By Nouriel Roubini | 11 November 2007
|
In this regard, evidence is mounting that debt-burdened and savings-less US consumers— that until recently used their homes as an ATM and borrowed against their housing wealth— is now on the ropes and at their tipping point. Let us consider first the factors that will lead to a consumption slowdown, and then the evidence that a slowdown has already started.
First, there is the wealth effect on consumption of changing home values. Estimates of such a wealth effect range between 5% and 7% of the change in wealth (see my survey— with Menegatti— of this literature). Recent work by Mark Zandi (Moody’s Economy.com) suggests figures closer to 7%. The total wealth effect of housing on consumption also depends on how far home values will fall. Current estimates range between a consensus of at least a 10% price fall, with some suggesting as much as a 15% fall, and some— like myself— arguing that home prices will fall as far as 20% or more. According to Fed data, the market value of the US residential housing stock was $21.0 trillion at the end of the second quarter of 2007. Thus, the fall in housing wealth could be in the $2 trillion (for a 10% drop in home prices) to $4 trillion range (for a 20% drop in prices). At $2 trillion and with a 5% effect one gets a fall in real consumption of $100 billion; with $4 trillion and with a 7% effect you get a fall in consumption of $280 billion. Even the lower figure implies a significant, sharp reduction in consumption, let alone the larger one.
Second, there is the effect of home equity withdrawal (HEW) on consumption. There is some debate in the literature on whether the effect of HEW is a proxy for the wealth effect or an additional and separate effect. Again the literature has a variety of estimates ranging from 50% of HEW being consumed according to Greenspan-Kennedy to 25% of it being consumed according to other studies. The appropriate measure of HEW is also important: gross or net, overall or active. HEW peaked at $700 billion annualized in 2005 and has dropped to about $150 billion by Q2 of 2007. So, the fall in consumption— assuming unrealistically no further fall in HEW from now on— would be $275 billion based on the Greenspan-Kennedy estimates or about $140 billion according to the more conservative estimates. Evidence suggests that this effect of HEW on consumption occurs with lags; that is why we have not yet seen its full effects on consumption as late as Q3. Rather, we will see its effects in the next few quarters. Another interpretation— according to Zandi— is that HEW (measured in a different way) has started to fall only in the recent quarters; so again the effect of falling HEW on consumption will be observed mostly in 2008.
Third, there is the effect of the ongoing and worsening credit crunch on the ability and willingness of consumers to borrow. The subprime mortgage market is now effectively dead while the 'near prime' Alt-A market is also comatose. Strains are even showing in the prime mortgage market, as rates on jumbo loans are sharply up. But this credit crunch will soon spread to other components of consumer credit (especially credit cards and auto loans) as the default rates on these other forms of consumer debt are also rising. For consumers that are affected, e.g., with little or no savings and on average with savings rates close to zero (or already below), diminished borrowing will directly restrain consumption growth. The use of credit will diminish because its price will rise, the maximum amounts made available to the consumer will be lower, the qualifications for getting a loan will be higher, and the difficulty of finding a willing lender will be greater. The slowdown in consumer borrowing will be due both to demand and supply factors: the supply of credit will be reduced; but even in the presence of available credit, many households will decide to reduce their demand for credit in order to replenish their savings. And, indeed, the savings rate of households has started to pick-up, if very slowly.
Fourth, with about $1 trillion of ARMs resetting in the next 18 months— about 450,000 households facing ARMs resets every quarter from now until the end of 2008— we will see nearly immediate effects on consumption of this surge in mortgage servicing costs. The direct average aggregate effect on disposable income may seem small: with the average ARM resetting at 300bps above the initial rate you get a fall in disposable income of about $30 billion. But the aggregate average effects hide important compositional effects: for the households whose ARMs are resetting, the fall in disposable income will be disproportionately larger than for the average household who does not have an ARM. Also, many households— who will not be able to refinance their resetting ARMs at a reasonable rate— will be forced into distressed sales (or repossession) that will lead to a further increase in the supply of homes, further declines in home prices, and in everyone's home wealth. So the compositional effects are more important and lead to a greater impact than the aggregate effects would seem to suggest.
Fifth, with oil (and energy) prices rising higher and higher (oil is now close to $100 a barrel) the reduction in household disposable income will be significant. The sharp reduction in consumption growth in Q2 (to 1.4% SAAR) was certainly due in part to the sharp increase in oil price during that period. A similar shock to disposable income is now underway. And with no evidence that oil and energy prices will fall in the months ahead (as global supply is tight and global demand is still strong) this reduction in disposable income will be persistent.
Sixth, consumer confidence is now sharply down based on both the Michigan and Conference Board measure. The early November Michigan measure of consumer confidence fell to 75.0 from 80.9 in October, its lowest level since 1995 (with the exception of the temporary drop after the Katrina hurricane) While it has become fashionable among some observers to dismiss the effects of consumer confidence on consumption, evidence suggests that, in periods of economic slowdown and strain, effects of confidence on consumption are indeed significant. And with polls now suggesting that 60% of Americans are expecting a recession in the next 12 months, it is clear that consumer confidence is headed south and will have a negative effect on consumption.
Seventh, income generation and job creation have significantly slowed down in the last year. Growth of employment has dropped by half— from 2% to about 1%. The employment rate and job growth have also fallen significantly. The household survey showed almost no job creation in 2007 and falling employment in the last month. And even the establishment survey data are not as good as the headline: the birth/death model used by the BLS creates statistical jobs that do not exist [at upper turning points in the economy]; and the employment fall in construction is undoubtedly larger than reported, as many workers are undocumented aliens.
Based on the latest figures, real disposable income growth is slowing down and close to a mediocre 1% (y-o-y) in real terms; so income generation is clearly weakening. The slowdown in job growth will accelerate in the next few months as the economic slowdown, the home building carnage, the losses in the financial sector, the weakness of manufacturing, and the job losses in the retail sector all persist and accelerate.
Eighth, the net worth of the household sector is now falling for a variety of reasons. Note that the Fed figures on household wealth are distorted by the Fed use of the OFHEO index for home prices. This measure— compared to the more precise indicators coming from the S&P/Case-Shiller— overestimates the earlier increase in home prices and underestimates its recent fall.
Three factors are leading to falling net worth for households. First, the fall in home prices and in home values. Second, the sharp increase in the last few years in consumer debt: the ratio of households’ debt to income is now above 130%— sharply up from a ratio closer to 70% in the 1990s and 100% in the early part of this decade. The debt to income ratio is at an historical high; and with interest rates, credit spreads, and the average rates of resetting mortgages going higher, debt servicing ratios are now going higher.
Third, there is now the beginning of a correction of the stock markets that will further reduce households’ wealth. The wealth effect on consumption of financial wealth is smaller than that of housing wealth (about 5% as opposed to 7%); and since the distribution of equity wealth is more unequal than that of housing wealth, the effect of a falling stock market on consumption are smaller than those of falling housing equity. But in the next few months the fall in equity prices is likely to accelerate. The equity indices are still higher than their beginning of 2007 levels but— since their peaks earlier this summer (or earlier in October)— they are now significantly down as the financial market turmoil, slowing earnings, slowing economy and higher risk aversion are now taking a toll on the stock market.
It is highly likely that— in spite of the effective "Bernanke put"— equity valuations will head further south as the massive credit problems and losses in the financial sector are recognized and as the economy slows further. Thus, a stock market below its earlier peaks and further falling stock prices ahead will impart additional negative effects on net worth and additional negative wealth effects on consumption.
Finally, note that typically heard statements that households’ net worth is "at an all time high" are meaningless as: a) such net worth is now falling; b) what matters is the ratio of such net worth to the GDP (but which is also falling, albeit more slowly); c) the change in the net worth— now negative— is much more important for consumption than its level, as at the start of any US recession in the past such net worth was "at an all time high", i.e., falling net worth ratios are much more relevant for consumption than their absolute level.
The above are the main factors which will influence a slowdown in consumption in Q4 and over the next few quarters. While even the consensus agrees that Q4 growth will be particularly weak, such consensus believes that consumption growth will recover in 2008 [[the inevitable positive 'spin': normxxx]]. But the basis for such consumption recovery forecast is unclear: the bearish factors described above will increase their momentum in the next few quarters; so there is little basis for believing that consumption growth will recover.
And there is now clear evidence that the slowdown in consumption growth has started in earnest. Let us look at the evidence.
First, note that while private consumption grew at 3% SAAR (or a 1% actual for the quarter) most of the growth in consumption occurred in July and August. Real consumption spending grew 0.3% in July, 0.6% in August and 0.1% in September adding up to 1% for the quarter and 3% annualized. So, in September the slowdown in real consumption was already clear at a mediocre and close to stalling rate of 0.1%.
Second, evidence suggests that October was weaker than September as far as retail sales go. Very weak data from the major retailers and weekly data on same store chain store sales— from both Redbook Johnson Research and from the UBS/ICSC surveys— show a sharp slowdown of retail sales in October on a y-o-y basis and an actual fall in October relative to September in nominal terms that implies an even larger fall in real terms. Indeed, sales at chain stores increased 1.6 percent from the same month last year, the worst October since 1995, according to UBS/ICSC. Based on a Bloomberg survey overall retail sales (data due on November 14th) are expected to have increased only 0.2% in nominal terms in October after increasing 0.6% in September [[it did: normxxx]]. So, Q4 started on a weaker note than September for retail sales and September was already a very weak month with nearly stalled real private consumption. Note also that the 3 month average of nominal core retail sales growth that was as high as 8.5% at its peak in the middle of 2006 is now down to about 4% (that in real terms translates in a mediocre 1%).
Third, all the factors above and especially the housing factors, rising gasoline prices and falling confidence are now leading to the weakest retail holiday since the recession of 2001. November and December sales and consumption are expected to be very weak this year. Major retailers are now revising down their forecasts for sales during the holiday season almost monthly. Both the National Retail Federation (NRF) and the RNS Retail are forecasting the weakest holiday sales growth since 2002. The NRF forecast is based on November and December expected purchases, while the RNS also includes October.
Fourth, retailers stock prices are now down sharply from their peaks earlier this year and they have fallen more than stocks in the financial sector. And the fall in earnings of the "consumer discretionary" sector in Q3 relative to the same period in 2005 is now 21%, even larger than the 17% for the financial sector. Thus, earnings and the stock markets are signaling that the retail and consumption sectors are weakening.
Fifth, there is evidence that durable goods consumption— the part of consumption that is most sensitive to expectations about the business cycle— is also weakening. Of course, most of this consumption is closely related to housing— building materials, furniture, home appliances— but the whole consumption segment is now falling thanks to the effects of the still worsening housing slump. Other components of durable goods consumption— especially autos/motor vehicles— are now under pressure. Auto sales were boosted in August by fire sales aimed at getting rid of the excess inventory of 2007 models before the launch of the 2008 models. Vehicle sales per capita have now fallen for all of 2007 and much more sharply this year in the states where the housing recession is most severe (e.g., California, Nevada, Arizona, Florida).
Finally, you don’t need consumption to actually fall (negative growth rate) to get a recession. With residential investment still in free fall, commercial real estate now showing some signs of strain, and capex spending by the corporate sectors remaining sluggish, it is enough for consumption to slow to 1% SAAR to trigger an economy wide recession. The evidence that real private consumption has not fallen since 1991 is irrelevant for two reasons: first, we can get an economy-wide recession if consumption slows down for a few quarters to 1% or below; second, the headwinds hitting the US consumer are the most severe since the early 1990s. A sharp slowdown in consumption growth will likely be the last straw. Expect Q4 growth to be 1% or below and this growth further to accelerate into negative territory by H1 of 2008.
Normxxx
______________
The contents of any third-party letters/reports above do not necessarily reflect the opinions or viewpoint of normxxx. They are provided for informational/educational purposes only.
The content of any message or post by normxxx anywhere on this site is not to be construed as constituting market or investment advice. Such is intended for educational purposes only. Individuals should always consult with their own advisors for specific investment advice.
No comments:
Post a Comment