The incentive schedule is a laminated sheet and I have carried a copy in my bag for eleven years. It lists what a female community health volunteer in Nepal is paid for, per event. A referral for an institutional delivery. A newborn visit within twenty-four hours. A completed immunisation card. Attendance at the monthly meeting at the health post.
Every item on it is a real thing that matters and I am not going to pretend the schedule was designed badly.
What is not on it is the four hours I spent one afternoon in a kitchen with a woman who was not going to go to the birthing centre, whose mother-in-law was in the room, and who eventually did go, eleven days later, for reasons that had nothing to do with anything I said that afternoon and everything to do with my having come back twice.
The referral is on the schedule. The eleven days are not.
I want to describe how that works in practice, because from a district office it looks like an incentive scheme with a coverage indicator attached and from a kitchen it is something else.
A volunteer with a schedule like this learns very quickly which activities produce a line in a register. The ones that do are the ones with a discrete moment: an event happened, on a date, to a named person, and somebody can verify it. The ones that do not are the ones made of repeated presence — going back, sitting, being seen at a funeral, knowing which household is in a dispute with which other household so you know who can be asked to speak to whom.
That second category is not soft. It is the entire mechanism by which anything reaches anyone in a ward of ninety households. The referral is the last four minutes of a process that took a month, and the schedule pays for the four minutes.
So what happens is that the coverage numbers rise, which is real — they have risen enormously in the years I have been doing this, and I have watched it happen to people I know, and anybody who wants to be cynical about performance-based incentives should first explain the immunisation figures for this district in 2011 against 2024. That is the counterargument and it is not rhetorical. Before the schedule, there was an honorarium that did not depend on anything, and there was no register, and I could not have told you what the coverage was because nobody could.
The schedule made the work visible and it made it countable and both of those were genuine gains.
The cost shows up somewhere else, and it took me years to be able to name it. It is in who gets reached. If my payment depends on completed events, then my time flows toward the households where an event is likely — the ones who were probably going to go anyway, who can be signed up in one visit, whose cards get finished. The eleven-day household is expensive. It is also, obviously, the household the whole programme exists for.
Nothing in the schedule tells me to neglect her. I do not neglect her; I went back twice and I would go back again. But I am one person with a ward and a family and a small shop, and every hour I spend in that kitchen is an hour that produces nothing in the register, and I am aware of that while I am sitting there, and I would be lying if I said the awareness has never moved my feet.
The fix I would want is not more money and not a longer schedule. It is one additional column: a record of visits made, not outcomes achieved, in households flagged by the volunteer herself. Not paid per visit, which would invite the obvious abuse. Simply recorded, so that when a ward's coverage is compared to another ward's, somebody at the district can see that one of them contains fourteen households that take eleven days and the other contains none.
At the moment those two wards look identical on paper and one volunteer looks lazy.
She had the baby at the birthing centre. The referral is in my register with a date on it, worth what it is worth, and next to it there is nothing at all.