The McNamara fallacy: the economy runs on work nobody measures
Updated: 5 days ago
"So, do you work?"
It's one of the first questions we ask when we meet someone.
For parents at home with young children, the answer is often "no".
But if they did the same work for another family, it would be a job. Childcare, cooking, and cleaning are all things people get paid to do. They say no because most people understand "work" to mean paid work.
Official data draws a similar line. In Aotearoa New Zealand, someone caring for their children full time is counted as "not in the labour force". Yet someone who spends an hour a week helping out, unpaid, on a relative's farm is counted as employed.
So why does official data leave this work out?
Part of the answer is the McNamara fallacy.
What is the McNamara fallacy?
Robert McNamara was the US Secretary of Defense from 1961 to 1968. He tracked the Vietnam War's progress through numbers, and body counts (the number of opposing fighters killed) became a key metric.
But this metric hid a lot. Body counts couldn't show whether local people supported the fighters, and that support affected whether new fighters kept arriving. The counts were also often inflated, as can happen when a metric becomes a target.
In 1971, social scientist Daniel Yankelovich named this approach the McNamara fallacy: measure what's easy, and disregard what's not. He warned that it leads people to treat whatever can't be easily measured as if it doesn't exist.
Work that goes into something bought and sold is easy to measure, because it shows up in wages, hours, and sales. Work done for a household produces nothing for sale, so it's left out of the numbers, regardless of how much it matters to the people doing it and the people they care for.
How unpaid work is left out
Unpaid work does get measured, but not often, and it's left out of the numbers that most decisions rely on.
In Aotearoa, this work has its own survey: the Time Use Survey. From what I can tell, it has been run twice: first in 1998/99, and again in 2009/10. (In contrast, the Household Labour Force Survey, which measures employment, publishes results every quarter.) Using data from the first survey, Stats NZ estimated that unpaid work in 1999 was worth about $40 billion, equal to 39% of gross domestic product (GDP), the main measure of a country’s economy.
That figure covers three kinds of unpaid work:
In the home. GDP doesn’t count unpaid work at home, like cooking, cleaning, and childcare, as production.
In the community. In 2018, 89% of New Zealand’s non-profits had no paid staff. Volunteers across this sector did work worth an estimated $4 billion that year, none of which was included in GDP.
In health and disability care. When someone leaves hospital, their stay is recorded: how long it lasted, what treatment they had, when they went home. The help they get at home afterwards, often from family, usually isn't part of any health record. Together, unpaid family carers of people who are ill, disabled, or older provide care worth an estimated $17.6 billion a year.
Unpaid health and disability carers outnumber practising nurses about 8 to 1.
Whose unpaid work is it?
The work of raising children, caring for people, and keeping households running is known as social reproduction. This work raises the next generation of workers, keeps today's workers going, and looks after people who can't care for themselves.
Women do most of it. In the most recent Time Use Survey, from 2009/10, women spent about 4 hours 20 minutes a day on unpaid work, and men about 2 hours 32 minutes.
Capitalist societies tend to split work into two kinds. Work that produces goods and services to sell counts as part of the economy. And unpaid work at home that is treated as something people do out of love or duty.
It’s easy to think of paid work as what holds up the economy, but what if it’s the other way round? Every paid worker was raised by someone. Many still rely on others to cook for them, look after them when they’re sick, or care for their children while they work. Without social reproduction, there would be no economy.
Why measure unpaid work?
We need decisions about how we live together to account for all the work that keeps society going, including the work that's never bought or sold.
To plan for care. As people live longer, the need for care is growing, and much of it is provided by family. If that care isn't recorded in our data, it's harder to see the gap between what's needed and what's available. Measuring it would show where support like respite and home help is needed most.
To see what policy changes do. When public services are reduced, the work often moves to families. In 2024, for example, New Zealand's disability support agency tightened the rules on what support funding could be spent on, citing rising demand. Measuring unpaid care would show whether a change like that saves public money, or simply moves the cost onto families.
To design support around people's lives. People move between paid work and caring over a lifetime, but many support systems are only built around paid work. Women's KiwiSaver balances are on average 25% lower than men's, and the Retirement Commission links part of that gap to years of unpaid caregiving. Better data would show where support falls short, and who carries the cost when it does.
What to do about it
The McNamara fallacy is easy to fall into, because the numbers we have are usually the ones that were easiest to measure. So, next time you're working with data about people, here are four things you can do:
Ask what your data can't see. Before you report a number, ask what it leaves out. For example, wages are easy to count, but the unpaid work that makes them possible isn't.
Use the unpaid work data that exists. The Census asks whether people look after children, care for someone who is ill or disabled, or volunteer. Stats NZ rated the 2023 results as poor quality because of low response rates, but they still tell you more than grouping everyone into "not working".
Give unpaid work a place in your data. If you design forms or surveys, give people a way to record caring and other unpaid work. If "not working" is the only option, that's what carers will become in your data.
Say what's missing. If your data can't show unpaid work, or the best data on it is years old, say so in your findings. How often something gets measured can show how much it's valued.
What counts as work is something we decide together, in the questions we ask each other and in the data we create.
I write about data and power, so if you'd like to follow along, the best place to do that is my newsletter. This article is part of the Things We Tell Ourselves series. Each article looks at an everyday saying, what the data shows, and what it means if you tell stories with data.
Kat Greenbrook is a data storytelling consultant, author, and workshop facilitator based in Wellington, New Zealand. She is the founder of Rogue Penguin and the author of The Data Storyteller's Handbook.
