If we use Newspoll for our polling data, and the Westpac/Melbourne Institute Consumer Sentiment Index for our consumer confidence data, we’ve got a reliable set that goes back to 1985.

Between 1985 and 1993, there was a strong and significant relationship between government political support and consumer sentiment. We can see that with a simple line chart and  scatterplot with regression line for that period’s data:

graph11

However, as the economic recovery of the ’90s gathered pace, the strength of the relationship started to weaken. If we do the same thing again, but this time for 1994-2010 data, what we see is no relationship at all.

graph22

The relationship didn’t suddenly stop in 1994, it started to subside at about 1.3% a month in strength, so by 2001 the relationship had pretty much ceased to exist — which is where we find ourselves today.

Statistically, at least, anyway.

What is weird is that since 2007 there has actually been a strong, significant, opposite relationship where as consumer sentiment goes up, the government vote goes down — but it’s too soon to tell if that actually represents some real world effect or if it’s just one of those weird statistical artifacts that can pop up in sub-samples some times.