press gallery...
21 Jun 2018

Footy fans and false correlations

Property investors are always on the lookout for the causes of house price growth. The reason is obvious – if we know what leads to price rises, we can identify potential boom areas before the growth starts and then sell out before it stops. One well tested way to do this is to find links between cause and effect, which is a process called correlation.

When we find such correlations and use them to predict housing market behaviour we need to be sure that a particular cause is linked to a certain outcome. This is because some apparent correlations can actually be accidental and they will totally mislead us as result.

One commentator recently claimed that your footy code preference may well influence where you end up living. The writer had analysed interstate migration trends over the past twenty years and discovered that there was a significant correlation between footy codes and preferred destinations.

The researcher found that people in New South Wales and Queensland prefer to move between those two States and that Rugby League (NRL) is the predominant footy code in both, while people in Victoria and Western Australia prefer to move between those States and that Aussie Rules (AFL) is the dominant footy code played in them.

This footy code correlation led the researcher to the amazing finding that it is your footy code preference that determines which State you will move to. It is certainly impressive stuff at first glance but does not stand up to any serious scrutiny at all, because the supposed connection between interstate movements and footy codes simply does not exist.

Most people do not go to football games or watch televised games except for World Cup Soccer when Australia is playing, so there is no obvious connection between interstate movers and footy fans. But even if there were, do footy fans really care enough about a particular footy code to allow that to influence which State they move to? It’s a big stretch to claim that because someone follows AFL they will prefer to move to a State where AFL is the main footy code.

The researcher also wrongly assumes that footy fans follow a particular code, but the reality is that they barrack for a team, which is quite different. In fact, many AFL teams are in NRL States and vice versa, such as Melbourne Storm, who are a Rugby League team playing in Victoria, and the Sydney Swans, an AFL team located in the NRL heartland of Sydney. Other examples include the Brisbane Broncos, West Coast Eagles, Gold Coast Titans, Adelaide Crows, Brisbane Lions, Fremantle Dockers, Port Adelaide Power, Greater Western Sydney Giants, Gold Coast Suns and the North Queensland Cowboys.

Further, in States such as South Australia and Tasmania, AFL is the predominant footy code but these States weren’t covered in the research, probably because the correlation stops occurring if you include them in the mix.

CORRELATIONS MUST BE BASED ON CAUSE AND EFFECT

If we make such false correlations, we can arrive at interesting but completely incorrect conclusions. For example, there is a strong correlation in Australia between the amount of ice cream which is eaten each month and the number of accidental drownings that occur, as this figure shows.




Based on the evidence, you might decide that ice cream eating somehow causes more drowning accidents, but both ice cream eating and drowning accidents are effects, and their behaviour is dictated by another cause altogether - the changing seasons. Ice cream tends to be consumed more in the hotter months of the year which also happens to be when more people take to the water. This example demonstrates the same error made if we try to correlate interstate movements and footy codes. It is changes in our personal, family or employment circumstances that cause such movements and dictate where we are most likely to go, not the football code we follow.

The search for causes of housing price changes is similar. We commonly hear that gentrification and urban renewal are causes of housing price growth, while a rise in distressed sales leads to falling prices. Gentrification and urban renewal do not cause rising prices, but both are the effects of increased demand for inner urban living in new well-appointed apartments close to entertainment, employment and recreation features. This demand is generated by upgrader households and investors which ten leads to price rises and the further gentrification of other inner localities.

Increased numbers of distressed sales are caused by rising interest rates and increases in unemployment which hit first home buyers much harder than other property owners. Their exposure to any worsening of their financial position is much higher than it is for well-established families who have been in their homes for years, or investors, who can often pass on higher costs by raising rents. Distressed sales and falling housing prices are the unfortunate effects of this unhappy situation, not the cause.

When you read about research that can claims to predict housing market performance based on correlations involving footy codes, you should ask yourself whether they are examples of cause and effect, or if there is another cause altogether.


Acknowledgement: Photo courtesy of Simon Braid