United States: A Tale of Two Tiers

David Greenlaw | New York

Over the past few months, the two-tiered nature of US economic activity has become increasingly apparent. The goods sector has displayed significant softness — primarily concentrated in the homebuilding and motor vehicle industries. Meanwhile, the service sector looks to be cruising along at a healthy growth clip. To be sure, the results of the Institute for Supply Management (ISM) surveys covering the manufacturing and service sectors in November highlighted the sharp divergence. However, there now appear to be indications of a near-term bottoming in motor vehicle assemblies as well as a possible moderation in the pace of decline in home construction.

The motor vehicle industry — accounting for about 3% of overall GDP — has certainly undergone a gut-wrenching correction over the course of 2006. In an attempt to improve long-run profitability, low margin fleet sales have been pared and legacy costs have been written down. The downsizing has been significant. From 2002 to 2005, domestic vehicle production averaged 12.1 million units annually — with very little variation around that pace (specifically, output was 12.3 in 2002, 12.1 in 2003, 12.0 in 2004 and 12.0 in 2005). Over the course of 2006, assemblies were cut to about an 11.0 million unit pace. Based on the Federal Reserve’s seasonally adjusted data, the downshift in vehicle production played out gradually over the course of this past year. Indeed, after troughing at 10.4 million units (annualized) in October, current assembly schedules point to sequential upticks in both November and December, followed by a flattening out in the first quarter of 2007.

Is such stabilization reasonable? We think it is. Our latest US economic forecast shows overall light vehicle sales (including imports) running near 16.1 million units in both 2007 and 2008. This represents a further slowing relative to the 16.5 million units sold in 2006 and the 16.8 average pace recorded during 2002–2005. Most importantly, current inventory levels appear to be in reasonably good shape. Indeed, at the end of November, stockpiles were 3.5% below last year and the lowest for that particular month in the past five years. So, with domestic production having been shaved by more than 1.0 million units relative to the 2002–2005 run rate and with sales likely to decline by a somewhat smaller amount — even after allowing for some pickup in imports — the industry appears to have already reached a production equilibrium. Thus, the powerful economic headwind associated with the downshift in motor vehicle production may now be behind us.

One other quirk involving the motor vehicle sector deserves mention. In 3Q, the statisticians at the Fed came up with a dramatically different estimate of seasonally adjusted motor vehicle output than seen in the GDP data. Specifically, the Fed’s IP figures showed a sharp decline in assemblies — enough to subtract 0.6 percentage point from GDP growth. Meanwhile, the GDP accounts showed motor vehicles adding 0.8 percentage point. While there is always some divergence between these two measures, due largely to differing seasonal adjustment factors, the gap evident in 3Q is unprecedented. We expect to see an offsetting swing in the respective measures in 4Q and have built this into our GDP estimate. However, the Fed’s data series is cleaner and certainly fits much better with the widespread indications of a significant pullback in vehicle production during 3Q. Down the road somewhere, we wouldn’t be at all surprised to see the Commerce Department revise its motor vehicle data in a manner that brings it into better alignment with the Fed series.

What about the other major identifiable headwind confronting the US economy — housing? As my colleague Dick Berner laid out in a recent analysis, while there have been some encouraging signs of late — in particular, a noticeable upturn in weekly mortgage application volume — it is still far too early to call a bottom (see “False Dawn for Housing Demand?” December 8, 2006). But, it does seem clear that progress is being made. The accompanying chart shows the NAR’s measure of housing affordability. The affordability gauge is a relatively simple metric that can be used to help value the housing market. It’s based on only three variables: home prices, mortgage rates and median household income. The higher the index the better — that is, a high reading implies high affordability and vice versa. Over the course of much of the past decade, affordability remained elevated despite skyrocketing home prices. Obviously, this was largely a reflection of declining mortgage rates. Only in the past year and a half did affordability start to show signs of becoming increasingly stretched as home prices continued to rise as mortgage rates bottomed out. By mid-2005, the affordability gauge was pointing to a fundamental misvaluation in the housing market. And the market now appears to be undergoing a price correction that will eventually restore a reasonable degree of affordability. Indeed, the figure shows historical data plotted through October with an extension of the series going forward based upon the following assumptions: (1) a 5% decline in home prices over the next year, (2) steady mortgage rates, and (3) a trend rate of growth in household incomes. In such a scenario, affordability is restored to an equilibrium level within a year or so.

Obviously, such an outcome does not necessarily mean that prices won’t go down by more than 5% in some markets. As seen in the figure, while affordability in the West (dominated by California) is consistently more stretched than for the nation as a whole, a 5% price drop would not be sufficient to restore the index to its 1995–2005 average. Indeed, certain regions of the country already appear to be experiencing significant price declines in response to severely stretched affordability. But this is all part of the adjustment process. As long as mortgage rates don’t rise too much, we expect the price correction nationwide to be roughly in line with that experienced in the 1990 episode. In that instance, real home prices, as measured by the OFHEO index, declined by about 6% over a 1-year timeframe.

What would such a price swing imply for the consumer? With the household sector’s holdings of residential real estate valued at a shade over $20 trillion as of end-3Q, a 5% decline in home prices would lead to about a $1 trillion loss of wealth. Applying a standard wealth effect of .04 (that is, a 4 cent impact on consumer spending for every dollar of change in wealth), implies a $40 billion hit to consumer spending in a static sense. This is significant, representing nearly 0.5 percentage point of consumer spending. However, it actually pales in comparison to the potential short-run impact associated with the recent plunge in gasoline prices. Through much of the spring and summer, the national average price of regular gasoline hovered around $3/gallon. Over the past few months, the price dipped to about $2.25/gallon. With gasoline and fuel oil accounting for 4% of overall consumer spending, such a swing in prices frees up roughly $90 billion of discretionary spending. In our view, this is one factor — in conjunction with generally stimulative financial conditions — that has helped to prevent the spillover of the housing market correction to the rest of the economy.

Of course, the sharp drop in homebuilding activity experienced during recent quarters has been a major direct hit to the overall economy. Indeed, our latest estimates suggest that Fed Chairman Bernanke was spot on when he indicated during a Q&A session following an October 4 speech that the decline in residential construction activity would shave about 1 percentage point from GDP growth during the second half of 2006. However, as the inventory of unsold new homes begins to respond to the cutback in new construction, the drag on the overall economy from reduced homebuilding should begin to ebb as we head toward mid-2007.

Setting the stage for 2007 growth. In sum, we appear to be at the end of a major correction in the motor vehicle sector and within a quarter or two of experiencing a deceleration in the pace of decline in residential construction activity. This should set the stage for the economy to resume growth at (or even a bit better than) trend in the second half of 2007.