Friday, December 17, 2010

Fannie Mae, Freddie Mac, the CRA, and the Financial Crisis

When politics and economics meet, the result looks like what happens when the Arctic air mass (sometimes know here as "the cold air from Canada") meets the warm air from the Gulf of Mexico in the Mid-West. You get some thunder and lightning, and it's hard to see what's going on.

As should not surprise anyone, the financial crisis has created opportunities for virtually everyone to promote their own ideas, by blaming the people they don't like for causing the crisis or not seeing it coming.

What did Fannie Mae (the Federal National Mortgage Association), Freddie Mac (the Federal Home Loan Mortage Corporation) and the CRA (the Community Reinvestment Act of 1977) have to do with the financial crisis, and why should we care? Some unhappy campers - the Republicans on the Financial Crisis Inquiry Commission (FCIC) - claim here that Fannie, Freddie, and the CRA had a lot to do with the financial crisis. This guy and Paul Krugman (here and here), among others, claim that the unhappy campers are liars, and that their ideas were debunked long ago. Obviously some people are not agreeing. Let's see if we can learn something here.

The CRA was certainly well-intentioned. Enacted in 1977, the Act was intended to correct problems of discrimination in lending by commercial banks. However, the law does not deal explicitly with discrimination, but instead works on promoting lending by commercial banks in the communities in which they draw deposits. There are no explicit penalties for bad behavior. Instead, the idea is to work through regulators - the Fed, the FDIC, the Comptroller of the Currency, for example - to monitor compliance with the law. Compliance is somewhat vague, and the regulators appear to have some discretion to determine what compliance means. Basically, the thrust of the law is to increase lending by commercial banks to poor people.

Now, the CRA seems like a poor piece of legislation. Straightforward Econ 101 tells us that, if the goal is to help poor people, it is more efficient to do this by simply giving them money rather than coercing someone else to help them. Of course, we know why governments tend to prefer mandates (No Child Left Behind comes to mind) rather than actually putting up the money, as mandates do not show up on the government's balance sheet. In the case of the CRA we have created a substantial bureaucratic cost. For example, every Federal Reserve Bank has a group of people, typically lodged in a community affairs department, whose sole purpose is to monitor compliance with CRA. These people take up space and are paid salaries. Further, it is not clear that CRA is effective. To the extent that banks would otherwise be compliant, the CRA just imposes costs on them - time wasted filling out forms. Further, if complying with CRA means engaging in unprofitable lending practices, this could cause commercial banks to close up shop in poor neighborhoods.

Now, what about Fannie and Freddie? These "government-sponsored enterprises" (GSEs) have a long history. Fannie Mae was created in 1938 as a governmental institution, but became "private" in 1968. Freddie Mac, in turn, was established in 1970. Today, Fannie and Freddie function essentially identically, and are of similar size. They are financial intermediaries that purchase residential mortgage loans from originators and either hold the loans in their own portfolios or package them as mortgage-backed securities (MBS) to be sold. Each institution issues debt (agency securities), and can purchase mortgage related assets, for example MBS created by other financial institutions. Fannie and Freddie receive special treatment, in that their debt is implicitly guaranteed by the federal government (a guarantee made explicit in September 2008), they pay no state or local taxes, and they are exempt from some securities regulations, among other things.

Is there a sound economic rationale for the existence of institutions like Fannie and Freddie? I don't know of any other high-income country where such institutions exist. Did those other countries somehow miss the boat? There are two issues here. The first relates to the role of securitization, and the second to whether securitization is necessarily a government activity. I like to use Canada as an example here, partly because I know something about it. In Canada, there is some securitization of mortgages, but most mortgage lending is done by chartered banks with the loans held in the banks' portfolios. Canadian banks do not securitize because they are large and geographically diversified. The reason securitization is important in the US is that we still have a large number of small banks which, if they held loans backed only by local real estate, would be poorly diversified. But do we need the government to sponsor the securitization of mortgages? Absolutely not. There is no good reason why private financial intermediaries cannot perform the same function. Of course, financial crisis experience highlights in a dramatic way the incentive problems associated with securitization. Indeed, this suggests why a banking system with large banks that hold loans in their portfolios (as in Canada) has advantages over a system with many banks and securitization.

Of course, Fannie and Freddie do not represent only an attempt by the government to replicate an activity that would otherwise be carried out by the private sector. The implicit government guarantees and favorable regulatory treatment are subsidies to the housing sector. In the United States, the housing sector has long been a sacred cow, and receives other subsidies, including the mortgage interest tax deduction. Is there an economic rationale for these subsidies? I have heard some people argue that there is a positive externality associated with home ownership. The idea seems to be that homeowners somehow take a greater interest in their communities than do renters. Baloney. First, Canada has achieved a higher home ownership rate (67%) than in the US (65%) without a mortgage interest tax deduction or any mortgage market intervention on the scale of Fannie and Freddie. This suggests that US housing subsidies show up more in the the quantities of resources allocated to the housing sector in each country rather than home ownership rates. Further, Germany has a home ownership rate of 43%, but I doubt that Germans are less community-minded than Americans. Indeed, people in other countries typically characterize Americans as being much more interested in their large houses than in public amenities like parks.

Now, how are Fannie, Freddie, and the CRA tied together? The Federal Housing Enterprises Financial Safety and Soundness Act of 1992 indirectly affected CRA compliance by requiring (note: another federal mandate) that Fannie Mae and Freddie Mac devote some of their activities to affordable housing. This of course is one of the dangers of financial institutions that are closely allied with the government. The temptation always exists to use these institutions as instruments of policy, once again by using mandates to accomplish policy goals rather than using tools that affect the government's bottom line.

While the CRA seems wrongheaded, the unhappy campers on the FCIC appear to be barking up the wrong tree in attributing an important role to CRA implementation in the financial crisis. I'm sure the banking industry would love to get rid of CRA, and I can see why attaching a goodbye-to-CRA element to a financial industry reform bill would be legislatively practical, but I don't think this one flies.

What about the role of Fannie and Freddie in the crisis? Controversy over this has been brewing for some time. One aspect of that, a few months ago, was Krugman's review of Raghuram Rajan’s book, and Rajan's rebuttal. Here is Krugman's bottom line:
Was government policy entirely innocent? No, but its sins were more of omission than commission. Fannie and Freddie shouldn’t have been allowed to go chasing profits in the late stages of the housing bubble; and regulators failed to use the authority they had to stop excessive risk-taking. But as much as conservatives would like to put soft-hearted politicians at the center of this story, they don’t belong there. And Rajan’s endorsement of the conservative story line, without even an acknowledgment of the problems of that line, comes across as slippery and evasive.
There are two story lines. First, the government-as-bad-guy story is that Fannie and Freddie, encouraged by the government, lax regulation, and their implicit subsidy, played a central role in encouraging incentive problems in the mortgage market, thus playing a crucial role in pushing up the prices of houses. This then laid the groundwork for the subsequent crash. The second story line is the private-sector-as-bad-guy story. Private mortgage originators, in league with the shadow banking system and the credit rating agencies, caused the incentive problems, and Fannie Mae and Freddie Mac went along with this to a small extent (so as to retain market share).

The private-sector-as-bad-guy story-tellers give the following arguments:

1. Fannie and Freddie's market share fell during the period leading up to the crisis. See this for example. Figure 1.2 shows a drop in the flow of new MBS by Fannie and Freddie from about 2004-2007, as subprime activity takes off.

2. The bust happened in commercial real estate as well as residential. Fannie and Freddie can only deal in residential-mortgage-related instruments. If the problems were more widespread, it can't be Fannie and Freddie causing the problem.

3. Other countries experienced housing booms and subsequent busts during the same period. Since Fannie and Freddie do not have any direct influence on housing markets in other countries, something else is going on. See (2).

There are counter-arguments to all of these points. With regard to (1), Rajan points out, in his rebuttal to Krugman's review of his book, that looking at MBS issues leaves out an important part of the story, as Fannie and Freddie both purchased substantial quantities of risky loans from originators that sat in their own portfolios, and purchased MBS from other issuers. On (2), residential and commercial real estate are clearly complementary. For example, a new housing development comes with a strip mall. In principle, high demand for housing fueled by indiscriminate lending encouraged by Fannie and Freddie also increases the demand for commercial real estate. On (3), this is neither here nor there. The large runups in house prices in other countries happened for various reasons. Whether those runups happened or not has no bearing on how I think about the role of Fannie and Freddie in the financial crisis. However, consider (again) Canada as an example. This is useful, as real GDP followed a similar path in the US and Canada during the recession, and we can sometimes think of Canada as the 51st state or the the 13th Federal Reserve District. However, the financial systems in Canada and the US are very different. During the financial crisis, there were no incentive problems in Canadian mortgage markets, no banks were severely stressed, no banks failed, and there were no bailouts. The chart shows housing prices in Canada and the US, from 2000 to the present (St. Louis Fed's housing price index for the US, Statistics Canada New Housing Price for Canada). As you can see, prices in Canada rose at a lower rate, on average, prior to 2006, when the housing bust begins in the US. Canada experiences only a dip in housing prices during the recession, with prices in Canada currently back to their peak. What does this tell us? It's certainly consistent with the idea that incentive problems in the US mortgage market drove up prices, it runs counter to the arguments of the private-sector-as-bad-guy story-tellers, but this does not tell me anything about what to attribute to Fannie and Freddie.

Where are we then? Here are my conclusions:

1. Forget about good guys and bad guys. If there was anything that the financial crisis taught us, it is that risks and incentives in the financial market have to be evaluated systemically. For example, a regulator may have considered a bank's holdings of MBS as low-risk, since they were insured by AIG, without taking account of the fact that AIG would in fact not be able to pay out on all its claims in some states of the world. Fannie and Freddie, like all other large financial institutions, played important roles in the financial crisis. Can we quantify those roles? Because of the complicated relationships among financial institutions, I think this is impossible.

2. We need financial reform that will unwind Fannie and Freddie. It has been clear for a long time that Fannie and Freddie are corrupt, inefficient institutions with no important economic role. Fannie and Freddie are currently under government conservatorship and on the receiving end of a flow of bailouts from the federal government. The only question here is how the unwinding should be done.

3. It's all about the government. As Walt Kelly would have said, "we have met the enemy and he is us." Financial crises are not a given. We experience these periodic episodes because of how our financial system was designed, and by virtue of how the regulations were set up. We cannot blame the financial crisis on Wall Street, any more than we can blame children for burning down the house when they are left at home with a box of matches and a can of gasoline. The cause of the problem was faulty design and faulty regulation. The government - i.e. you and me - did it, and we need to fix it.


  1. "Now, the CRA seems like a poor piece of legislation. Straightforward Econ 101 tells us that, if the goal is to help poor people, it is more efficient to do this by simply giving them money rather than coercing someone else to help them"

    This is one of the problems with Econ 101; students are never typically told that that depends on perfect information, perfect self-discipline, perfect expertise, no psychological disutility of choice, no externalities, etc., etc.

    Unfortunately, this isn't usually taught until second or third year economics which few people take.

    As a result of these things excluded from Econ 101, often people, and certainly society, are better off when they're not just given money.

  2. That said, the CRA in particular looks bad to me. Owning a home is not a good thing if you can't reasonably afford the payments, if you're not going to stay put for long, and other reasons. Owning a home is not always good, and neither is having access to credit, especially at high or exorbitant prices.

  3. Your number 3 has a semantics problem and can be misleading. When Republicans say it's the government's fault, what they mean (or what they want you to think) is that these problems occurred because of what the government did, and everything would have been great if the government did nothing and just let the free market be 100% free.

    What others mean when they say it was the market is that these problems occurred from government not doing enough to stop the harmful things the unfettered free market would do.

  4. Steve,

    First, thanks for following this up. I still think your conflating the arguments of "Fannie & Freddie need to go" with "Fannie & Freddie did it," but overall I thought your post was pretty even-handed.

    A couple comments:
    1) Re the CRA, most of the private-sector mortgage originators/securitizers weren't subject to the CRA, so it's even harder to blame that act (again leaving aside other possible flaws with it).

    2) Sub-prime mortgages that were subject to the CRA actually performed better than average.

    3) Re your number 3, and following up on Richard's comment, I think you could've pointed to things like Gramm-Leach-Bliley or Congress prohibiting the CFTC from regulating derivatives as useful counter-examples to the "too much evil guvmint" line.


  5. "The cause of the problem was faulty design and faulty regulation."

    1. Clarify this statement. What design, and what regulation? If by securitization design you mean the deregulatory 1984 Secondary Mortgage Market Enhancement Act, that just left the design up to Wall Street.

    2. What's your take on the role of the Canada Mortgage and Housing Corporation in the stability and extension of credit in the Canadian mortgage market? CMHC throws off some of your narrative, no?

  6. First, Canada has achieved a higher home ownership rate (67%) than in the US (65%) without a mortgage interest tax deduction or any mortgage market intervention on the scale of Fannie and Freddie

    Baloney. Can-Am homeownership rates have been relatively equal for generations and it is straightforward for a Canadian to make mortgage interest tax deductible for investment purposes.

    The CMHC book of business in the Canadian residential mortgage market is substantially more interventionist than Frannie and the FHA/VA. As of 2009, CMHC MBS guarantees and in-force insurance accounts for 85%+ of outstanding mortgage credit. Canadian chartered banks and other financial intermediaries are for the most part mere servicers, not at-risk investors, in both the MBS and portfolio mortgage market.

    The Big 5 banks in Canada put their fingers in the U.S. Federal Reserve Bank $3.3 trillion cookie jars during the panic.


  8. Thanks for this useful characterization of the debate. BTW, the following post has a nice discussion of one aspect of the 'Federal Housing Enterprises Financial Safety and Soundness Act of 1992' you mention:

    Also, I look to forward to seeing your response to the points raised wrt the CMHC.

  9. The real lesson from Fannie and Freddie's role in the crisis is that private risk taking and public guarantees are a combustible combination. In other words, moral hazard.

    Risky loans had higher profit margins, and the GSE's bought them to eek out a higher spread and generate higher returns on assets. The GSE's were essentially hedge funds with giant balance sheets, only they were allowed to have more leverage than any hedge fund, and the tax payer was put at risk for any losses. IMO, this was the outcome of a corrupt lobbying process rather than any ideological pursuit (plenty of Republicans were on the GSE contribution "payroll").

  10. "Clarify this statement. What design, and what regulation?"

    I have written about this here:

    The design problems go back to the 19th century at least. The US financial system got off on the wrong foot. We keep trying to fix it, and keep creating more problems.

    Baloney Anonymous:

    This is complete b.s. Simon Johnson had an ax to grind about big banks, and said some things like this that were less extreme, but still gross exaggeration. See this:

    1. You may find a way to deduct the mortgage interest on a rental property in Canada, but no homeowner can deduct mortgage interest on a principal residence in computing personal income taxes.

    2. CMHC is akin to the FHA. It does some direct lending, and it insures mortgages. Its role in the Canadian mortgage market is small relative to what is going on in the US.

    3. The notion that Canadian chartered banks act essentially as mortgage servicers is ludicrous.


    Yes, exactly.

  11. Actually, your graph looks like it could be bad for Canada.

    Higher (real) housing prices do far more bad than good over the long run. Who benefits? Only the small percentage of the population that will be moving from a bigger house to a smaller one (or relatively more expensive to relatively less expensive), and some investors.

    Everyone else has higher expenses and more financial stress, or far higher. In fact, increased housing costs are one of the biggest reasons why the percentage of income spent on fixed expenses by American families has skyrocketed over the last generation, and thus there has been an epidemic of financial distress. See here for details:

    I'd be happy if the average home price dropped to one cent. In the long run the gainers would swamp the losers.

    A house, for most people, is more of an expense (albeit usually a necessary one) than an investment, like a car (the average return is a smidge above inflation on a house, but is negative when adjusted for risk and opportunity cost). Like a car I'd be happy if its cost went the lower the better, even to one cent.

  12. ... no homeowner can deduct mortgage interest on a principal residence in computing personal income taxes.

    You are in error. Purchase-money mortgages are not the only type of mortgages in Canada. Learn to read more skillfully.

    CMHC is akin to the FHA. It does some direct lending, and it insures mortgages. Its role in the Canadian mortgage market is small relative to what is going on in the US.

    You are in error. 80% market share is not small. And the CMHC has public policy mandates similar to the CRA.

    The notion that Canadian chartered banks act essentially as mortgage servicers is ludicrous.

    If a bank off loads credit risk on "full faith and credit" CMHC and has limited unhedged duration risk due to short mortgage terms, it is just a spread servicer. Service my spread if the customer pays on time and collect from CMHC if the customer defaults.

    Triple baloney.

  13. "Purchase-money mortgages are not the only type of mortgages in Canada. Learn to read more skillfully."

    OK. Tell me about it.

    "You are in error. 80% market share is not small. And the CMHC has public policy mandates similar to the CRA."

    See my next post. You have a point, but I have no idea where your 80% comes from. This is market share of what? If you mean insured mortgages as a percentage of the total, it's too high. If it's MBS as a fraction of total mortgages outstanding, it's way too high.

    On chartered banks as servicers:

    1. You're only talking about insured mortggaes, obviously your argument does not apply to conventional mortgages, which are still a large (though shrinking) fraction of the total.

    2. On the insured mortgages, you're exaggerating. In the US, for example, typically an originator screens the loan, the loan either ends up in some financial institution's portfolio or is bundled into mortgage-backed securities, and some third party may collect the loan payments and maybe pay property taxes and insurance. The third party is the servicer. If a Canadian chartered bank holds an insured mortgage, it is the originator (it screened the loan), the holder of the mortgage, and the servicer. Further, I'm sure that in the event of default, it's not just a matter of collecting from CMHC. Presumably the arrangement with CMHC is that the bank seizes the collateral and goes after the borrower's other assets. Since you claim to be an expert on this, maybe you can enlighten me?

    Don't stretch the truth. It does not help you make your point. It just makes me and others inclined to discount what you're saying.

  14. Another thought for Baloney Anonymous:

    What is your point exactly? If you asked me to guess, I would say that you work for Fannie or Freddie, and you are trying to make the case that the Canadian mortgage market looks much like the US. This is certainly not the case.

    1. You may be able to find loopholes that allow you to deduct mortgage interest, but the majority of Canadian homeowners can't do it. If you can give me the details about how to do this, let me know so that I can tell my relatives living in Canada. They would love to hear about it. The characterization that mortgage interest is not deductible in Canada is a fact. Everyone knows this.

    2. CMHC's role in the Canadian mortgage market is important, and growing, and a lot of it is not good. That said, Canada does not have the severe incentive problems associated with Fannie, Freddie, and subprime lending in the crisis. The outcomes during the crisis bear witness to that view.

  15. Regarding market share. The CMHC insures about $475 billion in mortgages, probably 50% of Canadian mortgage debt outstanding. The federal government guarantees the CMHC's insurance.

    Privately owned Genworth insures about $225 billion in mortgages. The government also guarantees Genworth (although only to 90% of outstanding). The net result is that the government guarantees 65-70% of Canadian mortgage debt, which is the amount of mortgage debt currently carrying a Genworth/CMHC stamp of approval.

  16. I mean looking at that graph again, real housing prices are over 50% higher in Canada in the last decade. This is nothing to be thrilled about especially for Canadian families, and especially if houses weren't that inexpensive to start with. They might be a lot better off if they didn't have their housing run-up not deflating, like in the US.

    Of course, deflation of a housing bubble can lead to a plunge in AGGREGATE DEMAND, yeah I said it, but there are better ways to maintain aggregate demand, like high return government investment of the kind the free market grossly underprovides due to externalities, etc., etc.

  17. JP,

    I'm curious about what happens with an insured mortgage contract.

    1. Does the insurance premium vary with the type of loan? For example, is it different for a 5% down payment vs. a 10% down payment?

    2. What happens in the event of default? Who actually forecloses and goes after the borrower's assets?

    Say the Bank of Montreal holds the mortgage, I'm interested in the incentives that BMO has to exercise due diligence when they make the loan, and what the incentives are in the default stage. Do you think the contracts and insurance are set up in such a way that they have mitigated the moral hazard problem, or will BMO behave well because it is in a repeated relationship, and potentially faces increases in insurance premia if it does not behave well?

  18. For Baloney Anonymous:

    My favorite source of information, Wikipedia, tells me:

    "Canadian federal income tax does not allow a deduction from taxable income for interest on loans secured by the taxpayer's personal residence. But homes used in businesses as a landlord who owns a rental residential property can deduct interest as any other reasonable business expense. The difference being the deduction is allowed only when the property is not used for the taxpayer's personal use but is used as in any other type of business. However, there may be additional exclusions for passive activity losses.

    An indirect method for making interest on mortgage for personal residence tax deductible in Canada is through an asset swap, whereby the homebuyer sells his existing investments, purchases a house in full or in part by the sale, gets a mortgage on the house, and finally, buys back his investments with the money from the mortgage. The Supreme Court of Canada has ruled in 2001 in the Singleton v. Canada case [1] that transactions in the asset swap are to be regarded as distinct, thus rendering the interest on home mortgage acquired as part of the asset swap tax deductible.

    The home ownership rate in Canada is about the same as in the United States,[2] but Canadians have about 70%[3] equity in their homes on average (i.e., 30% mortgage debt), compared to only 45% average home equity in the United States."

  19. Stephen,

    1. Yes, for instance, see pg 10 of, a good paper by Eric Lacelles on the Cdn mortg market.

    2. My understanding is that banks are insured against mortgage default, but until default is officially declared, banks are held responsible by the CMHC for topping up unpaid balances ie. they face significant risks (collection and enforcement) as insured mortgages go into arrears and stay that way. I'm sure there are all sorts of court fees and legal frictions involved in moving a mortgage from "arrears" to "default".

  20. " and as Koning points out, the private sector via GE Capital (Genworth)insurance makes it close to 100%."

    No, I pointed out that it was 65-70%. The remaining 30% is uninsured and probably on banks' balance sheets. As an aside, Anon, ad hominems and misquotations don't make for good copy.

  21. JP,

    Thanks for all the information. My conclusion is that, while mortgage insurance with government participation has grown to actually exceed conventional mortgage lending, and MBS activity has also grown in Canada (also propelled by the government), this has not shown up in a significant increase in risk borne by the Canadian taxpayer or by the financial system. However, the trend is troubling. The mortgage and housing markets seemed to be working fine in Canada, say in the 1990s, and presumably an increase in the relative price of housing (particularly in Toronto and Vancouver) propelled a social policy of increased government participation to lower the cost of home ownership for low-income people. Of course you only have to look south of the border to see what the problems can be.