Three decades ago, few economists would speak against free trade or automation. Those who raised concerns about job displacements were met by replies that price reductions would compensate in the short run while higher incomes elsewhere in the economy meant new (and better) opportunities for everyone in the medium-run.
This is no longer the case. Many economists, most notably Nobel laureate Angus Deaton, have soured on free trade and automation. They argue that “big shocks,” like the rapid expansion of China (and its significant role in international trade) or the arrival of industrial robots, have heavily affected a large subset of the population — a subset whose incomes were below or at the median. They further argue that the effect was so large that their children now face fewer opportunities, resulting in less income mobility across generations. Those born to parents in the bottom 10 percent or 20 percent of the population before these “shocks” are increasingly locked into their socio-economic status.
In other words, “big shocks” hurt income mobility. The extra fear is that, by making society appear unfair, the reduction of income mobility will cause democratic erosion and a withering of the liberal democratic order.
These concerns should not be swept aside carelessly. They are genuinely important.
The problem is that they always consider these “big shocks” in a form of institutional vacuum. Regulatory burdens, the security of property rights or tax rates are rarely discussed. Even less frequently mentioned is the possibility that the adverse effects from the “big shocks” like automation or trade liberalization may be conditional on having a particular set of institutions.
Consider an example involving automation which is expected to adversely impact workers who are a substitute to industrial robots. Imagine that there are two islands that experience automation. In the first, labor markets are heavily regulated with stringent hiring and firing laws, high minimum wages, widespread closed-shop unions, costly occupational licensing and there are high tax rates on labor income. The second island has none of these features. On which island would you expect it would be harder for people to adjust to the shock created by the arrival of industrial robots? The answer, obviously, is the first one. The heavy hand of the government can rigidify markets and make it harder for people to adjust to the unexpected. In that case, it would be unsurprising to see that workers displaced by robots will be left worse off for so long that their children will be affected as well.
In a recent working paper co-authored with Pradyot Sharma and Alicia Plemmons, I explored whether regions exposed to industrial automation, resembling the first hypothetical island described earlier, experienced greater challenges in income mobility compared to those resembling the second hypothetical island. We used a dataset of intergenerational income mobility for children born in the 1980s for 600 “commuting zones” in the United States that combined with another dataset that measured the degree of exposure of each of these zones to industrial automation during the period. Then, we focused on one particularly egregious labor market regulation: occupational licensing.
Occupational licensing — the costly requirement of a license to be engaged in a particular profession — has grown massively in recent decades. Many of the new regulations fall on low- and medium-income professions.
If the “big shock” of industrial automation affected the poor more, it is reasonable to consider how occupational licensing blocked them from occupying jobs that were closer comparisons to their previous ones. If states are more aggressive in increasing occupational licensing on low- and medium-income professions, then they are more like the first imaginary island described above. Those that regulate less (or even move in favor of deregulation) are more like the second imaginary island.
Our findings reveal that states with less stringent occupational licensing regulations were able to mitigate 49 percent to 85 percent of the adverse effects associated with greater exposure to industrial automation. While these lesser-regulating states did experience some negative impacts, they were significantly less severe than those in states with heavier regulatory burdens.
It’s important to note that our focus here is on a single policy area — occupational licensing. This is a relatively narrow scope, and it’s plausible that incorporating policies that promote entrepreneurship, reduce taxes to boost labor demand, or deregulate in ways that lower the prices of goods disproportionately consumed by the poor would amplify these mitigating effects. Maybe even reverse them entirely!
While we must remain concerned about how major disruptions might affect our societies in the long term, particularly regarding intergenerational income mobility, it’s crucial to recognize that such concerns may be exacerbated by governments inadvertently turning these “big changes” into actual problems through prior policy missteps.
