Housing Policy, The Morning After
by lawrence j. white
lawrence j. white is a professor of economics at NYU’s Stern School, specializing in regulation. As a member of the defunct Federal Home Loan Bank Board, he helped oversee the operations of Freddie Mac.
Published May 2, 2016.
— David Frum
For better or worse, housing policy is no longer the topic du jour for cable news.
On second thought, make that for better. After falling by about one-third between mid-2006 and 2012, housing prices have generally stabilized, even rebounded a bit. With help from Uncle Sam, the mortgage market has cleaned up much of the debris left by defaults, and the default rate today is relatively modest.
But that doesn't imply anybody's truly minding the store. Federal housing-finance policy remains in limbo – and nowhere is that more evident than in the fate of Fannie Mae and Freddie Mac, the giant government-sponsored enterprises that pumped credit into the housing bubble until it burst and then became wards of Washington in September 2008.
At the time, pretty much everybody expected that these enterprises would be wound down and replaced with some new institutional arrangement designed to stabilize the residential mortgage market and make it more efficient. But that hasn't happened. Indeed, they have become even more important market-makers for home mortgages than they were in the frenzied years of the bubble.
What could and should be done about housing policy? It's best to add some perspective before asking what we've got now and how we got here.
More is ... More
Cutting to the chase, the United States' current housing policy is easily summarized in a few words: housing is good; more is better. And the way to get more is to subsidize it.
Start with the advantages of homeownership. Buyers can deduct the interest paid on mortgages from taxable income, while the state and local property taxes they pay as a consequence of homeownership are similarly deductible. And for most people, any capital gain they realize when they sell is exempt from income tax as well.
Meanwhile, Fannie Mae and Freddie Mac indirectly subsidize mortgage interest and reduce down payments by shifting credit risk from borrowers to Uncle Sam. What's more, they have company; the Federal Housing Administration, the Veterans Affairs Department and the Department of Agriculture all provide similar subsidies in their own niches of the mortgage market.
Renters get a piece of the action, too – though here, the goodies are spread around less widely. Federal, state and local programs provide vouchers so low-income households can rent from private landlords. And government agencies still shelter 1.2 million low-income households in public housing built in earlier times.
Billions here and billions there still add up to real money. The mortgage-interest deduction alone currently translates into forgone federal tax collections of about $70 billion a year. The property tax deduction and the capital gains exclusion together lead to another $50 billion in forgone federal revenues. And the credit guarantees offered by Fannie Mae, Freddie Mac and the FHA probably have an implicit cost of another $20 to $30 billion. On the rental side, the annual Department of Housing and Urban Development budget – which consists mostly of rent subsidies – is currently around $45 billion. Note, too, that none of these estimates includes unrelated state and local policies that favor housing.
One result of making homeownership more affordable may be a modest increase in the number of people living in their own dwellings. But there's a less-celebrated consequence: people buy more house. In 1950, the average new house provided about 1,000 square feet of living space; by 1980, the figure had reached 1,700 square feet. And in 2014, the average new house was over 2,600 square feet. Although much of this increase surely reflects higher incomes, some of it also reflects the impact of subsidies.
That's good, right? Not necessarily. Buying more housing because it's cheaper sucks capital from other investments. Thus, money that might have gone to repairing bridges, developing cures for cancer or financing higher education is devoted to larger great rooms, central air-conditioning and patio fire pits. Arguably as important, ill-targeted housing subsidies have contributed to suburban sprawl and led to the loss of open space, even as that sprawl exacerbated traffic congestion, air pollution and greenhouse gas emissions.
So, why is house ownership such a priority? One reason: it was touted as the symbol of the American Dream after World War II and hasn't been rethought since. Actually, there are also more tangible reasons.
Ownership allows households to escape the whims of landlords and encourages investment in maintenance and improvements. And it is widely seen (largely too optimistically) as an opportunity for long-term wealth accumulation and a hedge against inflation. Moreover, there is some evidence that ownership generates broader societal benefits in the form of attachment to community that creates support for good government, public education and other socially desirable goals.
There is also an element of opportunism on the part of antipoverty advocates, who believe that if they support subsidies for the middle class, low-income renters will be asked along for the ride. Moreover, the focus on housing creates natural allies with the construction unions, which remain a force in urban areas, where many poor people live. In this context, it's also worth noting that voucher-based subsidies may be the path of least political and social resistance to achieving increased neighborhood diversity.
Whatever the origins of subsidy-based housing policies, they've amassed an interest-group juggernaut to defend them. The fortunes of tens of thousands of businesses, along with tens of millions of workers, are affected. Obviously, that includes the construction firms and mortgage intermediaries. But it also includes enterprises ranging from heavy-equipment manufacturers to appliance makers to insurance companies – enterprises that are not strangers to organized lobbying.
Where the Windfalls Land
How effective are subsidies? As was mentioned above, they certainly encourage households to occupy more house. But do we get much bang for a buck in terms of broader ownership? Hardly. The tax preferences are sharply biased to favor higher-income households. First, they are structured as tax deductions. Thus, middle-income households that take the standard deduction are out of luck. And among households that do itemize, the value of the tax breaks rises along with their incomes. Consider, too, that higher-income households are far more likely to carry disproportionately large mortgages, meaning they will be able to deduct more dollars as well as get more of a tax break per dollar of interest paid.
High-income households are favored in subtler ways, too. The government-sponsored enterprises have historically tended to favor more expensive houses. Although the median existing house in the United States in 2015 sold for around $220,000, Fannie Mae and Freddie Mac will guarantee mortgages as large as $417,000 in most areas and up to $625,500 in the high-cost coastal regions. So the subsidies strongly favor higher-income households, who would likely buy anyway.
Rental subsidies, by contrast, are more narrowly targeted. Old-fashioned public housing sets tough income limits, while rental vouchers are distributed solely to lower-income households.
Mortgage Finance for Dummies
Well, OK; you're not dummies, and you may already know much of this. But a refresher will set up the meatiest part of this article. Lenders face two sorts of risks: credit risk – the chance that the loan won't be repaid in a timely fashion – and interest-rate risk – the chance that interest rates will rise, making the loan less valuable in the secondary market.
To protect themselves against credit risk, lenders must do their homework, evaluating the prospective borrowers' ability to meet their obligations, a process that goes by the name of underwriting. Lenders often require borrowers to collateralize loans with an asset that can be seized to cover a default. In real estate, it's a matter of course; with residential mortgages, the houses serve as collateral and the down payments reduce the prospect that the collateral will prove inadequate.
Indeed, lenders typically do even more to protect against credit risk. They may require a guarantor (a co-signer) as a backup, especially if the loan collateral is problematic. And they typically require home mortgages to be self-liquidating, with monthly payments that return a small but growing bit of the principal along with the interest.
In the postwar years, underwriting was generally quite conservative, in part because the loans were typically of very long duration (30 years), and in part because repayment guarantees could be obtained from the FHA or government-sponsored enterprises only if mortgages conformed to conservative standards. However, beginning in the late 1990s, the growing belief that housing prices would always increase eroded these normal underwriting precautions. If housing prices would always increase, the quality of underwriting hardly mattered: sufficient collateral would always be there to make the lenders whole. Meanwhile, the growth of the secondary market for mortgages and the securities created by bundling them made it easy to unload risky loans.
Lenders (and the buyers of mortgage securities) generally prefer adjustable-rate loans, with the rate floating along with some measure of general interest rates. This pushes the aforementioned interest-rate risk on to the borrower. By contrast, borrowers usually prefer the certainty of a fixed interest rate. And since the 1930s, American housing-finance markets have tilted toward fixed-rate mortgages, in part because those mortgages were championed by federal policymakers.
But one additional contract provision made fixed-rate mortgages irresistible to borrowers: the discretion to repay the loan before the due date without any fee for doing so, which frees those loans of interest-rate risk. So when market rates decline, borrowers can refinance at lower fixed rates, effectively denying the lender any benefit. But when interest rates rise in the economy, they can sit tight, paying below-market rates. Heads I win, tails you lose.
So What Could Possibly Go Wrong?
Before the 1990s and the creation of a national (and then global) market for residential mortgage-backed securities, lenders put to work the money of their own depositors, profiting from the difference between the interest rates they paid and the interest they collected on the mortgages. Their success thus depended on how well they assessed the creditworthiness of borrowers and how efficiently they serviced the loans. Banks paid for their own mistakes unless they received a guarantee from a government-sponsored enterprise or the FHA, which protected themselves by insisting on conservative underwriting standards.
But one factor was beyond the lenders' control: their income was limited by the fixed rates collected on long-term mortgages (typically for 30 years), while their interest expenses for deposits fluctuated with the market. This was an especially severe problem for savings institutions, which were often prevented by their charters from diversifying their portfolios into other kinds of assets or even from originating adjustable-rate mortgages. When interest rates rose sharply – as they did in the late 1970s and early 1980s – these institutions were caught between the proverbial rock and hard place. Thereafter, Washington gave the savings-and-loan industry the discretion to diversify investments.
The savings-and-loan associations gradually withdrew from the businesses of holding and servicing mortgages. However, an alternative mortgage-finance structure filled the gap: securitization. Here, an originating institution bundled hundreds or even thousands of mortgages into a security – really, a bond – that could be sold to third-party private investors. The interest and principal repayments of the underlying mortgages were passed through to the security holders by a servicing agent.
Note that the once vertically integrated housing-finance industry quickly fragmented into specialized services: one institution marketed and originated loans to homebuyers, another bundled mortgages into securities and sold them and yet another serviced these mortgages on behalf of securities owners.
In large part, this was and still is good news. Mortgages can be financed by a far broader range of institutional lenders, many of which (think of insurance companies and pension funds) are uniquely suited to holding fixed-rate mortgages because their liabilities are both long term and highly predictable. Another advantage is that specialization allows for greater efficiency and may encourage economies of scale. It may be mildly unpleasant for borrowers to deal with faceless computers and with call centers in the Philippines, but it's a lot cheaper than dealing with the bean-counting banker in It's a Wonderful Life.
However, there's also a major drawback to de-integration: each handoff – from originator to packager to investor – puts more distance between the lender and borrower, and thus puts a greater premium on truly understanding the risks involved. In particular, how can the investor or lender be sure that the originators have properly screened and vetted the borrowers and have issued mortgages only to creditworthy borrowers?
Consequently, it was no accident that the initial securitization of residential mortgages in 1970 was arranged by the Government National Mortgage Association – friends call her Ginnie Mae – an agency run by the Department of Housing and Urban Development. The mortgages being securitized were insured by the FHA (another government agency) while Ginnie "wrapped" FHA's insurance with its own promise of prompt payment to investors in the event that the underlying mortgage borrower defaulted. Investors in these securities thus didn't need to know much about the creditworthiness of borrowers – that was Uncle Sam's problem.
It was also no accident that the next enterprises to arrange mortgage securitizations were Freddie Mac (1971) and Fannie Mae (1981). Unlike Ginnie, their debt securities were not formally the liabilities of the U.S. government. But investors didn't believe Washington would ever leave either of them twisting in the wind (correctly, it turned out). Indeed, by the end of 2007, the government-sponsored enterprises' mortgage-backed securities accounted for over $3.5 trillion (no misprint) in mortgages when the overall face value of residential mortgages outstanding was around $12 trillion. In addition, because they could borrow cheaply (again, because everybody believed Uncle Sam would back them in a pinch), they had bought another $1.5 trillion in mortgages and held them as direct investments on their balance sheets.
Although there had been sporadic initiatives to arrange private-mortgage securitizations starting in the mid-1980s, these efforts did not gain momentum because of market concerns about credit risk. In the late 1990s, however, the market took off when the mortgage-bond securities were structured in "tranches," in which the senior creditors would have first dibs on the cash flow. Credit-rating agencies were corralled into giving the senior tranches the top ratings that insurance companies and other regulated institutions needed in order to buy them. That left the owners of junior tranches at the end of the line. But these junior securities were successfully marketed like high-yield bonds, with bundlers playing down the risk and touting the promised returns. Or they were recombined into new bundled securities – collateralized loan obligations – and the slicing and dicing could begin anew.
Sometimes there were as many as 15 to 20 tranches carved from the same bundle of mortgages, with each tranche (starting with the most junior) providing a buffer for the next-most-senior one. The risks buried in such complex securities were hard to analyze. And, apparently, few buyers (typically institutions) bothered. Once again, the conviction that housing prices would always increase led most market players to complacency.
Mortgage originators took advantage of this belief to extend credit to ever-shakier borrowers, making the pitch that borrowers would always be able to refinance or sell their houses for a profit. Even mighty Fannie Mae and Freddie Mac – which had previously maintained tight creditworthiness standards with respect to the mortgages that they would securitize or hold in their own portfolios – succumbed around 2004 and lowered their underwriting standards as well.
Of course, the music eventually stopped, with housing prices peaking in mid-2006 and sliding thereafter. Delinquencies and then defaults followed, while the annual flow of private-label mortgage securitizations that lacked a government guarantee fell from hundreds of billions of dollars' worth annually as late as 2006 to zilch in 2008. There has been almost no revival of private-label securitization since.
Although government-sponsored enterprises' securitizations continued largely unhindered (thanks to the comforting backstop of Uncle Sam), by the summer of 2008 Fannie Mae and Freddie Mac were rumored to be insolvent. Thereafter, their government-connected status was not sufficient to convince private investors to buy their IOUs. And in early September 2008, the two cash-short entities were put into government conservatorships, where they remain. They're still in business, though, thanks to their access to federal guarantees for their securities. Indeed, they are the primary suppliers of residential mortgage credit in the country.
Housing Policy 2.0
It's hard to create a coherent housing policy without a clear sense of what we'd like to achieve. My own first priority would be to reduce subsidies – especially the subsidies that are defended as promoting ownership. Yes, there are societal benefits to homeownership that go beyond the benefits to the homeowners. But there are also social costs, not the least of which is putting the savings of the struggling middle class at the mercy of the nice folks who brought you the mortgage crisis of 2007. Equally to the point, there are better ways to spend the hundreds of billions of taxpayer money that is now tossed into the maw of the housing behemoth each year.
If politics requires that some form of ownership subsidy remain in the tax code, at the very least we should transform the interest deduction into a refundable tax credit given to anyone who owns or rents. Or if one believes that ownership has positive spillovers worth the price, one could limit the subsidy to a first-time-buyer's tax credit – say, $5,000 per year for each of the initial five years of ownership. In either case, the tax break would have an equal net value to low-income and high-income households.
By the same token, my second priority is that renting should be seen as a respectable alternative to ownership, and not just a refuge for the young and the poor. Indeed, once one realizes that housing has not been nearly as good a long-term investment as the stock market – the figures are startling – it's obvious that homeownership is not for everyone. Houses are volatile, illiquid investments with outrageously high selling costs. Moreover, ownership impedes the ease with which households can move to take advantage of job, school and lifestyle opportunities.
And rent subsidies? As a general matter, unrestricted cash transfers are likely to be the most efficient way of making low-income households better off. Indeed, Brazil and Mexico have made big inroads in poverty with cash handouts conditioned only on an insistence that children in the household are vaccinated and go to school. But Americans have long preferred in-kind transfers – food, housing, preschool education, medical care – to cash. And some of the recent findings of behavioral economics support the idea that targeted subsidies may yield beneficial outcomes for low-income households. So, let there be rental subsidies, but in the form of vouchers that give renters maximum flexibility and narrowly target low-income families.
There would still be ways to help the struggling middle class. For starters, regulations that raise house prices should be examined with an eye to comparing costs and benefits. Some are no-brainers: restrictions on lumber imports from Canada ought to be dismantled, as should restrictions on the import of cement. More important, local building codes that increase costs without commensurate benefits in safety or reliability should be dismantled. Most important, suburban land-use zoning restrictions that prevent the construction of rental housing (or small-lot single-family houses) ought to be eliminated.
The third leg of housing policy reform is, of course, finance. Addressed in very broad terms, policy should encourage efficient allocation of capital in which mortgage rates accurately reflect risk and the opportunity cost of diverting resources from other productive uses. Unfortunately, the discussion usually revolves around code words with imprecise meanings – in particular, affordability and accessibility.
We certainly want adequate housing to be affordable and accessible, and we want to create potent incentives to encourage neighbor hood diversity. But too often those words serve as code for requiring lenders to reduce credit standards or to cross-subsidize low-income households by providing credit at below-market rates. The housing-finance system should not be a vehicle for internalizing the external societal benefits of homeownership.
In this spirit, the FHA, a government agency whose budget is part of the overall federal budget, ought to be the locus of efforts to make it easier for low- and moderate-income households to become homeowners. For example, the FHA should provide explicitly underpriced guarantees on their mortgages. But allowing people in this group to buy with only small down payments is not a good direction for policy because it puts them at greater risk of default. Down payments of 20 percent – or at least 10 percent – ought to be the target, as compared to the 3 to 4 percent currently allowed under FHA programs.
Then there is the question of limiting the size government mortgage guarantees. Currently, the FHA can guarantee mortgages as large as $625,500 in high-cost areas. A ceiling of 90 percent of the national median value of houses sold (currently around $220,000) seems like a reasonable target for FHA guarantees, regardless of whether an area is a high-cost one or not.
Next, we have to confront the conundrum of the 30-year fixed-rate mortgage. It's clear that this is an extremely popular financing instrument – both politically and in the marketplace. But it is also clear that the 30-year fixed-rate mortgage exposes lenders to a lot of interest rate risk. If banks and other depository institutions provide the credit, there is the severe maturity mismatch between their long-maturity mortgage loan assets and their short-maturity deposits. On the other hand, if bond investors provide the finance through mortgage-backed securities, the door is opened to the sorts of problems that brought such private-label securities crashing down in 2007. Further, under either model, allowing the borrower to have a fee-free option for refinancing exacerbates the interest-rate risk of the lender.
The popularity of the 30-year fixed-rate mortgage and its costs needs to be acknowledged. Let's start with an easy way to reduce those costs. If lenders were permitted to charge fees for prepayment, the competitive rate of interest would fall by an estimated half a percentage point.
© Bill Varie/Alamy
Unfortunately, prepayment fees have generally been termed "prepayment penalties," which has made them seem unfair to borrowers. But they are a normal element in commercial loan arrangements, including com-mercial real estate mortgages. They should become normal in residential real estate mortgages as well. And the word "penalty" should be banished from the lexicon of mortgage-finance policy discussions.
Then there's the question of the role depository institutions play in a market that remains dominated by fixed-rate mortgages. In 2007, banks and other deposit takers directly or indirectly held 30 percent of outstanding residential mortgages. That percentage is smaller today. But it would be reasonable to expect a return to that level if these institutions find a way to hedge against the risk of the maturity mismatch between deposits and mortgages.
Recall that insurance companies (especially life insurance companies) and pension funds have long-lived predictable obligations. It would seem a natural for them to profit by taking on some of the mismatch risk by selling derivatives to the banks – options, futures, swaps and the like – that can limit risk for the buyer. And it would be in the interest of regulators focused on the stability of the capital markets to encourage the development of these instruments. Indeed, it ought to be a priority.
The remainder of mortgages will have to be securitized. And a simple two-part tranche structure – a senior tranche that is protected by a junior tranche and by a minimum size, last-to-be-paid equity slice added by the securities packager – would seem to be a natural. Investors who seek safety (and are not afraid of long-maturity assets) would gravitate toward the senior tranche; again, this would seem to be the domain of insurance companies and pension funds (which are a ripe market, since they currently invest far less than 10 percent of their assets in mortgage securities). Hedge funds and mutual funds specializing in high-risk securities would be the likely investors in the junior tranche. Memories of what went wrong in the 2000s (along with greater regulatory oversight) should help keep the parties honest.
A big puzzle today is why private-label securitization hasn't made a comeback. Right now, the only securitization games in town are run by Fannie Mae, Freddie Mac and Ginnie Mae. After 2008, it initially seemed that they were underpricing their guarantees and thus making competition from the private sector difficult. Prior to the crash, they had charged guarantee fees of approximately 0.2 percentage points per year on the unpaid balance of mortgages. Since the interest-rate differential between mortgages backed by government-sponsored enterprises and those of equal quality that were too large for those enterprises to securitize was around 0.25 percentage points, it seemed that a small increase in the guarantee fees would level the playing field and lead to expanded private-sector securitization. But fees charged by the GSEs have since been raised by an even larger amount – they are now around 0.6 percentage points annually – and yet the private-sector securitization has not revived.
Perhaps the current policy uncertainty has created a chicken-and-egg problem. The private sector doesn't want to invest in securitization platforms that might be legislated out of existence in some future reform of mortgage-market regulation. But until a private-sector alternative is clearly viable, policymakers are unlikely to reduce the borrowing power of the government-sponsored enterprises; doing so would risk hobbling the mortgage market. Or perhaps potential institutional investors in private-label mortgage-backed securities have learned the lessons of the 2008 debacle too well – that originators and private securitizers (and the credit-rating agencies) are simply not to be trusted.
In any event, we seem stuck with a federalized Fannie Mae and Freddie Mac for now. And we don't seem to be paying much of a price, since they are tightly regulated and are not engaging in the slack underwriting that helped to inflate the bubble and drove them into the arms of the government. Indeed, they have become quite profitable.
One very real danger in letting them operate as usual, though, is that they could be used for politically inspired efforts to subsidize housing. In that vein, at the end of 2015, the Federal Housing Finance Agency proposed regulations imposing a "duty to serve" on the government-sponsored enterprises with respect to low- and moderate-income households. If the new regulations prove to have bite, these enterprises will be obliged to serve those market segments even if they can't cover their costs. And since they are currently highly profitable (with their profits going entirely to the U.S. Treasury) the government would end up funding this subsidy without Congress's assent.
Now, this may prove to be a good use of the people's money. But it is a bad precedent to fund the program by robbing Peter to pay Paul within the budgets of the GSEs, and thus not subject to routine oversight by Congress.
One positive post-crisis development has been a directive from the Federal Housing Finance Agency requiring Fannie Mae and Freddie Mac to share some of the credit risk on their mortgage-backed securities with the private sector.
In essence, Fannie and Freddie must buy insurance against limited amounts of the credit losses on the pools of mortgages that underlie their mortgage-backed securities. In a few instances, they really have bought insurance. But over 90 percent of these transactions nowadays involve the government-sponsored enterprises' issuance of so-called catastrophe bonds in which creditors would receive less than all their money back in the event that the mortgages underlying a specific government-sponsored-enterprise's security take a big hit.
The cost of managing credit risk this way is reflected in mortgage interest rates (unless the government-sponsored enterprises are persuaded to dip into another pocket to cover it). But that is the point: mortgage rates should reflect the full costs of the system that supports housing finance.
With time, one can hope that investors become more comfortable with mortgage credit risk, which would open the door to private-label mortgage-backed securities that provide some competition for Fannie Mae and Freddie Mac. This process would be aided by a phase-down in the maximum size mortgage that the entities can buy or securitize.
These are hardly radical changes. But they would push the needle in the right direction, making housing policy more transparent, as well as better aligning mortgage interest rates with costs and encouraging the private sector to compete with the government-sponsored enterprises.
Yes, We Might
Housing markets have largely healed from the deep wounds left by the crisis. But housing policy is still a mishmash of costly incentives that help the affluent more than struggling households. And the mortgage market is still largely dependent on federal guarantees to function. All this is understandable: organized groups representing almost everybody who deducts mortgage interest from taxable income is reluctant to rock the boat. In the words of Pogo, Walt Kelly's long-departed comic strip character, "We have seen the enemy, and he is us."
But another aphorism, originally attributed to Winston Churchill (and more recently to Rahm Emanuel), is still worth remembering: "Never let a good crisis go to waste." The sense of urgency for change created by the 2008 meltdown is rapidly fading. It would be a pity if we failed to use the residual room for maneuvering to create a fairer and more efficient system for financing housing.