their net worth- are both wrong. Standard accounting measures are almost always wrong because they are (generally) too slow to adapt to changing economic conditions. Hence, while accounting net worth of the two has worsened as they have set aside reserves for future losses, accounting rules do not do this fast enough to anticipate the full extent of future losses, so accounting (GAAP) net worth is probably too big.
The other measure, “mark to market” net worth, which estimates the market value of assets and liabilities of the two and takes the difference as their net worth, errs in the other direction. Right now mark to market net worth is around zero or negative for the two companies, primarily because the mark to market value of the non Agency securities has fallen sharply. This has been the genesis of much of the recent news about the two. It is certainly the case that if FF had to liquidate their portfolios they would be in trouble. However, they don’t have to; they are in a position of being able to hold them and fund their assets. Normally the mark to market value would also reflect the value of holding the assets as well as the value if sold; that is how competition in the market among buyers of securities works. But that only works if the market is working.
The problem is that there has been a liquidity crunch in the “non Agency” markets for mortgage-backed securities that is worse than the one we know about and can measure in the “Jumbo” market. The market for buying these has more or less evaporated, so it is hard to get the kind of clear read on the value of the securities that we would get in a market like the one for Treasury bonds. This is what is most controversial: the argument is that the senior, would-be AAA, subprime pieces are being priced at discounts that are way above any reasonable estimate of default loss. If that is the case, then holding the securities and using the proceeds to pay off the debt that funded them has a good chance of working—certainly it is better than selling the securities into a very thin market and trying to use the proceeds to pay off the debt.
Should we worry about this? Of course, but we should not overdo. Stress test tests run by FF and by outside analysts suggest that these securities are certainly not AAA, but they are not junk bonds either, and losses will be manageable. This is probably why the Congressional Budget Office opined recently that there probably will be no cost to the recently passed “bailout” bill.
But the “probably” part matters. Things could turn south, and that could be costly. Indeed, a problem right now is that FF would get most of the upside if things improve, but Treasury (taxpayers) would be stuck with a large part of the downside. This is the balance part. Guarantees, both for GSEs and banks, give financial institutions incentives to take risks because they get the upside benefits, but not all of the downside costs, and the market does not make them pay for it. On the other hand, the situation we are facing now, a possible financial meltdown, happens infrequently, but when it does it can cause a great deal of damage, a small glimpse of which we are seeing in the Jumbo market.
The short run response, in the recent Housing Bill, to this has been to shore up the perception of a guarantee, so that it is more or less certain that FF can continue to raise money. That is what the so-called line of credit is about (Ditto for opening the Fed discount window). It gives Treasury authority to extend its ability to make secured loans to FF (at Treasury’s discretion, not FF’s). Right now it is not necessary; FF are liquid and raising money easily, but knowing that the line is there probably makes it less likely that it will be used (for the same reason that the existence of deposit insurance mitigates bank runs). A separate part of the Bill allows Treasury to buy FF stock. This makes less sense. From a public policy perspective the issue is keeping the market open, and the line of credit can do this without helping shareholders. If FF fail, the line of credit will be the main vehicle for paying off the implicit guarantee.
Longer term, the balance will most likely come from limits on risk-taking and on capital reserves. The risk-taking of FF has not been a major issue, but the capital reserve has. Reforms going forward will probably raise the minimum levels of capital. The stress tests have not been mentioned much. That is unfortunate for two reasons: One is that as house prices fall and losses get paid out the ability to withstand the stress test going forward will diminish and there will be built in need to raise capital. Second the stress tests can be revised. The 1992 law that set up FF regulation required taking the stress test from the worst regional downturn in the data. We shall surely have a region (the southwest? Florida?) from the current period that will best the oil patch states. As a result simply updating the stress test will increase required capital, and it will do it the right way—related to the risk the institutions take.
If you want to get an idea of how the two are doing, forget the accounting net worth and the mark to market net worth and look at what is happening with stress tests.
A dimension of the problem that has not been seriously addressed is that FF, like banks, do not pay for the insurance they get. A way of solving this is via user fees. Charging fees would really solidify the guarantee and take it out of the conjecture box. To some extent this is a question of the extent to which taxpayers want to subsidize mortgage rates. It is not likely to be as good a way of controlling risk as are capital requirements and stress tests.
More broadly: there are four likely ways that housing finance will get done in the U.S:
1. GSEs.
2. Non Agency (“private label”) securitization.
3. Banks (and S and Ls)
4. Government owned institutions like FHA and Ginnie Mae
These all have benefits and costs. The second is the only 100% private (without guarantee) model. It runs the risk of fragility—probably not as bad, going forward, as the subprime debacle, but at least like the current Jumbo market. The third has many similarities to the GSEs. In principle the GSEs and banks are hard to separate; in practice the GSEs have won the market, but innovations like “covered bonds” can allow banks to do almost the same thing as securitization. Both have incentives for risk-taking and risk of bail out. The fourth presents the problems of government management and inflexibility (For instance, pricing by both FHA and Ginnie Mae are fixed by statute), and it is not clear that there is less risk.
Going forward, there is no doubt that there are risks, but not the sorts of catastrophes that have been floating around the press, blogs and newsrooms. The current zero or negative mark to market net worth is more a figment of a broken market than a judgment about future prospects. In any event the last thing we should want to do now is take away the liquidity role of FF. Around 90% of all adults are homeowners at some time in their life and almost all homeowners take out a mortgage at some time. The mortgage market is important, and keeping it open is important. Sometimes you need someone to bring the punch bowl back to the party when the guests are threatening to leave.