San José State University
Department of Economics
& Tornado Alley
Private Investment is the most volatile of the major components of aggregate demand. This is because when business investors switch from believing they will need additional productive capacity to believing that they will not need it the adjustment is from some positive amount to essentially zero. Private investment then goes into free fall dropping like a rock. This is illustrated in the recent recession. Below is the graph of the percentage change, at annual rates, for real GDP and private investment.
What this suggests is that the immediate reason real GDP declines is because of a collapse of private investment. This certainly was the case for the Great Depression of the 1930's. Between 1929 and 1932 real private investment fell by 90 percent. This loss of demand subsequently resulted in the unemployment rate rising from 3 percent to 25 percent.
To demonstrate the special role of private investment demand in the declines in read GDP the data for 1947I to 2010III were obtained from the Bureau of Economic Analysis. The quarter-to-quarter percentage changes in real GDP and private investment were computed and then sorted into the quarters with declines in read GDP and those without declines. The percentage changes in private investment were then regressed on the percentage changes in real GDP. Two separate regression lines were computed; one for negative changes in real GDP and one for positive changes. The regression lines were constrained to go throught the origin. Thus the overall relationship was linear with a bend at the origin.
The regression coefficient for the 41 cases of declines in real GDP is 0.91780 whereas for the 212 cases of increases it is 0.32943. What this seems to mean is that when real GDP declines about 92 percent of the decline is due to the decline in private investment. On the other hand when real GDP increases only about 33 percent of the increase is due to the increase in private investment.
The graph of the data and the regression estimates, shown in magenta, is given below.
The requirement that the regression lines pass through the origin seems a bit restrictive. Another regression was estimated that allowed for a bend where the percentage change in real GDP is zero but put no constraint on the intercept of the equation. The graph for the data and this regression line is shown below.
The regression coefficient for the negative changes in real GDP is 0.78341 and for the positive changes 0.47123, indicating that when real GDP declines about 78 percent of the decline is due to the decline in private investment and when it increases about 47 percent of the increase is due to the increase in real investment. The t-ratio for the difference in the slopes of the relationship between the change in investment and the change in GDP for the cases of declines and and increases in real GDP is 3.5, thus confirming that the difference is statistically significant at the 95 percent level of confidence.
The immediate cause of a recession, a decline in real GDP, is largely (78 percent) due to the decline in real private investment.
HOME PAGE OF Thayer Watkins