Housing in America
How America accidentally nationalised its mortgage market
THE most dramatic moment of the global financial crisis of the late 2000s was the collapse of Lehman Brothers on September 15th 2008. The point at which the drama became inevitable, though—the crossroads on the way to Thebes—came two years earlier, in the summer of 2006.
That August house prices in America, which had been rising almost without interruption for as long as anyone could remember, began to fall—a fall that went on for 31 months (see chart 1).
In early 2007 mortgage defaults spiked and a mounting panic gripped Wall Street. The money markets dried up as banks became too scared to lend to each other. The lenders with the largest losses and smallest capital buffers began to topple. Thebes fell to the plague.
Ten years on, and America’s banks have been remade to withstand such disasters. When Jamie Dimon, the boss of JPMorgan Chase, talks of its “fortress” balance-sheet, he has a point. The banking industry’s core capital is now $1.2 trillion, more than double its pre-crisis level. In order to grind out enough profits to satisfy their shareholders, banks have slashed costs and increased prices; their return on equity has edged back towards 10%. America’s lenders are still widely despised, but they are now in reasonable shape: highly capitalised, fairly profitable, in private hands and subject to market discipline.
The trouble is that, in America, the banks are only part of the picture. There is a huge, parallel structure that exists outside the banks and which creates almost as much credit as they do: the mortgage system. In stark contrast to the banks it is very badly capitalised (see chart 2). It is also barely profitable, largely nationalised and subject to administrative control.
That matters. At $26 trillion America’s housing stock is the largest asset class in the world, worth a little more than the country’s stockmarket. America’s mortgage-finance system, with $11 trillion of debt, is probably the biggest concentration of financial risk to be found anywhere. It is still closely linked to the global financial system, with $1 trillion of mortgage debt owned abroad. It has not gone unreformed in the ten years since it set off the most severe recession of modern times. But it remains fundamentally flawed.
The strange path the mortgage machine has taken has implications for ordinary people, as well as for financiers. The supply of mortgages in America has an air of distinctly socialist command-and-control about it. Some 65-80% of all new home loans are repackaged by organs of the state. The structure of these loans, their volume and the risks they entail are controlled not by markets but by administrative fiat.
No one is keen to make transparent the subsidies and dangers involved, the risks of which are in effect borne by taxpayers. But an analysis by The Economist suggests that the subsidy for housing debt is running at about $150 billion a year, or roughly 1% of GDP. A crisis as bad as last time would cost taxpayers 2-4% of GDP, not far off the bail-out of the banks in 2008-12.
America’s housing system has always been unusual. In most countries banks minimise their risk by offering short-term or floating-rate mortgages. American borrowers get a better deal: cheap 30-year fixed-rate mortgages that can be repaid early free. These generous terms are made possible by the support of a housing-finance machine that funnels cheap credit to homeowners and, in doing so, takes on the risk, thereby shielding both the borrowers and the investors.
For decades lightly regulated thrifts did most of this lending. But in the 1980s they blew up due to a mixture of risky lending, inadequate capital and bad bets on interest rates. Between 1986 and 1996, over 1,000 thrifts were bailed out at a cost to taxpayers of about 3% of one year’s GDP.
The vacuum left by the thrifts was filled by the new technology of securitisation, which seemed, for a while, to make the risk vanish altogether. There are several steps. Mortgages are originated, or agreed, with millions of homeowners. The loans thus underwritten are then spruced up to look more attractive or realise some profits; for example sometimes insurance may be taken out against defaults, or the rights to “service” loans (collect interest payments) sold off. Next the loans are guaranteed and securitised. The bundles of bonds thus produced are then flogged to investors. After all this, derivatives contracts are created whose value is linked to these bonds.
The machine blew up in 2006-10 for a host of reasons, the most important of which was wild and sometimes fraudulent underwriting. There was a run on mortgage bonds and on the firms that issued or owned them. There have since been three big changes.
The trouble with Gosplan
The second big change is that the government’s improvised rescue of the system in 2008-12 has left it with a much bigger role (see chart 3). It is the majority shareholder in Freddie Mac and Fannie Mae, mortgage companies that were previously privately run (though with an implicit guarantee). They are now in “conservatorship”, a type of notionally temporary nationalisation that shows few signs of ending. Other private securitisers have withdrawn or gone bust. This means that the securitisation of loans, most of which used to be in the private sector, is now almost entirely state-run. Along with Fannie and Freddie, the other main players are the Veterans Affairs department (VA), the Federal Housing Administration (FHA) and Ginnie Mae, which helps the FHA and VA package loans into bonds and sell them.
In all, these five bodies own or have guaranteed $6.4 trillion of loans: a book of exposure three times larger than Mr Dimon’s balance-sheet. The FHA, an agency tasked with promoting home ownership, has tripled its guarantee book since the crisis. The mortgage bonds into which these entities bundle their loans are perceived by investors to be almost as safe as Treasuries; though they charge a fee for this protection, it is far lower than that which private companies that do not benefit from the backing of the state would have to charge if they were taking on the same risks. Thus they face no competition.
The last big change is the withering of the derivatives superstructure. The baroque instruments of the 2003-07 bubble, such as CDOs, CLOs and swaps on the ABX Index, have been stripped back after huge losses: trading activity has fallen by 90%. The mortgage machine is safer as a result. But even shorn of this amplifying mechanism, the machine is still connected to the broader world of global finance. American banks own 23% of all government mortgage bonds.
American officials who served during the crisis tell war stories about trying to persuade their counterparts in China and elsewhere not to dump all their mortgage bonds. As a result of their efforts foreign central banks, private banks and financial firms still hold 15% of all mortgage bonds; Barclays’ mortgage-bond holdings are worth 22% of the bank’s core capital. The rest are mainly owned by domestic investment funds and the Federal Reserve which, due to its asset-purchasing scheme, holds $1.8 trillion of government mortgage bonds, or 27% of the total.
This new credit machine has plenty of flaws. Almost everyone in the business worries that regulation of the new mortgage originators which funnel loans to the government-guarantee firms is too loose, for example; supervisors are looking at tightening up. But the biggest issue is the danger that sits with the state-run securitisers that magically transform risky mortgages into risk-free bonds. With a dearth of reliable market signals and a diminished profit motive, the risk appetite of the mortgage system is now entirely controlled by administrative fiat. There are at least 10,000 relevant pages of federal laws, regulatory orders and rule books.
These are meant to prevent another blow-up by screening out undesirable loans before securitisation. They stipulate the profile of the borrower (a debt-servicing-to-income ratio of more than 43% is a poor lookout) and, indeed, the dimensions of the house (if prefabricated, it must be at least 12 feet, or 3.6 metres, across). They define the documentation required. They specify the design of mortgages: balloon payments (whereby repayment of the principal is pushed back to the end of the loan period) are a no-no, as are some fee structures. They impose rules on counterparties: mortgage insurers, for example, must have over $400m of assets at hand. Although there are no government quotas for the volume of new loans there are soft targets.
Like water through cracks, risk still finds a way in. Federal law is silent on loan-to-value limits for borrowers, so this is one area where risky lending is booming, with a fifth of all loans granted since 2012 having LTV ratios of 95%, meaning homeowners are underwater if house prices fall by more than 5%. Most of these sit with the FHA. One big bank admits that it is selling at face value high-risk loans to the government that it expects will make a 10-15% loss due to homeowners defaulting.
My indecision is final
All the new rules are silent on the mortgage system’s purpose. One potential justification is simply to facilitate a liquid mortgage-bond market. By acting as a common guarantor, the state can ensure that mortgage bonds are homogenous and easy to trade ($220 billion-worth change hands every day). Another is to subsidise home loans for a broader political or social purpose.
In the absence of a grand design or clear political direction, the mortgage machine has assumed both roles.
One response to the new mortgage system is to leave it be. After all, the previous approach, in which private securitisers played a bigger role, was a disaster. Household debt is relatively restrained at the moment; measured by debt-service-to-income ratios it is 10% below the long-term average. Based on the post-war experience, housing-debt crises come only every 25 years or so; it is not yet time to worry about another one.
Leaving aside its fundamental irresponsibility, a course of inaction carries hard-to-quantify costs in the form of subsidies for borrowers. The securitisation industry believes there are reasons for not holding it to the same standard as the banks. But imagine that it were: that it had to carry the same level of capital as banks do and to make an adequate (10%) post-tax profit on that capital. The higher costs entailed give a sense of the scale of the current distortion. On this basis The Economist calculates the subsidy on mortgages to be running at $150 billion a year, 1% of GDP. (This estimate includes the impact of the Fed’s bond-buying on interest rates and the cost of tax breaks on mortgage-interest payments.)
And the status quo also means that, in the event of another crash, taxpayers would be landed with a big bill. How big? Consider a spectrum of scenarios. At one end, the cumulative mortgage-system losses are 10%, the same as the actual losses in 2006-14 according to estimates by Mark Zandi of Moody’s Analytics. At the other, cumulative losses on all mortgages are assumed to be 4.4%—the level the Fed used in its stress tests of the banks in May 2016. Adjusting for the pockets of capital in the system, and the profits made by some parts of it, both of which can help absorb losses, this means that the total loss for taxpayers if another crisis strikes would be $300 billion-600 billion, or 2-4% of GDP. Most of this would fall on Fannie, Freddie and the FHA, which would need to draw money from the government to pay out on the insurance claims made by investors.
Such a bill would hardly bankrupt America. But it would enrage it again. It is similar in size to the $700 billion TARP bail-out that Congress reluctantly passed in 2008. Lawmakers might be unwilling to pay for a repeat performance, especially with some of the benefit going abroad—and the mere possibility of their not stumping up would set the world’s financial markets a-jitter. If Congress signed off, a populist president might still be able to scupper the deal; the credit line through which Fannie and Freddie would be paid is governed by a contract between the Treasury and their regulator that comes under the executive. The catastrophic impact that a mortgage-bond default would have on the markets would almost certainly serve to ensure that the politicians did, indeed, act. But the capacity of American politics to disregard what used to seem almost certain is on the up these days.
How to waste a crisis
It would have to recapitalise and raise its fees enough to offer an acceptable profit on that capital. The subsidy would fall. Administrative controls could be eased. The risk of loss could be passed into private hands, either by privatising the mortgage-securitisation firms or by allowing them to shrink, with private banks and insurers now able to compete on a level playing field. Using the same approach as the Fed’s bank-stress tests, the system would need about $400 billion of capital. The cost of American mortgages would rise by about one percentage point.
There are various proposals for reducing the government’s role in the system; the White House floated several in 2013, and there is a range of reform bills floating around Congress, the best of which is known as Corker-Warner. But no one is in a hurry to pass reforms that would result in higher mortgage rates at a time when the middle class is struggling. A lot of policy discussions obfuscate the basic issues, assuming either that mortgages are now much safer than they were in the past or that the mortgage-guarantee firms can be safer than the banks even though not subject to the same stringent capital rules.
The government has pragmatic reasons to procrastinate. The coupons it gets on money loaned to Fannie and Freddie count as income but their debt doesn’t end up on its books; that provides a nice fillip for the accounts. The status quo also lets it avoid confronting a noisy group of hedge funds taking legal action over the treatment of Fannie’s and Freddie’s shareholders in the bail-out. If the government were to recapitalise or restructure the mortgage firms, it would probably need to reach a settlement with the hedge funds or defeat them.
To be fair, some parts of the mortgage system are trying to find ways to push risks on to the private sector. Fannie and Freddie have written new “risk sharing” deals that take a slice of the risk on about $850 billion of bonds, and package it into securities that are sold to investors or swap contracts with reinsurance firms. But even if these measures did not look a little too like some of the opaque instruments that blew up in 2007-08 to be entirely comforting, they would be no substitute for proper reform.
So the trigger for the most recent crisis remains the part of the global financial system that has been least reformed. Mortgages are still the place where many of America’s deepest problems meet—an addiction to debt, the use of hidden subsidies to mitigate inequality, and political gridlock. In the land of the free, where home ownership is a national dream, borrowing to buy a house is a government business for which taxpayers are on the hook.