The Washington Post
Why Chips and child care are a poor policy mix
We Americans love a good twofer. We hunger for a free lunch. Two (or three) birds are always better than one, and so on. Unfortunately, this approach has infected our industrial policymaking, too, sticking taxpayers with lots of big, expensive, larded-up programs that try to achieve everything, a little bit. However noble the aims of President Biden’s economic agenda, virtually every ambitious program gets saddled with too many other unrelated objectives to do any of them well.
The latest example involves a (well-intentioned but ham-handed) attempt to bundle computer chips with child care.
The bipartisan Chips Act appropriated $52 billion in subsidies to the U.S. semiconductor industry. It’s one of several planks of the country’s turn back toward “industrial policy,” the idea that government should exert more influence in deciding which sectors grow. This particular law supposedly has a narrow national and economic security objective: making the United States less reliant on other countries — particularly China — for key inputs necessary to produce cars, computers, phones and other everyday technologies. This need became particularly clear during the pandemic, when the disruption to global semiconductor supply chains left downstream U.S. manufacturers struggling to operate.
This strategy has hit some snags, however.
There’s a shortage of skilled construction workers to build complex semiconductor factories (called “fabs”), as well as all the other infrastructure and energy projects that Congress has simultaneously funded in its industrial-policy binge.
For example, the Columbus, Ohio, area alone is the planned home for two new fabs and a battery plant. These yet-to-be-built projects will require around 10,000 construction workers.
Sounds great, right? Biden thinks so. It’s a sort of triple dividend — access to chips, batteries and more middle-class jobs. Except the country is already running at close to full employment, and there aren’t enough available skilled construction workers in all of Ohio to fill these positions. Other nearby states need workers for their own projects. And there are scant plans to scale up the construction labor force (such as, say, through immigration) anytime soon.
Semiconductor projects underway elsewhere also face serious logistical and budgetary challenges.
In a recent earnings call, the company building two new fabs in Phoenix said construction costs alone were four to five times (!) more expensive in Arizona than in its native Taiwan, not only because of labor costs but also the permit process and other obstacles. These are common complaints in the United States; it’s more expensive to build almost anything here than it is in other countries, even those with robust labor and environmental protections.
All of which is to say that the U.S. government should be working to get costs down, or otherwise find ways to get more bang for the taxpayer’s buck. If U.S. policymakers want to rapidly expand semiconductor manufacturing — or improve our infrastructure, or accelerate the clean-energy transition — that requires incentivizing leaner, more efficient operations, not creating make-work projects.
Instead, U.S. officials have attached additional expensive conditions and objectives to these and other initiatives.
This week, the Commerce Department announced a slew of new requirements for those applying for Chips Act funds. Some relate to actually building quality products. Others cram in unrelated social goals. For example, larger Chips Act grant recipients must guarantee child care for all workers who build or operate their plants.
I’m very much in favor of expanding access to child care, via a strategic, comprehensive, well-thought out plan. But piggybacking rushed child-care initiatives onto unrelated semiconductor manufacturing objectives is likely to diminish our ability to do either thing well.
After all, chip manufacturers and construction firms probably have a good sense of what kinds of compensation and benefits are most effective at attracting scarce workers; if spending an additional dollar on child care rather than salary attracted hires, they’d provide it.
We should be optimizing for the objective we’re supposedly pursuing. I’d rather see a well-designed chips program, and a separate well-designed child-care program (ideally one that doesn’t lock workers to a specific employer to maintain access to care). But instead — presumably because it’s so hard to get any given bill through Congress — virtually all these industrial policy programs take a similarly Christmas-treed, all-of-the-above approach. Biden’s marquee infrastructure bill and climate law have also been weighed down with expensive requirements — in this case “Buy America” and other protectionist provisions that mean any given amount of money will buy taxpayers fewer bridges, broadband or electric cars.
There’s already some skepticism about the wisdom of the U.S. pivot on industrial policy. This is a chance for proponents to prove their case — and they appear to be blowing it.