North Carolina and South Carolina Unemployment Rates: a Forecast for the Years 1999 and 2000
College of Business
Western Carolina University
College of Arts and Sciences & College of Business
Western Carolina University
This paper forecasts the unemployment rates in North Carolina and South Carolina for the years 1999 and 2000. The paper takes advantage of the Phillips Curve relationship between inflation and unemployment. Variables used include Consumer Price Index and unemployment rates from 1989-1998. The unemployment rates in both North Carolina and South Carolina are forecast to rise slightly over the next year and then decline. North Carolina unemployment is projected to rise to 3.7%, and then fall to 3.6%. South Carolina unemployment will rise to 4.1% and fall to 3.9%. The government, the Fed, and firms should continue present measures to maintain the low unemployment level. (JEL: E24)
Part 1. Introduction

This paper forecasts the unemployment rate for the years 1999 and 2000. The explanatory variable is the consumer price index (CPI). This paper estimates the relationship between unemployment and the CPI using the Phillips Curve approach. The estimate is based on data from the years 1989 through 1998.

When the demand for labor is high and unemployment is low, it is expected employers will increase wages sharply; also, firms and industries should offer higher wage rates to attract the most suitable labor from other firms and industries. On the other hand, it appears workers are reluctant to offer their services at less than the prevailing rates when the demand for labor is low and unemployment is high, which causes wages to fall slowly. This kind of wage stickiness can cause prolonged unemployment.

The remainder of this paper is organized as follows: Part 2. presents the data used to forecast unemployment for the years 1999 and 2000; Part 3. presents the theoretical basis for the approach adopted in forecasting unemployment; Part 4. presents forecasts of unemployment for 1999 and 2000; Part 5. evaluates the importance of the forecast for the economy; and Part 6. discusses conclusions for economic policy.

Part 2. Data

Data are taken from the web site of the Bureau of Labor Statistics. The measures of North Carolina unemployment, South Carolina unemployment, and Consumer Price Index for all urban consumers are Bureau of Labor Statistics variables LAUST37000003, LAUST45000003, and CUUR0300SA0. These variables are published monthly, but are not seasonally adjusted. Therefore, annual data values were used. The sample period is 1989 through 1998. The forecast horizon is two years into the future.

The Phillips Curve implies a relationship between the Consumer Price Index and unemployment rates. Unemployment should respond more directly to wage inflation than the CPI or overall price inflation, but average wage rates in North Carolina and South Carolina could not be found. The only data for wage rates available were based on certain individual industries, and using these figures would not produce a representative picture of North Carolina and South Carolina labor markets. Phillips [1958] found a negative relationship between unemployment and inflation.

It would have been desirable to use seasonally adjusted monthly data for a more detailed forecast. Because the monthly data were not seasonally adjusted, annual averages were used. No other data transformations were performed.

Part 3. Economic Theory:
The Phillips Curve as a Forecasting Instrument for Unemployment

Average annual unemployment rates are used to project future unemployment for both states for 1999-2000. This forecast assumes unemployment is a function of inflation:

u = f(inflation)

This Phillips Curve relationship is a simple model appropriate for forecasting short-term unemployment rates. Since this forecast uses the Consumer Price Index and unemployment as variables, we must use the equation:

inflation = %D CPI = [CPIt-CPIt-1]/CPIt-1

  Substituting the definition of inflation into the Phillips Curve gives:

ut = f(CPIt)

Lagging the right-hand-side of the modified Phillips Curve by two years gives a forecasting equation which can be used to predict unemployment two years into the future:

ut = f(CPIt-2) (1.
Estimates of Forecasting Equation 1 using the U.S. CPI and unemployment for North and South Carolina are presented in Part 4.

The work of A.W. Phillips suggests that as the Consumer Price Index (inflation) rises, unemployment should decline, so unemployment is a negative function of the Consumer Price Index. The Phillips curve is a good theoretical model for our purposes because it predicts an inverse relationship between the nation’s inflation rate and its unemployment rate (Phillips, 1958, p.245).

The relationship being estimated is a short-run Phillips Curve (SRPC), which is appropriate for a short-run forecast based on annual data. Phillips original paper studying unemployment in the U.K., also used annual data.

Part 4. Empirical Results:
Forecasts of North Carolina and South Carolina Unemployment Rates for 1999-2000
Forecasting Equation 1 was estimated for North Carolina by an ordinary least squares regression for the period 1991-1998. The R-squared of the estimate is 0.9107. The estimate of the equation is (with t-statistics in parentheses) as follows:
NC ut = 16.03(10.93) - 0.08(-7.82)CPIt-2

The data was evaluated, with no lag, for the years 1989-1998, although the R-squared was not as high as with lagged data. This regression had an R-squared of 0.0963 and t-statistics of 2.21 and -0.92, for the intercept and slope, respectively.

The regression on lagged data was estimated for the years 1991-1998 because this is the period after the most recent recession. Typically unemployment rates decline after a recession, and inflation rates rise. This sample period avoided the high unemployment rates of the 1990-1991 recession.

Forecasting Equation 2 was estimated for South Carolina by an ordinary least squares regression for the period 1991-1998. The R-squared of the estimate is 0.5816. The estimate of the equation is (with t-statistics in parentheses):

SC ut = 18.2(4.19) - 0.09(-2.88)CPIt-2

The lower R-squared is probably due to the long persistence of high unemployment in South Carolina following the recession. South Carolina unemployment remained as high as 6% as late as 1996. The greater noise in the South Carolina data helps explain the low R-squared.

A regression was also estimated without lags for the years 1989-1998. As with North Carolina, the R-squared statistic was very low, 0.0538, and t-statistics of 1.85 and -0.67, were also disappointing.

The forecast unemployment rate is equal to the sum of the coefficients multiplied by the lagged Consumer Price Index. In order to forecast for the years 1999 and 2000, the RHS variable is lagged two years. Therefore, the lagged equation is:

ut = a+b(CPIt-2)

Table 1 reports the forecast of unemployment rates in North Carolina and South Carolina for the years 1999 and 2000:

Table 1
Forecast Unemployment for North Carolina and South Carolina, 1999-2000
Unemployment Rate
(North Carolina)
Unemployment Rate
(South Carolina)
Part 5. Forecast Implications:
Unemployment Decreases in the Near Future

Unemployment forecasting is important and necessary to measure the performance of the economy as a whole. Household welfare increases when unemployment is low. As the unemployment rate declines, firms hire more workers. These workers earn wages which are recycled back into the economy as consumption and investment, helping firms earn profits and stay competitive.

With the lower unemployment rate, the state governments will have higher tax revenues. This will give the government more flexibility to spend on assorted programs. There will also be lower state expenditures on unemployment compensation because more people will have jobs. In addition, a decline in the unemployment rate allows individuals to believe that their jobs are secure and therefore increase their standard of living. This can also increase the state’s budget surplus, or lower the deficit. Since more are employed, fewer people will be demanding welfare/transfer payments from the state.

Part 6. Policy Conclusions
The unemployment rates in both North Carolina and South Carolina, which are currently low by historical standards, are projected to rise slightly and then decline in the next two years. In North Carolina the increase will be to 3.7% for the year 1999, and then it will decline to 3.6% in the year 2000. South Carolina’s unemployment rate will increase to 4.1% in 1999, then decrease to 3.9% the following year.

Assuming the forecast turns out to be correct, firms, government, and the Fed should continue current policy measures. This will keep the unemployment rate low. Due to the fact that South Carolina is more vulnerable to a higher unemployment rate, North Carolina firms could consider recruiting workers across the state line. This may be a logical and economically sound option because workers could commute or make a short move.