Can The U.S. Afford A Massive Wage Subsidy? A Macroeconomic Simulation
One of the most popular policy ideas to reduce rising inequality and automated job loss is the expansion of the government's current wage subsidy program, the Earned Income Tax Credit. I recently advised one member of Congress, Ro Khanna, on a bill to massively increase the EITC, and increase its payout to $3,000-6,000 per year at a cost of $1.4T over a decade.
But, how would the government afford such a large expansion? And, how much would this method of payment hurt economic or job growth?
To answer these questions, I* recently commissioned a macroeconomic simulation from the Open Research Group, which has built a new open-source platform for modeling the potential effects of large-scale economic reforms.
The full report is here and I’ll note highlights below:
A $1.4T wage subsidy can be paid for in two ways. One way is to increase the top two marginal tax rates to 63% and 57% respectively. The other is to defund other federal programs and modify tax laws (such as decreasing military spending and ending the home mortgage tax deduction).
Under either payment method, the impact to economic growth is moderate, but not large. In total, over a decade, economic growth (GDP) would decrease by about 2% (total, not per year), as more resources are put into labor rather than capital investment (among other reasons). Economic growth takes a bit of a hit, but it also significantly increases equality and savings accounts.
The lower-middle and middle-class would see a significant bump in income. In the figure below, about 1/3rd of tax filers in the 3rd income quintile (somewhere between $35-65k a year) would see a boost of between $500 to $1,500.
Generally, the results from the simulation are in line with results from a recent New York City randomized field experiment of an EITC expansion, Paycheck Plus, which found that wage subsidies significantly reduce the negative aspects of poverty and, in some cases, boost employment. Paycheck Plus offered up to $2,000, but the average recipient never ended up making enough in income to receive the full amount.
This is good news for our simulation: no model could ever be more accurate than a real-life experiment like the one in New York. But, we show that if something like Paycheck Plus (or larger) was expanded nationally, the US economy could likely absorb it without a significant negative downtown.
I think this simulation is significant because this kind of “dynamic scoring” has only recently been used by policymakers to consider tax reform; until recently, the government would just do a “static score” of what it would cost the state and ignored all the changes in productivity, saving, and investment that happens as a result of reform. When the federal government began to consider dynamic scores, it was for a major bill (like the 2018 tax reform bill) and not something like a change in welfare.
Dynamic scoring has previously been very expensive and often utilized closed-source software that was not available for public scrutiny. OpenRG’s new platform, run by the University of Chicago’s Richard Evans, is available to the public.
To my knowledge, this is both the first time someone has forecasted the implications of large-scale cash welfare and made the model available to the public on an open platform.
And, the model itself shows that while there are significant tradeoffs to pay for a cash welfare policy, it is affordable.