Are confident consumers good for the government?

This is an expanded version of something I wrote for the Crikey daily mail last week.

Does increasing consumer confidence boost the government vote? Orthodox wisdom would have us “yes” – good consumers are good government voters. If people are happy with their economic lot, if they’re happy with their economic expectations about the future, then they will – so the theory goes – reward the government for overseeing a fortuitous economic environment that boosts prosperity and their material well being.

However, the actual answer is somewhat different. Once upon a time there was indeed a strong relationship between consumer confidence and government political stocks -back in the boom/bust of the late 80’s and into the 1990’s recession period – and that probably explains why this myth is so enduring.

But a myth it generally is – at least in terms of the strength of the relationship – as we shall see.

To have a proper look we’ll need to go and get some data. For consumer confidence we’ll use the monthly Westpac-Melbourne Institute Consumer Sentiment Index.It doesn’t really matter what consumer sentiment metric we use here since they all move together anyway.For those that want the actual data, you can download it from the RBA as an Excel file here.

For our polling data, we’ll use Newspoll – but we’ll take the monthly average of the Newspoll figures to keep the Newspoll data chronologically consistent with our monthly consumer sentiment data, and it also knocks out a little of the polling noise. Combined, these give us a series of data running from December 1985 through until February 2010. Throwing them in a simple comparative line chart to see what we’re dealing with gives us:

The consumer sentiment is on the left hand axis, the government primary vote is on the right. From the beginning, what stands out is the early correlation in the series up until around 1993 and the way they lost their connective dynamics around 1994 and never really regained it in any substantial manner. If we just run this chart again, but split it into the two periods of pre and post 1993, it becomes a bit more obvious.

Similarly, if we run a simple scatter plot and regression line through these series, we find a solid relationship in the first period and no relationship at all post 1993:

What’s that Skippy? “The relationship between government political stocks and consumer sentiment looks like a function of economic downturns and the early recovery period?”

Perhaps, my marsupial friend – but there’s a little bit of sub-sample data that throws a spanner into the works.

If we look at how consumer sentiment and the government primary vote relationship operated over the 2006-2010 period where the last economic cycle peaked and turned, we get:

The relationship ran the wrong way – where government popularity increased as consumer sentiment decreased. Obviously other things were driving the vote.

Which is pretty much where we’ve been for 20 years – living in a political world where other things are driving the vote. Some periods the relationship will hold one way, other periods it will run the opposite – suggesting that there’s no overall causative relationship actually at play that can be witnessed or separated from those “Other Things”.

While there was certainly a nice solid relationship between consumer sentiment and government popularity 20 years ago, that relationship appears to have washed out of the system in any real meaningful way. We can get as nerdy on this as we want – knocking out the autocorrelation, looking for relationships between the first differences and lagged first differences of the respective series – but it all ends up the same way. What once was no longer appears to be.

Something to keep in your thought orbit the next time you hear about a small rise or fall in consumer sentiment being good/bad for the government of the day. It just doesn’t really matter if we look at even the not-so-recent history.