Last week, UVA professor Sarah Turner posted a piece at Brookings on the regressive nature of the federal student loan payment pause titled “Student loan pause has benefitted affluent borrowers the most, others may struggle when payments resume.” It’s a worthwhile read, and focuses on the source of the bulk of the forgiveness that I have tracked in AEI’s Student Debt Forgiveness Tracker .

At the outset of the pandemic, then-President Trump suspended payments and interest accrual on federal student loans. This has become known in some quarters as the “pause.” Congress extended the pause through September 2020. Trump subsequently extended it twice through the end of his term, and President Biden has extended the pause six additional times, and it will run through the summer. No student loan payments have been required in 37 months, and counting.

The payment pause is expensive, though some of my close friends on Twitter have derisively argued this is not really a cost—just a suspension. However, the U.S. Treasury is forgoing billions each month that would have been otherwise accrued or paid. That is a cost, and I include it in my estimates of forgiveness in the tracker because it is what would have been owed, but will never be collected. That cost has been forgiven.

Turner is correct that the pause benefitted more affluent borrowers. Her article benchmarks those costs at about $5 billion per month, a similar measure used by the Committee for a Responsible Federal Budget (CRFB). I have different dynamic estimates of the pause cost based on quarterly government reports on the student loan portfolio, and find the monthly cost is growing with the size of the portfolio, from about $5.29 billion at the outset of the pause to $5.91 billion according to the most recent data. I prefer my estimates because a $600 million differential eventually amounts to real money. All told through this month, the pause will have cost over $213 billion, a nearly incomprehensible sum.

Turner adroitly explains that the pause benefits higher income households, and uses a simple graph to tell the tale. As she describes the figure: “it shows estimates of the distribution of households with student loan debt (the dark blue bars) and the value of total student loan payments (the light blue bars) by decile of family income from the year before payments were paused.” Since wealthier deciles pay more per month, and have more debt whose interest is forgiven, their benefit is greater than lower deciles.

While Turner notes that higher-income Americans enjoy greater benefits from the pause, she never estimates the size of the differential. I used data on student debt held by income decile from a post on the Liberty Street Economics Blog by the Federal Reserve Bank of New York to estimate the relative share of the benefits of the pause accruing across income categories. My estimates have some imprecision, but they are representative of the real differences in benefits.

The chart above shows the differences in benefits by thirds of income, and a nearly $22 billion differential between the lowest and highest third by income. Put another way, the highest third of American by income received 33 percent more relief from the pause than the lowest third by income. The middle third received 20 percent more than the lowest third. Intentions can be debated, but the regressive result is clear.

This brings me back to the argument that the pause is not a cost. That’s fine to argue—even though I think it’s mistaken—but it does not get around the fact that the benefits from the pause save higher income Americans far more than their lower income peers. Whatever you call it, the real question is: Can you support a policy that is this regressive three years after the pandemic started?

This article originally appeared in the AEIdeas blog and is reprinted with kind permission from the American Enterprise Institute.