Comment 55: USA(1) Americans are moderating the consumption of food and gasoline, but are substantially increasing spending on housing? (May 15, 2008)

Note: This series of comments will replace “USA_Updates” housed in a different directory

Many financial analysts, observers and economists have anticipated that the US economy has already entered the period of contraction in the first quarter of 2008. To their surprise the growth rate of real GDP in the first quarter of 2008 was announced to be +0.6% (annualized rate) by The Bureau of Economic Analysis on April 30, 2008. This has prompted us to look closely into the real personal consumption expenditure that comprises more than 70% of the real GDP. We report here the peculiar behavior of the real consumption of owner occupied nonfarm dwelling since the beginning of 2007. If this component is adjusted to a reasonable level, the growth rate of real GDP in the first quarter of 2008 will be pulled down to +0.2% from the official estimate of +0.6%.

Before going into the detailed discussion about the real consumption of owner occupied nonfarm dwelling, we need to make clear what the word “real” means. Superficially “real” means “inflation adjusted”. Though all the numbers in the “real” GDP report are expressed in dollars, they are actually implying the quantity of consumptions like how many gallons of gasoline, or how many pounds of beef are consumed. To make this point clear we will consider the following fictitious example of “real consumption of meats” as an illustration: The current real GDP report uses the prices of various categories of year 2000 as the standard so that the dollars appearing in the real GDP report are called 2000 chained dollars. In our fictitious example we also use the prices of 2000 as the standard. We assume that the meats consist of three varieties, they are, beef, pork and chicken. The input data and the calculated results are summarized in the following table:

Example: Real Consumtion of Meats
Varieties of Meats Year 2000 Year 2005
Nominal $ Price/pound Pounds Consumed Real Cnsumtion
in 2000 chained $
Nominal $ Price/pound Pounds Consumed Real Cnsumtion
in 2000 chained $
Beef $40 million $4/pound 10 miilion $40 million $84 million $7/pound 12 million $48 million
Pork $50 million $2/pound 25 million $50 million $60 million $3/pound 20 million $40 million
Chiken $50 million $1/pound 50 million $50 million $120 million $2/pound 60 million $60 million
Total $140 million - - - - - - 85 million $140 million $264 million - - - - - - 92 million $148 million

The columns “Nominal $” and “Price/pound” under years 2000 and 2005 are input data to the example. The column “Pounds consumed” is calculated by dividing “Nominal $” by “Price/pound”, and the column “Real consumption in 2000 chained $” is calculated by multiplying “Pounds consumed” by “Price/pound” of year 2000. Let us take the case of “Beef” of year 2005 as an explicit illustration. “Nominal $” spent on beef in 2005 is given as $84 million, and the “price/pound” is given as $7/pound. Dividing $84 million by $7/pound, we get 12 million pounds as “Pounds consumed” in 2005. Multiplying 12 million pounds by “Price/pound” of year 2000, that is, $4/pound, we get the “Real consumption of beef of 2005 in 2000 chained $” as $48 million. When adding up similarly calculated “Real consumptions in 2000 chained $” of 2005 for pork and chicken, we get the “Real consumption in 2000 chained $” of meats for 2005. If we look only in the series of beef consumption and take the ratio between the real consumption of 2005 with the real consumption of 2000, we get the ratio of 12 million pounds over 10 million pounds, that is, the ratio between the pounds of beef consumed in 2005 over the pounds of beef consumed in 2000. Thus by tracing the change of the real consumption of beef as time goes by, we are observing the time change of the quantity of beef consumed in pounds. However, when we add up real consumptions of beef, pork and chicken, the multiplication of “Price/pound” of 2000 on the “Pounds consumed” in 2005 for each category gives a proper weight to each category so that one pound of expensive beef will be weighted four times than one pound of less expensive chicken. By this method we get a composite index of real meat consumption based on the quantity consumed but expressed in 2000 chained dollars. It is also due to the conversion of quantity of consumption into 2000 chained dollars so that we can add very different kinds of consumptions like real consumptions of meats, autos, heart surgeries, fees paid to lawyers and so on into a giant composite quantity based consumption index called “real personal consumption expenditure” expressed in 2000 chained dollars. Similarly by adding real personal consumption expenditure with real consumption of net exports, real consumption of private fixed asset investments and real consumption by Governments, of course, all expressed in 2000 chained dollars, we get the real GDP.

In the “Service” sector of the real personal consumption expenditure, there is a component called the “real consumption of housing”, and within it there is a sub component called the “real consumption of owner occupied nonfarm dwellings”. It is this real consumption of owner occupied nonfarm dwellings that is showing very peculiar behavior since the beginning of 2007. In order to understand what this real consumption of owner occupied nonfarm dwellings is, we first look at another sub component within the real consumption of housing, called the “real consumption of renter occupied nonfarm dwellings”. This sub component contains the real consumption of rental apartments and the real consumption of rental homes. The real consumption of rental apartments is obtained by multiplying the number of apartment units rented out by the rents of year 2000, and the real consumption of rental houses are calculated similarly. The number of rented apartment units and rented out houses will vary as economic condition changes, so the real consumption of renter occupied nonfarm dwellings will change accordingly. In the case of the real consumption of owner occupied nonfarm dwellings, the statistics assumes that if an owner lives in his own house, he is paying himself a rent equal to the rent such a house can command if rented out in the open market. Thus the real consumption of owner occupied nonfarm dwellings becomes the number of owner occupied nonfarm dwellings multiplied by the fair market rent of year 2000. Unlike the number of rented out apartment units and houses that varies with economic conditions, the number of owner occupied nonfarm dwellings is very difficult to decline. Even the attrition from fire and natural disasters cannot dent the number very much since most of the destroyed houses will be rebuilt quickly. The major cause of decline of the number of owner occupied nonfarm dwelling is foreclosure and the willful abandonment by the owner, and the major reason of the increase of the number of owner occupied dwellings is new home sales. In the graph at the right the annualized monthly % changes of real consumption of owner occupied nonfarm dwellings are plotted as the black curve, and the annualized monthly sales of new homes are plotted as the red curve respectively. The data of the real consumption of owner occupied nonfarm dwellings are not very accurate up to the middle of 2000, causing the monthly % changes to jump violently from one month to the next. Nevertheless we can observe a general trend that rise from the nadir of 1995 to a high around 1998, and then plateau out until the middle of 2000. From the middle of 2000 the data has become a statistical disaster. There are stretched periods that data points in the black curve moves smoothly but then suddenly make sharp jumps whether up or down. Most of those sharp jumps occur at the turn of the calendar year. It seems that the data have serious normalization problems. Especially bizarre is the sudden jump of the black curve at the beginning of 2007. The monthly changes of the black curve are synchronized with the new home sales data of the red curve through 2004 to 2006. By the end of 2006 the annualized monthly % change of the real consumption of owner occupied nonfarm dwellings had dropped to +1.3% already. However, the black curve jumped to more than 4.0% at January of 2007, and has stayed around 4.0% until the first quarter of 2008. This 4.0% rate is not so different from the rate registered in 2004 and 2005, the peak of the housing and the mortgage bubble. The black curve is telling us that US consumers has started a buying spree of new homes since the beginning of 2007 so that the number of owner occupied dwellings is growing strongly for more than a year. Considering the steadily increasing number of foreclosures and the sharp fall off of new home sales, it is purely incredible that the number of owner occupied dwellings is growing so briskly since the beginning of 2007. We believe that the true monthly % changes of real consumption of owner occupied dwellings is falling since the beginning of 2007 along the black line labeled as “A”, extending the trend already established in 2006 in synchronization with the new home sales of the red curve. If we use the black line of “A” to recalculate the real consumption of owner occupied nonfarm dwellings since the beginning of 2007, the growth rates of real GDP in the fourth quarter of 2007 and in the first quarter of 2008 are both pushed down to +0.2% from the official estimates of +0.6% for those two quarters respectively. The adjusted result here shows that the growth rate of real GDP is indeed crawling near zero, and with small fluctuations it can easily fall into the negative territory.