Where the Voters Aren’t

You could be for­giv­en for think­ing that all Wel­ling­to­ni­ans are obsessed with polit­ics. Advance vot­ing star­ted today, and already some voters are proudly sport­ing their “YES I have voted” stick­ers. Yet thou­sands of Wel­ling­to­ni­ans have still not enrolled to vote. Who are these dis­en­fran­chised people, and why have they not enrolled?

Luck­ily, the Elect­or­al Com­mis­sion has data, broken down by elect­or­ate and 5‑year age bands. The inter­act­ive charts there are pretty good, but I decided to down­load the data and put it togeth­er in dif­fer­ent ways to explore it fur­ther. Here’s the enrol­ment pat­terns by age across the Wel­ling­ton elect­or­ates (arranged approx­im­ately from the out­er regions to the centre):

As has been widely noted, it’s the young­er voters who are less likely to be enrolled. But the details of this pat­tern vary by elect­or­ate. In the out­er elect­or­ates, by the age of 30 nearly every­one is enrolled; but when you get to Wel­ling­ton Cent­ral, the lack of enrol­ment goes right up to 40. It’s not just the youth who haven’t got around to enrolling: in some cases there is not­able dis­en­gage­ment among Gen X as well as Gen Y. Is this pat­tern com­mon across the coun­try?

I wanted to look at the nation­al trends, but with a twist: I inver­ted the age axis so that it showed the approx­im­ate year in which each cohort turned 18 and was hence eli­gible to enrol. Was there a time when there was a dis­tinct shift in enrol­ment pat­terns, or have people just gradu­ally been becom­ing less inter­ested? Let’s have a look.

The pat­tern is strik­ing: up until about the mid-90s, the pro­por­tion of non-enrolled people was very gradu­ally increas­ing, but it stayed below 5%. From the mid-90s, the slope changes sud­denly, and while the young­est people (18–24) are still the least likely to be on the roll, if any­thing the rate of increase has been eas­ing off. I thought about what might have triggered this: Rogernom­ics? Stu­dent tuition fees? House prices? But one major, rel­ev­ant event stood out; a fun­da­ment­al change to the vot­ing sys­tem, with the first MMP elec­tion in 1996.

This sur­prised, and to be hon­est, dis­mayed me. Surely a pro­por­tion­al sys­tem, which should be fairer and give people a great­er say, would encour­age enrol­ment rather than dis­cour­age it? But maybe all the quirks and appar­ent gam­ing of the sys­tem has put people off. Thresholds. Waka-jump­ing. Over­hangs. Epsom. Win­ston.

I pos­ted this to Twit­ter, and it gen­er­ated a lot of dis­cus­sion and some help­ful objec­tions. The data is a snap­shot in time rather than lon­git­ud­in­al, so it could be that it’s about age rather than gen­er­a­tion, and per­haps New Zeal­anders have always put off enrolling until they were older. People do go on and off the roll over time, so it may not just be about people “stop­ping enrolling” after some date. But that hypo­thes­is is con­sist­ent with the over­all decline in enrol­ment and turnout.

Oth­er pos­sible explan­a­tions included: the WWII and Boomer gen­er­a­tions see­ing vot­ing as a hard-won right; a range of oth­er wor­ries affect­ing young people (stu­dent debt, uncer­tain careers); and a sense that the polit­ic­al and eco­nom­ic nar­rat­ive is so dom­in­ated by older prop­erty-own­ers that it has become irrel­ev­ant to young­er people. One of my oth­er thoughts was that people who were first eli­gible to vote in 1996 would barely remem­ber a time before Roger Douglas, and that since then both major parties had espoused vari­ous degrees of neo­lib­er­al­ism, remov­ing the sense that there’s any chance to fun­da­ment­ally change the eco­nom­ic sys­tem.

After that dis­mal thought, I went back to look­ing at geo­graph­ic vari­ations. Why do some elect­or­ates have a great­er age vari­ation than oth­ers? To make map­ping easi­er, I used the over­all age pat­tern to hone the com­plex pat­tern down to a single index: the dif­fer­ence between the per­cent­age of over-35s enrolled and the per­cent­age of 25–35 year-olds enrolled. I ini­tially looked at 18–35 year-olds, but those maps were clearly dom­in­ated by stu­dents (e.g. the dif­fer­ence was massive in Ilam and Duned­in), and the mobil­ity and oth­er com­plex­it­ies of stu­dent life seemed to drown out oth­er pat­terns. I made a map, with zoomed-in sec­tions to make the cit­ies more legible (click for a high­er-res­ol­u­tion ver­sion):

While there are some inter­est­ing vari­ations between provinces, the aspect that most strikes me is that there’s a gradi­ent from cent­ral cit­ies to sub­urbs to rur­al areas. The more urb­an a place, the wider the enrol­ment gap between under- and over-35s. There’s a host of pos­sible reas­ons for this: mobil­ity, eth­nic diversity, recent immig­rants who are unsure of their eli­gib­il­ity, the dis­tance from fam­ily and their vot­ing tra­di­tions. There’s room for a lot more research on this.

One thing we can say for sure is that in the cent­ral cit­ies, young people have giv­en away their voice. They will spend more time in the future, but are hav­ing less say in it. People between 18 and 35 make up more than half of the eli­gible voters in Wel­ling­ton Cent­ral, but only 41% of the cur­rent roll. If the nearly 10,000 eli­gible but unen­rolled 18–35 year-olds in this elect­or­ate enrolled and voted, you could make a real dif­fer­ence to the way Wel­ling­ton is rep­res­en­ted in Par­lia­ment.

So, what are you wait­ing for? Enrol now.

[Data source: Elect­or­al Com­mis­sion enrol­ment stat­ist­ics as at August 2014, with estim­ated pop­u­la­tion stat­ist­ics as at 30 June 2014 (pro­vi­sion­al) using 2013 census data. Charts and maps cre­ated with open-source tools: R, QGIS and Lib­reOf­fice.]