Student debt is again an urgent topic of debate in Washington. President-elect Joe Biden has proposed that Congress pass a bill canceling $10,000 in federal student debt per borrower, although many Democrats want him to be more aggressive: Some want him to cancel $50,000 in debt per borrower, through an executive order. Others would go even further: Biden “must cancel all federal student debt on Day 1 of his administration,” Rep. Ayanna Pressley (D-Mass.) has declared.

Given the stakes — both for students and the budget — it is important for policymakers and the general public to understand the effects of different forgiveness options. Many economists have already pointed out that loan forgiveness would disproportionately help people on the upper end of the earning scale: Two scholars with the Brookings Institution, for instance, note that 60 percent of educational debt is owed by households in the top 40 percent of earners — those, that is, with an annual income of $74,000 or more. This skew is hardly surprising, as many Americans do not go to college and only a fraction go to graduate or professional schools. And those who do go to college earn substantially more, in part thanks to the degree: Over their lifetime, the median couple with two bachelor’s degrees earns $1.8 million more than the median couple with two high school diplomas.

But our research finds that loan forgiveness would benefit upper-income people even more than is generally understood. That’s partly because few realize that a large loan forgiveness program for low-income borrowers already exists: Many low-income borrowers are already enrolled in income-driven repayment plans that not only peg payments to wage levels but forgive loans after some period, typically 20 or 25 years.

To get a better sense of who benefits from loan forgiveness, we analyzed data from a 2019 survey of consumer finances conducted by the Federal Reserve Board of Governors, including 5,777 households. Using the “present value” of loans (which lowers the figure if part of the loan is eventually likely to be forgiven) changes the picture, we found: If a comprehensive loan-forgiveness program were passed, we calculate that the average person in the top 10 percent of earners would receive $5,944 in forgiveness, while the average individual in the bottom 10 percent of earners would receive $1,070.

That average includes both borrowers and non-borrowers, as well as college dropouts. (The average loan size for the 845 borrowing households in the sample is $41,400, with wealthier families owing more.)

Overall, households in the top 30 percent of the earnings distribution would receive almost half of all the dollars forgiven. To put it even more starkly: Erasing all federal student debts would benefit the top 10 percent of earners just as much as the bottom 30 percent combined. Those calculations are for wiping away all debt, but forgiveness is even more regressive if it is limited to $10,000 or $50,000.

It’s important to understand the distributional consequences of a $1.6 trillion debt jubilee before a decision is made in part because erasing college debt won’t reduce the average debt of Americans. Rather, it will transfer private liabilities to the general public by increasing the federal debt. The long-run consequence will be an increase in taxes or a decrease in public spending. The redistributional effects will be immediate.

Even though income-driven repayment plays are underutilized — some borrowers don’t know about them — roughly half of people with student debt are in such plans. A typical IDR program requires that a borrower pay 15 percent of any “discretionary income” — defined as income above 150 percent of the poverty line. After 20 or 25 years, the balance is forgiven, regardless of the amount due.

To see how this plays out, compare two theoretical borrowers in an income-driven repayment plan: “Lara” and “Tonya.” A single mom living with two kids, Lara earns $50,000. Under the most recent IDR rules, her payments — regardless of loan size — would not exceed $1,742 per year. Assuming that interest accrues at 4 percent, and that she is 15 years away from forgiveness, Lara will fully repay her loan only if her balance is at or below $19,400. After that, any remaining debt will be forgiven; it makes no difference whether she borrowed $20,000 or $50,000.

As a result, Lara doesn’t benefit much from the $10,000 in loan forgiveness that Biden proposes. If she happens to owe $28,400, and $10,000 is forgiven, thereby dropping her balance to $18,400, she will save $1,000. If she owes more than $29,400, she won’t benefit from debt forgiveness at all — she’ll end up paying the same proportion of her income for 15 years.

Meanwhile, Tonya, a married mother of one, owes $50,000 and earns twice as much as Lara, $100,000. Tonya would pay $6,742 annually, under an income-driven repayment plan. Because her higher income lets her pay off her entire loan, she would benefit from every dollar of forgiveness the government provides.

Too much of the debate over student-loan forgiveness has focused on reducing the nominal balances without considering whether they are on track to be repaid. Our analysis suggests some policy alternatives: Nudging more people into income-driven repayment plans would achieve more forgiveness for low-income borrowers, for instance. So would making those plans more generous.

We looked at what would happen if every borrower who would benefit from an income-driven repayment plan were put into one — specifically, a plan that required a 10 percent payment for all income above 150 percent of the poverty line, and forgave loans after 20 years (the terms of the more generous current plans). In that case, people in the bottom decile of earners would get four times the amount of debt forgiveness as those in the top decile. The bottom half of earners would get more than half the gains.

The middle of the income curve would also get more help than under universal forgiveness: People from the 40th to 80th percentile would get nearly 70 percent of the benefits.

Making IDR plans more generous could also lead to more forgiveness for low earners. We calculate what would happen if the IDR threshold were increased from 150 percent to 300 percent of the poverty line. (In other words, no repayment until someone made three times poverty wages.) Such a policy would benefit the bottom third of borrowers more than would blanket forgiveness of $50,000 in debt, we found. Yet, the overall cost to taxpayers would be nearly three times smaller because it would barely reduce the payment of high earners.

The current plans to erase $10,000 or $50,000 or more for every student borrower are highly regressive: They mainly help the better-off. Of course, that’s not the only factor to consider in evaluating such programs. Student debt might warp career choices, or dampen entrepreneurship, for example. More research is necessary on these and other aspects of the issue. But surely policies that benefit high-income families more than low-income families deserve careful analysis.