Is Care Work Infrastructure?

(And how did "infrastructure" become the sneaky way of saying "good")

Current events have overtaken our reading this month! I invited Jeff Spross to discuss Atul Gawande’s 2017 essay on “The Heroism of Incremental Care.” And it turns out it was the perfect reading for us both as the country considers various family policy bills and whether they count as “infrastructure.”

At Tiny Book Club, I’m committed to providing you something you may not find anywhere else—a demand that people (me included) define what they mean by infrastructure. It was a load-bearing discussion, so I’ve added some bolding (all mine) throughout.

Leah: Jeff, thank you for joining me this month. Gawande’s piece has been a good complement to the discussion of whether care work counts as infrastructure. And, aside from the object-level question, I’m fascinated by the way there seems to be consensus that infrastructure is good (even if we don’t consistently fund it). 

Discussing whether childcare is infrastructure seems like a way of sneaking in the claim “childcare is good” or “childcare is important” without having to talk as explicitly about values. It reminds me of how, throughout the pandemic, there’s been a certain amount of politicians punting responsibility by saying “trust the science” or “trust the scientists.” Science can tell us about risk, but deciding how to weight it is a value judgement. 

Why do you think “infrastructure” has this aura of authority?

Jeff: Leah, thanks for asking me to participate! I agree the Democrats seem not entirely confident of winning a straightforward argument that “childcare is good.” So they’re trying to import it under the umbrella of “infrastructure.” As to why “infrastructure” does inspire that confidence, I have a few roundabout thoughts.

We have an idealized, deeply individualistic narrative in America about how we organize our collective economic activity. We don’t coordinate or cooperate; we compete. And we don’t rely on government; we rely on markets and the profit motive and self-regulating price signals. But now we have all these interlayered crises in our society for which that narrative seems obviously inadequate. Childcare and eldercare are two examples. And I thought Gawande’s piece really highlighted how healthcare can be another.

The dedicated, incremental care he advocates really cries out for a society that is more generous to everyone in terms of money and time. Patients must be able to sit with their symptoms, learn them and chronicle changes, while waiting on solutions that can take months or years to reveal themselves. Doctors must be able to do the same, while getting to know a smaller number of patients as nuanced individuals with specific bodies and specific lives. All that must somehow result in a flow of revenue that is both adequate for the doctors, and reasonable for their patients to shoulder. Gawande’s stories illustrated that both providers and insurers can be indifferent to that sort of approach, as they’re focused on the mammoth revenues that can be extracted from technologically intensive, one-off interventions; but also how just everything in our economic lives conspires against that sort of care, from the demands of work schedules to how much people are paid.

So we have all these failures to address, and if individualistic competition and markets aren’t working, government is really the only other tool on hand. And to get back to your original question, my sense is that “infrastructure” is one of the few realms of economic activity where everyone mostly agrees the “markets and small government” narrative doesn’t hold; where robust intervention and direction by government is still instinctively and widely accepted as legitimate. 

Part of that is probably because any Economics 101 college course will tell you that “traditional” infrastructure – roads, bridges, pipes, wires, airports – are often circumstances of “natural monopolies” where market forces fail. (You’re not going to have multiple companies building multiple physical systems of highways or water mains, all competing to serve the same city or neighborhood.) Part of it is also the class and gender dynamics Anne-Marie Slaughter pointed to in the piece you linked: Most of the decisions about what economic activity our society will undertake are made by wealthy white men. By dint of their privilege and freedom to just throw their personal wealth at problems, such people can ignore concerns like childcare and eldercare if they wish. As you put it to me, Gawande is essentially calling for concierge care for the poor and vulnerable. But we associate concierge care with the rich precisely because the rich are the customer base that can actually support that form of care out of their own pockets. When it comes to caring for children or aging parents, if the “infrastructure” most people rely on is failing, the wealthy and privileged can escape into their own parallel, boutique versions of that “infrastructure.” But they must still drive on roads and rely on trains and cargo ships to move their product. Even Donald Trump must still fly through LaGuardia with the rest of the commoners, and suffer its inadequacies alongside them.

So physical infrastructure, by its very nature, forces a kind of collective and egalitarian mindset, in that there is simply no escaping it. I suspect all that is why, even if “infrastructure” seems a poor conceptual framework to shove all these concerns under, it’s the conceptual hill Democrats are most comfortable fighting on.

---

Leah: Health care is definitely a space where patients don’t have the power we usually attribute to “consumers.” I’m not the one paying my doctor, my insurance is. I’m (usually) not the one choosing my insurer, my employer does. I’ve loved buying insurance independently on the Obamacare exchanges, because I can pick an insurer based on the service they provide—not what package they offered my employer without my input. 

The theory is that my insurer winds up pushing for services and systems that benefit me—they save money if I’m healthy, but I wonder how much that holds up as I change jobs and insurers regularly. My present day insurer isn’t going to reap the benefit of my long-term health. 

At every level, I have no way to reward good actors or effectively punish bad ones.

Jeff: Gawande’s piece also left me pondering what sort of broader health care system would best provide that kind of dedicated, incremental care to everyone. I agree, “shopping” for health care is certainly not like shopping for shoes or coffee makers. And in experiments where researchers test how consumers handle buying health care directly, without the insurance middle man, the results are often uninspiring. Health care is fraught with undercurrents of morality and mortality; and as you say, the benefits are often long-term, which is a problem for you the consumer as much as your insurer. How do you judge doctors, treatments, or providers if the results come years or decades down the line?

Ideally, yes, insurers would act as agents on our behalf, helping us navigate amongst care options. But that raises a classic principal-agent problem, since insurers can compete on who finds the best care and the best value, or they can “compete” on who best avoids the customers posing the greatest insurance risk. (And once employers are buying insurance on our behalf, you’ve layered on an additional principal-agent problem!) An embarrassing factoid is that preventative care doesn’t usually save insurers or government much money: what you don’t spend on a small number of expensive interventions is washed out by what you do spend on everyone’s prevention, or by what you spend on routine care simply because everyone lives longer. When Gawande cites all those studies to back up his case for incremental care, the good outcomes have to do with people’s quality of life rather than with cost savings. So it’s not clear to me that the market framework of efficiencies and profits actually captures the values we’re trying to achieve in health care all that well.

Yet in theory all these things should be fixable while maintaining a basic market structure. You regulate the insurers to solve the principal-agent problem, such as Obamacare’s rule that you can’t deny someone coverage or overcharge them for a pre-existing condition. I suppose you could even use regulation to require all insurers cover Gawande’s incremental care, though a lot of devils may be in those details. Then you provide aid so everyone can afford the higher premiums that result when you stop insurers from avoiding risk. (Obamacare’s aid was always much too low, and it’ll be interesting to watch now that the American Rescue Plan has significantly—albeit temporarily—increased it.)

But that’s all also an extremely complex policymaking endeavor, with a lot of things you have to get just right to drive providers and for-profit insurers in the right behavioral directions. And Obamacare was in many ways poorly designed and compromised. Switzerland is arguably a country that’s done a much better job at this approach; basically Obamacare on steroids. Yet even the Swiss still rely on big, direct government coordination like all-payer rate setting.

So I think it’s pretty reasonable to conclude that a genuinely market-based health care system is a kind of utopian dream, forever over the next horizon, while something like Medicare-for-All is the more practical and (ironically) small-c conservative approach: A setup where we can just decide as a democratic polity that, say, Gawande’s incremental care is something we want to provide universally, without having to worry about cajoling or corralling a bunch of private firms in the market. But then, of course, the quality of your governance institutions, and especially the quality of your local democracy, becomes crucial. If the general practitioner for a particular neighborhood isn’t working out, can the neighbors get together, vote the doctor out, and get a replacement? If they need an additional doctor, can they make that happen? Does the system respond rapidly and effectively to requests? Ultimately, socialist planning and capitalist markets are two different institutional toolkits for solving the same problems: Not just coordinating resources, but enforcing accountability for how resources are used, and promptly recognizing and correcting mistakes. And I suspect we’d have much healthier and constructive politics if we just went in with the assumption that neither toolkit is “right,” but that both are better suited to some forms of economic activity than others.

---

Leah: I appreciate your definition of infrastructure above. I haven’t really thought of it in terms of natural monopolies (childcare doesn’t fit that definition!). If I had to define the fuzzy category we’re talking about, I might wave my hand in the direction of “technical debt.”

That’s a programming concept about the long term costs you rack up when you take cheap shortcuts now. It might mean people not adding comments to their code, using hard to parse variable names, or relying on a kludge that can't take a heavier load. You don't have to have debt, it comes from being rushed or by not taking the long view that you're building something people will rely on for some time. Part of how I answer “is it infrastructure” is by thinking “is it the opposite of technical debt?” Which I’ll admit is not a standard way of putting it!

Infrastructure is an investment in the future, particularly when we, like the designer of the London sewer system, plan for the needs of generations to come, not just our own. It’s adding a third tube from the start to your subway system, so you can do maintenance without shutting down regular service (one of the many problems hamstringing DC). 

And when we push to fund it collectively, it’s partially because we think that investment will pay off in a huge way—but the people who will benefit eventually can’t float the cost now. They won’t have a surplus until after we invest. That’s why we do it together, through the government, rather than just through private effort.

Jeff: I like the analogy of “technical debt” to infrastructure: that basically infrastructure is whatever process we’ve designed sufficiently well that it will run reliably in the substructure of our lives and communities without us having to correct it or worry about it much. No, it’s not the traditional definition. But I bet it melds pretty well with a lot of non-wonky Americans’ instincts.

Childcare definitely isn’t a natural monopoly, nor would I say it has the same sort of peculiarities as health care that don’t play well with markets. Yet it also clearly isn’t working for most people. And we have to worry about it constantly! I get where someone like Brad Wilcox is coming from when he argues that government-provided childcare is social engineering that pushes people towards a particular family model, so we should just give everyone cash instead. But I think that neglects how much the crisis of childcare is shaped by the particular nature of our particular economy.

The same places where people can get good jobs are the places where costs-of-living are skyrocketing, which feeds back into more expensive childcare. Meanwhile, the places where costs-of-living are cheap also tend to be places where few if any good jobs are to be had. It’s a nasty paradox: where you can get good employment is where you’ll pay through the nose for childcare, and where childcare is affordable it doesn’t really matter, because work opportunities are poor. So inequality, at both the individual but especially the regional level, is itself a form of “technical debt.” As is our long failure to reach and sustain anything like full employment, and our failure to make sure good jobs are abundant across all geographies, and the decades of stagnating incomes that have forced more and more people into the two-working-parents model. If we lived in a world where we’d addressed all that “technical debt,” childcare probably could be solved by just giving everyone a child allowance and letting the market sort out the rest. But in this world, universal provision by the government becomes attractive as a way to just cut through the morass.

Of course, I suppose the “technical debt” definition of infrastructure falls into the trap of, well, anything could be infrastructure then! In which case nothing is infrastructure. Personally, I’m fine with defining infrastructure as “anything that enables people to engage in broader economic activity,” but that has the same problem: it’s not just roads and bridges and ports, it’s childcare and eldercare and healthcare, and it’s food and clothes and really anything that isn’t clearly “entertainment.” I’d also add that markets can be very finicky things, insisting on just the right set of circumstances to work properly. So the problem of private actors neglecting long-term, future-oriented investment doesn’t just happen in the context of natural monopolies, but rather can happen across a range of sectors whenever the world takes an unexpected turn.

Which all just kicks us back to the original question of, why do we care so much what is and isn’t “infrastructure?” Why does that have to be the only category of problem we’re willing to seriously commit state capacity and the public coffers to solving? I could rant about the silliness of deficit hawkery all day, but a lot of this seems to be driven by the unspoken dream that politics just has to address a few discrete areas of life, and the rest everyone can sort out on their own. But whether that vision is actually possible to achieve, the country we’re in now – and the many path dependencies it’s now built up – really isn’t cooperating.