Global
By Stephen Roach | New York
March 23, 2007
SourceThe bursting of two bubbles seven years apart – dot-com and housing – holds the key to the macro outlook. While different in many respects, these sharp swings in asset markets share one thing in common – the initial belief that any spillovers would be limited and that the rest of the economy and financial markets would remain unscathed. Just as that view turned out to be wrong in the early 2000s, I fear a similar outcome today.
There are two ways to look at spillovers -- contagion within an asset class and repercussions from one sector to another in the real economy. Both are obviously important. Seven years ago, the asset-market excesses were most acute in the so-called pure Internet plays – a collection of over 350 companies that had a combined equity market capitalization of over $1.1 trillion in late 1999, or 6% of the total value of US equities. When the dot-com bubble burst in early 2000, most were confident that the remaining 94% of the equity market would be relatively well insulated. In the end, of course, nothing could have been further from the truth. The broad S&P 500 index tumbled some 49% in the ensuing two and a half years.
Today it’s subprime mortgages – a relatively small segment of the home loan market that accounts for only 11% of total outstanding securitized mortgage debt. The consensus view is that the other 89% of the mortgage market is in good enough shape to weather any storm. On the surface, the relative dispersion of delinquency and default rates seems to support this contention. According to the Mortgage Banker’s Association, subprime delinquencies rose to 13.3% in 4Q06 – the highest reading since mid-2002. By contrast, for prime loans, the past-due portion stood at just 2.6% in late 2006 – the highest since mid-2003 but literally one-fifth the delinquency rate in the subprime sector. In terms of the asset effects, the key issue is whether the credit problems will spread up the quality spectrum – reminiscent of the contagion from dot-com to broader equities some seven years ago. It’s obviously too soon to know with any certainty, but the latest results of the Fed’s Senior Loan Officer Survey on Bank Lending Practices are not exactly comforting. In early 2007, the portion of respondents that were tightening overall mortgage-lending standards rose sharply – exceeding the readings hit in the 2000-01 recession and returning to levels last seen in the 1991 recession. Moreover, this latest tally represents sentiment as of January 2007 – before the full force of the subprime carnage broke out into the open. This is a fairly clear indication, in my view, that the problem is spreading.
The contrasts between the spillovers in the real economy in the aftermath of the two bubbles could well be of even greater importance to the macro call. Seven years ago, the spillover risks were concentrated in business capital spending – a sector that made up about 12.5% of the US economy in late 1999 and early 2000. In the aftermath of the bursting of the equity bubble, business capital spending fell 16% from its late-2004 peak to its early-2003 trough; by contrast, over the same nine-quarter period, the rest of the economy increased by 5%. Today, the potential spillover risks are concentrated in personal consumption – a sector that makes up about 71% of real GDP, or nearly six times the size of the capex spillover sector that was the defining characteristic of the last post-bubble shakeout.
Therein lies the case for a post-bubble macro contagion that could end up being a good deal worse over the next year than it was seven years ago. What’s especially worrisome about the current situation is that real GDP growth has already slowed to just 2% over the past three quarters – far short of the 3.7% annualized pace of the previous three years. Yet this downshift is largely an outgrowth of a steep recession in homebuilding activity, together with collateral impacts of a recent downtrend in business capital spending. By contrast, the American consumer has barely flinched, with average gains of 3.2% in real consumption since mid-2006 representing only a modest downshift from the astonishing 3.7% growth trend of the past decade. Should the consumer move into a more meaningful period of consolidation – precisely the risk as equity extraction from residential property now slows in a post-housing-bubble climate – then macro contagion could become an increasingly serious problem.
The forecasting landscape has long been littered with carcasses of those who have been dumb enough to bet against the American consumer. From time to time, there have been unconfirmed sightings of my skeletal remains in that heap. The lesson for the bruised and battered forecaster is to pick your spots carefully in betting against the American consumer. I strongly believe this is one of those times. For over a decade consumers have been spending well beyond their means, as those means are delineated by the US economy’s income generating capacity. Over the 11-year 1996 to 2006 period, annual growth in real personal consumption expenditures has averaged 3.7% -- the fastest 11-year period in post-World War II history and fully 0.5 percentage point faster than the 3.2% average growth in real disposable personal income over the same period. The income shortfall has been more than offset by wealth effects from surging asset markets – first equities and more recently housing.
With the property-related wealth effect now going in the other way in a rapidly softening housing market and with labor income generation hardly picking up the slack – private sector compensation in the current expansion remains over $400 billion below the profile of a typical five-year expansion – consumers are under mounting pressure to pull back. That pressure is compounded by record debt burdens, record debt service ratios, and negative personal saving rates – the latter an indication of the extent of the disparity between excess consumption growth and subpar income generation. In a framework that allows for consumers to draw support from both income and wealth effects, the implications of a bursting of the housing bubble are painfully simple: Consumers should respond to diminished wealth-based saving by rebuilding long-depleted income-based saving rates. And there is almost no way for that to incur without a meaningful retrenchment of the growth in personal consumption expenditures.
Just as the capex spillover was the defining feature of the post-bubble climate seven years ago, the extent of any consumer retrenchment in today’s economy will undoubtedly be the defining characteristic of the current phase. Spillover risk is also likely to dominate the Fed policy debate and bear critically on the state of world financial markets. At a minimum, I believe that consumption growth will have to slow 0.5 percentage point below the pace of income generation – sufficient to enable the personal saving rate to start rising once again. If growth in disposable personal income holds to its 11-year trend of 3.2%, that would imply a 2.7% growth pace for growth in real consumption expenditures – enough to knock 0.7 percentage point off underlying GDP growth over the next year. However, if income growth weakens – a more likely outcome in a post-housing shakeout – the consumption hit on GDP growth could be in the 1.5 percentage point range. That could push GDP growth toward the 1% threshold – easily into growth-recession” territory and not all that far from an outright downturn. Given the likelihood of meaningful consumer spillovers, I would place about a 40% probability on an outright recession scenario in late 2007 and early 2008.
The Federal Reserve now seems to be preparing itself for just such a possibility. By changing its risk assessment at the March policy meeting, it is in better position to provide support to the cumulative weakening imparted by a consumer spillover. In adopting such a posture, many have concluded, as I noted last week, that “the daisy chain of the Greenspan put” is alive and well (see my 17 March dispatch, “The Great Unraveling”). Not surprisingly, the Fed is finding it very hard to break this chain. As bubble follows bubble, it has become exceedingly difficult to disentangle fluctuations in asset markets from shifts in asset-dependent real economies. To the extent an inflation-targeting Fed uses growth risks as a proxy for inflation risks, a monetary easing may seem to make a good deal of sense if spillovers come into play. Yet in the end, of course, a monetary easing that is tied to a deteriorating growth outlook for an asset-dependent economy also turns out to provide support to the underlying asset class itself.
That’s the “Fed put” in a nutshell – a de facto targeting of asset-dependent growth risks. It would take a Volcker-like toughness to bring this insidious process to an end. Yet both Greenspan and Bernanke seem to be cut from a very different cloth..