Many historians have a love-hate relationship with official statistics. The love comes from the fact that statistics can really help buttress an argument. If you’re arguing that a community has become more ethnically diverse over a finite period of time, census statistics showing that change are more convincing than an observation that there are fewer white faces on the streets than there used to be. The hate comes from the fact that the type of statistics that are collected often changes over time. Some census information that is collected in the present was not collected in the past and vice versa, making it hard to make direct comparisons over space and time.
This was brought home to me when I was compiling tenure data for the Te Ara housing entries. I wanted to know the proportion of renters and home owners over time. My starting point was the 1916 census, the first time housing tenure statistics were collected. As I trawled through the succeeding censuses it was soon apparent that it was going to be no easy task to complete. The variables kept changing: old ones like ‘buying on time payment’ (a government scheme, distinct from paying a mortgage) disappeared and new ones like ‘rent free’ were introduced. In 1956 rent free became ‘free dwelling with job’ and in 1971 this was subdivided into ‘free dwelling with job’ and ‘provided free, not with job’ before changing yet again. The 2006 introduction of the family-trust variable added further complexity to the picture. In the end I divided the data into three variables: renting, owned and other.
I thought I had done a good job. The result showed a long term trend of declining home ownership that was supported by other evidence. But in March this year I got an email from an analyst at Statistics New Zealand, Angela, asking me how I had calculated the figures because their 2006 figure for home ownership was higher than mine. After much to-ing and fro-ing we discovered that we had used different denominators and had placed some variables in different tenure categories for different reasons. Angela recalculated the historical figures according to her methodology – our respective calculations were not too dissimilar until 1996, when our home ownership rates began to diverge – and we agreed that Te Ara would adopt these so we were consistent. The result underlined the productive relationship between Te Ara and Statistics New Zealand. It also highlighted the flexibility of web-based production in being able to make rapid content changes.
Interestingly, in the Dominion Post on 20 May 2014, the director of Auckland University’s Retirement Policy and Research Centre, Michael Littlewood, suggested that census data showing declining rates of home ownership was questionable. He argued that if home ownership rates were really falling, we would see a strong increase in the number of renters, and this was not happening. More needs to be known about the role of family trusts in the whole equation. Littlewood provocatively concluded: ‘We cannot draw any particular conclusion about home-ownership trends from census data.’ This is unfair, but it shows that debate over housing statistics is set to continue and the way Te Ara’s housing tenure graphs are calculated may change again in the future.