When reading this paper, keep in mind that it was written in the early 1980s. The most recent interesting macroeconomic phenomenon at that time was the "stagflation" of the 1970s, which coincided to some degree with the disruption of the Arab oil embargo in 1973.
"Granger causality" is an econometric technique that, under certain circumstances, can give us clues about the causal relationship between two variables. In its simplest terms, Granger causality asks the following questions about two variables y and x: Given knowledge of the behavior of y in the recent past, does knowledge of the recent behavior of x improve the ability to forecast current y? If so, then we say that "x Granger causes y." It is possible to have four Granger causality relationships between x and y: (1) neither Granger causes the other, (2) x Granger causes y but y does not Granger cause x, (3) y Granger causes x but x does not Granger cause y, and (4) both x and y Granger cause the other.
In the Granger causality tests of Table 2, a p-value less than 0.05 implies that we reject the hypothesis of NO Granger causality.
Questions for analysis
Explain Hamilton's Hypotheses 1 through 3 on page 230 in your own words. Which hypothesis does he end up favoring? What evidence does he provide against each of the others?
The most zealous advocates of the real-business-cycle model argue that supply shocks are the dominant source of economic fluctuations. Does Hamilton's evidence support this claim? Explain.
How convincing is Hamilton's evidence for the time period he studies?
Hamilton's paper was written in the early 1980s. Since then, we have had recessions from July 1990 to March 1991, from March to November 2001, and from December 2007 to June 2009. Look at the behavior of oil prices prior to these periods and make a quick assessment of how well Hamiton's results would carry forward?