How general elections affect the property market

It is a question that I have often heard asked. As well as being a regular explanation made when laying the blame for a period of quieter sales leading up to the general election.

So the question is – are there any facts that can be brought to bear to substantiate or dispel this belief? I have never seen any factual analysis – that is until now.

There have been seven general elections held over the past 22 years for which accurate property sales statistics have been kept by the Real Estate Institute. That should be sufficient data to provide some insight.

The question is then how to evaluate the period running up to the election as compared to a normal period to see if there is an effect? A further and broader question: is when is there ever a normal period in the property market with so many variables at work? In my view looking at year-on-year sales volume variance is not robust enough; while it deals to the seasonal factor, it is open to the influence of different stages of the property market cycle.

The measure I have come up with is a seasonal comparison. It uses the three preceding months leading up to each general election date and calculates the representation those months were of the total sales for that year. I then compared that percentage representation as a single figure against the norm for the same three months of the year based on a larger set of preceding data going back to 1992.

In other words, to take the example of the last general election in November 2011, I took the sales for the months of August/September/October of 2011 and calculated that against total sales in 2011, which was 8.84%. I then compared that to the normal average of the months of August/September/October across all years from 1992 to 2010, which was 8.59%. So, in this particular case, I would say that that election in 2011 saw no negative impact on property sales in the lead up to the election. In fact sales were slightly ahead of normal.

Here is the result of this analysis for each of the eight general elections since 1993.

What to make of these results?

You could say that on average general elections depress property sales as five of the eight elections caused property sales to decline as compared to normal. However, there is no real consistency. The completely and significant opposing variance in the results for 1993 and 1996 are too significant to ignore.

I did look at the issue of political leaning of elections as a factor. National won the 1993 and 1995 elections and the 2008 and 2011 elections which saw varied outcomes, whereas Labour's impact in winning the other elections, all of which led to falls in sales, so maybe the political factor is key?

Without a convincing answer I reflected on the impact economic sentiment has on the property market at the time of each election. I plotted these variance to the norm for the three months run up to each election against the GDP trend from Reserve Bank data.

Now in my view this correlation makes sense and aligns the election to the cycle of GDP as follows:

1993 - GDP on a surge, economic optimism = 14% rise in relative sales

1996 - GDP declining, economic pessimism, compounded by first MMP election = 19% decline in relative sales

1999 - GDP starting recovery from Asian crisis, economic caution = 7% decline in relative sales

2002 - GDP cautious recovery from post 2001 falls, economic caution = 2% decline in relative sales

2005 - GDP declining, economic caution = 1% decline in relative sales

2008 - GDP collapsing, economic pessimism = 5% decline in relative sales

2011 - GDP recovery, growing economic confidence = 3% rise in relative sales

It would be safe to say that, not surprisingly, the impact of an election on the property market is more a reflection of the economic confidence at the time than any across-the-board view that elections dampen property markets.

As to September 2014, well, with one month's data in the system, I am sorry to say for all those looking to blame the election for a dampening of sales results – it doesn't look like it.