So what scale of losses would trigger the apocalypse this time round? I don’t know. Worse than that, nor does anyone else. It would be nice to think that there’s an office someplace—in the IMF, in the Federal Reserve, in the Bank of International Settlements, maybe somewhere in Brussels—where some wise officials have all the necessary data to hand. But there is no such office. Remember that Citigroup was astonished by its own insolvency. Senior officials at the bank were amazed to find themselves chopped off at the knees. Dick Fuld never thought Lehman would go under.5 James Cayne of Bear Stearns thought the same of his firm (and said so, in language that would make a lumberjack blush6).7 What we can say with confidence is that if there is a wave of losses that looks large in relation to bank capital—let’s say a trillion or so—and if one large and visible lender goes under, we will be even worse off than we were in September 2008: facing a nuclear chain reaction, where each new detonation threatens to cause a still larger explosion down the chain. And this time, there are no bunkers.
So let’s take a look at the losses that are potentially coming our way. We start with the grim probabilities of sovereign default.
Sovereign defaults
By this point in the book, you may be starting to suspect that I have a slightly negative view of the world. Personally, I’d disagree with that contention. I think I’m realistic. In addition, as we’ll see in due course, I think there are solid reasons for optimism in the medium to longer term. Nevertheless, I want to keep my personal view strictly to one side, so we’ll focus here instead on what the financial markets themselves expect, using the most recent data available.
The data we’ll use rely on credit default swaps, which as we’ve already seen are essentially a way to buy insurance against the risk of default. Because these swap prices are publicly displayed, it’s possible to see what insurance premiums are being demanded and paid. Clearly, the higher the premiums charged, the higher is the implicit risk of default. Indeed, it’s possible to use credit default swap pricing to estimate the ‘cumulative probability of default’ for all sovereign debt markets. The cumulative probability of default (CPD) is exactly what it sounds like. It’s the chance that a default happens at some point within a given period—in our case, we’re looking at the next five years.
Table 22.2 presents these data. The cumulative probability of default data has been crunched by an outfit called CMAVision, and is available online. Government debt data are sourced from the IMF and are also publicly available. The third column simply multiplies the risk of default by the amount of government debt outstanding.
Table 22.2: The cost of sovereign default
Source: Debt statistics from IMF, World Economic Outlook, April 2011; default data from CMA Global Sovereign Credit Risk Report, 3rd quarter 2011 (full details available from www.cmavision.com).
A few comments about these figures. First, I’ve tried to be kind. I’ve ignored all non-European borrowers whose public finances are questionable. I’ve also ignored all European sovereigns (notably France) where the risk of default is currently considered to be under 15%. Those are big concessions to unreality. France’s debt is already the subject of market gossip, and the table above simply assumes that the country’s debt is unchallengably strong. That’s not the case, but I’m assuming that it is. Additionally, there are numerous non-European sovereign borrowers that would be at risk in the context of a global credit meltdown—Venezuela, Argentina, Pakistan, Ukraine, Dubai, Iraq, Lebanon, El Salvador, the Dominican Republic, Vietnam, Egypt, and the State Bank of India, to mention only those countries with a risk of default at 15% or greater.
Secondly, the way I’ve presented the data tends to suggest, wrongly, that the risks are uncorrelated. In practice, the risks will be very closely correlated. If Portugal goes under, the pressure on Spain will become almost irresistibly strong. If Spain goes under, the likelihood of default by France and Spain will rocket to levels where the markets will freeze, funding ceases, and insurance becomes all but unbuyable. This point is both hopeful and scary. It’s hopeful in the sense that if the creaking institutions of the eurozone do manage to get to grips with the countries currently on their most endangered list, the prospects for everybody else immediately improve. (Though since intense financial pressure is the only thing forcing reform in Italy, for example, any relief on that front is of questionable long-term value.) But it’s also scary. Thus far, European institutions have proved hopelessly inadequate to deal with the mounting crisis. Unless they show a lot more leadership very soon, the chances of Europe-wide contagion will rapidly increase.
Finally, I don’t want to pretend that these figures are more precise than they are. For one thing, the scale of losses, as finally realized, will be significantly lower than the total scale of default. So, for example, a Greek sovereign default would likely affect all $473 billion of outstanding debt, but the affected creditors would nevertheless expect to get some portion of their money back. (At the moment, the ratings agencies are predicting between 30 and 50 cents on the dollar.) You could therefore argue that the overall figure of more than $2 trillion is unrealistically gloomy. To some extent I can agree with that, except that if a bank holds the debt of a government in default, that asset becomes almost worthless until some clarity is reached on final payouts. The debt can no longer be used as collateral. It won’t be tradable, or at least certainly not in any sizeable volume. For a significant period, the asset will be effectively worthless, no matter what the finally realized value may one day prove to be.
Besides, although the data in this table are the best available, that doesn’t make them good. Because there are huge vested interests seeking to minimize the scale of the crisis (I would number among them every European government, the British and American governments, and every European, American, and Japanese bank), the markets themselves are highly subject to manipulation. After all, some readers will have wondered why I’ve made things so complicated. Why look at the cumulative default probability as predicted by spreads in the CDS market? That sounds like technobabble. Why not take the face value of European government bonds, then subtract the current market value in order to derive the total destruction of value? That would be a beautifully direct and unanswerable calculation.
Indeed it would. Unfortunately, however, there are no dependable prices for these government bonds available anywhere. If I call a broker today and request a bid/offer price for, let’s say, a ten-year Portuguese bond, I will be quoted two prices: a price at which that broker is prepared to buy bonds and a price at which they’re prepared to sell. But if I actually seek to take them up on that bid, I’ll often find that they’re prepared to trade a maximum of one single bond. In the context of the government bond markets, that’s a little like a grocery store only being prepared to sell rice by the grain. And let’s suppose, for the sake of argument, that I’m happy to trade bonds in that way. If I actually seek to close a deal, I’ll often find that brokers suddenly declare themselves unable to complete a trade. They’ll make up some kind of story why their bid prices have disappeared, but it’s not the story that matters—it’s the disappearance. Indeed, if there’s one thing more scary than a grocer willingly to sell rice only by the grain, it’s a grocer who doesn’t even have the guts to do that.
It’s issues of that kind that have pushed me into the slightly convoluted calculation mechanism used above—but although the CDS markets are less prone to manipulation, that doesn’t mean they’re not suffering from it. If there were a transparent market in these things, I’d expect the figures emerging to be very significantly worse than those I’ve just given you.
But you’ve already got me tagged as a pessimist and I don’t want to add any further to your suspicions. So we’ll take the market data as being objective fact and simply note that, according to current prices for credit default swaps, the financial markets are braced for possible defaults of around $2,111 billion. If dominoes start to fall, tha
t number will increase very rapidly indeed, so we might want to add an extra trillion or so into our worst case scenario. (Though in fact, a real ‘worst case’ would be substantially worse than that.)
As we saw in table 22.1, the total capital in the European bank markets is just $1,297 billion. If sovereign defaults affect some $2–3 trillion worth of bonds, the destruction of value could well be in the region of $1–2 trillion. Or, to put these facts in plain English, the loss of value in the European sovereign debt market could more or less destroy all the capital available in the European banking sector.
The US housing market
The size of the US mortgage market is $13.7 trillion. Of that total, some relates to non-residential mortgages (including farms). Now, we’ve already discussed the fact that the commercial property sector is in serious difficulties, but we’re simply going to set that aside for the time being—all $2.4 trillion of it—and simply examine the $11.3 trillion residential mortgage sector.8
We saw in an earlier chapter that the US housing market is in serious trouble. Almost a third of all home sales are triggered by financial distress. Almost a quarter of all homeowners are suffering negative equity. The noted economist Robert Shiller suggests that further house price falls of 10–25% are perfectly feasible.9
The truth is that house price falls of 25% simply don’t bear thinking about. If prices sank that low, countless homeowners would seek to walk away from their mortgages, or sell their houses, sooner than service their debts. It would be appropriate and orderly to do so. Mortgage companies could hardly even offer a threat of repossession, because to repossess something is pointless unless you figure you can sell it, and under the scenario we’re discussing the market would be all but bombed out.
But still, at least the math is simple. A loss of 25% applied to a mortgage market of $11.3 trillion implies total losses of an additional $2.8 trillion. Since the total quantum of capital in the US banking sector is just $1 trillion, on this scenario the US banking sector has also disappeared.
I think it’s fair to argue, however, that $2.8 trillion seems harsh. Not all the loss in housing value will end up impairing mortgage values. Equally, a 25% fall in house prices is at the bottom end of Shiller’s range of estimates. Perhaps it would be fairer, then, to assume that the loss of value arising from the US housing market would be more contained at, let’s say, $1–2 trillion. The US banking sector has still disappeared, but at least there’d still be a handful of genuinely well-managed banks poking out above the waters, like Ararat after the Flood.
European housing markets
As we saw in the chapter on housing, however, US house prices are sensible compared with those now prevalent across parts of Europe and much of the rest of the world. Some European countries with large mortgage markets (Germany, the Netherlands, Italy) have fairly sober house prices. They’re either minimally overvalued or, in Germany’s case, actually undervalued. (I’m using The Economist’s figures, which use the long-term ratio of house prices to rents to derive a ‘fair’ valuation.10) We can therefore ignore those happy markets and restrict ourselves to focusing on the three markets—Britain, France, and Spain—where the volume of outstanding mortgages is large and where the extent of house price overvaluation is extreme.
Table 22.3: The coming European housing crash
Source: Mortgage data from European Mortgage Federation.
The potential for losses in these markets can be summarized as shown in table 22.3.11 These data, in line with my practice throughout this chapter, are deliberately optimistic. I’m ignoring all the smaller mortgage markets where prices are crazy (Sweden’s 36% overvaluation, Ireland’s 23%). I’m assuming that markets simply revert to their long-term mean levels, without overshooting on the way down as they overshot on the way up. If house prices fell to just 20% below their long-term value—which is well within the range expected by Robert Shiller for the US market—the losses wouldn’t be $1.4 trillion, but more like $2.25 trillion. But still. I’m an optimist. You can be gloomy if you like; me, I’m prepared to settle for losses of $1.4 trillion … which would almost precisely wipe out the European and British banking sectors … except that, as we’ve seen, those have already been blown out of the water by sovereign debt problems, so there’s nothing much left to be wiped out. Which is good news, I guess.
Nonperforming loans
So far, we’ve restricted ourselves to two sectors: lending to governments and lending against houses. Naturally, however, banks are involved in a far wider range of activities than just that. They lend to individuals via personal loans and credit cards. They lend to companies, large and small. They offer firms mortgages collateralized by commercial property. In times of economic distress, naturally both firms and individuals may struggle to service their debts. The more extreme the economic turmoil, the greater the likely rate of default.
Unfortunately, we lack any easy way of putting numbers to those likely defaults. Remember that banks need to fund their operations through extensive interbank borrowing and use of the various securities markets. If lenders and investors start to scent that a bank is in trouble, they’ll immediately start to withdraw funds or demand higher rates of interest. Either of those outcomes will severely hinder the chances of a weak bank making it through a difficult time unscathed. The result is that banks, particularly the weak ones, have a strong incentive to hide any troubled debt. The good news is that auditors and regulators do require disclosure of any ‘nonperforming loans,’ which means that any problems should, in theory, be out in the open.
Only they’re not. The bad news is that it’s easy to disguise a bad loan as a good one. Let’s say you’ve lent $10 million to a borrower who’s struggling. That borrower needs to pay back $3 million this year to stay within the terms of your original agreement, but he calls to say that the $3 million simply isn’t there. What do you do? Well, roughly speaking, you have three options. The first is the nuclear one. You call in your loan. The borrower can’t pay. You put the company into liquidation and an administrator tries to salvage what he can, right down to selling the company’s office furniture, if need be. You won’t get your money back, but you’ll get something.
Option two is honesty. If the borrower reckons he can find about $6 million in total over the next few years, you could choose to accept that $6 million and tell your auditor to book a loss of $4 million for the money that you will never see again. You don’t put the company into administration (an action which in itself is quite likely to impair the company’s value). Instead, you let the company go on managing its business, get what money you can, fess up to the losses, and everyone moves on. No one likes option two, but it’s still by far the best and most honest alternative.
The final option, however, is commonplace. There’s probably a technical term for it somewhere but I call it ‘deceit’ or, if you prefer something a little catchier, ‘extend and pretend.’ Here’s how it works. When your borrower tells you that he simply doesn’t have the $3 million available to service his loan this year, you tell him not to worry. You say he can pay you next year. Or the year after. You rewrite the loan agreement, in fact, to defer all those payments he can’t make and will never be able to make to sometime in the future. In cold, hard, economic terms, nothing has improved. You’ve still lost money on the loan. You know it, your borrower knows it. But crucially, the loan is no longer in default. Because you’ve rewritten the loan agreement, the borrower now looks like a model client who’s made every single payment on time.
Perhaps you think such a thing would never happen, or certainly not on any real scale and not with any banks of substance and reputation. But you’d be wrong. A study, conducted by Barclays Capital and reported by the Financial Times, took a group of twenty-seven banks and examined those loans which seemed to be nonperforming (e.g. where interest wasn’t being paid or where maturities had been extended). For the first quarter of 2011—and bear in mind that only twenty-seven
banks were involved in this study—the total of nonperforming loans was some $200 billion as compared with the reported total of $129 billion.12 That’s potentially as much as $70 billion smuggled away in games of extend and pretend.
Misrepresentation on that scale is somewhat alarming, to put it mildly. On the other hand, in the context of the figures we’ve been looking at so far, $70 billion sounds manageable. But don’t be fooled. The IMF estimated the total losses arising from the first phase (2007–10) of the credit crunch (more on this at the end of the chapter). Excluding sovereign losses (which didn’t play a part in that phase) and residential mortgages (which we’ve already covered), the IMF estimated that losses in the corporate and other consumer sectors amounted to approximately $2.2 trillion, of which about $1.4 trillion ended up pounding bank balance sheets. If there is a new global recession unfolding now, or soon, I’d expect it to be more severe than last time, simply because the effects of the previous collapse were cushioned so heavily by governments borrowing and printing money. Since governments are now out of cash, there’s no more cushioning left. A cautious estimate of possible losses would therefore be for the same again—another $2.2 trillion. A worst case scenario would add at least a trillion to that, or let’s say $3.5 trillion for the sake of a round number. As before, not all of these losses would afflict the banks themselves. Some would be borne by other parts of the financial system, but all the same, the losses could be of a system-crushing scale.
Planet Ponzi Page 28
