Why Might Alan Greenspan Have Thought That He Could Have Influenced The Outcome of Elections?

In my previous post on the Fed and elections, I showed that for whatever reason, in the four consecutive close elections during the Greenspan era, there were unusually large changes in the levers that the Fed controls. One might call it coincidence, or one might note that these movements seemed designed to benefit the Republican candidate.

For instance, the 12 months through November 1992 saw the second largest percentage increase in real M1 per capita in the Fed’s history (in any 12 month period ending in November). 1996 – the single largest percentage decrease in real M1 per capita in the Fed’s history. 2000 saw the third largest percentage decrease in real m1 per capita in the Fed’s history. After three elections in which the Republican candidate came in second in the popular votes, 2004 saw a new approach: the third lowest real Fed Funds rate (i.e., the Fed Funds rate less the real inflation rate) in the Fed’s history. Truly, Greenspan liked historical extremes and even records around election time.

Now, say you’re cynical. You might think to yourself: “Why should Greenspan have thought that he could cause enough of a change in real GDP per capita to influence elections? And why didn’t he?

In earlier posts, I looked at correlations between real GDP per capita from 1959 to 2006. However, Bill Polley and PGL have suggested that there was a structural break in the 1990s – that the proliferation of new payment technologies meant that real M1 per capita became less and less important. I think they’re right. And I think there’s another structural break in the sample – after the collapse of the Bretton Woods System.

Table 1 below shows the correlation between real M1 per capita and real GDP per capita, using quarterly data for two periods: 1974 – 1991, and 1992 – 2006.

Through 1991, the movements in the money supply tended to follow changes in the economy. Put another way, the Fed reacted to changes in the economy. However, there was no reason to assume that the Fed couldn’t, well, move first. Since the correlation was positive, goosing the money supply when a friend was in office, and dropping it when an enemy was in office would seem to be a nice way to move the economy a bit, perhaps enough for some votes.

However, because of the structural break that occurred in the 1990s, the effect was precisely the opposite that someone who would have wanted to influence the election would have anticipated. That might well have led him to, well, push harder, so to speak, leading not only to slight attempts to goose things in the desired direction, but to the unusually large changes that were actually observed. Since he was normally in the business of reacting to, not leading change, it might have taken a while for him to realize what was going on. Perhaps even three tries. Put another way – in 1992, 1996, and 2000, if Greenspan wanted to influence the election, he might have, in fact, caused inadvertent damage to his own cause.

By 2004, enough time would have elapsed for someone working at the Fed to have realized that the world had changed – hence, the use of a new lever, namely a reduction in the Fed Funds rate. (There was also a bit of a decrease in the real M1 per capita for good measure.)

Now, if this is truly what happened, if Greenspan truly tried to influence elections and failed (or even got the opposite result than he wanted), does it excuse his attempts? Well, no. After all, its not the job of the Fed Chair to influence elections. Relying on incompetence or a structural break to even out the Fed’s bad behavior seems to me to be a risky strategy, and what’s to say the Fed today doesn’t have better information leading up to 2008? Enough people were incensed that the 2000 elections were, in effect, decided by the Supreme Court. Do we want the 2008 elections decided by Ben Bernanke?


Several readers have noted in comments and in e-mail messages that even unprecedented changes in real M1 per capita such as occurred in 1996 are not necessarily “unusual” and might have more to do with the Fed’s policy. The Taylor rule came up a few times. So… I’m collecting data now to check how the fed funds rate matches up with the Taylor rule.


Monthly M1
Monthly CPI
Quarterly population data which I then linearized into monthly
Quarterly real GDP per capita


As always, let me know if you want my spreadsheet.