## Info

Y = nominal GNP M = money stock (M1) E = high employment expenditures P = GNP deflator (1972 = 100) PE = relative price of energy X = output in 1972 dollars XF = potential output (Rasche/Tatom) RL = corporate bond rate U = unemployment rate UF = unemployment rate at full employment DUM1 = control dummy (1971-III to 1973-I = 1; 0 elsewhere) DUM2 = postcontrol dummy (1973-II to 1975-I = 1; 0 elsewhere) Source: Federal Reserve Bank of St. Louis, Review, May 1982, p. 14.

estimated by OLS. Equations (1), (2), and (4) were estimated using the Almon distributed-lag method with (endpoint) constraints on the coefficients. Where relevant, the equations were corrected for first-order (p1) and/or second-order (p2) serial correlation.

Examining the results, we observe that it is the rate of growth in the money supply that primarily determines the rate of growth of (nominal) GNP and not the rate of growth in high-employment expenditures. The sum of the M coefficients is 1.06, suggesting that a 1 percent (sustained) increase in the money supply on the average leads to about 1.06 percent increase in the nominal GNP. On the other hand, the sum of the E coefficients, about 0.05, suggests that a change in high-employment government expenditure has little impact on the rate of growth of nominal GNP. It is left to the reader to interpret the results of the other regressions reported in Table 20.9.

784 PART FOUR: SIMULTANEOUS-EQUATION MODELS

EXAMPLE 20.4 (Continued) TABLE 20.9

IN-SAMPLE ESTIMATION: 1960-I TO 1980-IV (Absolute Value of t Statistic in Parentheses)

(1) Yf = 2.44 + 0.40Mt + 0.39Mt-1 + 0.22/W(-2 + 0.06/W(-3 - 0.01/W(-4

(2.15) (3.38) (5.06) (2.18) (0.82) (0.11) + 0.06Et + 0.02Et-i - 0.02Et-2 - 0.02E(-3 + 0.01 E(-4 (1.46) (0.63) (0.57) (0.52) (0.34)

(2) P = 0.96 + 0.01 PEt-1 + 0.04PEt-2 - 0.01 PEt-3 + 0.02PEt-4

- 0.00(Xt - XF*) + 0.01(Xt-1 - XFm) + 0.02(Xt-2 - XF*_2)

+ 0.02(Xt-3 - XFt-3) + 0.02(Xt-4 - XFt-4) + 0.01(Xt-5 - XFt-5)

+ 1.03(PAt) - 0.61(DUM1t) + 1.65(DUM2t) (10.49) (1.02) (2.71)