Industrial Policy Stifles Progress
When Congress picks winners, we all lose.
The once-beleaguered CHIPS Act has finally passed and will soon receive President Joe Biden's enthusiastic signature. The big ticket item in that legislation is $52 billion worth of subsidies for computer chip manufacturers, but once the bill's passage looked inevitable, it was stuffed full of additional spending. The CHIPS and Science Act's cost has now ballooned to $280 billion. And emboldened Democrats have already moved on to another spending spree with the Inflation Reduction Act, a slimmed-down version of Biden's "build back better" initiative.
Both bills reflect a cross-party shift toward embracing industrial policy—the idea that the government should jump into the economy with both feet and have fun getting wet. Facetiousness aside, the neoliberal era from the late 1970s through the 1990s—when economic thinking carried more political sway and resulted in massive deregulation of airlines, railroads, and interstate trucking and the privatization of the internet—is far behind us.
This change can be seen in former President Donald Trump's embrace of tariffs, Biden's continuation of many of those tariffs, and officials' apparent reluctance toward even the minimal deregulation which would have quickly solved the ongoing infant formula crisis. But we don't need to infer anything. Consider this recent quote from a senior administration official: "Part of our effort is to create space again for very serious people to really go to bat for the idea that the government has a rightful role to play" when it comes to industrial development. Yikes.
The political zeitgeist has been moving back toward industrial policy, seemingly coincident with rising populism since the 1990s—despite abundant evidence that central planning is poisonous to innovation. Whether it's encouragement via subsidies or constraint via regulation, using the government to guide the economy is akin to thinking that just a little bit of cyanide won't hurt.
To armchair economists, industrial policy seems like a solution for the country's economic woes: "Infuse money into Industry A, add trade protections for Industry B, protect workers in Industry C from automation, and the economy will soar! New technology will arrive sooner, domestic firms will outcompete foreigners, and steady employment will ensure a chicken in every pot." That indeed was the thinking behind Depression-era policies which extended that crisis by seven years.
Economies are not deterministic like physics or chemistry. You can't pull a lever to achieve a particular effect. A better analog is biological or ecological systems, where there are second- and third-order effects to any given stimulus.
Think about the reintroduction of wolves to Yellowstone National Park: Increased predatory pressure keeps elk herds on the move, leaving more young willow trees for beavers. Growing beaver populations dam more waterways, altering the habitat and spurring additional difficult-to-predict effects. That's economic policy: You must plan for unexpected downstream effects (pun intended).
That thinking has been missing in Congress this past month. I don't know what microchip subsidies or a mistitled inflation-fighting bill will ultimately do, but neither do our elected officials.
Compounding the problem is that people, not some agnostic supercomputer, determine which industries and companies are considered worthy of a boost. Humans are subject to influence and pressure, turning industrial policy into a contest of who can secure the most government favoritism—a political game of Hungry Hungry Hippos.
Policies protecting companies from competitive pressure, like subsidies or tariffs, allow them to take their eye off the ball. This "X-inefficiency" means they're less efficient and pay less attention to customers' desires.
Intel, the definitive winner of CHIPS Act subsidies, understands this all too well. After decades as the market leader of commercial computer chips, a series of fumbles and failure to focus on customers caused the company to lose its edge over Advanced Micro Devices (AMD) and Apple, which were working hard to design a better mousetrap.
X-inefficiency leads to "dynamic inefficiency"—the lack of motivation to adapt to changing market conditions. The result is reduced innovation, slowed economic development, and increased vulnerability to socioeconomic shocks.
Given enough time, the politics surrounding the CHIPS Act—and politicians' inability to reverse mistakes—will almost surely lock some companies into less effective decisions. Intel has doubled down on manufacturing—understandable if it's getting subsidies for it—while AMD and Apple have seen success by focusing on design and outsourcing production. We'll see which approach works best, but the political baggage will make it harder for Intel to adapt if needed.
Lastly, industrial policy motivates "unproductive entrepreneurship." Some of the best and brightest minds inevitably withdraw from productive activities premised on voluntary exchange, and instead use their skills to find autocratic mechanisms to extract political payoffs—the entrepreneurial version of the dark side. Their skill grows with experience, meaning the effect increases the longer this continues.
The government can encourage particular results during times of crisis—as it did with Operation Warp Speed—but any such policy should focus on allowing markets to function and removing obstacles rather than trying to predict the future. The crystal balls policy makers peer into are easily clouded by charlatans, and we all lose when they win.