How bad can a US downturn possibly get?

First, the good news. The US housing slump is unlikely to drag the US economy down very much. It is true that up to now, the major source of weakness in the US economy has been the housing sector. At the end of Q2, 2007, the Federal Reserve Board reported the market value of the US residential housing stock $21.0 trillion and single family mortgage debt at $10.1 trillion. About 14 percent of the mortgage debt, $1.3 trillion, say, is sub-prime. According to the Case-Shiller data, since the end of 2006, the average US house price may have fallen by 5 percent or so, knocking about one trillion dollars of the value of the US housing stock. The brain behind these data, Bob Shiller of Yale University, believes a further decline, over a period of years, of 15 percent is in the cards.

The reason for the good news is that there is no first-order wealth effect from a change in house prices on private consumption; a decline in house prices is a redistribution from home owners to consumers of housing services, that is, from landlords to renters. The fundamental value of a home equals the present discounted value of the flow of housing services its yields, now and in the future. This is also what those who live in the home expect to pay for it, now and in the future, either as rents or, if they are owner-occupiers, as the opportunity cost of the imputed rental income they forego by living in their homes. The FRB/US econometric model used recently by Governor Frederic S. Mishkin of the Federal Reserve Board in his Jackson Hole Conference paper  to analyze the US housing market erroneously attributes the same wealth effect to housing wealth as to stock market wealth, and thus assigns it a marginal propensity to consume of 3.8 percent rather than zero, which would be the number suggested by economic theory.

The good news is qualified because housing wealth is collateralisable wealth and can therefore help to relax liquidity and cash-flow constraints faced by households. The scope for financing consumption through mortgage equity withdrawal is reduced when house prices fall and this may cause a temporary fall in consumption. To allow for this collateral effect, Mishin raises the marginal propensity to consume from 3.8 percent to 7.6 percent.  Emulating this, I will raise my estimate of the combined wealth and collateral effect of housing price changes from zero to 3.8 percent. The impact on annual consumption of the 5 percent decline in house prices would then be $38 billion. With a $ 12 trillion annual GDP, this is a decline in demand of just over 0.31 percent of GDP – not quite the stuff recessions are made off. Even if, over the next three years, house prices were to continue to decline at a 5 percent rate, this would give us a further cumulative decline in demand of just over 0.9 percent GDP.

The decline in home prices, and the increased cost and reduced availability of finance for homebuilders and home purchases will no doubt further depress the demand for new housing. However, construction is a smallish sector in the US – only about 4.5 percent of GDP in 2007. Since 1975, the share of construction in GDP has not fallen below 3.7 percent (in 1992 and 1993). It seems unlikely that the construction sector will decline by as much as an additional one percent of GDP.

The sub-prime crisis, like any financial crisis, is first and foremost a distributional question

The sub-prime crisis, the write-downs by commercial banks and investment banks of CDO and other ABS exposures, the ABCP meltdown and related financial kerfuffles are first and foremost a redistribution of financial wealth from creditors to debtors. All these derivative claims are ‘inside’ financial claims – for every creditor there is a matching debtor. These write downs and write-offs do not in and of themselves destroy any net wealth.


The only assets for which there are no liabilities – ‘outside assets’ – are, in a closed economic system, the stocks of natural resources, physical capital, human capital and goodwill. We can take the value of equity as the best approximation to the value of the stocks of privately owned physical capital, of goodwill and of natural resources owned by private corporations. The value of the housing stock owned by households can also be measured reasonably accurately. Publicly owned infrastructure capital and human capital cannot be valued using readily available market prices.

All other financial instruments, including sub-prime mortgages, ABCP and all asset-backed securities are inside instruments for which changes in valuation do not change aggregate wealth but just redistribute it. That does not mean that such changes (and the precise circumstances under which they occur) don’t matter for aggregate demand or supply. In general they will not be neutral. But it is important to recognise that for every unhappy banker who writes off $200,000 of mortgage debt, there is one happy mortgage borrower who now will no longer have to service that debt.

Where the redistribution involved is from investment banks (and ultimately from the shareholders in these investment banks) to sub-prime borrowers the net redistributional effect on aggregate consumer demand may well be positive. Of course, defaults, personal bankruptcy, foreclosures and other reassignments of property rights involve real resource costs. A single foreclosure on a sub-prime mortgage has been estimated to cost as much as $50,000. With as many as 2.2 million families with a subprime loan made from 1998 through 2006 who expected to lose their home to foreclosure in the next few years, a real resource cost of $110 billion will be incurred.

Furthermore, the losses suffered by the banks will lower their regulatory capital. So will the need to take back on balance sheet a whole range of illiquid assets that had been parked in a variety of off-balance sheet vehicles like sivs and conduits. Ratios. How large could these numbers be? Banks have currently written down sub-prime backed assets to the tune of about $40bn, and assorted pundits estimate this number will rise to $60 billion soon. Ben Bernanke, during his testimony before the Joint Committee of the House and Senate on Thursday, November 9, gave a guestimate of $150 bn for the total losses suffered eventually by the financial system because of the subprime debacle; earlier this summer he had mentioned a figure of $100 bn. Even $150 bn seems low. New subprime issues in 2004, 2005 and 2006 were $363, $465 and $449 bn respectively. By 2004, any attempt at quality control in subprime origination had already gone out of the window, so I would expect that most of these loans will default before long, unless there is a major bail-out by the Congress. How much of the almost $1.3 trillion of subprime lending during these three years is covered by the value of the properties held as collateral is unknown, as there is no prior experience of lending to such a large subprime population. A loss of $150 bn would be just be under 12 percent; that seems low, and one could easily conceive of it being as high as 20 or 25 percent. In addition, losses will be made on subprime loans made before 2004 and after 2006, and on higher-rate loans, including Alt-A and prime loans. A $250bn to $300bn eventual loss on all mortgage-related exposure would seem to be in the ball park.

At the end of October 2007, the net worth of commercial banks in the US (as reported by the Fed) stood at just under $ 1.1 trillion (against assets of $10.7 trillion). Tier 1 capital stood approximately at $964 bn. While quite a significant share of the mortgage-related losses will be born by financial institutions other than commercial banks, such as investment banks, commercial banks’ capital will take a significant hit. In addition, US commercial banks have, through unused liquidity commitments, obligations of up to $350bn to sponsored conduits that used to fund themselves with ABCP; they also are exposed to the risk of having to take back and hold up to $140bn of loans taken out for failed leveraged buy-out type deals, and are warehousing, at a loss. an indeterminate but significant amount of loans and other assets that were intended for securitisation, or packaging into other complex structures. The combination of losses and unintended asset accumulation may depress the banks’ capital ratios to the point that dividends and share repurchases are threatened and even rights issues may have to be contemplated. All that does not do much for their willingness to engage in new lending, including to the real economy.

Possibly more serious than the objective magnitude of the losses that banks will ultimately have to face is the uncertainty, indeed ignorance, about where the losses are likely to hit. The opaqueness of many of the new financial instruments and the lack of transparency and minimal reporting obligations of many of the new financial institutions, is that we still don’t know who is exposed to what – who owns what and who owes what and to whom.

If all banks were required to mark to market or mark-to-model their assets and off-balance-sheet exposures using a common, verifiable methodology, we would not have the current situation where everyone is either trying to pass possible badly impaired illiquid hot potatoes on to a less well-informed counterparty, or trying to delay recognising the losses on those assets they cannot get rid off, in the hope that something will turn up and make all the bad things go away. It has even led to a proposal to create a kind of private market maker of last resort (the Single Master Liquidity Enhancement Conduit aka ‘Superfund’ proposed by Citigroup, Bank of America and J P Morgan Chase). There is a real risk that a facility of this kind would permit on a much larger scale the reciprocal taking in of each other’s dirty laundry at sweetheart prices that is rumoured to take place already among some banks. There is a grey area, with a thin line hidden somewhere in the middle, between wanting to prevent panic sales at fire-sale prices and trying to manipulate the market to as to avoid the recognition of the true magnitude of the losses that have occurred. In the mean time, lack of trust combines with ignorance about the true value of what’s being offered for sale to produce a de-facto buyers strike.

If the progressive reduction we are seeing in the willingness and ability of banks to lend, extends to lending to the non-financial sector (households and non-financial corporations), there could be further effects on the real economy. I believe households in the US are vulnerable to this, because they are highly indebted by historical standards and have been running financial deficits for many years. Non-financial corporates, however, are in rather good financial shape, both as regards the leverage in their balance sheets (low) and as regards their financial surpluses (adequate). In addition, non-financial corporates have the option of bypassing the banking system altogether and going to the domestic and international capital markets directly. I expect to see more disintermediation by the non-financial corporate sector at the same time that we see re-intermediation of the off-balance-sheet non-financial vehicles into the banking sector.

Other credit risk mis-pricing problems

Subprime mortgages were not the only area of credit where lending and borrowing discipline collapsed. Similar reckless behaviour could be observed in unsecured consumer lending, including credit card lending, and with car loans. Both credit card receivables and car loans have been securitised on a large scale, and indeed both assets have been combined with mortgage loans in CDOs and other complex structures. 

Good news and bad news from the rest of the world 

The major source of demand strength is the US economy is the external sector. This is not surprising, as the rest of the world is growing faster than the US and the real effective (that is, trade-weighted) exchange rate of the dollar has dropped like a stone. Exports are a much more important source of demand in the

US (12.0 percent of GDP in 2007Q3) than they were in 1975, for about 8.5 percent of GDP or in 1965, when they were 5.2 percent of GDP. From its peak in 2002Q1 the broad effective real exchange rate of the US dollar had depreciated by 25 percent by 2007Q3. The precipitous decline of the nominal and real exchange rate of the dollar since September can easily have added another 5 percent to the cumulative real depreciation rate. In real terms, the dollar today is as weak as it has been at any time since 1970.

It is therefore not surprising that the growth rate of real exports has been pretty spectacular recently (9.6 percent in 2007 Q3 on a year earlier). There is no doubt that, if the dollar stays weak (let alone weakens further) and if global growth slows down only modestly, the growth of export demand can easily more than compensate for the decline in residential investment. 

Since the last quarter of 2003, the US terms of trade (relative price of exports to imports) has declined by eight percent, that is about two per cent per year. The US trade balance deficit averaged about 5.4 percent of GDP. As a result of the terms of trade deterioration since 2003 Q4, the US has suffered an annual real income loss equal to two percent of 5.4 percent of GDP, that is 0.11 percent of GDP, which is tiny.

An index of the real oil price (West Texas Intermediate divided by CPI) peaks at 48.8 in the first half of 1980, following the second oil price shock. The latest official data, for September 2007, have the index at just over 38.4, after a trough of 6.9 at the end of 1998. Since September, the oil price has risen by a further 25 percent while the increase in the CPI has been negligible. The real price of oil today therefore again stands at around 48 – its all time peak. This oil price shock will have a marked negative effect on potential output, and will increase future inflationary pressures through the output gap channel. This adverse potential output effect of an increase in the relative price of oil is over and above any general price level effect from an increase in the dollar price of oil. If potential output weakens more than aggregate demand, the ugly word ‘stagflation’ will have to be dusted off. Inflationary pressures in the US have for a long time been higher than acknowledged by the Fed. Headline CPI inflation for the 12-month period ending September 2007 was 2.8 percent, while core inflation, the will-0’-the-wisp that used to be viewed by the Fed as a good predictor of headline inflation in the medium term, was 2.1 percent. There is no doubt that the collapse of the dollar and the explosive increase in the dollar price of oil will give further momentum to US inflation, just at the time that the growth rates of both actual and potential output are declining.

For the Fed, these are interesting times indeed. 

A final feature of the contribution of the rest of the world to the economic prospects for the US is that growth in the rest of the world, including the key emerging markets of China, Russia, Brazil and to a lesser extent India, is much more fragile than is generally appreciated. The mis-pricing of credit risk extended to the emerging markets. While sovereign risk in China and Russia is virtually absent, credit risk and other risk attached to private financial instruments is high. Neither China nor Russia benefits from the rule of law. There is no minority shareholder protection, creditor or bondholder protection, nor do we find any of the other institutions and fora for the resolution of property rights disputes taken for granted in the advanced countries. Administrative or regulatory expropriation has been a common phenomenon in Russia, Kazakhstan, Venezuela, Argentina and may also become a recurring phenomenon in resource-rich African countries and in other South-American countries. In my view these non-sovereign, private risks are significantly underestimated and underpriced.

There is also more sovereign risk than is priced in by the markets affecting emerging markets that do not have huge foreign exchange reserves and whose external surpluses are either vulnerable or absent altogether. Argentina and the Philippines are examples, and even a reasonably well-managed country like Turkey is more vulnerable to internal and external shocks than its ratings suggest.

The risk to global growth outside the US is therefore higher than generally recognised, and skewed to the downside.


The destruction of value and wealth thus far in the US as a result of the housing sector crisis is manageable. Its effects are mitigated and could well be more than offset by the strength of the export sector. However, the sub-prime crisis is but the tip of the credit risk mis-pricing iceberg. Unsecured consumer loans and car loans, and the large stock of ABS backed by credit card receivables, are waiting to join the credit risk-repricing party.

The single best thing that could happen would be for the true magnitude of the losses suffered by banks and other exposed parties to be revealed and put in the P&L. Until what happens, fear of getting stuck with the hot potato makes banks unnaturally unwilling to extend credit against the kind of collateral that they would not have thought about twice accepting at the beginning of the year. 

Continued global economic growth and dollar weakness are a necessary condition for the US to avoid a serious slowdown, or even a recession. While both may continue to materialise, the risks to global growth are higher than generally recognised and rising.


Frederic S. Mishkin (2007), “Housing and the monetary mechanism”, Finance and Economics Discussion Series, Divisions of Research & Statistics and Monetary Affairs, Federal Reserve Board, Washington, D.C. 2007/40, August.


Maverecon: Willem Buiter

Willem Buiter's blog ran until December 2009. This blog is no longer active but it remains open as an archive.

Professor of European Political Economy, London School of Economics and Political Science; former chief economist of the EBRD, former external member of the MPC; adviser to international organisations, governments, central banks and private financial institutions.

Willem Buiter's website