The care paradox

As capitalism matures, how do we stay human?

This is the second essay in a three-part series for a project with the Joseph Rowntree Foundation, Social justice in a digital age. The first essay is here.

When we imagine a high technology future, we picture things that are shiny and new. From AI to Teslas to crypto, the future is about disruption and innovation. We’re not wrong about this, but we’re only half right. Because just as mountains are inseparable from valleys, so a digital society gets its character from the peaks of high-tech and from the troughs of low-tech in between.

What stays low-tech in an age of high-technology? A big part of the answer is care. Whether we’re helping an elderly person to wash themselves with dignity, relieving the dread of a person who is terminally ill, or helping a child find their way in the world, care is irreducibly human; no robot or AI can replace it.

This irreducible humanity has an important implication: it’s hard to make care more productive. Or at least harder than other activities. In fact, to the extent that care’s human touch — the time, love, and attention it takes — is its value, care can’t be sped up.

In this fact lies a force of geological power and insistence: as technology drives innovation in gadgets and gizmos, care falls ever further behind. It gives rise to a phenomenon I call the care paradox: the more that capitalism matures, the more we squeeze and sideline the most human aspects of life. It’s a phenomenon that is coming to define the character of a 21st century society, yet it’s one we barely acknowledge, let alone proactively shape.

In this essay I want to spend some time with the care paradox — to understand what drives it and see what it would take for us to regain a sense of control. It’s a story of two halves, the first focused on economics and the second on politics. And it all starts somewhere surprising: a world not of care but of music.

Part 1: An insistent economic force

At Cafe Oto in east London, folk musician Peggy Seeger sits at the piano. Over the next hour or so she has the audience eating out of the palm of her hand. She sings melodies from her childhood and songs by her mother, avant garde composer Ruth Crawford Seeger. It’s a moving performance, heavy with the passage of time. Seeger is 87; her first album was released 68 years ago.

It was the rising price of live performances like this one that first led the economists William Baumol and William Bowen to a phenomenon that became known as Baumol’s cost disease. Puzzled as to why tickets kept getting more expensive, faster than average inflation, the economists developed a theory that Baumol would later call, with uncharacteristic immodesty, “the longest valid forecast ever to emerge from economic analysis”.

The quickest way to understand the Baumol effect is to watch a performance like Seeger’s and ask a question that feels out of place in the context of music: is she productive?

As Seeger moves through her setlist, we get the answer; each track takes just as long to perform now as when it was first composed, often decades before. In the nearly 70 years since Seeger’s first album, as average labour productivity rose 300%, the productivity growth of a live music performance has been zero.[1]

The theory that Baumol and Bowen drew from this observation was summed up in a 1967 article:

If productivity per hour rises cumulatively in one sector relative to its rate of growth elsewhere in the economy, while wages rise commensurately in all areas, then relative costs in the nonprogressive sectors must inevitably rise, and these costs will rise cumulatively and without limit.

In plain english, the logic runs like this. Over time, technology makes some tasks more productive; think, for example, of intensive agriculture or assembly line manufacturing. This pushes up wages for workers in these sectors, which means that workers in other sectors, who after all live in the same economy, see their wages rise too. The result is that, in sectors without productivity growth, prices must rise in order to balance things out.

Baumol avoided forecasts but the implication of this theory was clear: as capitalism matures, the price of services that are performed in-person will rise. The chart below, for the US economy, shows the prescience of this insight, and the pattern in other countries is the same: goods have gotten cheaper, while personal services with a heavy element of what Baumol called ‘handicraft’ have become more expensive.

In the performing arts, where Baumol first focused, ticket prices have consistently outpaced inflation. In the US, a typical theatre ticket cost 5 cents in the early 1900s. If that had grown in line with inflation, we’d pay $1.75 to see a play today. In fact, cheap seats on Broadway or the West End now average $26 (£23), rising to $157 (£140) at the upper end. The Rolling Stones’ early London gigs cost 4 shillings, equivalent to £5 today. On their latest tour, standing tickets went for £700, an 18-fold real terms increase.

It’s care, though, that has turned out to bear the most socially significant consequences of the Baumol effect. Services that are care-intensive, or rich in human contact, share the irreducible humanity of a live music performance. And so, like concert tickets, the price we pay for care has risen year after year.

We see this in how much we spend on care-intensive services like healthcare. In 1958, the average person in the US spent $134 a year on healthcare; by 2012 the average American spent $8,953 a year. The price of higher education — another service that relies heavily on in-person contact — has also soared, with annual tuition at Harvard rising from $800 in 1953 to $47,900 in 2019. The story is the same in social care, where in the UK a place in a residential care home now costs an average £34,940 a year and £45,000 in the southeast of England.

To see how these economic trends have changed our lives, we can compare prices to wages. Thanks to the falling price of goods, an average US worker today needs only three days’ wages to buy a toaster, a TV, and a portable music player, while the closest equivalents in 1958 — with far worse functionality — cost 4.6 weeks of wages to buy, a time cost drop of 86 percent. In the same period, the time cost of washing machines and tumble dryers fell 83 percent. So while we think of the 1960s as the golden age of consumerism, with respect to the price of goods these were austere times versus today. For the price of a home entertainment system in 1962 you could now buy a laptop, a GPS, a digital camera, a home theatre system, a plasma TV, an iPod, a Blu-ray player, a TV recorder, an iPhone, a camcorder, and a digital radio, and have money left over. Except you wouldn’t need to, since the iPhone alone would do most of those things.

For care-intensive services, we can reverse all of that: despite wages rising, we feel more squeezed. In 1958, the average wage in America was $1.98 an hour, yet it took only 15 days’ work to cover average healthcare spend for the year. By 2012, the median wage was $12.80 an hour and it took 58 days to cover a year’s healthcare costs. From childcare to social care, the story is the same: we’re richer yet we have to work longer to pay for care.

We see these trends manifest in culture. If you speak to an elderly relative about their childhood, you get a sense not just that people were poorer two or three generations ago, but of how our attitude towards things and human services has reversed. The cliche that people used to look after things better is true for a reason: things were expensive. If your watch broke, you fixed it; if your shoes became scuffed, you took them to the cobblers. People darned socks. Yet services were cheap compared to now. Even families on quite ordinary incomes would have, if not servants, then at least someone — typically ‘a girl’ — to help around the house. In 1931, domestic service was the most common occupation for women, accounting for one in four women in paid work.[2] In the long-run — despite periodic spikes in inflation, like the one we’re living through now — we’ve shifted to a disposable culture. Who darns socks any more? Yet if a friend had a maid today you’d think they were auditioning for Downton Abbey.

To what extent are these trends explained by the Baumol effect? The answer is complicated and I’ll link in the footnotes to some explainers that go deeper than I have space to do here.[3] TL;DR — the Baumol effect isn’t the only reason that the cost of care-intensive services has risen; in health and education in particular there are other forces at work.[4]

Still, underneath a lot of complexity, the Baumol effect is there, doing its steady, inflationary work. And there are two care-intensive sectors, childcare and social care, that provide a particularly pure environment in which we can see these forces at work. An hour spent caring for a child or an elderly person today is pretty much the same as it was in 1870. And so as we ponder the implications of cost inflation, we’ll find they’re hardest to escape in childcare and social care.

All of which is to say: Baumol was onto something. But also, isn’t there something missing from this story?

After all, when we think of care today, and especially childcare and social care, we hardly picture wages rising to push up prices. Instead, the image that springs to mind is of an ever-intensifying squeeze. Care is rife with low pay, burnout, and a struggle to protect even the most basic levels of human dignity and respect.

So what’s going on? To answer that, we need to move to the second half of our story: the role of institutions.

Part 2: Stubborn institutions

Like many economic theories, it’s easy to interpret Baumol’s work in an overly simplistic way. There’s a version of his theory — not, to be fair, one he endorsed — in which the logic feels neatly hydraulic. Productivity rises in progressive sectors, which pushes up wages for these workers. This in turn pulls up wages in stagnant sectors, so prices in these sectors must rise.

It’s as if the whole thing balances nicely; the economy finds a new level. Really, though, economies are more political than this. Economic forces don’t flow freely; they hit up against human constraints. And it’s from this interaction of economics and politics that the care paradox is born.

So for a more rounded sense of how this all plays out, we need a mental model that accommodates politics, and in particular a model that conveys the power of institutions.

I’ve come to think about it this way. Maybe Baumol was right that stagnant productivity in care-intensive services creates an inflationary or expansionary pressure. But maybe the consequence of this pressure — whether it pushes prices up, or instead oozes out in another direction — depends on the institutional surroundings. Maybe stagnant productivity makes care behave like expanding foam, or like magma oozing from the earth. The expansionary pressure is insistent but the substance is malleable or fluid; it can’t be stopped, but it can be shaped. And so institutions act like moulds, determining the social realities into which the economic force of uneven productivity cools and hardens over time.

Let’s take this mental model and apply it to the recent history of care. Since care-intensive services tend to be funded, subsidised, regulated, or delivered by the state, we’d expect institutions to play a decisive shaping role. So what has happened as an insistent economic force has pushed up against stubborn political institutions?

The first part of the answer would be familiar to Baumol: as stagnant productivity has inflated the unit cost of care, the resultant pressure has pushed up public spending, which has inflated the size of the state.

In 1950, the NHS cost £460 million a year. By 2020 this had risen 340 times to £158.4 billion, a 10-fold rise after inflation and a doubling from 3.5% to 7.1% of GDP. We see the same across mature economies; healthcare now accounts for 6% to 8% of GDP in most developed economies (and roughly twice that in the car crash of US healthcare), typically reflecting a 10- to 20-fold increase after inflation since 1970. Alongside growth in public spending on other caring services, this accounts for a major part of the state’s expansion in the 20th century.

Alongside rising public spending, we spend more privately too. From healthcare to childcare to eldercare, we all spend more of our budgets on care than previous generations.

We see this again in a care-intensive sector like healthcare, where private spending has risen fast, even as public spending has grown. Private spending has tended to rise quickest at the edges of a service like the NHS, as if the inflationary pressure from care has oozed out the sides. In the UK, for example, private healthcare spend rose from £14.5 billion in 1997 to £40.3 billion in 2019, around 40 percent faster than inflation, with rapid growth in services like physiotherapy and counselling.[5]

In another care-intensive sector, childcare, we’ve seen a similar rise in both public and private spending. Even as governments have increased subsidies for childcare, in the average OECD economy the cost of childcare for two children has risen to account for 30% of the median women’s wage (a stat that in itself says a lot about the unequal division of labour), while in Britain full-time nursery for a child under two accounts for more than half the average wage. In the US, childcare spending grew 2,000 percent from the 1970s to the late 2000s, so that for many families childcare now costs as much as buying a new car every year.[6]

So far, this is all pretty consistent with what Baumol had in mind. But the political institutions that sit around care have also redirected the ever-intensifying pressure in other directions.

For one thing, workers in caring sectors have become more and more squeezed. The most acute examples of this are in social care and childcare where, contrary to Baumol’s original expression of his theory, wages have fallen behind those in the wider economy, reaching unsustainably low levels.

In the UK, for example, care workers now earn an average £15,000 a year, while more than three quarters of frontline care workers earn below the real Living Wage. And since care workers are often not paid to travel between appointments, many earn below even the legal minimum wage.[7] Pay for care workers isn’t just low; it’s falling, even relative to the UK’s weak average wage growth. In the decade since 2012, average care worker wages fell behind retail checkout assistants for the first time. In the same period, only hairdressers — another irreducibly human activity — fared worse.

If we look at data on staff wellbeing, we see that this isn’t a case of the economy finding a sustainable new level; wages in care-intensive sectors are being pushed down to crisis point. In the UK, around a third of care workers now leave each year, twice the national average turnover rate, while vacancy rates run at over 10%, creating critical staff shortages.

As well as compressed wages, we’ve seen a squeeze on workers’ time. From healthcare to social care, frontline care work has become ever more frantic, creating a crisis of burnout among nurses, teachers, and childcare professionals. Within these care-intensive jobs, it’s care itself — from the time it takes to reassure a worried patient, or to lift the spirits of an isolated elderly person — that has been pushed out.

One study of nursing in 12 developed countries showed that in nine countries, even before the pandemic, more than a quarter of the workforce was burned out, while a US survey found burnout affects more than half of all physicians. In the UK, the proportion of NHS staff feeling unwell due to work-related stress rose from 28% in 2008 to 37% in 2016 and double the number of staff left due to poor work-life balance in 2015 than in 2011.

In social care, meanwhile, nearly two in three domiciliary care workers now say they often have 15 minutes or less for a visit with a patient; three quarters feel this compromises the dignity of the people they care for; nine in ten say they have no time for a conversation with the often elderly and isolated people they are helping; and three in ten are too rushed to help people wash or go to the toilet. As a result, people who receive care at home often suffer the indignity of an hours-long wait for the next carer so that they can go to the toilet, not uncommonly soiling themselves before help arrives. Meanwhile, within complex services like healthcare, it’s once again the caring aspects of the job that have been squeezed hardest, with 65% of nurses saying they have no time to comfort their patients.

In another context, like manufacturing, we might think of speed as a good thing, equivalent to higher productivity. In care, however, things are different. When a farmer yields more food per acre or a factory worker makes more widgets with a new machine, that’s productivity growth. But when a domiciliary care worker squeezes an hour of care into 15 minutes, it’s more like asking Peggy Seager to play her folk songs at four times their original tempo. Would this make Seeger more productive? Maybe, in the narrowest possible sense. But even the most ardent Taylorist would agree this misses the point.

Finally, the effect of stagnant productivity has been shaped by one other distinctive characteristic of care: most of the care that takes place in society happens in the informal economy, unpaid and unrecognised in GDP statistics, and all-but invisible.

As a result, as we’ve failed to absorb the rising pressure from care into formal care-intensive sectors, we’ve seen an oozing out into unpaid care, which now takes up more of our time.

The last few decades have seen rapid growth in unpaid care in all demographics but the people most affected have been the sandwich generation; people in late middle-age who care for both children and elderly parents. In the UK from 2001 to 2015, there was a 16.5% increase in carers in their 50s and 60s and a 33% jump in the number of people who care for 50+ hours a week.[8]

For many parents, it now feels like every waking minute is spent shuttling kids from A to B even as the laundry and dishes pile up, marking the end of each day with an exhausted collapse onto the sofa. For a growing number of people, the collapse is interrupted by a call from elderly parent who needs help of their own — or help navigating caring services that are themselves struggling to cope.

It all helps to explain a paradox of modern life: we think of technology as labour-saving, yet as we get ever more technologically advanced, life only seems to get more frantic.

And the rise of unpaid care is also a stark reminder that, at root, this is all a story about power and gender.

Even now that the postwar rise of women’s work has all-but closed the gender employment gap (although not the gender pay gap), women still do 75% of unpaid childcare and 60% of unpaid adult care. If women were paid for this care it would cost over £100 billion a year, around three quarters the cost of the NHS.

And since women also still do the majority of paid care, that squeeze on wages in care-intensive sectors keeps the gender pay gap stubbornly wide. Women make up 82% of the adult social care workforce and even more in the lowest paid frontline domiciliary care jobs, while 89% of nurses, 93% of the early years workforce; and 98 percent of nursery nurses are women.

As Madeilene Bunting writes in her book about care, Labours of Love, “care is the feminist issue”.

What does this all add up to? Certainly it makes me think the care paradox has a lot of explanatory power, helping to make sense of why life in a mature capitalist society feels the way it does.

It helps to explain why our society seems simultaneously abundant in certain possessions, like technology and clothes, and yet at the same time impoverished when it comes to care. It explains how taxes can be historically high as a share of GDP even as public services are on their knees. It explains why our labour market seems so upside-down, rewarding 25-year-old coders with wages that are four times what an experienced care worker can hope to earn. And it explains why technology seems not to deliver us lives of leisure but of stress and burnout.

The more I dug into these dynamics, the more I came to think of the care paradox as dark matter; a powerful yet invisible force that makes sense of phenomena we’d previously observed.

So I was struck, reading Bunting’s wonderful book, to hear Anne Marie Rafferty, Professor of Nursing Policy, describe care in the same way. “Eighty percent of the universe is dark matter but scientists can’t find it”, she tells Bunting, “it’s the same with that ineffable part of care.” Care seems to run through our society unseen, holding everything together.

So what can we do about all of this? If institutions have such power to shape the social implications of uneven productivity growth, surely we could use this power in a more proactive and intentional way?

Part 3: Policy implications

The more time I spend in the politics of care, the more our behaviour reminds me of a toddler. We spend half our time with our hands over our ears, screaming that we don’t want care costs to rise. We spend the rest of our time lulling ourselves to sleep with fairy tales that will make it all OK: the magic potion of childcare ratios, or the fairy dust of another NHS reorg.

If we could start by acknowledging the forces at work, we could then focus on how best to shape them. So where would that take us on policy? Although there’s no simple answer, I’ll end with a few reflections on the choices we face.

First, there’s no escaping that we’ll need to invest far more in care in the future. This is a simple consequence of uneven productivity growth and it creates an important role for the state. When productivity grows unevenly, we need to redistribute the gains, smoothing the rising mountains of progressive sectors with the deepening valleys of stagnant sectors like care.[9] And because there are such powerful reasons to socialise the costs of care, a lot of this smoothing will need to happen through the tax and spend mechanisms of the state.

We shouldn’t be too glum about this; if we interpret the Baumol effect in the right way, we see that it’s really a good news story.

My favourite exposition of the Baumol effect comes from the economist Joan Robinson, who wrote to Baumol after reading his 1965 article, sharing her take on his work. Robinson’s central point was that Baumol’s cost disease isn’t really a bad thing at all. Yes, productivity growth varies by sector; of course it does. But falling prices in progressive sectors more than cancel out rising prices in stagnant sectors. That’s why, on average, we get richer over time. So, although care gets relatively more expensive, it also gets, paradoxically, more affordable. The trick is to avoid spending all of our productivity gains on cheaper stuff; we need to save enough for care.

Robinson’s point is illuminating because it reminds us that the Baumol effect isn’t really a question of economics; it’s a question of politics.

The problem isn’t that care gets less efficient and therefore more expensive; the problem is that care feels more expensive because we get much better at making everything else, so other things get cheaper. So in a sense, being squeezed by care is just what progress feels like. And because so much of care is funded through the state, this squeeze feels political; it’s not prices that are rising, it’s taxes. Worse, because care doesn’t get better (in fact, because we tend to underinvest in care, it gets worse) we feel like we’re paying more for nothing. All of which explains why our relationship with care feels so fraught and why it seems to get more fraught every year, as capitalism matures.

One role for public policy, then, is to channel a rising share of GDP into care in a way that is economically and politically sustainable.

If that sounds hopelessly radical, remember that it’s essentially what we did In the 20th century, when we banked productivity growth is three main ways:

1. We earned more; real wages rose sixfold from 1870 to 2020, so that we were able to buy more food, clothes, and technology with each hour of work.
2. We worked less; average working hours roughly halved over the same 150 year period.
3. We pooled a rising share of our prosperity to subsidise care and to make access to care more equitable.

The radical mechanism we used to pool a rising sharing of GDP for care is called tax. We created income tax to pool a portion of wage growth and we created sales and profit taxes, like the UK’s VAT and corporation tax, which made goods in progressive sectors more expensive than they would otherwise have been, and care more affordable.[10] In the UK, we did this to the tune of over £300 billion a year, or nearly 15% of GDP, reflecting the budget for healthcare, social care, education, and childcare. So while we tend to think of tax as a mechanism to redistribute from rich people to poor people, it also redistributes from progressive sectors to stagnant sectors like care.

It’s instructive to think what our society would look like today if we hadn’t made these choices. Imagine, for example, if we had kept the state at its size in 1920. We would live in a caricature of the predicament I described above: we’d have cheaper things, from clothes to technology; there would be higher inequality between carers and coders, and between men and women; access to care would be less equal; and doctors, nurses, and teachers — not to mention parents — would be more burned out.[11]

This historical experience should guide the choices we make now. To my mind, it suggests there’s not much point arguing for a smaller state; unless we want to desocialise care-intensive services like healthcare, unravelling our 20th century consensus that these services should be funded or heavily-subsidised by the state, care cost inflation entails a bigger state. A better use of our time is to argue over the pace at which we expand care-intensive services, the scope of these services, the balance of public and private funding, and the funding mechanisms to use (e.g. taxation vs. social insurance, and the most sensible forms of tax).[12]

I’m conscious that this first implication is hard to stomach for people on the right of politics, so let’s turn to the second implication for policy, which is harder for people on the left.

For most of this essay I’ve taken a fairly deterministic line on care cost inflation, talking as if the Baumol effect is inescapable. I’ve done this on purpose because I think we’re in a state of denial about the economic fundamentals of care; we talk as if we could free ourselves from care cost inflation with just one more push on public sector efficiency, which isn’t true.

So we can’t escape care cost inflation. But we can slow its rate of growth. To do this, we would need to be much bolder about public service reform and think about reform in different ways.

When we talk about public service reform, we tend to focus on bureaucratic or process-oriented reforms like NHS reorganisations or rejigs to exams or curricula, or we debate how to apply management methods like choice and competition. At other times, we get distracted by digital gimmicks like giving iPads to nurses, or excited about AI-driven diagnostics, thinking that a specific technology might be the answer.

If we want to move the dial on care cost inflation, we would need to engage in much deeper and more patient work to change the delivery model for care-intensive services, making these services more scalable. This is at root about digital transformation, since it’s about upgrading the mindsets and technologies that underpin caring services to make better use of internet-era capabilities.[13]

It seems likely that these approaches have huge potential in health and education, where internet-era methods can be used to deliver services in a more scalable way. New delivery models probably have less potential, at least in the near-term, in childcare and social care, where Baumol’s original insight is so hard to escape; there’s only so much you can do to scale up caring for two year olds or helping a frail person to get dressed.

In practice, reform can help partly by making sure frontline professionals spend as much of their time as possible on activities that are irreducibly human. Every minute a nurse or doctor spends on admin amplifies care cost inflation. This means good management is important, because the last thing we want is to displace admin onto nurses, teachers, or childminders, or for nurses to waste time sending faxes or filling out paper forms.

More profoundly, though, reform is about changing how we deliver services, for example by expanding remote and group delivery models in healthcare, to protect in-person 1–2–1 time for when it’s most valuable.

In the last year I’ve seen for myself the power of technology in care. At the start of the year, I injured my hip and it became chronically inflamed. At first, I saw two GPs in-person, each time getting kind but generic and rushed advice: ‘rest up’. When that didn’t work, I searched on Youtube, watching videos from the world’s leading orthopaedic surgeons. (The top video had 2.4 million views, which, if delivered as in-person appointments, would take five full-time GPs 40 years.) Later in the year, I turned to the NHS for help with another condition and I ended up on a group video call with 10 other patients chaired by an NHS dietician. Far from degrading the quality of care, the group call — a replacement for an in-person 1–2–1 appointment — made the care better. It was more physically accessible; it was reassuring to meet people in a similar position; and I learned a lot hearing about people’s coping strategies.

This speaks to the power of new delivery models. But from my day job leading tech at a big frontline charity, I also know that technology-led public service reform has limitations and risks.

When technology evangelists — almost always well-off, able-bodied, men under 50 — extol the virtues of digital services, they tend to underestimate the extent of digital exclusion. But they’re also often naive about what care really is. They talk, for example, as if a GP appointment is an exchange of information, when it’s just as much an act of reassurance, respect, and compassion, and even diagnostics rely on things you can only pick up through physical presence, like body language and tone. Much of the value of care doesn’t translate to Zoom.

That said, there’s naivety on the other side of this debate too. It’s wrong to assume that digital services are less accessible than in-person services. For many people — people with physical disabilities, for example, or with severe anxiety, or insecure jobs, or for that matter caring responsibilities — a good digital service is more accessible than travelling to a service in person.

Plus, new delivery models — especially technology-enabled community healthcare — are better suited to the lifestyle illnesses, chronic conditions, and mental illnesses that now make up most of the burden of ill-health. Better, at least, than a 1–2–1 clinical delivery model that was designed for an earlier age of acute, communicable physical disease. So, done well, new delivery models can make services better and more caring even as they also mitigate the effect of care cost inflation. In any case, we know from soaring GP waiting times that we’re kidding ourselves if we think the way to protect 1–2–1, in-person care is to retain it as the default model; we need to focus resources more tightly.

So overall, my sense is that, if we’re radical enough about reform, and stop distracting ourselves with fiddles and reorgs, there’s huge untapped potential to make care-intensive services more scalable. If we do this well, we can also make services more caring, not less, particularly by combining technology with more relational and community-based approaches to care.

The third implication for policy relates to unpaid care. This speaks to the fact that most care is informal, not showing up in stats like GDP or in formal care-intensive sectors like the NHS. In this respect, the currency of care isn’t just money, it’s time.

I’ve long been fascinated by the changing politics of working time. If you go back 150 years, working time wasn’t just a political issue, it was the political issue. At the first ever meeting of Britain’s Trades Union Congress, it wasn’t wages that were the first topic of discussion, it was working hours. So it feels apt that, as we move deeper into our own technological revolution, the politics of working time are heating up again.

In the last decade, the idea of a four-day working week has become only the most emblematic example of a renewed debate about work intensity and working hours. And when you think of the pressure people are under from care, plus the rising intensity of the time we spend at work, this makes sense. The more that our time is squeezed, the more precious it seems. And because women still do far more unpaid care than men, working time is also a central issue for gender equality.

Here’s one thing I find interesting about the renewed interest in working time: In the UK at least, the state is so far missing from the debate. When you look back to historic debates about working time, we made progress from a three-way push; there was pressure from civil society, especially from unions standing up for their members; there was innovation from forward-thinking businesses who saw that exhausted workers weren’t productive workers; and there was leadership from the state, as politicians encouraged new norms to form and, over time, parliaments codified these norms into law.

In recent years, many governments, including in the UK, have pursued incremental reforms to address gender inequalities in unpaid care, introducing policies like shared parental leave or the right to request flexible working hours. Only in a few countries, though, from the Netherlands to Denmark, has there been a more intentional push to reduce working hours as a way to help people care — and to get men to play their part.

If you look at long-term international trends in working hours, historic declines have stalled in countries like the UK and the US but from Denmark to the Netherlands the decline in average working hours has continued, so that workers in these countries now work far fewer hours than workers in the UK and US. Average annual working hours are now 1,349 in Germany, 1,363 in Denmark, and 1,417 in the Netherlands, versus 1,497 in the UK. This is the equivalent to 19, 17, and 9 fewer days of work a year.

Evidence suggests that in countries with lower working hours people aren’t just less stressed, and care isn’t just more equal, productivity is also the same or higher. And since these economies also tend to be more equal, disposable income for a typical family tends to be higher too; in the Netherlands, for example, median household income is $39,000 versus $35,300 in the UK. People work less and earn more. So, far from paying the price for shorter working hours, there’s reason to think shorter hours could be part of a care-centred strategy for growth.

As I’ve spent time exploring the care paradox, I’ve come to think it’s one of the oddest of the many peculiarities of life in a mature capitalist economy. With each year that passes, our technologies get ever more powerful, yet we choose to use these powers on ever more mundane things. We reward people handsomely for helping us optimise how to share videos of cats while we squeeze, sideline, and starve the most human aspects of life.

We don’t, in any true sense of the word, choose this outcome; it arises from the combination of two facts. The first is the economic fact that irreducibly human activities like care are hard to make more productive. The second is the political fact that we have good reason to socialise the cost of care, and we don’t like paying tax.

Those two facts together, one axiomatic and the other deeply ingrained, give rise to the care paradox. And this, in turn, helps to explain why life now feels the way it does. There are the queasy juxtapositions of a society in which technological abundance sits alongside poverty and indignity; there are the perennial policy puzzles, from the baffling cost of childcare to the way public services eat ever money while always being starved; and there’s the oddity that technological progress doesn’t seem to make our lives easier but is, if anything, just speeding the treadmill up.

If we keep failing to acknowledge the forces driving these outcomes, the future will be a nasty caricature of the present. It would be better to accept the forces, try to understand them, and build a 21st century state that is capable of shaping them. To do that, we’ll need to develop a policy agenda in three parts.

First, we’ll need to accept that the cost of care will keep rising as a proportion of GDP. Rather than pretending we can stop this, and doing ourselves an injury in the process, we should seek to mitigate our tendency to underinvest in care. We need economically and politically sustainable ways to spend a rising share of GDP on care, especially social care and childcare. We should spend less time pretending we can cap the size of the state and instead spend our time debating the most economic and socially sustainable ways to fund care into the future.

Second, in addition to funding, we need to reimagine how we deliver care, especially in education and health, to make full use of internet-era technologies. This won’t prevent care cost inflation or negate the need for steady increases in funding, but it can help. This entails far-reaching reform. It doesn’t mean bureaucratic reorgs or digital gimmicks, it means replacing the pre-digital mindsets and technologies that still underpin these services with mindsets fit for today’s technological age.

Third, we should resuscitate the politics of working time. A four-day working week is only one emblematic example of what a more purposive agenda on working time could look like, and long-term trends suggest it’s achievable within a generation. More widely, we should double down on policies that seek to balance out unpaid caring responsibilities between women and men.

Underlying all of this, we’ll need a cultural shift: we must value care more highly. We should measure our success not in our ability to squeeze care into ever shorter time slots but in our ability as a society to protect more time for care. We should see care in a high-technology world not as a distraction and a drag on productivity but as the last sanctuary of our most human moments. We should evaluate public policy in a way that captures care’s rich emotional value; a heartfelt conversation; an act of respect; a space for vulnerability, connection, and dignity.

As ever, those are just some thoughts and I’d love to hear responses and critiques.

This essay is part of a project exploring social justice in a digital age. The care paradox is one of three forces that I think can help us make sense of the character of contemporary capitalism and delineate what we need from a 21st century state.

I explored the first force, related to the way digital platforms are changing our lives as consumers and workers, in an earlier essay, The Invidious Hand.

The third force will be covered in a final essay, Unequality, to be published in January, exploring how we generate value today and how inequality changes in an intangible age.

In the meantime, to keep up with the work you can follow me on Medium or Substack. And for the story behind all this, from Victorian sewers to digital dragons, there’s my book, End State.

Footnotes

1. US labour productivity grew 299% from 1950 to 2018. In their paper on the Baumol’s effect, economists Eric Helland and Alex Tabarrok go back further, to the first performance of Beethoven’s String Quartet No 14, in 1826. The piece took four people 40 minutes to play, the same 2.66 person hours that it takes now, 200 years on. Average real wages meanwhile rose from $1.14 to $26.44, meaning the effective cost per performance has leapt 23-fold from $3.02 to $70.33.
2. p.16, Labours of Love, Madeleine Bunting
3. With apologies for lack of diversity, see this long treatment, focused on the US, from Eric Helland and; Alexander Tabarrok; a UK-centric analysis from Diarmid Weir; a more critical take from Scott Alexander; and this op-ed from Tyler Cowen. I would love to hear of any other sources.
4. In health, for example, on top of the Baumol effect there’s the rise in chronic conditions; our insatiable appetite for quick access to the latest medical innovations, which means innovation drives up costs rather than driving them down; and of course our ageing society. In education, meanwhile, prices are pretty dysfunctional. For one thing, there’s the power of signalling — a degree from Harvard isn’t valuable because it teaches you a lot; it’s valuable because you can say you went to Harvard. This means innovation doesn’t tend to push down prices, since new providers can’t compete with the cache of established institutions. Plus education is partly positional; to get a top job, you don’t need to be impressive; you need to be more impressive than the next person. So education becomes the cultural equivalent of a peacock tail.
5. One takeaway from these trends is that we tend to misframe our public debate about healthcare spending. We argue over the extent of healthcare spending, but really there’s no future in which we don’t spend far more than we do today. The more honest debate is over whether we want the additional spend to be public or private. For the same reason, we also tend to misframe the debate about NHS privatisation. We talk as if NHS privatisation is an untouchable political idea but, thanks to cost inflation, the NHS has an in-built self-destruct mechanism: if funding doesn’t rise fast enough, it privatises itself, replacing public with private provision.
6. Why childcare is so ridiculously expensive, Derek Thompson, The Atlantic (2019)
7. Social Care 360, The Kings Fund (2022)
8. p.12, Labours of Love, Bunting
9. A note on productivity. It would be understandable to read all of this and think I’ve missed the point. Surely today’s major problem isn’t uneven productivity growth but, rather, no productivity growth? My short answer is that we need both: an agenda working on aggregate productivity growth and an agenda on care as an aspect of productivity, since the two are interlinked. For one thing, we won’t be able to spend more on care unless we can restore decent productivity growth in progressive sectors. For another thing, a care crisis drags on productivity since (a) a population that is chronically ill, burned out, and stressed, is unlikely to be productive. And (b) care-intensive sectors are an increasingly important contributor to productivity, since one implication of the Baumol effect is that a growing share of the workforce is employed in care. So a sustainable strategy for care must be part of any strategy for productivity growth.
10. One thing that makes this all harder today is that a lot of the 20th century growth in public service spending was funded by a big one-off reduction in defence spending and, in Britain at least, cuts to capital/infrastructure spending which it would be impossible or ill-advisable to repeat.
11. It’s harder to assess the dynamic effects of these choices. For example, if we’d had lower taxes and invested less in care throughout the 20th century, would we have had higher or lower productivity growth in progressive sectors?
12. Note also that last time around the answer wasn’t to raise the horse tax or the servant tax; instead, we created new taxes that were more relevant and economically sensible, and it seems likely that we’ll need the same now too.
13. This isn’t about using frontier technologies like artificial intelligence, which take up far too much of our debate about technology-enabled reform. The average organisation — including most corporations and certainly most public service institutions — hasn’t yet made full use of technologies from 2002 or 2012, let alone 2022. So most of this is about applying the basic principles of digital delivery.