6
Fannie and Freddie
The goal of this strategy, to boost homeownership to 67.5 percent by the year 2000, would take us to an all-time high, helping as many as 8 million American families across that threshold.… I want to say this one more time, and I want to thank again all the people here from the private sector who have worked with Secretary Cisneros on this: Our homeownership strategy will not cost the taxpayers one extra cent.1
—President Bill Clinton
We want more people owning their own home. It is in our national interest that more people own their own home. After all, if you own your own home, you have a vital stake in the future of our country.2
—President George W. Bush
The federal government’s role in the housing market goes back at least to 1938 with the establishment of the Federal National Mortgage Association (which later became Fannie Mae) and the deductibility of mortgage interest, which is as old as the income tax.3 But the federal government’s role changed fundamentally in the 1990s, when it (along with state governments) pursued a wide array of policies to increase the national homeownership rate. I focus here on the most important change—the expansion of the role of Fannie Mae and Freddie Mac, particularly their expansion into low down payment loans.4
Some, such as Paul Krugman, argue that Fannie and Freddie had nothing to do with the housing crisis. They were not allowed to make low down payment loans; they were not allowed to make subprime loans. They were simply innocent bystanders caught in the crossfire.5 Krugman has also argued a number of times that Fannie and Freddie’s role in housing markets was insignificant between 2004 and 2006: “they pulled back sharply after 2003, just when housing really got crazy.” According to Krugman, Fannie and Freddie “largely faded from the scene during the height of the housing bubble.”
In fact, from 2000 on, Fannie and Freddie bought loans with low FICO scores, loans with very low down payments, and loans with little or no documentation—Alt-A loans.6 And between 2004 and 2006, Fannie and Freddie didn’t “fade away” or “pull back sharply.” As I show below, they still bought near-record numbers of mortgages, including an ever-growing number of low down payment mortgages. And while private players bought many more subprime loans than the GSEs, the GSEs purchased hundreds of billions of dollars of subprime mortgage-backed securities (MBS) from private issuers, holding these securities as investments:
Fannie and Freddie bought 25.2% of the record $272.81 billion in subprime MBS sold in the first half of 2006, according to Inside Mortgage Finance Publications, a Bethesda, MD–based publisher that covers the home loan industry.
In 2005, Fannie and Freddie purchased 35.3% of all subprime MBS, the publication estimated. The year before, the two purchased almost 44% of all subprime MBS sold.7
The defenders of Fannie and Freddie are right that Fannie and Freddie’s direct role in subprime lending was smaller than that of purely private financial institutions. But between 1998 and 2003, Fannie and Freddie played an important role in pushing up the demand for housing at the low end of the market. That in turn made subprime loans increasingly attractive to other financial institutions.
IT’S ALIVE!
The word “conduit” is often used to describe Fannie and Freddie’s role in the mortgage market. A conduit is a tube or pipe. Just as a tube or a pipe carries water to raise the level of a reservoir, so Fannie and Freddie added liquidity to the mortgage market, increasing the level of the funds available so that more could partake. That additional liquidity steered by Fannie and Freddie to increase loan availability above and beyond what it would be otherwise seems to be a free lunch of sorts—a way to overcome the natural impediments of timing and risk facing banks and thrifts at very little cost.
Fannie and Freddie increased liquidity to the mortgage market by buying loans from mortgage originators. Banks were happy to sell their loans and give up some of the profit because this meant they wouldn’t have to worry about lending money today that wouldn’t return for years, with all the risks of default, interest rate changes, and prepayment. Fannie and Freddie financed their purchases of loans by issuing debt. They also bundled the mortgages into securities, selling those to investors. Eventually, Fannie and Freddie also used their profits to buy the mortgage-backed securities and collateralized debt obligations issued by other players in the market.
Fannie and Freddie did indeed make homeownership more affordable and accessible. Joseph Stiglitz, in his book The Roaring Nineties, argued that the original incarnation of Fannie (as an actual government agency before it was semiprivatized in 1968) was a classic example of fixing a market failure:
Fannie Mae, the Federal National Mortgage Association, was created in 1938 to provide mortgages to average Americans, because private mortgage markets were not doing their job. Fannie Mae has resulted both in lower mortgage rates and higher home-ownership—which has broader social consequences. Homeowners are more likely to take better care of their houses and also to be more active in the community in which they live.8
But Fannie and Freddie (created in 1970) were not the textbook creations of economists. At some point, Fannie and Freddie stopped acting like models in a textbook and became something more than conduits. Politicians realized that steering Fannie and Freddie’s activities produced political benefits. And Fannie and Freddie found it profitable to be steered.
Fannie and Freddie had always had certain cost advantages that were not available to purely private players in the mortgage business. They were not subject to the same Securities and Exchange Commission disclosure regulations when they issued mortgage-backed securities. They were not subject to state and local income taxes. Both Fannie and Freddie could tap a credit line of $2.25 billion with the Treasury. The amount of capital they were required to hold was much smaller than that required of private firms.
But the most important advantage for Fannie and Freddie was the implicit government guarantee, embodied in the first letter of their names, the letter “F” for Federal. Fannie Mae’s original name was the Federal National Mortgage Association. Freddie Mac was the Federal Home Loan Mortgage Corporation. Investors believed correctly that the federal government stood behind Fannie and Freddie, which were, after all, called GSEs: government-sponsored enterprises. At the same time, Fannie and Freddie were publicly traded corporations with stockholders.
The business model at Fannie and Freddie was very simple. Because of the government guarantee, they could borrow money cheaply. They could then earn money by buying mortgages that paid a higher rate of interest than the rate Fannie and Freddie had to pay to their lenders. It was a money machine that was incredibly profitable (see figure 4).9 There was only one constraint—the government didn’t let Fannie and Freddie exploit this opportunity fully.
Figure 4. Combined Earnings of Fannie and Freddie, 1971–2007 (in billions of 2009 dollars)
Source: Author’s calculations based on data from 2008 OFHEO Report to Congress (author’s conversion to 2009 dollars).
Because the government might be on the hook for any losses, Fannie and Freddie operated under a regulatory regime where they could only buy what were called “conforming loans”—loans with at least 20 percent down, loans no bigger than a certain amount, and loans with adequate documentation. These restrictions limited Fannie and Freddie’s ability to expand and take advantage of the implicit guarantees from the government. There are only so many borrowers who can put 20 percent down.
But beginning in 1993, these restraints began to loosen.10 Fannie and Freddie faced new regulations requiring minimum proportions of their loan purchases to be loans made to borrowers with incomes below the median. In 1993, 30 percent of Freddie’s and 34 percent of Fannie’s purchased loans were loans made to individuals with incomes below the median in their area. The new regulations required that number to be at least 40 percent in 1996.11 The requirement rose to 42 percent in 1999 and continued to rise through the 2000s, reaching 55 percent in 2007.12 Fannie and Freddie hit these rising goals every year between 1996 and 2007.13
These requirements seemed like such a good idea at the time. Why not spread the benefits of homeownership more widely? Why not take advantage of the spread between the interest rate at which Fannie and Freddie could borrow and lend? Why not increase Fannie and Freddie’s profits? It seemed like such a magical free lunch: more homeowners, more profits, and more politicians who could claim they were helping people.
Which brings us to one other group sitting at the table playing with other people’s money: politicians. Politicians are always eager to spend other people’s money. It’s what they do for a living. But it’s an even better deal for politicians if they can hide the fact that they’re spending other people’s money or delay when the bill comes due. That’s what they hid with Fannie and Freddie. The politicians told Fannie and Freddie to be a little more flexible with their guidelines. As a result, more people got to own houses and the politicians got to take the credit without having to raise taxes or take away any politically provided goodies from anyone else.
Fannie and Freddie’s increases in loan purchases, especially loans to low-income borrowers, helped inflate the housing bubble. That bubble in turn made the subprime market more attractive and profitable to lenders. It also set the stage for the collapse. Housing policy interacting with the potential for creditor rescue pushed up housing prices artificially. When it all fell apart, the taxpayer paid (and is still paying) the bill.
Figure 5. Home Purchase Loans Bought by GSEs, 1996–2007
Source: Author’s calculations based on data from HUD, “Profiles of GSE Mortgage Purchases,” Tables 10a and 10b, http://www.huduser.org/portal/datasets/gse/profiles.html. Data from 1998 received from FHFA via personal correspondence. Tables 10a and 10b received from FHFA in personal correspondence.
In the crucial years of housing-price appreciation, between 1997 and 2006, the number of loans bought by Fannie and Freddie expanded dramatically. Figure 5 shows the number of home purchase loans bought by Fannie and Freddie. Home purchase loans are loans used by borrowers to purchase homes (rather than to refinance homes). The number jumped by roughly 33 percent in 1998, then by another 25 percent in 2001, and by another 20 percent in 2005. The annual number of loans they purchased doubled between 1997 and 2006.
As figure 6 shows, Fannie and Freddie’s purchases of home purchase loans made to borrowers with incomes below the median grew even more quickly. These purchases more than doubled.
Figure 6. Total Home Purchase Loans Bought by GSEs for Below-Median-Income Buyers, 1996–2007
Source: Author’s calculations based on data from HUD, “Profiles of GSE Mortgage Purchases,” Tables 10a and 10b, http://www.huduser.org/portal/datasets/gse/profiles.html. Data from 1998 received from FHFA via personal correspondence. Tables 10a and 10b received from FHFA in personal correspondence.
Fannie and Freddie’s purchases of low down payment loans (loans with a down payment of 5 percent or less, or a 95 percent loan-to-value ratio) increased at an even faster rate (see figure 7).
But if Fannie and Freddie could only buy conforming loans—with at least 20 percent down, loans no bigger than a certain amount, and loans with adequate documentation—how did the opportunities available to Fannie and Freddie expand so incredibly? With the encouragement of politicians from both parties, Fannie and Freddie relaxed their underwriting standards, the requirements they placed on originators before they would buy a loan. They called it being more “flexible.”14
Figure 7. Total Home Purchase Loans Bought by GSEs with Greater than 95% Loan-to-Value Ratio, 1998–2007
Source: Author’s calculations based on data from HUD, “Profiles of GSE Mortgage Purchases,” Tables 10a and 10b, http://www.huduser.org/portal/datasets/gse/profiles.html. Data from 1998 received from FHFA via personal correspondence. Tables 10a and 10b received from FHFA in personal correspondence.
For loans made to low-income borrowers, they created special partnerships, using new criteria to determine whether they would buy a loan from an originator.15 They partnered with some of those originators, the traditional lenders—local and national banks—to develop new products with more “flexible” standards and terms.16 And they got fancy with technology.
Around 1995, both Fannie and Freddie unveiled automated software for originating loans: Desktop Underwriter and Loan Prospector, respectively. The software made assessing the riskiness of loans more “scientific” by using credit scores. Fannie and Freddie claimed that based on statistical analyses of the relationship between credit scores and default rates, loans that were once considered too risky were now actually fine.17 These software programs allowed Fannie and Freddie to do an end run around the traditional lenders, creating a cottage industry of mortgage brokers who originated loans for Fannie and Freddie. The software made it cheaper to originate a loan. That was a good thing. But it also allowed more “flexibility” in lending standards, which ended up being a very bad thing.
A Christian Science Monitor article from 2000 discussed the impact of automated underwriting:
So for borrowers with good credit, the automated system allows higher debt-to-income ratios than conventional underwriting. That means a borrower might qualify for a larger loan than someone with the same income and poorer credit.
Some other advantages of automated underwriting:
“By analyzing the credit assessments done by Desktop Underwriter, we found that lower-income families have credit histories that are just as strong as wealthier families,” said Fannie Mae chief executive Frank Raines in a speech to the National Association of Home Builders. As a result, 44 percent of Fannie Mae’s business is now conducted with low- and moderate-income families. Mr. Raines added that having a strong credit history could offset the need for a large down payment.18
The most important change at Fannie and Freddie, however, was their approach to the down payment. In 1997, fewer than 3 percent of Fannie and Freddie’s loans had a down payment of less than 5 percent.19 But starting in 1998, Fannie created explicit programs where the required down payment was only 3 percent. In 2001, they even began purchasing loans with zero down. With loans that had a down payment, they stopped requiring the borrower to come up with the down payment out of her own funds. Down payments could be gifts from friends or, better still, grants from a nonprofit or government agency.
These changes weren’t secret; executives and politicians bragged about how Fannie and Freddie were buying riskier loans. Frank Raines, the CEO of Fannie Mae at the time, testified before the US House Committee of Financial Services in 2002:
For example, a down payment is often the single largest obstacle preventing a family from purchasing a home. Fannie Mae was at the forefront of the mortgage industry expansion into low-down payment lending and created the first standardized 3-percent-down mortgage. Fannie Mae financing for low down payment loans (5 percent or less) has grown from $109 million in 1993 to $17 billion in 2002.
We’ve also used technology to expand our underwriting criteria, so that we can reach underserved communities. For example, our Expanded Approval products make it possible for people with blemished credit to obtain a conforming mortgage loan. And we’ve added a Timely Payments Reward feature to those loans, enabling borrowers to lower their mortgage payment by making their payments on time. These mortgage features have been crucial tools in reaching into communities that were previously underserved. The mortgage market today has a wider variety of products available than ever before, and therefore is better poised to meet the individual financing needs of a broader range of homebuyers.20
Between 1998 and 2003, the absolute number of loans purchased by Fannie and Freddie with less than 5 percent down more than quadrupled (see table 1). Also by 2003, 714,000 loans, 28 percent of Fannie and Freddie’s total volume of home purchase loans, were loans with less than 10 percent down.21
When the down payment was less than 20 percent, Fannie and Freddie required private mortgage insurance (PMI). On a zero-down payment loan, for example, the borrower would take out insurance to cover 20 percent of the value of the loan, protecting Fannie and Freddie from the risk of the borrower defaulting. But starting in the 1990s, an alternative to PMI emerged—the piggyback loan, a second loan that finances part or all of the entire down payment. The use of piggyback loans grew quickly beginning in the 1990s through 2003 and even more dramatically in the 2004–2006 period.22 For example, in a study of the Massachusetts mortgage market, the Warren Group found that in 1995, piggyback loans were 5 percent of prime mortgages. The number grew to 15 percent by 2003. By 2006, over 30 percent of prime mortgages in Massachusetts were financed with piggyback loans. For subprime loans in Massachusetts, almost 30 percent were financed with piggybacks in 2003 and more than 60 percent by 2006.23
There are no public data yet available on how many of Fannie’s loans with 20 percent down were really piggyback loans with zero down—loans where the borrower had no equity in the house. Suffice it to say that Fannie and Freddie contributed to the zero or low down payment frenzy with their support of 3 percent down and eventually no money down loans. The full extent of Fannie and Freddie’s involvement in low down payment loans is unclear because of the piggyback phenomenon. Maybe we’ll find out down the road.
WHAT STEERING THE CONDUIT REALLY DID
Whether one measures by the total number of loans or by dollar volume, Fannie and Freddie took the originate-and-sell model of mortgage lending through the roof. What was really going on? Individuals, institutions, and governments were lending money to Fannie and Freddie, who used that money to buy loans from originators, who gave that money to people, who used that money to buy homes. Fannie and Freddie were conduits for investors to make loans to homeowners. Fannie and Freddie did so in wildly increasing amounts even as the quality of the loans deteriorated. Perhaps they did it in blind exuberance. But they were encouraged to be blind. When the government implicitly backed Fannie and Freddie, it severed the usual feedback loops of a market system.
The fees that Fannie and Freddie paid their originators made origination extremely profitable. Because there was no feedback loop that punished bad loans, many more bad loans were made. Not only did people borrow money as a lottery ticket, but surely originators encouraged potential homeowners by deceiving them about the financial products they were buying.24 The implicit guarantee of Fannie and Freddie and the housing mandates removed the normal restraints of prudence on homeowners and originators.
Consider an investing odd couple: the Chinese government on the one hand and my father, a cautious investor in his 70s, on the other. Both invested in Fannie and Freddie bonds because they paid more interest than Treasuries and were probably just as safe. They weren’t paying attention to what was going on with Fannie and Freddie’s portfolio of loans because they didn’t need to. They counted on the implicit guarantee. It was a free lunch for my father and the Chinese—a good return without any risk. We know investors weren’t paying attention because between 2000 and 2006, even as Fannie and Freddie took on more and more risk, Fannie and Freddie’s borrowing costs stayed constant or even fell relative to Treasuries. The market viewed bonds issued by Fannie and Freddie as almost interchangeable with Treasuries. Alas, the market was right.25
The American taxpayer ultimately paid for that “free lunch.” And a few trillion dollars flowed from the Chinese and my father and other investors into new houses and bigger houses because the Fannie and Freddie conduit offered such an attractive mix of risk and reward. That flow of money was terribly costly: channeling precious capital into housing meant it didn’t flow into other areas that were more valuable but that were artificially made less attractive. So we got more and bigger houses and less of something else—less money going to fund new medical devices, or cars that get better gas mileage, or more creative entertainment, or something else creative people could have done with more capital.
NOTES
1. William J. Clinton, “Remarks on the National Homeownership Strategy,” June 5, 1995, http://www.presidency.ucsb.edu/ws/index.php?pid=51448.
2. George W. Bush, “Remarks on Signing the American Dream Downpayment Act,” December 16, 2003, http://www.presidency.ucsb.edu/ws/index.php?pid=64935.
3. Roger Lowenstein, “Who Needs the Mortgage-Interest Deduction?” New York Times, March 5, 2006, http://www.nytimes.com/2006/03/05/magazine/305deduction.1.html?_r=2&pagewanted=print.
4. I want to thank Arnold Kling, who helped me understand the workings of the banks, the housing market, and the rationale for Fannie and Freddie in this podcast: Library of Economics and Liberty, “Kling on Freddie and Fannie and the Recent History of the U.S. Housing Market,” http://www.econtalk.org/archives/2008/09/kling_on_freddi.html.
5. See, for example, Paul Krugman, “Fannie, Freddie and You,” New York Times, July 14, 2008, http://www.nytimes.com/2008/07/14/opinion/14krugman.html?_r=2&oref=slogin.
6. See Theresa R. DiVenti, “Fannie Mae and Freddie Mac: Past, Present, and Future,” Cityscape: A Journal of Policy Development and Research 11, no. 3 (2009). Between 2001 and 2005, Fannie and Freddie purchases of single-family mortgages hit all-time highs—over $900 billion in each year and over $2 trillion in 2003. In each of those years, 5 percent of Fannie Mae’s volume was loans with credit scores below 620. Another 10 percent or more were between 620 and 660. Freddie Mac’s numbers were almost as large. In 2003, Fannie and Freddie purchased $285 billion of single-family loans with credit scores below 660. By 2008, Fannie Mae alone was holding $345 billion of Alt-A loans. See Maurna Desmond, “Fannie’s Alt-A Issue,” Forbes, May 6, 2008, http://www.forbes.com/2008/05/06/fannie-mae-closer2-markets-equity-cx_md_0506markets50.html. Below, I detail Fannie and Freddie’s involvement in low down payment loans.
7. Alistar Barr, “Fannie Mae Could Be Hit Hard by Housing Bust: Berg,” MarketWatch, September 18, 2006, http://www.marketwatch.com/story/fannie-mae-could-lose-29-bln-in-housing-bust-hedge-firm.
8. Joseph Stiglitz, The Roaring Nineties: A New History of the World’s Most Prosperous Decade (New York: W. W. Norton and Company, 2004), pp. 104–105, http://books.google.com/books?id=yxhV44nSN_4C&printsec=frontcover&dq=Joseph+Stiglitz,+The+Roaring+Nineties:+A+New+History+of+the+World%E2%80%99s+Most+Prosperous+Decade&source=bl&ots=O9HdPQven7&sig=95Fh7OkZa4eDK5E7ax1y6kRCZ10&h1=en&ei=SZKyS8O6BInQtAO098HMBA&sa=X&oi=book_result&ct=result&resnum=1&ved=0CAYQ64AEwAA#v=onepage&q=&f=false.
9. Office of Federal Housing Enterprise Oversight, 2008 Report to Congress (Washington, DC: OFHEO, 2008), http://www.fhfa.gov/webfiles/2097/OFHEOReporttoCongress2008.pdf.
10. US Department of Housing and Urban Development (HUD), “HUD Prepares to Set New Housing Goals,” U.S. Housing Market Conditions Summary (Summer 1998), http://www.huduser.org/Periodicals/ushmc/summer98/summary-2.html#note5.
11. HUD, “Overview of the GSEs’ Housing Goal Performance, 1993–2001,” http://www.huduser.org/Datasets/GSE/gse2001.pdf.
12. HUD, “Overview of the GSEs’ Housing Goal Performance, 2000–2007,” http://www.huduser.org/Datasets/GSE/gse2007.pdf.
13. Neither GSE reached the 2008 goal of 56 percent: the party was over.
14. Federal National Mortgage Association (Fannie Mae), 2002 Annual Housing Activities Report, March 17, 2003, p. 12, http://www.fhfa.gov/webfiles/382/HMG_MAE_-_2002_AHAR.pdf.
15. Jay Romano, “Your Home: Lowering Mortgage Barriers,” New York Times, October 20, 2002, http://www.nytimes.com/2002/10/20/realestate/your-home-lowering-mortgage-barriers.html?pagewanted=all.
16. “CitiMortgage and Fannie Mae Announce $100 Billion Affordable Housing Alliance,” Business Wire, October 29, 2003, http://www.allbusiness.com/banking-finance/banking-lending-credit-services/5774550-1.html.
17. See John W. Straka, “A Shift in the Mortgage Landscape: The 1990s Move to Automated Credit Evalutions,” Journal of Housing Research 11, no. 2 (2000).
18. Gary Crum, “Get Fast Loan Approaval,” Christian Science Monitor, July 3, 2000, http://www.csmonitor.com/2000/0703/p16s2.html.
19. HUD, “HUD Prepares to Set New Housing Goals,” http://www.huduser.org/Periodicals/ushmc/summer98/summary-2.html.
20. Franklin Raines, testimony before the House Committee of Financial Services, Hearing on H.R. 2575, the Secondary Mortgage Market Enterprises Regulatory Improvement Act, 108th Cong., 1st sess., 2003, http://financialservices.house.gov/media/pdf/092503fr.pdf.
21. These figures on the loan-to-value ratio are taken from HUD, “Profiles of GSE Mortgage Purchases,” Tables 10a and 10b, http://www.huduser.org/portal/datasets/gse.html.
22. One reason that piggybacks supplant PMI during this period is that the piggyback loan’s interest is tax deductible. Starting in 2007, PMI became tax deductible.
23. See Eric Rosengren, “Current Challenges in Housing and Home Loans: Complicating Factors and Implications for Policymakers,” paper presented at the New England Economic Partnership’s Spring Economic Outlook Conference, Boston, MA, May 30, 2008, figure 11, http://www.bos.frb.org/news/speeches/rosengren/2008/053008.pdf.
24. For on-the-ground examples of the incentives facing lenders and homebuilders, see Alyssa Katz, Our Lot: How Real Estate Came to Own Us (New York: Bloomsbury USA, 2009). Katz also gives an excellent overview of the myriad political forces pushing homeownership.
25. Sam Eddins, director of research at IronBridge Capital Management, pointed out to me that the cause of the spread between GSE bonds and Treasuries was not so much due to the uncertainty over whether the government would indeed rescue the GSEs in the event of default, but rather the differential tax status of Treasuries versus GSE bonds. Treasuries are exempt from state and local taxes while GSE bonds are not.