AI can lead to a fix of this broken government program

America’s unemployment insurance system was designed for the workers of the 1920s, has been growing financially weaker since the 1970s, and has become less generous since the 1980s. It was so utterly inadequate during the pandemic — the most sudden and severe job loss the U.S. has ever seen — that Congress had to build a whole new program even as it was propping up the old one.

Now Americans are afraid of the impact AI will have on their jobs and careers, and a lack of leadership is feeding that fear. If government and elected officials want to be responsive to their constituents, they need to address this anxiety head on. The most direct way to do that is to build a new program to help workers who are looking for a job.

It shouldn’t be that hard.

The current program has deep flaws that offer some straightforward lessons to incorporate in the new version. One obvious problem is that the current system is actually a confederation of 53 state and territory programs, with no benefit or tax coordination among them. It is difficult to guarantee a minimum level of security to all Americans if it comes with the caveat, "If you lose your job, don’t live in North Carolina."

The next unemployment system, like America’s economy, should be national.

Another easy fix concerns the way the system is paid for. Unemployment insurance benefits are funded through a payroll tax paid by the employer, which means workers’ benefits are a function of their past wages, not their current income. This is a very good thing — you don’t have to prove you’re poor to get unemployment, you just had to have lost a job.

Very good, but could be better still. There are a lot of workers in the U.S. who are not officially on the payroll, and thus ineligible for unemployment benefits. The system should include them, which in turn would establish a broader base of funding. And if the payroll tax were expanded so both the employee and employer paid it, as the case with Social Security, it would not only broaden the base of funding, but promote worker knowledge of the program, something the current version struggles with.

From there, things get trickier. In the current system, a worker needs the permission of their former employer to receive a benefit — the employer must verify that the job loss was through no fault of the employee. People who quit or are fired are not eligible. Should this carry over to the next system?

Any benefit that gives money to a person will have a work disincentive. If workers are too comfortable in unemployment, they’ll stay unemployed longer, which could lead to lower levels of employment. By this logic, unemployment benefits should err on the side of parsimony.

Or not. For one thing, workers receiving unemployment are less likely to grow discouraged in their search and drop out of the labor force. Unemployment benefits can’t get someone a job, but they can keep people looking. Giving up on search leads those discouraged workers to means-tested benefits or into disability programs, which cost even more money. In fact, studies show that the use of unemployment benefits decreases the use of welfare benefits, while more generous unemployment benefits decrease the use of food stamps in the short term and disability in the long term. Researchers at the Federal Reserve Bank of San Francisco examined the effects of the very generous pandemic unemployment benefits and found that even $600 extra a week did not discourage job-taking; the value of a time-limited unemployment benefit is tiny when compared to a permanent job.

Put another way: The pandemic recession holds the distinction of being the downturn with the largest job loss, the most generous unemployment benefits — and the fastest labor market recovery. Notably, employer attestation of no-fault job loss was waived. To the extent that one can view the pandemic as an experiment to let anybody claim unemployment, it’s hard to argue the economy’s recovery was worse off for it.

Functionally, forgoing employer permission simply means giving workers who quit or were fired an unemployment benefit. Normatively, it means giving up trying to distinguish between good unemployed workers, who are the victims of hard economic luck, and bad unemployed workers, who are lazy and unproductive. Far be it from me to pass judgment on those who pass judgment — but I would point out that, from an economic standpoint, there are more useful and productive distinctions to make.

Distinctions like, reemployment prospects. Consider the varying duration of unemployment. Most unemployment spells are short, under 10 weeks. Some are longer, closer to six or nine months. And some approach a year, or even longer. The upshot: Most workers get reemployed quickly, some take longer but are still successful, and a small share will struggle to find a job. The current program, however, gives a flat benefit for 26 weeks (in most states).

Rather than focus on what mechanism sent workers into unemployment, or whether they are at fault, this new system would focus on how quickly they are likely to get out of it. The program can be designed to triage the various kinds of unemployed workers, moving unemployed workers through successive tiers of benefits designed to match their spell of unemployment.

—In Tier 1, they get generous benefits for a short period, say six to eight weeks. This is basically a "no questions asked" benefit — you show up, say you’re unemployed, and get cash.

—In Tier 2, they get less generous benefits for a longer period, maybe up to six months. They have meetings with an employment counselor to get labor market information and search advice.

—In Tier 3, which kicks in after six months, they enter a long-term unemployment program that mixes cash benefits with training or education programs.

The length of these tiers can vary with labor market conditions, both macro and individual. Reemployment is harder when the unemployment rate is high, but it can also be harder for older or more tenured workers. A 30-year-old who spent three years with a company is going to have a different unemployment experience than a 50-year-old who was there for more than a decade.

The design can be more creative still. It’s not clear that workers gain much from getting a weekly benefit rather a lump sum, for example. After all, it’s not like mortgages and utility bills have to paid weekly. Nor does tuition, for training classes or a degree. Nor do moving costs, if someone wants to go to a new city to look for work.

In sum: A new federal unemployment system would have a broader tax base, easier access to benefits, and be designed so those benefits reflect what we know about reemployment.

Building this new system will no doubt reveal growing pains — but as long as we are committed to maintaining a strong unemployment program, it can be successful. And they’re relatively achievable goals that will allow us to focus on the real challenge and the scarier prospect, that final phase of triage: long-term unemployment. We know how to design a system to help workers find a job. It’s much harder to design a system to help workers who don’t find one. That comes next.

This guest essay reflects the views of Kathryn Anne Edwards, a labor economist, independent policy consultant and co-host of the Optimist Economy podcast. She wrote this for Bloomberg.