A Quarterly Forecast of U.S. Investment Using Per Capita Income from a Simple Aggregate Expenditures Perspective for 1999-2000
 
DEVIN ANDERSON and AGGRESTER BELL JR.
College of Business
Western Carolina University
 
Abstract
 
U.S. investment growth has ranged (1-4%) from 1992 to 1998. Quarterly data are used to forecast U.S. real investment expenditures by regression. This forecast is based on the Keynesian Aggregate Expenditures Model. As Gross Domestic Product (GDP) and interest rates increase, investment is found to increase directly. Investment appears to experience cyclical variations: approximately every two years investment experiences a one-quarter decline. As soon as that quarter is over, investment rises again. The data collected show long-term interest rates and population are poor explainers of investment. Investment is projected to increase through the next two years, suggesting "If it ain't broke, don't fix it!" (JEL: E22, E47, E66).
 
Part 1. Introduction
 
This paper forecasts U.S. investment for the years 1999-2000. The explanatory variables are population, GDP, and the long-term 30-year bond rate. The approach adopted by this paper is based on the Keynesian Investment Function using real data for 1992-1998. The simplicity of the Keynesian approach makes it easy to apply and interpret.
 
Investment is a part of GDP. Investment also expands the capital stock, making it possible to produce more GDP in the future. Investment is expected to be a negative function of interest rates because interest is the opportunity cost of investment.
 
The rest of this paper is organized as follows: Part 2. presents the data used to forecast investment - real data for investment, GDP, population, and long-term interest rates; Part 3. presents the theoretical basis for the approach adopted in forecasting investment; Part 4. presents the forecast of investment for the years 1999-2000; Part 5. evaluates the importance of the forecast for the economy; and Part 6. discusses conclusions for economic policy.
 
Part 2. Data
 
All variables are taken from the Federal Reserve Bank of St. Louis Federal Reserve Economic Data (FRED). The measures of investment, GDP, population, and interest rates are FRED variables GPDIC92, GDPC92, TPC92, LTRC92, are all real variables. Investment and GDP are given in billions of chained 1992 dollars at seasonally adjusted annual rates (SAAR). Population is given in thousands of people and interest rates are given in percent. The data is observed quarterly. These values are used in the forecast to achieve a more exact forecast. Interest rates and population are observed monthly. The monthly data is put together in quarterly form by averaging. The data cover the period 1992.1-1998.4.
 
GDP is a measure of national output used in the aggregate expenditures model. The exclusion of government investment gives the model total emphasis on private sector investment. The long-term rate is used instead of the federal funds rate because the model is dealing with firms and households instead of government, although the Fed does control the federal funds rate, the model focuses on personal income and the behavior of private firms.
 
Part 3. Economic Theory
 
The forecasting approach being used is very appropriate for the purpose of forecasting investment. The model states that GDP, population, and interest rates all help determine investment. This model depends on the Keynesian Aggregate Expenditures Model. The aggregate expenditures model states that GDP is a function of investment, along with other variables. For the purpose of this model, it will be assumed that this relationship can be switched around. This model assumes that if GDP depends on the level of investment, investment is also a function of GDP. In addition, the forecast depends on the Keynesian aggregate investment function.
 
The Keynesian investment function can be written as:
 
I = f(i,Y)
 
where I is investment, i represents interest rates, and Y is GDP, or some other measure of national income.
 
Using quarterly data, the forecasting equation can be written as:
It= a + b It-2 + c Yt-2 + d Popt-2 + e it-2
 
where Pop is the Total Population. GDP increases due to the rise in population. The rise in population puts more people in the work force, which generally leads to greater GDP and greater investment. The increase in both GDP and population provides an increase in investment.
 
Part 4. Empirical Results
 
Regression results are reported in Table 1.
 
Table 1
Regression Estimate of investment Forecasting Equation
Dependent variable Estimated Coefficient T-Stat
Intercept 381.9347 .0452
Investment -1.0239 -3.5054
GDP .5738 2.3304
Population -.0030 -.2320
Interest rates 20.3982 1.1406
R-Square=.9776   F-Stat=164.3657
 
Equation:
Investment = 381.9347 - 1.0239(Investment) + .5738(GDP) - .0030(POP) + 20.3982(INT.)
 
The R-Square is 97.76%. The independent variables in the model can account for 97.76% of the variation in investment. This is a relatively high percentage. There are other factors that influence the investment rate which are neglected by this model, which analyzes only four factors. The F-statistic is 164.3657. This indicates very strong rejection of the null hypothesis of zero slopes. The evidence is strong enough to conclude that at least one of the independent variables has an influence on the dependent variable.
 
The coefficient for lagged investment is -1.0239. For each additional dollar increase in investment for the same quarter two years before, investment for the current quarter decreases by (-)$1.0239. At first this seems counterintuitive. However, after analyzing the data year by year it is apparent that there is a trend forming. Investment declines every two years. This cycle is the reason for the negative coefficient on this variable. More investment today indicates less will be required two years into the future; less investment today indicates more will be required.
 
The coefficient for GDP is .5738. For each additional dollar increase in GDP, investment increases by $0.5738, fifty-seven cents. This is realistic. It makes sense that as GDP increases, investment also increases. The coefficient for population is -.0030. For each additional unit, population goes up, investment goes down by .0030. This variable has very little effect on total investment. The coefficient for interest rates is 20.3982. For each additional one-percent increase in interest rates, investment will increase by 20.3982 billion chained 1992 dollars.
 
An absolute value T-statistic greater than 2 indicates the variable is good for explaining the forecast target. The independent variables investment and GDP have T-statistics greater than 2, rejecting the null hypothesis that those variables can be removed from the investment forecasting function.. Population and interest rates have T-statistics less than 2. These measures are not significant in the regression. They were kept in the regression model under the assumption that they contribute useful information.
 
By using the model, a forecast for the years 1999-2000 is constructed. Results of this forecast are presented in Table 2.
 
Table 2
Forecasts of Investment 1999-2000
(Measured in billions of chained 1992 dollars)
Quarter Forecast
1999.1 1071.6
1999.2 1069.7
1999.3 1113.7
1999.4 1118.8
2000.1 1111.2
2000.2 1156.2
2000.3 1177.5
2000.4 1219.4
 
This forecast shows a significant drop in investment between the last year of data (1998.4) and the first year of the forecast. This can be explained by the negative coefficient on the lagged investment variable. Investment levels soared in 1997. This sharp rise has a significant effect on the forecast. The right-hand-side variables are lagged two years. The investment explosion that occurred in 1997 has a tremendous negative impact on the forecast for the quarters of 1999-2000.
 
Part 5. "If It Ain't Broke, Don't Fix it"
 
With interest rates currently holding steady, and GDP rising steadily, everything seems to be going well. Although there seems to be a drop-off in investment every two years, investment has been on an upward long-run trend during the decade. Investment is a leading indicator of economic conditions. High investment means high capital maintenance and accumulation, that the country is doing a lot of things right. Generally speaking, it means more and better jobs, low inflation, higher per capita income, strong currency value, etc. This is why it is important to maintain high levels of investment. To continue to grow, we simply cannot keep the capital stock at current levels. However, things seem to be going pretty well, so major changes are not needed.
 
Part 6. Policy Conclusions
 
Investment is predicted to drop in the first quarter of 1999. After this initial fall, investment will consistently rise for the next two years. This initial fall is nothing to worry about. When looking at the overall picture, it is apparent that this will only be a stepping stone to higher investment figures in future years. If this forecast comes true, industries should expect continued high economic growth. Evidence suggests that for now the economy is only going to get better and stronger. There will be room for growth for all industries here in the country and into foreign markets. However, these industries should also be aware that nothing lasts forever. These industries should also take advantage of this situation by stabilizing themselves for possible hard times down the road.
 
The government should basically leave everything alone for now. However, the government should make sure that interest rates continue along their current path. Although a rise in interest rates would help stimulate investment, (according to the regression estimated in this paper,) an increase in interest rates would surely cause other problems for the overall economy. If the government becomes over-involved, it could have an adverse impact on the economy. The right policies appear to be currently in place for the economy to prosper.
 
Reference