Two messages about my own paperwork land in the same fortnight, and only one of them is trying to help me. The first asks me to check that I am still enrolled to vote at the address I actually live at. It takes a minute, I am, and the office that sent it banks nothing either way, because a name on the roll is not money. The second message is not a reminder at all. It is the amount I already owed with a late fee bolted onto it, sent by an office that had the due date in its system long before I went looking for it.
The only place they come together
Neither task was hard. Renewing a rego is not hard. Paying rates is not hard. Checking the roll takes a minute. What is hard is that every obligation sits in a different agency, behind a different login, on a date chosen by someone who has never met the others, and the only place in the entire country where they all come together is inside your head. The transport agency does its job. The council does its job. Internal Affairs does its job, the Electoral Commission does its job, and each of them hands you the identical unbilled task: hold this, and hold it until exactly the right week.
The obvious reading is that this is ordinary bureaucracy, and that the fix is one of two things. Better websites: a nicer portal, one login for all of it, another round of digital transformation. Or better you: a calendar, a system, a bit more discipline. Both are fair and both have been tried, and the letter still arrives to say the thing you meant to do is now late. Which leaves the narrow version of the obvious reading, the one that is genuinely hard to argue with. The second office should simply have done what the first one did, and sent a text.
The best-evidenced favour in government
That is right, and it is not a small point, because a well-timed reminder is close to the best-evidenced intervention in public policy. When New York City texted people before their court date, the best-performing set of messages cut failure to appear from 38% to 28%, an estimated 3,700 fewer arrest warrants in a single year, at under a cent a message. (Alissa Fishbane 2020)
When the United Kingdom’s Behavioural Insights Team reworded the standard tax letter, payment rates rose across more than 200,000 taxpayers, the strongest single version lifting them by about 5 percentage points. (Michael Hallsworth 2017) Nothing changed except the words on the page. Timing is the team’s wider claim rather than that particular paper’s finding, and it points the same way: their own distillation of the whole field is four letters, EAST, for easy, attractive, social, and timely. (Behavioural Insights Team 2014) A well-built text reminder is all four at once. It arrives where you already are, at the moment you can act, with the next step one tap away.
None of that evidence is from here. The strong trials are American courts and British tax debt, and nobody has run one against a New Zealand rego, warrant, or rates bill, so the size of the effect on this particular stack is a wager rather than a measurement. The direction of it is not seriously in doubt. Reminding people is cheap, humane, and it works, so a decent state should remind people more.
I stopped there for a while, and it was too early to stop. A reminder is not weather. Somebody decides to build it, times it, words it, and chooses whether it goes out at all, and that somebody has interests of their own.
The reminder has an author
Richard Thaler and Cass Sunstein gave the person who makes those decisions a name, the choice architect, and treated the role as basically benign: a designer arranging the furniture so the tired and the busy tend to land somewhere good. (Richard H. Thaler and Cass R. Sunstein 2008) Jeremy Waldron, reviewing Sunstein in 2014, refused to leave it there. (Jeremy Waldron 2014) His objection was not that nudges fail. It was that the books had almost nothing to say about trust. Much of the choice architecture we move through, he pointed out, is arranged by people whose interests are not necessarily ours, marketers and salesmen among them, and the enthusiasts kept skipping the obvious question. Who is watching the architect.
Waldron left that as a worry about dignity and manipulation, a general unease that a hand is on the furniture and you cannot see whose. Put a revenue meter on it and the unease turns concrete.
When the architect keeps the fine
In the United States there is a name for what happens when a government funds itself from the penalties it collects: taxation by citation. The Institute for Justice has documented the mechanism across hundreds of municipalities. (Institute for Justice 2019) Once fines and fees become a line in the budget, enforcement drifts toward the money and away from the purpose it was meant to serve, and the body doing the enforcing acquires a standing interest in the offence continuing to happen. A town whose court is paid for by the fines that court levies does not, at some institutional level nobody has to say out loud, want fewer fines.
Now hold the two messages from the start of this side by side. Both are nudges in Thaler’s sense, both features of an environment that change what a busy person does, and they differ in one respect that decides everything. The first office loses nothing when I comply. The second office loses the late fee. A text that lands the day before the deadline is, for a body that has come to lean on the penalty, a message that costs it money every single time it works. The incentive to build that reminder is not neutral. It is negative.
I cannot tell you which New Zealand offices are in which category, and that is the uncomfortable part of writing this. One verified pair, the office that texts and the office that does not, would settle the argument in a paragraph.
The reminder that is never built
This is the part that makes the whole thing so hard to see, and it is why Waldron’s question has no comfortable answer here. You can count the reminders a state sends. You cannot count the ones it decided never to build. A missing text leaves no trace, files no report, and prompts no complaint, because the person who would have complained never knew the reminder was possible in the first place. The failure is an absence, and absences do not turn up in an audit.
Where compliance and revenue point the same way, the text gets written, and it is a genuine kindness. Where they diverge, nothing gets written, nobody experiences a manipulation, and the fee arrives looking exactly like a fact about the person it landed on. They do not think a message was withheld. They think they are disorganised. The difference between a state that reminds you and a state that fines you was never competence, or budget, or better software. It is whose ledger the deadline happens to sit on, and the citizen is the only party in the arrangement who cannot see which one it is.