Economic Impacts of Student Loan Forgiveness
The Biden administration has announced the long-discussed partial student loan forgiveness. Briefly, the program intends to forgive $10,000 of student debt for people who earned less than $125,000 in 2020 or 2021.
A significant debate arose around this topic covering a wide range of issues, including the fairness of the program, the impact this program will have on inequality, and whether the forgiven amount is sufficient. Here, I will focus mainly on the impact this debt forgiveness will have on the eligible people, along with what macroeconomic consequences it may have. Indirectly, this will also contribute to the discussion on the ‘fairness’ of this policy.
As the student loan forgiveness program currently appears to be a one-off policy (a one time reduction in student debt balances), rather than a wider change to college-funding or college costs1, the impact of student loan forgiveness will be primarily restricted to the current debt holders that will see relief in the coming months. Thus, the question of what is the overall impact of student debt on the economy would need to be addressed in a separate post. Focusing on this post, the issue of quantifying the impact to the individual and to the wider economy of this student loan forgiveness problem is not an easy question due to the fact that student debt is unique relative to other debt. Student debt is not collateralized like auto-loans or mortgages; it is not based on current income like personal loans and is not closely controlled (credit score requirements, credit limits, high interest rates) like credit-card debt.
One recent study that has given an estimate at the potential impact that could be expected from student debt relief is a working paper by Di Maggio, Kalda, and Yao: “Second Chance: Life Without Student Debt”. In this paper, the authors use a unique set of circumstances that resulted in what can be called a natural experiment2: National Collegiate Student Loan Trusts held around 800,000 private student loans, but as they were not able to prove the chain of title, and thus could not demonstrate that these loans existed, some of the loans were legally discharged. This created a natural experiment - certain debt holders had their debts randomly cleared due to clerical issues by the Trusts (“Experiment Group”), while others had their debts maintained (“Control Group”). As who had their debts cleared was random (the Trust originally bought these loans from the market, not individuals), any changes in outcomes of the two groups can be attributed to this debt discharge.
The results show that the impact of this debt discharge is significant - the student loan discharge of $7,400 resulted in approximately $4,600 lower non-student debt balances, a $3,000 increase in income and a $1,000 decline in delinquency amounts. The impact of the debt relief is greater (altogether $8,600) than the debt relief amount ($7,400), which is referred to as a multiplier greater than 1 (i.e. $1 of debt relief leads to a greater than $1 benefit to the economy). The results from this paper generally match what other researchers (for example: Rothstein and Rouse, Goodman, Isen and Yannelis) have found on the topic of student debt.
Firstly, the debt discharge led to an overall reduction in indebtedness. The authors showed that the group impacted by debt relief reduced their credit card debt and also originated smaller mortgages. This suggests that student debt relief results in deleveraging (approximately 60% of the discharged amount was used to deleverage), better financial management and increased financial resilience in case of adverse events. Similar behaviors have been shown in other instances of debt reduction (Bellon, Cookson, Gilje and Heimer). Secondly, the observed $3,000 increase in income shows that having student debt negatively impacts the holder’s quality of job, measured in terms of income. Students graduating with student debt straight out of college typically have a weaker bargaining position and thus are forced to accept worse job offers, as they cannot wait for better ones due to the specter of debt repayments.
This study demonstrates that the impact of the student debt cancellation is expected to be net positive to the economy. One key difference between this natural experiment and the Biden administration plan is the targeted income groups. In the natural experiment discussed above, all the loans were already in default and the debt holders were not paying their interest or principal. Furthermore, the average income in the experiment was around $2,000 per month, well below the Biden administration threshold amount of $125,000 per year (c. $10,000 per month). Whether the effect of the Biden student debt relief will be more pronounced or slightly dampened than the results of the natural experiment is not clear3. Although the sample of the natural experiment is on average less wealthy than the group being impacted by the Biden administration, the natural experiment does not have any direct income effects4 (i.e. with a debt reduction, your interest payments will go down, but since these loans were already in default, the interest payment reduction did not materialize), just a wealth effect (i.e. the impacted people were ‘wealthier’ as they did not have the debt anymore). Interestingly, other research - Dobbie and Song - has shown that the wealth effect is the key driver, not the income effect. By looking at credit card debt, the authors found that debt discharge which only occurred after several years, had significant positive effects, while a payment plan or repayment reduction had no positive impact. Though this question has not been empirically analyzed in a student debt setting, the research would suggest that income effects might not be as large.
The Wider Economy
What about the other potential impacts on the economy? Given the temporary and targeted nature of the Biden administration plan, it is difficult to conclude from economic research whether there will be permanent measurable other benefits. One study by Ambrose, Cordell, and Ma, “The Impact of Student Loan Debt on Small Business Formation,” found that increasing student debts from 2000 to 2010 resulted in a 14% drop in the formation of small businesses with 1 to 4 employees. The driving channel is that student debt limits the ability to tap into other forms of debt to start businesses. Another study (Bahadir and Goricheva) found that states that saw larger increases in student debt to income levels saw reduced consumption growth in the medium term, which negatively affected growth in those states. Thus, given these studies we could expect temporary increases in business formation and stronger consumption, which would stimulate the economy.
However, as the administration did not address the issue of financing higher education, student debt burdens will remain elevated and all the effects mentioned above will be unwound as a new generation of graduates are saddled with debt. The prospect of needing debt to finance college has been shown to deter pursuing higher education by poorer households (Callender and Mason); student debt impacts students’ choices of major and job in favor of higher paying jobs instead of lower paying ‘public service’ jobs (Rothstein and Rouse); student debt reduces early career retirement savings (Rutledge, Sanzenbacher and Vitagliano); student debt amplified the 2008 financial crisis (Amorim. Eberly and Mondragon).
Is the Biden Plan Regressive?
There has been a lot of talk about whether the Biden plan or a blanket forgiveness of $10,000 for all students may be regressive - benefitting higher income individuals more than lower income. This was argued by Catherine and Yannelis in “The Distributional Effects of Student Loan Forgiveness” and widely discussed online. Although I do not focus on this question in this post, the research cited above appears to contradict these findings. Catherine and Yannelis argue that instead of debt forgiveness, we should focus on an income driven repayment plan, as it would be the most progressive and least expensive. However, this does not appear to address the research (Dobbie and Song) that has shown that only debt discharge showed positive effects for debt holders, whereas payment reductions did not have any material impacts. Given that the Catherine and Yannelis model of income repayment is theoretical, rather than based on data, the effect of income driven repayment schemes might be overstated.
Conclusion
Empirical evidence appears to show that the student debt forgiveness will be positive for the wider economy - the boost to the economy will outweigh the cost of enacting this policy. The evidence also appears to suggest that it is forgiveness rather than changing payment plans that generates this positive impact. Furthermore, this forgiveness may repay itself quite quickly - with the study above suggesting that incomes go up by $3,000 for a $7,400 forgiveness - assuming a linear trend that would be about $4,000 higher income for $10,000 forgiveness - a marginal federal tax rate of 25% would yield an extra $1,000 of revenues for the Federal government, repaying the forgiven amount in approximately 10 years. This is a low estimate, as it only taking into account the Federal government tax increase due to higher incomes, ignoring state tax revenues and any other economy wide impacts.
The main issue of the Student Loan Forgiveness is that it does not address the issue of higher education funding. Debt funding of college has many measurable adverse impacts on the economy and the allocation of resources, which I will describe in a different post.
The program does alter the current income driven repayment policy, benefitting debt holders, which may have wider funding impacts.
In the hard-sciences (e.g. biology, chemistry), an experiment is when we take two groups and treat one of them with an intervention (for example, a medicine) and argue that any difference of outcomes between the groups is due to the treatment. That is because there shouldn’t be any difference in the group prior to the treatment if the enrollment into the groups was random. In social sciences (e.g. economics, psychology), such experiments are usually not allowed for ethical reasons or not feasible. However, they tend to occur naturally due to laws and regulations that arbitrarily divide people into two groups. For example, two groups with no discernible difference between them: one that receives a government intervention and one that doesn’t.
Generally, it is not advised to extrapolate findings based on one group to other groups. Thus, since the Di Maggio, Kalda, and Yao research paper focused on people with incomes on average up to $30,000 a year, we might not see the exact same behavior of people who earn up to $125,000 a year. In economics and statistics, this is called the question of external validity.
The income effect is the impact of having a higher stream of income.