THE PRICE LEVEL: TESTING FOR FISCAL DOMINANCE 6
In summary, we find that the post war US data strongly favors the MD regime over the FD regime. The response of Liabilities/GDP one period after an innovation in Surplus/GDP is negative and significant. This is consistent with the MD regime (in which a surplus should pay off some the debt), and it is at odds with the FD regime, since we find no evidence that innovations in Surplus/ GDP are negatively correlated with future discount factors or future surpluses (apart form the feed-back from Liabilities/GDP that one would expected in a MD regime).
In the preceding analysis, we were implicitly assuming that there were no regime switches in the post war period. However, Figure 2 may cast some doubt on this assumption. Liabilities/GDP falls in the ‘50s, ‘60s and early ‘70s, it appears to level off in the late ‘70s, and it begins to rise again in the ‘80s and ‘90s. Kremers (1989) provides some formal evidence in support of this view; his statistical model predicts debt growth well from 1966 to 1981, but it significantly under predicts debt growth in subsequent years.
Is it possible that the “Reagan Revolution” produced a regime switch that is hidden in our full sample results? Is it possible that the US has been in a FD regime since 1980?
Figure 2 also suggests that there may have been a break in the Surplus/GDP series. To obtain formal evidence in support of this view, we regressed Surplus/GDP on a constant and tested for structural breaks. These Chow tests are reported in Table 2. The tests do indicate a break, but they suggest that the “Revolution” might have preceded Reagan, beginning as early as 1975.
To see whether the US might now be in a FD regime, we reestimated our VARs using more recent sample periods. Figure 5 shows impulse response functions from a VAR for 1975-1995. The results are surprisingly strong given the short sample period. The VAR for 1980-1995 looks much the same. An innovation in Surplus/GDP makes Liabilities/GDP fall one period later, and the fall appears to be significant. Moreover, there is no evidence that the innovation in Surplus/GDP is negatively correlated with future discount factors.
We conclude that a “Revolution”—if one occurred, and whether it was inspired by Reagan or predated him—may indeed have produced a new fiscal regime, but we find no evidence that the United States has switched from a MD regime to a FD regime.
Evidently, the financial markets anticipated a fiscal consolidation at some time in the future.