The subprime mortgage crisis in America pummelled the poor. Now, it’s about to hit the middle class too, says James Ferguson.
To paraphrase Jane Austen, it is a truth universally acknowledged that the US economy cannot recover until the housing market bottoms out. Like many such ‘truths’, there is in fact no evidence for this assertion. History shows that the stock-market finds a floor well ahead of the economy, that the economy (at least in terms of private-sector hiring and firing) bottoms out next, and that the housing market is the very last trend to turn. Therefore, any sign that house prices may soon level out would be very good news indeed for the wider US economy and even more immediately, the stockmarket. And so many US housing market statistics have now hit new lows that many commentators suggest we are near or even at the nadir. So are they right?
One encouraging indicator, say the optimists, is the US National Association of Home Builders (NAHB) survey. The survey, which has traced an almost unbroken line through the past three years, now gives somewhat mixed signals. Estate agents are more downbeat than at any time during the 25 years that the survey has been conducted – in fact, the headline NAHB reading has fallen to a level below half the low-point it hit at the bottom of the 1991 recession (see the chart). However, a small but significant minority of estate agents are now more optimistic about the prospects for the market six months ahead.
This is worse than the 1990s
Unfortunately, that optimism is not yet supported by the numbers of prospective new buyers coming through estate agents’ doors. They are staying away in record numbers. Still, some argue that this is an encouraging sign in itself. After all, things can only get better from here, can’t they? And indeed, as the chart shows, the rebound in confidence from the 1991 lows was swift and steep. However, there was no banking crisis back then, so lower interest rates fed through fairly smoothly to easier credit conditions at the banks. This time we’re facing a credit crunch – and it is biting harder than ever before.
Twice as many banks are tightening mortgage standards as did at the trough in the market in 1991, according to the latest Federal Reserve survey of senior loan officers. As for subprime mortgages, in a classic case of bolting the stable door after the horse is well over the hill, all the banks surveyed are now tightening lending standards. So unlike 1991, the Fed’s easy-money policies are not getting passed on by the banking system to Americans on Main Street. What’s more, even though US mortgage rates have been cut (specifically for qualifying new buyers) and affordability has recovered to its early 1970s highs, US banks report that mortgage demand not only remains very low but is still falling, to fresh 20-year lows. Banking crises don’t just desiccate the lenders; new borrowers dry up too.
The problem with US housing
The market is undoubtedly correcting. The lack of credit has seen the median new house price fall 21.4% from its March 2007 peak. Mortgage rates have fallen at the same time, boosting affordability from both directions – properties cost less, and so do the loans to buy them. Meanwhile, reflecting the gloomy sentiment and ultra-tight borrowing conditions (across the board, not just for home loans), house-building activity has seized up. New housing starts are running at about 250,000 a year less than even during the very worst days of the 1982 and 1991 recessions.
And this is the key to the market hitting rock bottom. The US housing market’s malaise may have been driven by the easy credit bubble and triggered by sub-prime mortgage defaults, but at heart it is an inventory-overhang problem – there are just too many houses. Just look at the ‘completed new homes for sale’ numbers. House-builders prefer to sell a new home off-plan before they’ve even broken ground. It’s much less risky that way. But most homes are sold during construction, which takes eight or nine months on average. Prospective buyers can wander around the show home before selecting their lot and choosing final fit details, like kitchen and bathroom fixtures. The worst-case scenario for the house-builders is if a new development is completed but the properties aren’t all sold. They’ve spent all the money building it, and the only way to make anything back is by selling it. So ‘completed new homes for sale’ reflect the entirely sunk-cost stock of inventory for the house-building sector. And it’s at a record high. Having never been above 125,000 since records began in the 1960s, it peaked at more than 200,000 units a year ago. That’s about $54bn of unsold stock clogging up the house-building sector’s inventory channel. Even with housing starts having nose-dived, completed stock still numbered 196,000 at the end of November – still about $50bn. The 22,000-home inventory clearance in December was seemingly only achieved by a 6% month-on-month fire-sale collapse in new house prices.
For much of the past decade, newly completed houses accounted for just 7%-8% of annual US housing sales (most people buy an existing home). But – partly due to the plunge in housing sales – the glut of new houses alone now exceeds six months of current annualised house sales. In other words, at this rate, US house hunters would have to buy nothing but new homes for the next six months to clear the overhang. If you then throw in ‘existing houses for sale’, including foreclosure auction properties (which account for more than half the home sales in many states, including California), then at current levels of demand, it would take 16 months to sell all the houses that are currently for sale. That makes the inventory problem four times as bad as the nine-year average from 1997 to 2005.
Now, there is an argument to say that the ‘existing houses for sale’ data may be inflated by sellers who are trying their luck, but don’t genuinely need to sell – if they don’t find a buyer, they’ll just take the house off the market. On the other hand, the flow of foreclosed properties is likely to stay steady, partly because of the way the US housing market works. In a key difference to the UK, US mortgages are largely non-recourse loans. That means homeowners can hand their keys back to the mortgage lender, which then has no further claim on the borrower’s assets – so they don’t have the same incentive as British homeowners to cling on to a house that’s plunging in value.
With foreclosure auction prices setting the pace, house prices are likely to keep falling until this over-supply has eased. So although plenty of market indicators are already at record lows, the idea that things can only get better from here is just wishful thinking – there are still new depths to be plumbed.
How long will prices fall?
The key to the US market bottoming out is the ‘completed new houses for sale’ data. Householders, assuming they have the resources, can pick and choose when to sell their homes. But house-builders are ultimately forced to sell – they can’t just sit on unsold stock forever. So even although US house prices have been falling for almost two years already, it looks likely that they will remain under pressure for at least three more years. That’s because, based on the rate that previous inventory peaks have been worked down, that’s how long it will take before unsold inventories reach their usual trough of around 70,000 units, or 10% of total home sales. That will require the inventory overhang to fall from more than six months of total housing sales to something below one month’s worth.
The worst could be yet to come
There’s also a whole new phase of this particular US housing slump yet to reveal itself. And it’s a problem that could be more widespread, more damaging and ultimately longer lasting than the sub-prime meltdown. To understand why, we need to look at two features of the US market that differ quite markedly from our own. As mentioned above, most US mortgages are what is known as ‘non-recourse’, meaning that the lender can only go after the borrower for the housing collateral and not for any other assets. So even if the house is sold off for less than the value of the mortgage, you don’t owe your lender any more money.
This is because Fannie and Freddie – the Government Sponsored Enterprises (GSEs) who are the dominant backers of mortgage lending in the US – do two things that no private mortgage provider could ever do. One: they take the housing collateral or the loan repayment – whichever is the lower. Two: they give borrowers 30-year fixed rates but allow them to reset if rates fall below that level at any time. Getting the worst end of the deal in these two ways, you have what amounts to a state-subsidised incentive to middle-American homeowners (voters). The fact that Fannie and Freddie briefly masqueraded as wholly private-sector entities was always nonsense.
The other major difference between the US market and ours is that there are no buy-to-let mortgages over there. What there were increasingly were ARMs (Adjustable Rate Mortgages – much like our Standard Variable Rate); option ARMs (which allow payment ‘holidays’); jumbo loans (for bigger mortgages); and self-certs (so-called ‘liar loans’), known collectively as ‘Alt-A’ mortgages.
Anyone speculating in the huge property bubble in the US, especially in the so-called ‘sand states’ (California, Nevada and Arizona), would have had to self-certify that they were owner-occupiers and that each property was their primary residence. Of course, none of these speculators, often buying off-plan, intended to hold these properties long enough for anyone to find out. Certainly the mortgage brokers didn’t seem to care. But demand evaporated very fast indeed. Pali International’s US housing analyst, my colleague Stephen East, reckons that there is now sufficient condo supply in Miami to satisfy nigh on 50 years of currently depleted demand. These condos are in many cases not just un-let but also still unfinished. Their owners will now be aware that they may have committed mortgage fraud. But the small print on the mortgage terms hides another, much more potentially disastrous revelation.
There is no cultural memory in the US of recourse mortgage loans for the reasons noted above. Fannie and Freddie guaranteed most mortgages historically, as very conservative structures, usually with minimum 25% downpayments, a maximum size and with fixed rates. However, the mushrooming of the shadow banking system led to a vast array of exciting new mortgage products. What few borrowers realised was that none of these new exotics were backed by Fannie and Freddie, but by the banks. And banks write their loans as recourse.
So America’s army of middle-class property speculators – already fretting about committing mortgage fraud – are about to find that if they default on any of their Miami condos or Vegas lifestyle plots, the bank can come after them for everything they own. So far much has been made about how the subprime crisis has hit America’s poor. But the mortgage disaster about to hit middle-class America could be even more devastating.
Meanwhile, in the UK…
Unlike America, the British property bubble is not a problem of over-supply. In the US, land is abundant and a property price boom spawns a house-building frenzy. The UK is different. Our housing stock is rather more fixed than in the US. This led most people to believe that, in the face of apparent growing demand from immigrants and from the trend towards smaller families in particular, that it was a supply shortage that was fuelling UK house-price growth. So much so that the latest housing report, chaired by Kate Barker, ex-member of the Bank of England’s Monetary Policy Committee, recommended that Britain start a huge new house-building plan – a plan which has only recently swung into action.
So it’s a shame that the government hadn’t looked a bit less politically and a bit more judiciously at what exactly was fuelling Britain’s house-price boom. For starters, we wouldn’t now be planning to build several new towns in the face of what will probably prove to be the worst downturn in UK house prices in living memory. To assess where and when we might expect UK house prices to stop falling, we have to look more closely at what made them go up in the first place.
The demand and supply argument never held much water. First, if there was a shortage of housing, why did rents fail even to keep up with the rate of inflation for five out of the last six years? Especially when you consider that immigrants are mostly economic migrants who often can’t open a bank account, let alone buy a house. The answer, as we have regularly pointed out at MoneyWeek, is that with demand almost unlimitedly high and effectively fixed, it’s the supply and price of credit that actually determines house prices. That’s bad news in a banking crisis, when credit is not only going to be tight but banks will want to withdraw loans they’ve already extended. Indeed, the Council of Mortgage Lenders already expects net mortgage lending to be negative in 2009.
Another issue for this country is that, just as in the late 1980s, our housing market appears to be six to 12 months behind the US. That means that their market has moved much further back towards equilibrium than the UK market. Not only is US affordability back at 1970s highs, but after much bigger price falls than we’ve sustained, the house price-to-income ratio is only 10-20% above long-term norms (see red line on chart). But in the UK, the house price/earnings ratio has only just begun to correct, suggesting we need to see significantly more downside before house prices are back in line with earnings. Taking the 1995 low for this ratio as our worst-case scenario and assuming zero growth in earnings (possibly a fair assumption in a deflationary recession), that would imply a further 41% downside for the Nationwide House Price Index.
Doubtless some of that fall will be negated by wage growth. But if bank crises normally witness a 36% real fall in house prices, as Harvard professor Kenneth Rogoff and his collaborator Carmen Reinhart found in a recent study, there’s no reason to think the UK will buck the trend this time around. As for how long it will take, past bank crises suggest house prices fall for six years. The last correction also saw prices fall every year from 1990 through to 1996. And at that time, there was no bank crisis and interest rates were cut each and every year. So it seems likely that, given the scale of both the distortions in long-run house price ratios and of the current banking bust, that our housing market will take at least a ‘normal’ amount of time to fix and will remain a good six to 12 months behind the US right the way through the cycle. So we appear to be looking at another five years of falling prices and a trough no sooner than 2013 or 2012 at best.