This paper updates through the 1996 election the equation originally presented in Fair (1978) explaining votes for president. Previous updates are in Fair ( 1982, 1988, 1990, and 1996). The specification of the equation has not been changed for the current update. The equation in Fair (1996) has simply been reestimated using the latest updated data. The updated data are discussed in Section II; the updated estimates are presented in Section III; and predictions for 2000 are presented in Section IV. The appendix contains a complete description of how the data were collected and a listing of all the data. The appendix allows one to duplicate the results in this paper if desired.
Since the last update, the Department of Commerce has issued GDP data in chained (1992) dollars back to 1929, and these data have been used. This is an important change, since the old Department of Commerce procedure of using fixed weights as far back as 1929 was problematic. Data prior to 1929 were obtained, as before, from Balke and Gordon (1986). The appendix discusses the splicing of the Balke and Gordon data to the Department of Commerce data.
The geometric average of the 320 quarterly growth rates of real per capita GDP between 1913:1 and 1996:4 is 1.65 percent using the previous data and 1.93 percent using the updated data. (All growth rates in this paper are expressed at an annual rate.) The data revisions thus increased the average growth rate by 0.28 percentage points over this period.
The notation for the variables is as follows:
The specification of the equation is :
Table 1 below presents five estimates of the vote equation. The first four use the updated data, where the estimation periods end in 1996, 1992, 1988, and 1960, respectively. The fifth estimate is taken from Table 1 in Fair (1996). It uses the old data, and the estimation period ends in 1992. The predicted values and errors are presented at the end of the table for each estimate. For the second estimate the predicted value for 1996 is outside sample; for the third estimate the predicted values for 1992 and 1996 are outside sample; for the fourth estimate the predicted values from 1964 on are outside sample; and for the fifth estimate the predicted value for 1996 is outside sample. All the predictions for the first four estimates use the updated data. The predictions for the fifth estimate use the old data except for the prediction for 1996, which uses the updated data.
__________________________________________________________________________
Table 1
Five Estimates of the Vote Equation
Estimate: 1 2 3 4 5
Previous
Sample: 1916-1996 1916-1992 1916-1988 1916-1960 1916-1992
a 0.472 0.469 0.467 0.463 0.468
1 (86.34) (83.59) (118.95) (77.56) (90.62)
a -0.016 -0.034 -0.018 -0.028 -0.034
2 (-0.63) (-1.19) (-0.90) (-1.24) (-1.26)
a .047 0.061 0.027 0.041 0.047
3 (1.96) (2.34) (1.36) (1.76) (2.09)
a 0.0070 0.0071 0.0074 0.0082 0.0065
4 (7.45) (7.69) (11.57) (7.73) (8.03)
a -0.0072 -0.0062 -0.0076 -0.0044 -0.0083
5 (-2.77) (-2.31) (-4.02) (-1.18) (-3.40)
a 0.0091 0.0106 0.0070 0.0066 0.0099
6 (3.86) (4.06) (3.39) (2.67) (4.46)
a 0.040 0.041 0.054 0.054 0.052
7 (3.25) (3.37) (5.98) (4.29) (4.58)
a -0.033 -0.029 -0.019 -0.018 -0.024
8 (-3.07) (-2.54) (-2.36) (-1.98) (-2.23)
SE 0.0215 0.0210 0.0145 0.0151 0.0190
2
R 0.947 0.951 0.978 0.987 0.960
No. obs. 21 20 19 12 20
V | V | V | V | V | V
Year act. |pred. error |pred. error |pred. error |pred. error |pred. error
| | | | |
1916 .517 |.507 .009 |.497 .020 |.508 .009 |.508 .008 |.495 .022
1920 .361 |.389 -.028 |.386 -.025 |.371 -.010 |.364 -.003 |.382 -.021
1924 .418 |.421 -.004 |.416 .001 |.430 -.012 |.425 -.007 |.419 -.001
1928 .412 |.427 -.015 |.426 -.014 |.423 -.011 |.425 -.013 |.427 -.015
1932 .592 |.608 -.017 |.606 -.014 |.592 .000 |.587 .005 |.607 -.015
1936 .625 |.643 -.018 |.641 -.017 |.636 -.012 |.636 -.012 |.629 -.005
1940 .550 |.560 -.010 |.558 -.008 |.567 -.017 |.554 -.004 |.553 -.003
1944 .538 |.529 .009 |.530 .008 |.537 .001 |.541 -.003 |.522 .015
1948 .524 |.505 .019 |.507 .017 |.515 .009 |.518 .006 |.518 .006
1952 .446 |.439 .007 |.439 .007 |.444 .002 |.438 .007 |.449 -.003
1956 .422 |.427 -.005 |.433 -.010 |.422 .000 |.425 -.002 |.418 .005
1960 .501 |.489 .012 |.490 .011 |.484 .017 |.483 .018 |.494 .006
1964 .613 |.612 .001 |.610 .003 |.601 .013 |.591 .022 |.617 -.004
1968 .496 |.496 .000 |.495 .001 |.490 .006 |.489 .007 |.504 -.008
1972 .382 |.402 -.020 |.406 -.023 |.393 -.011 |.380 .002 |.392 -.010
1976 .511 |.514 -.004 |.511 -.001 |.507 .003 |.486 .025 |.506 .004
1980 .447 |.456 -.009 |.452 -.005 |.448 -.001 |.455 -.008 |.446 .001
1984 .408 |.386 .023 |.384 .024 |.383 .025 |.370 .038 |.387 .021
1988 .461 |.476 -.015 |.474 -.013 |.472 -.011 |.466 -.005 |.489 -.028
1992 .535 |.491 .044 |.495 .040 |.459 .075 |.450 .084 |.501 .034
1996 .547 |.527 .021 |.514 .033 |.528 .019 |.523 .025 |.516 .031
_____________________________________________________________________________
The results in Table 1 show that the coefficient estimates are fairly stable across the estimation periods, given the small sample sizes, which is encouraging. Perhaps the most remarkable feature of the results is how well the equation estimated only through 1960 does. For this equation the eight coefficient estimates are based on only twelve observations. There are nine outside sample predictions for this equation, beginning with 1964, and the mean absolute error is only 2.4 percentage points for these nine predictions. If 1992 is excluded, the mean absolute error is only 1.65 percentage points. The 1996 prediction is outside sample by 36 years!
The predictions show that the most problematic election for the equation is 1992, where V was substantially underpredicted. Much of the discussion in Fair (1996) is concerned with trying to account for this error, and again this material will not be repeated here. The results for 1996, on the other hand, are fairly good. The actual estimated equation used before the election, estimate 5 in Table 1, has an error of 3.1 percentage points using the actual economic data. Since the results for 1996 were fairly good, there seemed little need to try to change the equation, and so, as mentioned above, the specification was not changed.
The equation estimated through 1996, estimate 1, is taken as the "final" equation. The coefficient estimate for g3 is .0070, which says that a one percentage point increase in g3 leads to a .70 percentage point increase in the vote share. The coefficient estimate for p15 is -.0072, which says that a one percentage point increase in p15 leads to a .72 percentage point decrease in the vote share. The coefficient estimate for n is .0091, which says that each good news quarter adds .91 percentage points to the vote share.
For the 2000 election I is 1, DPER is 0 (the incumbent is not running), DUR is 1, and DWAR is 0. The vote equation can thus be written
The October 10, 1998, Blue Chip consensus forecast for the inflation rate is 1.7 percent for 1998:4, 1.8, 1.8, 1.9, and 2.0 percent for the four quarters of 1999, respectively, and 2.3 percent for 2000. Given these forecasts and the actual values for the first seven quarters, the implied value of p15 is 1.7 percent. The Blue Chip consensus forecast for the growth rate of GDP (not per capita) is 2.4 percent in 1998:4, 1.9, 2.0, 2.2, and 2.3 percent for the four quarters of 1999, respectively, and 1.9 percent for 2000. Population is growing at about .9 percent per year, so .9 needs to be subtracted from the Blue Chip growth rates to put them on a per capita basis. None of the forecasted quarters is a good news quarter (i.e., a per capita growth rate greater than 3.2 percent), so n remains at 3. Assuming the 1.9 percent prediction for 2000 is the same in each quarter, the Blue Chip predicted value of g3 is 1.0 (= 1.9 - .9). Using these predictions for p15, n, and g3, the predicted value of V is .445 from the above equation. The estimated standard error of the equation is .0215, and so this prediction is a little more than two standard errors below .5.
If the November 3, 1998, economic forecast from my US model is used in place of the Blue Chip forecast, the results are similar. The US model forecast has implied values of 1.4 for p15, 3 for n, and .9 for g3. These predictions lead to a predicted value of V of .447.
The other estimated equations in Table 1 give a similar message. Using the Blue Chip forecast of the economy, the second estimated equation predicts V to be .435, the third predicts V to be .445, the fourth predicts V to be .438, and the fifth predicts V to be .432. These compare to .445 using the first estimated equation.
Both the Blue Chip forecast and the US model forecast have the economy growing only moderately between now and the election, and so even though inflation is low, the economy is not a big plus for the Democrats. More growth in 1999 and 2000 would be needed to bring the predicted value of V above .5. So based on the assumption of only moderate growth rates between now and the election, the vote equation says that the Republicans have an edge in the 2000 election.
The data that were used to estimate the vote equation are presented in Table A-1 below. Data are presented back to 1880 even though only data back to 1916 were used in the estimation work. Quarterly data on nominal GDP, real GDP, and population are needed to construct g3, p15, and n, and these data are presented in Table A-2. In Table A-2 Y denotes nominal GDP, X denotes real GDP, and POP denotes population. Let a subscript k denote the kth quarter of the sixteen-quarter period of an administration. Also, let Z = X/POP, per capita real GDP, and let P = Y/X, the GDP deflator. Then g3 and p15 are constructed as:
V is the Democratic vote divided by the Democratic plus Republican vote except for the 1912 and 1924 elections. For 1912, V is the Democratic vote divided by the Democratic plus Republican plus Roosevelt vote. For 1924, V is the Democratic vote plus .765 times the LaFollette vote divided by the Democratic plus Republican plus LaFollette vote. The voting data for 1880-1916 were obtained from U.S. Department of Commerce (1975), pp. 1078-1079. For 1920-1932 the data were obtained from U.S. Department of Commerce (1988), p. 232, and for 1936-1996 the data were obtained from U.S. Department of Commerce (1997), p. 271.
I, DPER, DUR, and DWAR are defined in the text. In the construction of DPER Ford is not counted as an incumbent running again, since he was not an elected vice president, whereas the other vice presidents who became president while in office are counted.
The data on nominal GDP were obtained as follows. For 1929-1945 annual data were obtained from Table 1, p. 147, of the U.S. Department of Commerce, Survey of Current Business, August 1998. For 1946:1-1996:4 quarterly data were obtained from this same table, pp. 148-150. Quarterly data for 1877:1-1945:4 are available from Balke and Gordon (1986), pp. 789-795. The Balke and Gordon values for 1877:1-1928:4 were used exactly, but the values for 1929:1-1945:4 were adjusted to take account of the new Department of Commerce annual data. For 1929:1-1945:4 each quarterly value for a given year was multipled by a splicing factor for that year. The splicing factor is the ratio of the Department of Commerce value for that year to the respective yearly value in Balke and Gordon (1976), pp. 782-783.
The data on real GDP were obtained in a similar way. For 1929-1946 annual data were obtained from Table 2a, p. 151, of the Survey of Current Business, August 1998. For 1947:1-1996:4 quarterly data were obtained from this same table, pp. 152-154. Quarterly data for 1877:1-1946:4 are available from Balke and Gordon (1986), pp. 789-795. The Balke and Gordon values were spliced to the Department of Commerce values. All the Balke and Gordon quarterly values for 1877:1-1929:4 were multiplied by the same number. This number is the ratio of the Department of Commerce value for 1929 to the 1929 value in Balke and Gordon (1976), p. 782. For 1930:1-1946:4 each Balke and Gordon quarterly value for a given year was multipled by a splicing factor for that year. The splicing factor is the ratio of the Department of Commerce value for that year to the respective yearly value in Balke and Gordon (1976), pp. 782-783.
The data on population were obtained as follows. For 1877-1928 annual data were obtained from U.S. Department of Commerce (1973), pp. 200-201, A114 series. Each of these observations was multiplied by 1.000887, a splicing factor. The splicing factor is the ratio of the A114 value for 1929 in U.S. Department of Commerce (1973) to the value for 1929 in Table 8.2 in U.S. Department of Commerce (1992). For 1929-1945 annual data were obtained from U.S. Department of Commerce (1992), Table 8.2. Quarterly observations for 1877:1-1945:4 were obtained by interpolating the annual observations using the method presented in Fair (1994), Table B.6. For 1946:1-1996:4 quarterly data were obtained from Table 8.2 in U.S. Department of Commerce (1992), Table 8.2 in U.S. Department of Commerce (1981), and updates. The latest data as of this writing are in Table 8.3 of the National and Income and Product Accounts.
_____________________________________________________________
Table A-1
Data Used in the Estimation
Election V I DPER DUR DWAR g3 p15 n
1880 0.4978 -1 0 -1.75 0 3.879 1.974 9
1884 0.5015 -1 0 -2.00 0 1.589 1.055 2
1888 0.5041 1 1 0.00 0 -5.553 0.604 3
1892 0.5173 -1 -1 0.00 0 2.763 2.274 7
1896 0.4776 1 0 0.00 0 -10.024 3.410 6
1900 0.4683 -1 -1 0.00 0 -1.425 2.548 7
1904 0.3999 -1 0 -1.00 0 -2.421 1.442 5
1908 0.4552 -1 0 -1.25 0 -6.281 1.879 8
1912 0.4529 -1 -1 -1.50 0 4.164 2.172 8
1916 0.5168 1 1 0.00 0 2.229 4.252 3
1920 0.3612 1 0 1.00 1 -11.463 16.535 5
1924 0.4176 -1 -1 0.00 0 -3.872 5.161 10
1928 0.4118 -1 0 -1.00 0 4.623 0.183 7
1932 0.5916 -1 -1 -1.25 0 -14.901 7.069 4
1936 0.6246 1 1 0.00 0 11.921 2.362 9
1940 0.5500 1 1 1.00 0 3.708 0.028 8
1944 0.5377 1 1 1.25 1 4.119 5.678 14
1948 0.5237 1 1 1.50 1 1.849 8.722 5
1952 0.4460 1 0 1.75 0 0.627 2.288 6
1956 0.4224 -1 -1 0.00 0 -1.527 1.936 5
1960 0.5009 -1 0 -1.00 0 0.114 1.932 5
1964 0.6134 1 1 0.00 0 5.054 1.247 10
1968 0.4960 1 0 1.00 0 4.836 3.215 7
1972 0.3821 -1 -1 0.00 0 6.278 4.766 4
1976 0.5105 -1 0 -1.00 0 3.663 7.657 4
1980 0.4470 1 1 0.00 0 -3.789 8.093 5
1984 0.4083 -1 -1 0.00 0 5.387 5.403 7
1988 0.4610 -1 0 -1.00 0 2.068 3.272 6
1992 0.5345 -1 -1 -1.25 0 2.293 3.692 1
1996 0.5474 1 1 0.00 0 2.918 2.268 3
______________________________________________________________