Predicting Kentucky
GSP Trends using Monthly Unemployment Rates: a Simple Keynesian Approach
Western Carolina University
The monthly unemployment rate in Kentucky is forecast to continue to fall throughout the years 20002002. The expected reduction is 33% further, from 4.25% in Jan. 2000 to 2.84% in the month of Dec. 2002. The forecast is based on the Keynesian Investment Function, using a derivative of the aggregate expenditures model. This equation is the Unemployment function. For the purposes of this forecast, government spending, net exports, and consumption were assumed to be constant. Investment is used as a proxy to prepare an equation to forecast unemployment. Since unemployment is tied to the interest rate, as is investment, the interest rate is used to help explain GSP as well as real disposable income. (JEL: E0, E2, E4)
This paper forecasts the monthly unemployment rate (U) in the Commonwealth of Kentucky for the years 2000 through 2002. The explanatory variables are real disposable income, threeyear lagged unemployment, and the interest rate. The unemployment forecast will then be used to predict Kentucky’s Gross State Product, or GSP. A Keynesian investment function is used to provide a relation to GSP with regards to unemployment. This approach benefits from simplicity and allows deduction of the role of unemployment in GSP.
Continued reduction in the monthly unemployment rate would mean a continued increase in Kentucky’s GSP. Conversely, an increase in the monthly unemployment rate would indicate a reduction in GSP. The forecast horizon of three years was chosen to minimize the possibility of external factors impacting the economy in an unforeseen way. The forecast horizon is short enough to avoid seriously overstating or understating the future monthly unemployment rate.
The rest of this paper is organized as follows: part 2. presents the data used to forecast unemployment, estimate real disposable income, and estimate the interest rate; part 3. presents the theoretical basis for the approach adopted in forecasting the monthly unemployment rate; part 4. presents the forecasts of the monthly unemployment rate for the years 2000, 2001, and 2002; part 5. evaluates the importance of the forecast for the economy of Kentucky; 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 unemployment and real disposable income are FRED variables KYUR, and DSPIC96, are all real variables, and real disposable income is given in billions of chained 1996 dollars at seasonally adjusted annual rates (SAAR). The Kentucky unemployment rate is seasonally adjusted and was provided by the Workforce Development CabinetDept. for Employment Services. The primary source is the U.S. Department of Commerce Bureau of Economic Analysis.
The interest rate data are the secondary market 3month Tbill rates given in FRED variable TB3MS, which is given as a percent discount. The primary source is the Board of Governors, U.S. Federal Reserve System. These are monthly variables. The sample period for the data used is from the first month of 1978 to the first month of 2000. The forecast horizon is three years into the future.
Unemployment is used as a substitute of GSP. When unemployment is low, GSP is assumed to be rising. Thus when unemployment is high, GSP is assumed to be falling. Real disposable income was used as an indicator of real GSP. Investment is positively related to GSP, going up when GSP rises and decreasing when GSP falls. Alternative measures were considered, including real consumption, which could have been used in this case, though it may not have provided equally good results. The 3month Tbill rate was used because it is the one interest rate the Federal Reserve System exerts very much control over, since the Fed is the primary buyer and holder of U.S. Treasury securities, which also strongly influences market interest rates. For example, the Fed also controls the discount rate, but that has less direct influence on market interest rates.
Part 3. Economic Theory:
Forecasting with the Keynesian Investment Function and the Derived
Unemployment Function
This forecast uses the Keynesian Investment Function as its base. This function is relevant to the forecast because it includes, or can include, both the interest rate and real disposable income (Y). The Keynesian Investment Function can be written as:
or as:
Where r is the interest rate, taken from the secondary market for three month Treasury Bills and Y is real disposable income. Investment (I) is part of the Aggregate Expenditures Function, written as:
AE = C+I+G+NX (3.
I is related to r and Y, therefore when I goes up, unemployment goes down. The relevant equation, the unemployment function, derived from the AE model, is written as such:
U_{t} = f(U_{t3},
Y_{t3},r_{t3}) (4.
Where U is unemployment, Y is real disposable income, and r is the interest rate. The estimate of equation 4 and forecast based on the regression are presented in part 4.
A possible shortcoming of this paper is that the monthly unemployment rate is below the natural rate of employment, which is 6%according to the Keynesian model. This trend, if sustained, could lead to inflation and a variance of the interest rate over the forecast period. This would cause a reduction in the GSP.
Monthly Unemployment, and Forecast Monthly Unemployment Rates
The unemployment function, equation 4, was estimated with 1978.12000.1 monthly data. The regression equation is written as follows:
U_{t} = a + bU_{t3 }+ cY_{t3 }+ dr_{t3}
The equation is solved and the results are found in Table 1, below:
Table 1 Regression Estimate of Consumption Forecasting Equation 1: 1992.11998.3 

Explanatory variable 
Estimated coefficient 
tratio 
Intercept 
15.90 
19.54 
U_{t3} 
0.019 
0.59 
Y_{t3} 
0.0022 
17.58 
r_{t3} 
0.19 
6.84 
R^{2} = 0.8467 
F (zero slopes) = 418.99 
Prob F = 0.00 
The adjusted R^{2} of the estimate is 0.847, indicating that 85% of the variation in U can be explained by the three variables of U, Y, and r. The tstatistic of U is 0.59, indicating a weak rejection of the null hypothesis that U=0. The tstatistic of Y is 17.58, indicating strong rejection of the null hypothesis that Y=0.
The tstatistic of r is almost 7, again justifying a rejection of the null. The same is true for the intercept, which has a tstatistic of almost 20. This means that the intercept is not found at zero.
The following table contains the results of the forecast monthly unemployment rate:
Table 2 Forecasts of Monthly Unemployment Rates for
Kentucky for the years 20002002


Month 
Forecast
Unemployment

Jan00

4.25 
Feb00 
4.21 
Mar00 
4.17 
Apr00 
4.15 
May00 
4.08 
Jun00 
4.02 
Jul00 
4.02 
Aug00 
3.98 
Sep00 
3.92 
Oct00 
3.88 
Nov00 
3.85 
Dec00 
3.82 
Jan01 
3.77 
Feb01 
3.73 
Mar01 
3.65 
Apr01 
3.62 
May01 
3.60 
Jun01 
3.53 
Jul01 
3.48 
Aug01 
3.43 
Sep01 
3.34 
Oct01 
3.18 
Nov01 
3.13 
Dec01 
3.17 
Jan02 
3.10 
Feb02 
3.06 
Mar02 
3.00 
Apr02 
2.99 
May02 
2.98 
Jun02 
2.88 
Jul02 
2.89 
Aug02 
2.86 
Sep02 
2.92 
Oct02 
2.81 
Nov02 
2.81 
Dec02 
2.84 
This forecast projects unemployment to reduce by 33% from the Jan. 2000 unemployment rate. The reduction will be very gradual. This forecast suggests that perhaps a new production possibilities curve is being reached as the employment rate continues to be well above the normal 94% employment, or natural rate of employment. This therefore entails an increase in GSP.
Unemployment forecasts are interesting to a broad range of people because unemployment is a leading indicator of GDP/GSP trends. The forecast predicts a continued drop in unemployment, leading to an increase in the GSP of Kentucky. The trend towards lower and lower unemployment tends to suggest a gradual increase in GSP of Kentucky, a higher standard of living, and opportunities for low risk investment could be possible in Kentucky.
The relatively low interest rates in Kentucky, combined with the low unemployment rates, tend to suggest that an increase in interest rates would lead to an increase in the rate of unemployment. This would lead to a reduction in GSP.
Monthly unemployment rates will continue to go down; a further reduction of 33% from the Jan. 2000 level is forecast. Therefore GSP will rise.
Assuming the forecast is correct, the increase in GSP due to the increased employment in the Commonwealth should benefit the entire society. With the high rate of employment, an eventual upward pressure on wages is a possibility. This could eventually lead to a rise in inflation and subsequent rise in interest rates.
The forecast assumes that consumption and government spending are not going to undergo any significant variations in the next three years. With dramatic increases in either, the downward trend in unemployment could reverse.
The Fed should attempt to hold interest rates at or near the current level to maintain the current level of industrial production alongside the level of employment.
Federal Reserve Bank of St. Louis, Federal Reserve Economic Data (FRED), http://www.stls.frb.org/fred/
Hazlitt, Henry, The Failure of the "New Economics," Lanham MD: University Press of America, 1983.
Mulligan, Robert F., “Measuring Economic Performance”, http://www3.wcu.edu/~mulligan/
measuringeconomicperformance.html