Americans owe more than $1.6 trillion in federal student loan debt, a financial burden that has been tied to delays in marriage, childbearing, and homeownership among younger cohorts of borrowers (Velez, Cominole, and Bentz Reference Velez, Cominole and Bentz2019). Moreover, ballooning debt has had especially pernicious effects for Black Americans, expanding the racial wealth gap (Seamster and Charron-Chénier Reference Seamster and Charron-Chénier2017). The stark consequences of mounting student loan debt accelerated calls for student debt relief, especially in the wake of the COVID-19 pandemic and the subsequent economic recession. Further, the extended pause on student loan payments during the pandemic offered a policy window in which wide-scale student debt relief became more politically feasible than ever before.Footnote 1 Many prominent Democratic lawmakers, policy experts, and activists called on President Biden to embrace a robust program of student debt forgiveness. During the 2020 presidential campaign, Biden offered support for a limited approach that included cancelling up to $10,000 of debt for federal loan borrowers—a pledge he then presented concrete plans to enact in August of 2022 (White House 2022). But at a town hall with voters in February 2021, he shot down calls for $50,000 in student loan forgiveness, stating “I will not make that happen.” In his justification, Biden argued that the government should not forgive debt for certain types of people, like those who chose to go to expensive, Ivy League schools.
Despite the fact that Ivy League borrowers represent only 0.3% of people with federal student loans (borrowers from selective colleges and universities more broadly represent only 12%), the notion that some borrowers do not deserve debt forgiveness because they made a choice to take out costly loans is consistent with the type of logic opponents of debt relief often employ. For example, Senator Tom Cotton (R-AR) tweeted his opposition to student debt forgiveness in April 2022, asking “Why should a trucker who didn’t go to college have to pay off a lawyer’s student loan debt?” And in the wake of the president’s new debt relief plans, Republican opponents have continued to push this line of reasoning. Arguments about who does and who does not deserve student debt relief are important because they inform policy debates about whether and how to provide help for the millions of Americans who will still have outstanding educational debt even if the new policy is implemented. But to what extent do voters find these different arguments compelling? How do members of the public evaluate deservingness in the context of student loan debt, and what does it mean for continued policy action on debt forgiveness?
There is ample scholarship on the role that perceptions of deservingness play in determining policy outcomes (e.g., Schneider and Ingram Reference Schneider and Ingram1993; van Oorschot Reference van Oorschot2000; van Oorschot et al. Reference van Oorschot, Roosma, Meuleman and Reeskens2017), but much of that literature explores deservingness in the context of targeted social welfare policies. While scholars increasingly view credit access and debt relief as part of the welfare state (Prasad Reference Prasad2012; SoRelle Reference SoRelle2020; Wiedemann Reference Wiedemann2021; Michener, SoRelle, and Thurston Reference Michener, SoRelle and Thurston2022), personal debt has a stigma and history of its own in the United States that might influence perceptions of student borrowers. Furthermore, while many social welfare programs benefit groups that can be relatively easily defined based on a set of shared, politically salient characteristics, the target population of “student loan borrower” is diffuse and ill-understood. This means we know little about how people evaluate deservingness in the case of student debtors, despite the fact that perceptions of deservingness are among the most prominent frames employed in policymakers’ arguments for and against student debt forgiveness.
In this study, we employ original conjoint and framing experiments to explore two interrelated questions about support for student debt relief: How do borrower characteristics condition perceptions of deservingness for student loan debt forgiveness, and do arguments about deservingness influence support for student debt cancellation? Drawing on insights from the social construction of target populations (Schneider and Ingram Reference Schneider and Ingram1993; Kreitzer and Smith Reference Kreitzer and Smith2018) and CARIN literatures (van Oorschot Reference van Oorschot2000, Reference van Oorschot2006; Meuleman, Roosma, and Abts Reference Meuleman, Roosma and Abts2020),Footnote 2 we explore how a variety of borrower traits shape people’s evaluations of deservingness for debt relief. We expect that, while support for student loan forgiveness will be generally high, certain borrower attributes may condition perceived deservingness with consequences for the politics of debt relief. Specifically, we hypothesize that borrowers 1) who are deemed to have greater need and 2) who are perceived to have taken action to pay off their debts will be considered more deserving of relief. By contrast, we expect those who are seen as having greater control over the decision to accumulate debt—for example, by choosing a more expensive educational institution or profession—will be deemed less deserving of debt forgiveness. Finally, we anticipate that the partisan and racial identity of both borrowers and evaluators may shape perceptions of deservingness.
In general, we find evidence that all of these attributes are at work to varying degrees in shaping perceptions of deservingness for student debt forgiveness. Borrowers are judged to be most deserving of debt forgiveness when they are seen as in need of debt relief (for example, due to their earning potential) and when they are seen as having “made an effort” to repay their debts. Overall, racial minorities were more likely to be judged deserving of student debt relief than white borrowers, but priming race in arguments for broader student debt relief diminished support among Republicans. And while some identity characteristics of survey respondents produced modest effects on judgments of which borrowers were deserving, men, Republicans, and higher-income respondents were less supportive of student debt cancellation overall. Perhaps the most important finding, however, was the generally high support for student loan forgiveness across treatments and borrower types. Our results shed light on the political dynamics for a highly salient post-COVID policy issue, while extending scholarly understandings of deservingness to a critical, and understudied, aspect of the American welfare state. They also offer insight into how the president’s new policy measures to relieve student loan burdens may unfold politically moving forward (see also, SoRelle and Laws Reference SoRelle and Laws2023).
Deservingness and Student Loan Debt Forgiveness
Public opinion about government social programs is often driven by the perceived deservingness of the target population at issue (Schneider and Ingram Reference Schneider and Ingram1993). Groups that are more positively constructed in public discourse—such as children, veterans, and the elderly—usually enjoy more generous policies passed with less opposition, while negatively constructed groups such as criminals or single mothers will rarely receive favorable policy treatment. But where student loan borrowers fall on this spectrum is less clear. Unlike many target populations that share an identifiable set of group characteristics, student borrowers are a diverse and varied cohort containing young, middle-aged, and, increasingly, older Americans; financially strapped, middle class—as well as wealthy professionals—and borrowers who vary widely with respect to their educational attainment and profession. About 15% of all adults report currently having student loans, and there are 46 million federal student loan borrowers total (Cilluffo Reference Cilluffo2019)—although this number may be reduced by as many as 20 million pending the successful implementation of Biden’s debt cancellation plans. An estimated 42% of borrowers went to public colleges and universities, while 23% attended private, non-profit colleges, and 20% attended private, for-profit colleges (Kantrowitz Reference Kantrowitz2020). Roughly 38% of debtors never received a four-year degree.
Because there is no settled social construction of student loan debtors, the perceived deservingness of the target population likely depends upon which of the many possible constructions of student loan borrowers is salient at a given time. A very different picture of deservingness might emerge, for example, if borrowers are imagined as low-income adults saddled by debt from predatory for-profit colleges, struggling middle-class parents still hampered by student loan payments a decade after graduation, or wealthy professionals with high-dollar graduate debt (but also high-dollar income). All characterizations are true for at least some portion of borrowers. Thus, the issue for evaluations of deservingness is not simply “how is this group constructed?” but which social construction is salient in any particular policy debate.
Scholars have examined the most common individual characteristics affecting perceptions of deservingness in social policy, adopting the acronym CARIN to represent the five primary mechanisms: Control (how responsible the person is perceived to be for their situation), attitudes (being likeable, well-behaved, and thankful for help), reciprocity (having “earned” support through hard work or contributions to society), identity (sharing group membership), and need (van Oorschot Reference van Oorschot2000, Reference van Oorschot2006). While the CARIN criteria have at times been criticized for containing overlap, the five deservingness principles have been shown to represent “distinct logics of social justice” that can be measured and tested independently (Meuleman, Roosma, and Abts Reference Meuleman, Roosma and Abts2020).Footnote 3 These criteria, in combination with the insights from the broader social policy literature, provide a helpful foundation from which to tease out hypotheses about evaluations of deservingness in the context of student debt forgiveness.
Need is the most basic component of the CARIN typology and is in some ways the core of notions of deservingness. If a person or group is viewed as being in need, they are often seen as deserving of government help. Studies employing the CARIN framework have found that certain groups—the elderly and disabled—are consistently viewed as most deserving of relief, largely based on their perceived need for assistance (van Oorschot Reference van Oorschot2006). But other elements quickly enter into judgments of deservingness, for example, whether the group or individual in question is to blame for their circumstances (control) and whether they have earned the right to help because of their contribution to society or hard work (reciprocity).
Control, or perceived responsibility for one’s hardship, looms large in shaping perceptions of deservingness. As Oorschot (Reference van Oorschot2006, 26) explains, “once the public feels that a person can be fully blamed for his or her neediness, other criteria become irrelevant.” The question of why a person or group is in need becomes important, especially for less positively constructed groups like the unemployed. For example, in the area of health deservingness, judgments often come down to whether the subject is seen as having caused their health problem through their individual behavior. If so, they are consistently viewed as less deserving, and priming such scenarios makes respondents less supportive of government health care or increased government spending on health (Gollust and Lynch Reference Gollust and Lynch2011; Gollust, Lantz, and Ubel Reference Gollust, Lantz and Ubel2010).
With respect to personal debt, notions of control are also likely key, as the idea of personal responsibility for one’s debt is particularly salient in the American context (SoRelle Reference SoRelle2022a, Reference SoRelle2022b; Calder Reference Calder1999). Despite the strong role of the American state in expanding credit access, the area of personal credit is largely depoliticized, with most Americans understanding their problems with financing in individual or market terms, rather than as an area to make political demands of government (SoRelle Reference SoRelle2020, Reference SoRelle2022b). This all suggests that Americans might be especially resistant to debt forgiveness, as borrowers may be deemed personally responsible for choosing to take out debt to spend beyond their means.
Yet within the broader arena of debt relief, there is reason to think that student loans might be more amenable to government proposals for debt forgiveness than, for example, those aimed at credit card debt or payday loan debt. Most student loans derive from the federal government, making demands for government action to relieve them potentially more logical than for debts initiated by private companies. While student loans historically operated in the hidden realm of the “submerged state,” the connection between the government and student loans has become clearer since the Affordable Care Act enshrined direct government lending for federal student loans (Mettler Reference Mettler2011). Perhaps most significantly, the connection between government and student debt was made explicit during the pandemic, when the swift, early, and long-lasting pause on federal student loan payments provided tangible and highly visible financial relief to Americans.
Furthermore, student loans may be less susceptible to the argument that borrowers “made a bad choice” than for credit card or other types of debt, or even some social welfare policies, as student debt comes from the socially desirable effort to pursue higher education, whose cost borrowers have no control over.Footnote 4 Though attitudes toward higher education have polarized somewhat in recent years, pursuing higher education is generally viewed as central to achieving the American dream, and college students are a positively constructed target population (Kreitzer and Smith Reference Kreitzer and Smith2018). Most Americans believe that parents and (to a lesser extent) government bear responsibility for paying for college, in addition to students, which suggests that burdensome debt may be viewed as unfair.Footnote 5 Taking on debts in pursuit of higher education is often viewed as a “good” debt—an investment in one’s future that people expect will be returned.
Reciprocity is another important component of deservingness, where those who have worked harder—or are otherwise seen as having “paid their dues”—may be considered more deserving of relief. There are clear connections between reciprocity and debt. In a country that values the individual contract as highly as the United States (Fraser and Gordon Reference Fraser and Gordon1992), not paying one’s debts has historically been a mark of personal failure and even immorality (Calder Reference Calder1999; Graeber Reference Graeber2011). Filing for personal bankruptcy—the most extreme form of federal debt forgiveness—is deeply stigmatized. Thus, proposals for outright debt forgiveness of any kind collide with deeply rooted values in American political culture that are related to, but not the same as, the bootstrap logic that in part drives public opposition to welfare.
Finally, identity and related assumptions about attitudes figure centrally both in the CARIN framework and the wider literature on evaluations of deservingness. In its broadest sense, the notion of identity suggests a relational evaluation based on the characteristics of both the target and the evaluator. In-group members are typically viewed as more deserving than out-group members. Perhaps nowhere has the effect of identity on perceptions of deservingness been more developed than with the case of race. In American social policy, notions of deservingness among beneficiaries are often driven by racist attitudinal stereotypes that serve to justify cuts or the imposition of stricter eligibility requirements (Gilens Reference Gilens1999; Sparks Reference Sparks, Schram, Soss and Fording2003; Schram, Fording, and Soss Reference Schram, Soss and Fording2003). When policy beneficiaries are perceived to be comprised primarily of racial minorities, support for robust government benefits is often lower. If views of student loan forgiveness follow this pattern, the positive or negative construction of borrowers may be particularly pronounced and negative if the policy is perceived to benefit racial minorities. There is some existing evidence to suggest such an outcome. Blame associated with the financial consequences of subprime mortgage borrowing is heavily racialized: it is seen as a result of bad personal choices when victims are primarily women and people of color, but it is viewed more commonly as a structural problem when white borrowers are also affected (Strolovitch Reference Strolovitch2021).
Hypotheses
What expectations then follow about how the public will view the deservingness of student loan borrowers who might receive debt relief? This is a complex policy space, a unique type of government support, and it affects a large, diffuse, and varied policy target. We anticipate that highlighting different attributes of student loan borrowers may condition judgments about whether they are deserving of government relief. First, we expect the pervasive finding that assessments of deservingness are altered based on evaluations of how much control a person has over their situation will apply to student loan borrowers. This is particularly likely given the emphasis placed on personal responsibility for debtors. Thus, we anticipate that assessments about borrowers’ control over their situation—including the choice to attend more expensive institutions or perhaps borrow for advanced degrees—will diminish perceptions of deservingness in the context of debt relief. Evidence of this logic has been present in debates over student debt cancellation proposals. For example, in criticizing a plan proposed by Senators Chuck Schumer and Elizabeth Warren, one Wall Street Journal op-ed argued that “the main beneficiaries of the Warren-Schumer write-down would be higher-income students who borrow more for expensive graduate degrees” (Wall Street Journal 2021). Thus, the control hypothesis stipulates:
H1 (Control): Borrowers perceived to have greater control over—or responsibility for—their student debt burden will be assessed as less deserving of debt forgiveness.
We also expect borrowers who are perceived to have worked hard or paid their dues to be judged more deserving of government relief for their debts. This follows the reciprocity component of deservingness, in which past contributions to society or other indications of having “earned” a benefit contribute to positive judgments of deservingness. We conceive of reciprocity along both a general and policy specific dimension. Borrowers might be judged based on their broader contributions to society (e.g., as taxpayers) or based on their contributions toward paying off their debt in this specific policy context. As such, we predict:
H2 (Reciprocity): Borrowers who are perceived to have “paid their dues” either generally to society or specifically to alleviate their debt will be judged more deserving of loan forgiveness than those who have not.
The debate around student loan forgiveness also has significant class dimensions that may shape perceptions, tapping into the need component of deservingness. For critics, one key objection to student debt cancellation has centered on its potential for upward redistribution, benefiting wealthier borrowers more than lower-income borrowers. Jason Delisle (Reference Delisle2020) of the American Enterprise Institute argues that “high income students would be the major beneficiaries of loan forgiveness, as would the countless lawyers, doctors and others with advanced degrees who account for 42 percent of all student debt.” This works against the idea that student loan borrowers really need the government’s help and may tap into widespread resentment of the rich (Piston Reference Piston2018). In the personal debt arena need may also be interpreted as “ability to repay.”Footnote 6 We anticipate that perceived need will be a predictor of deservingness in the context of student debt forgiveness.
H3 (Need): Borrowers perceived to have greater financial need—or less ability to repay—will be judged as more deserving of debt relief than those perceived to have less financial need.
Finally, we anticipate that intersecting identity and attitudinal assumptions might shape perceptions of deservingness in multiple ways. As described previously, racial identity is one of the most significant drivers of deservingness evaluations in American social policy. Thus, we expect the racial identity of borrowers (and participants) will be a significant factor. Those favoring student loan forgiveness proposals often argue that debt relief could help close the racial wealth gap, as racially minoritized groups (and Black borrowers in particular) are more likely to be saddled with high levels of student debt (Seamster and Charron-Chénier Reference Seamster and Charron-Chénier2017; Hamilton and Zewde Reference Hamilton and Zewde2020). While this has been a key argument by proponents, studies in a variety of social policy areas have found that highlighting minoritized groups as beneficiaries of a program often decreases support (Gilens Reference Gilens1999; Piston Reference Piston2018; Rigby et al. Reference Rigby, Soss, Booske, Angela and Robert2009; DeSante Reference DeSante2013). We thus derived the following hypothesis related to race:
H4 (Race): Minoritized borrowers will be judged as less deserving of debt relief by white respondents.
While the social policy literature has focused considerable attention on race, partisanship has emerged as another important feature of identity in the American context, one that may equal racial identity in importance (Iyengar and Westwood Reference Iyengar and Westwood2015; Westwood et al. Reference Westwood, Iyengar, Walgrave, Leonisio, Miller and Strijbis2018). Thus, respondents’ party affiliation may be a strong predictor of opinions, particularly in the area of higher education, which has become increasingly polarized in the last ten years (Parker Reference Parker2019). Further, two of the CARIN criteria—control and reciprocity—dovetail with partisanship in the American context in ways that lead us to expect differences between partisans. As observed in other policy domains (e.g., Hacker Reference Hacker2019; SoRelle Reference SoRelle2020), we anticipate that Republicans may be especially reliant on notions of reciprocity (getting something for nothing) and control (commitment to personal responsibility) in their evaluations of deservingness, making them less supportive of student debt forgiveness and more receptive to arguments against debt relief framed in those terms. In experimental studies in the health arena, for example, Republicans are more likely to assign blame to individuals for their health problems (Gollust and Lynch Reference Gollust and Lynch2011) and are more responsive to framing arguments that highlight people’s control over their situation (Gollust and Lynch Reference Gollust and Lynch2011; Gollust and Cappella Reference Gollust and Cappella2014). Thus, we anticipate:
H5 (Partisanship): Republican respondents will judge borrowers as less deserving in general and will be especially responsive to attributes that tap into “control” and “reciprocity” in their evaluations of deservingness.
We also consider the possibility that these two identity dimensions—race and partisanship—will coalesce to shape perceptions of deservingness, with Republicans less amenable to seeing racial minorities as deserving of relief when compared with Democrats. The increasing convergence of partisanship and racial identity has been demonstrated by a growing cohort of scholars (e.g., Mason Reference Mason2016; Mason and Wronski Reference Mason and Wronski2018; Westwood and Peterson Reference Westwood and Peterson2020), with considerable overlap between Republican and white racialized group identity. This racialized partisanship influences support for a range of issues from affirmative action (Bell and Lui Reference Bell and Lui2023) to support for D.C. statehood (Nteta et al. Reference Nteta, Rhodes, Wall, Dixon-Gordon, La Raja, Lickel, Mah, MacWilliams, Markowitz, Reid and Theodoridis2023) to name a few. We anticipate that this dynamic will reinforce the previous hypothesis, with Republicans especially responsive to efforts that frame opposition to debt relief in racialized deservingness narratives. Footnote 7
Data and Methods
We explore people’s attitudes about deservingness and student debt relief with two original experiments: a conjoint design and a framing experiment.Footnote 8 Conjoint, or discrete choice, experiments are an increasingly common tool used by political scientists to capture the distinct causal effects of specific factors in multidimensional preferences (Hainmueller, Hopkins, and Yamamoto Reference Hainmueller, Hopkins and Yamamoto2014). In our case, a conjoint experiment allows us to consider separately the effects of multiple student loan borrower characteristics on people’s evaluations of how deserving those borrowers are of debt relief. To gauge the effect of borrower characteristics on perceptions of deservingness, participants were presented with six successive pairs of hypothetical student loan borrowers with information about eight borrower attributes that both map onto our respective hypotheses and have featured prominently in debates over debt forgiveness: occupation, race, employment status, type of institution attended, level of education for which loan was taken, amount of debt remaining, time in repayment, and repayment history. Each of the attribute levels was fully randomized for each borrower profile, such that every possible borrower profile was equally likely. Table 1 describes the full set of attributes and possible levels for each, and it notes the reference category used in the following analysis. Table 1 also includes a description of how these attributes map onto the deservingness mechanisms we hypothesize are likely to shape people’s evaluations.
Borrower deservingness conjoint attributes and levels
With respect to our first hypothesis—control—we included information about the borrower’s current or most recent occupation, type of college attended, and level of degree (graduate or undergraduate) for which debt was generated. Each of these attributes are commonly used as tropes in the debate over deservingness (recall Senator Cotton’s remarks about Ivy-league lawyers) because they indicate that a person made a choice to take on more (or less) debt for a more (or less) costly education. To capture reciprocity, we include a description of employment status, which is commonly used in the deservingness literature as a general measure of reciprocity, as well as time spent repaying the loan and payment status to capture more policy-specific reciprocity. The most straightforward measure of need we include captures the amount of debt outstanding, but we also anticipate that participants may interpret occupation and employment status as potential indicators of need as well. Finally, we include race as a key identity marker. It is important to note that identifying specific measures for the primary CARIN attributes is a complicated exercise, with attributes being open to interpretation by the survey taker. Because some attributes may tap into multiple potential mechanisms of deservingness, we take advantage of open-ended comments from a retrospective probe immediately following the conjoint to help untangle how participants interpret the attributes with respect to our specified mechanisms.
After considering the borrower pairings, participants were asked to evaluate borrower deservingness in two ways. First, they were asked to choose “which borrower most deserves to have a significant portion of their outstanding student loan debt forgiven,”Footnote 9 which we will refer to as the forced choice response. Participants were then asked to rate the deservingness of each borrower separately on a scale from one to five, where one equals very undeserving and five equals very deserving of debt relief. By measuring our primary dependent variable—perception of borrower deservingness—in both ways, we not only replicate a common strategy in conjoint design (see Hainmueller, Hopkins, and Yamamoto Reference Hainmueller, Hopkins and Yamamoto2014), but we also capture two useful dimensions of support for debt forgiveness. While the forced choice question will help to illuminate the attributes that most matter to people when evaluating which borrowers deserve debt relief, the ability to rate borrower deservingness separately allows for the possibility that people may deem borrowers deserving (or undeserving) of debt relief irrespective of those attributes. For example, a participant might select Borrower A in the forced choice question because they prioritize time spent in repayment when faced with a discrete choice, but they may rate both borrowers as “very deserving” of debt relief in the following question because the lack of that attribute is not sufficient to make Borrower B undeserving of loan forgiveness in their view. Respondents were then asked to explain what they were thinking about as they made their choices. These open-ended responses provide rich context to help understand how respondents judged borrower attributes and how they interpreted the different characteristics offered.Footnote 10
While the conjoint experiment directly tests the degree to which different borrower characteristics influence perceptions of deservingness for student debt forgiveness, we are also interested in understanding whether arguments for or against student debt cancellation that are couched in terms of deservingness actually shift people’s support for political action to forgive student debt. We incorporate a separate framing experiment to consider this possibility. Participants were asked to “read the following brief excerpt from a recent news article about student loan debt.” They were randomly assigned to receive a control or one of six treatments described in table 2, each of which centers commonly-used deservingness arguments to either support or oppose student debt forgiveness.Footnote 11 After receiving the treatment, participants were asked “Would you support or oppose a plan to forgive federal student loan debt?”
Deservingness framing experimental conditions
The need frame emphasizes how debt forgiveness would help middle- and working-class Americans, while the racialized need treatment includes that language, then provides additional information about the racial wealth gap, which has become a common talking point among debt cancellation proponents. The control frame, by contrast, argues that it would unduly benefit wealthy elites, a common argument among opponents of cancellation. Because it is rare for people to be exposed to a single frame for any issue, scholars have demonstrated that both sides of an argument should be presented to maximize the external validity of information-based experiments like this one (Sniderman and Theriault Reference Sniderman, Theriault, Saris and Sniderman2004; Chong and Druckman Reference Chong and Druckman2007). Thus, the competitive frame combines the racialized need and control treatments into one frame. The last two endorsement frames explore whether the positive and negative deservingness arguments hold more weight when presented by partisan elites. Finally, participants were asked to provide standard demographic information, party identification, and information about their own use of student loans. These responses are used to subset our analyses to test those hypotheses that predict varied assessments of deservingness based on the race and partisanship of the evaluator.Footnote 12
Analysis and Results
Before we address the results regarding perceptions of deservingness and the politics of student loan forgiveness, it is useful to consider the broader patterns of support for student debt relief. We anticipate that such support will be generally high, although we expect it to vary across groups. Table 3 presents the percent of respondents by respondent demographic who expressed support for government assistance to student loan borrowers, which we capture using two different measures. The first, corresponding with column A, reports aggregated responses across all borrower profiles to the conjoint question asking participants to rate the deservingness of a particular borrower profile. For each of these questions, the percent in the table reflects those who deemed borrowers as somewhat or very deserving of government support/debt relief. The second measure, corresponding with column B, asked respondents “Would you support or oppose a plan to forgive federal student loan debt” after presenting them with the experimental control condition described above in table 2. The percent in the table reflects those who somewhat or strongly supported such a program.
Support for government relief for student loan borrowers
Irrespective of the measure we use, we find evidence to suggest that Americans generally report high levels of support for government intervention to provide relief to student loan borrowers, which is consistent with our expectations as well as public opinion polling on the issue. A majority of respondents from every category except self-identified Republicans support a government plan to forgive student loan debt. And perhaps remarkably given today’s level of partisan polarization, 40% of Republicans—including Republican-leaning independents—also voiced support for government debt relief. Similarly, all groups except Republicans agreed that a majority of the borrower profiles merited government intervention to relieve student loan debt, while that rate dipped to 37% for Republican respondents. These levels of approval for both measures were especially high among core Democratic constituencies—self-identified Democrats (85% and 77% respectively) and non-white participants (85% and 73%).
Thus, any differences we observe in what drives judgments of deservingness for student debt forgiveness occur in the context of fairly high support for the policy. Some of our respondents noted the difficulty of the conjoint task precisely because of their strong support for debt relief. As one respondent explained, “It was difficult to pick one over the other because I strongly believe that all student debt should be forgiven for all borrowers, unconditionally.” Indeed, in describing their choices, about 17% of respondents volunteered that they support student debt forgiveness for all.Footnote 13
Participant Characteristics and Deservingness Evaluations
While the primary focus of our analysis concerns how borrower characteristics shape evaluations of deservingness for student debt relief, it is also worth considering briefly how participants’ own traits correspond with overall evaluations of support. Figure 1 reports the predicted effect of respondent characteristics on two indicators of support for student debt relief: the borrower deservingness ratings from the conjoint experiment and responses to a question asking “In general, would you support or oppose a government program to alleviate [student loan debt]?”.
Relationship between respondent characteristics and support for student debt relief
Note: Points are coefficients from OLS regression (n=1,492, 741) and bars represent 95% confidence intervals.
As figure 1 shows, party identification is the strongest correlate both of support for student debt forgiveness and evaluations of deservingness. As a person becomes more Republican, they become less supportive of federal student loan debt forgiveness and less likely to evaluate a candidate as deserving of debt forgiveness (consistent with our partisan hypothesis). The other major predictor of support is whether the person has ever had student loans themselves. Having had student debt increases both support for government relief and perceptions of borrower deservingness for debt forgiveness. Age and income are also consistent predictors for both measures, with older and wealthier respondents less supportive of each measure.Footnote 14 Finally, there are two characteristics that influence only one measure of support: Gender is a meaningful predictor of support for government debt relief, with men less likely to support the measure than women; however, gender does not correspond with evaluations of borrower deservingness. And education negatively corresponds with evaluations of deservingness but not support for debt relief. Interestingly, a bivariate measure of respondent race (white equals one) does not correspond with overall attitudes toward debt forgiveness. In general, these results are consistent with the descriptive evidence presented earlier, with party ID and its demographic correlates producing the most dramatic differences in support on questions of student debt forgiveness.
Who Deserves Student Debt Forgiveness?
While we find that support for student debt relief is widespread, consistent with existing public opinion data on the subject, we know far less about how different borrower attributes condition that support. This is a critical question as politicians and the public react to the student debt relief plan announced by President Biden and debate policy details that include or exclude beneficiaries based on characteristics like income, type of institution attended, and other salient features. The following analyses report the results from the conjoint design, which isolates the effects of several different borrower traits on people’s evaluations of how deserving borrowers are of student debt relief.Footnote 15 The results reported in the following coefficient plots represent the average marginal component effects (AMCEs) estimated using OLS regression with standard errors clustered at the respondent level.Footnote 16 The AMCE captures the average change in the probability that a borrower’s profile will be selected when a particular attribute level is listed relative to a baseline category (Hainmueller, Hopkins, and Yamamoto Reference Hainmueller, Hopkins and Yamamoto2014). The baselines are included as the first listed attribute level for each category in the following coefficient plots.
Figure 2 presents the results for both the forced choice (left panel) and borrower deservingness rating (right panel) questions. There are several commonalities between the two. Occupation has a significant impact on deservingness evaluations. Relative to being a doctor, all three other occupation categories—small business owner, high school teacher, and restaurant server—made borrowers more deserving of debt relief in the eyes of participants. One possible explanation is that occupation is standing in as a proxy for income, and thus, of need, and people rate those in greater need as more deserving.Footnote 17 Another possible explanation for this finding is that occupation reflects an assessment of borrowers’ responsibility for their own debt. Doctors, who made the choice to pursue an expensive graduate degree, may be deemed as more responsible for choosing to incur debt than for borrowers with the other three occupations, tapping into the control dimension of deservingness.
Direct effect of attributes (AMCE) on borrower preference & rating
Note: Treatment effects from conjoint experiment. Points are coefficients (AMCEs) from OLS regression (n=5968, 5975) and bars represent 95% confidence intervals.
In the open-ended responses following the conjoint, over one-quarter (28%) of respondents mentioned occupation as one of their main criteria in judging which borrower was more deserving. These discussions predominantly framed occupation as a proxy for income (tapping into the need principle)—conceptualized by some as ability to repay. For example, one respondent explained, “I was weighing which would have an easier time paying off their loans in the future. This was mostly based on what their current or most recent job was: was it high- or low-paying?” Another stated that “I tried to pick the person who either had more debt or a lower paying job.” Another said, “Profession mattered—like a doctor shouldn’t have their loans repaid since they’ll make it back.” But in a few cases respondents thought of occupation as more indicative of the borrower’s responsibility for their debt. As one noted, “Why did you pick a profession that doesn’t [offer] compensation equal to your loans? Shame on a student for not doing research, going to community college for two years to save money, not saving for college, etc.”
When we consider some of the other attributes that might capture the personal responsibility (control) mechanism, we see mixed results. For example, the level of education (graduate or undergraduate) that a borrower incurred debt from did not affect people’s deservingness evaluations. This pattern is notable given the strong emphasis on graduate and professional debt by opponents of forgiveness, and it works against the notion that borrowers are being punished for making the choice to pursue “extra” education. On the other hand, relative to Ivy League colleges, attending all other institution types is associated with a significant increase in the likelihood that a borrower is identified as deserving of debt relief. This is consistent with the narrative some politicians, including President Biden, have articulated for placing limits on who gets student loan forgiveness, and it may indicate a viewpoint that people who make the choice to attend more expensive schools are responsible for their own debt and thus less deserving of relief. In the open-ended responses, one typical expression of this logic argued that “choosing to go to an Ivy League college is a choice, not mandatory for an education. The individual made an informed choice and chose to go into higher debt.” Another respondent reasoned that “the people that went to community/public colleges attempted to get their education with as little debt as possible. They tried. Those that went to Ivy League schools had no thought of the debt they would incur or how to pay their loans. [In my opinion] they shouldn’t be rewarded.”
We also see strong evidence that reciprocity is meaningful in shaping people’s perceptions of deservingness. Employment status, which serves as the general measure of reciprocity, matters in interesting ways. Relative to being employed, people who are unemployed but looking for work are not deemed less deserving; however, those who are unemployed and not looking for work—an indication that they are not “pulling their weight” or trying to improve their situation—are deemed less deserving of government debt relief.Footnote 18 Employment status was also the most commonly cited factor in the open-ended responses, with 37% of participants addressing that attribute. One typical response explained that “someone who isn’t looking for work and isn’t employed shouldn’t have priority over someone who is trying hard to get ahead.” Another more bluntly stated that “people not looking for work are deadbeats who do not deserve forgiveness.” In some cases, being unemployed was also taken as a proxy for need: “If they were unemployed I thought they should be forgiven more,” but this reasoning was far less common. The harsh judgment of the unemployed echoes findings in the CARIN literature that unemployment is an area where recipients are viewed as less deserving, or more conditionally deserving, depending on whether the unemployment status was within their control or not (Jensen and Petersen Reference Jensen and Petersen2017; van Oorschot Reference van Oorschot2006).
Relatedly, time in repayment, which we argue is a clear debt-specific metric for reciprocity, has a consistent effect on deservingness evaluations. Borrowers who have been repaying their student loan debt for any amount of time are viewed as more deserving of debt relief compared to those who have not begun repayment, and the magnitude increases for those who have been in repayment more than five years. As one respondent explained, they based their judgments mainly on the “length of loan payments and payment history. If someone has been paying diligently for a long period of time they are more deserving then someone who hasn’t made a payment.” Interestingly, whether a borrower has experienced trouble meeting their monthly payments (repayment status) does not consistently influence which borrower is deemed most deserving of debt forgiveness, but it does have a marginally significant negative effect on individual ratings of borrower deservingness. Open-ended responses indicate that those who were behind on payments or had never made a payment were usually seen as less deserving because they were not “making an effort.”Footnote 19 As one respondent put it, “If the person was paying back the loan I felt they were more deserving of forgiveness. At least they tried rather than shirked their duties.”
With respect to the primary measure of need, outstanding loan balance (the amount of debt borrowers have yet to repay) produces a shift in assessments of deservingness when participants are forced to choose the most deserving but not when they rate borrowers individually. Borrowers who owe more than $10,000 in student loan debt are more likely to be chosen as most deserving of debt relief relative to those who owe less than $10,000. This cutoff is especially notable given the President’s proposal to limit the amount of debt forgiven to $10,000. In open-ended responses, the amount of debt was mentioned as a key factor in about 25% of comments, most focusing on the attribute as a measure of need. As one respondent concluded, “People with more debt need it more.”
Perhaps unique to the credit context, however, many participants’ judgements related to need coalesced in assessments of the ability to repay. For example, one participant articulated “I tried to determine who would have a better chance of paying back the debt and then choose the other person.” This was often directly linked to occupation: “If someone is a high school teacher, it will be more difficult to pay back their loans than it is for a doctor.” Judgments of “ability to repay” are need-based, but they include factors beyond income, like the amount owed and perceived earning potential based on degree, reflecting the complexity of determining “need” in the area of personal debt.
While each of the stated attributes generally confirm our hypotheses and the larger expectations of the deservingness literature, the effect of race on evaluations of deservingness appears initially to be at odds with broader findings that racial minorities are viewed as less deserving of social policy benefits than their white counterparts. Instead, evidence from the conjoint suggests that, relative to white borrowers, borrowers of color are viewed as more deserving of debt relief when a choice is forced. Black borrowers in particular are evaluated most highly, and they are the only racial group that continues to get a marginally significant boost in deservingness evaluations when borrowers are rated individually. These findings may be part of a larger shift in views of deservingness of racial minorities. A recent paper on health-related deservingness similarly finds preference for racial minorities, which the authors suggest may be due in part to increased awareness of structural racism following the George Floyd killing and subsequent revival of the Black Lives Matter movement (Williamson et al. Reference Williamson, Trujillo, Willen and Walsh2021). Similarly, while racial cues historically dampen support for welfare in the American context, a more recent study found no effect overall, but rather racialization of welfare only resulted in lower support among respondents who expressed pre-existing racial biases (Harell, Soroka, and Iyengar Reference Harell, Soroka and Iyengar2016).
Taken together, these results offer evidence consistent with most of our hypotheses. Reciprocity, control, and need all appear to matter for evaluations of student loan borrower deservingness in ways we would expect, while racial identity runs counter to our hypothesis. Specifically, a borrower’s occupation, race, employment status, amount of debt, time in repayment, and type of institution all influence people’s evaluations of deservingness for student loan forgiveness. But it is important to again underscore that these differences emerge in the context of generally high support for student debt forgiveness. As one respondent expressed, “I think that college loan debt should be forgiven for everyone. It should not be determinant on income, job level, education level, what type of school you went to etc. I tried to pick the person who either had more debt or a lower paying job but I still think both candidates were always deserving of loan forgiveness.”Footnote 20
Identity and Assessments of Deservingness
The previous section considers how people generally perceive the deservingness of student loan borrowers. But to what extent do constituencies evaluate deservingness differently? This is an especially important political question given the reality that not all voters participate in equal measure, nor do they matter equally to elites of each party. It is also necessary to subset respondents to more fully address whether the identity mechanism drives perceptions of deservingness. One possibility is that voter preferences are distinct among members of different political parties—that partisan identity shapes evaluations of deservingness.
Figure 3 presents the AMCEs for the forced choice question by self-reported party ID (including leaners). While there are some notable differences, perhaps the most interesting outcome is the relative consistency across partisans in the factors that matter to evaluators.Footnote 21 The directional effects (and non-effects) for experience with student loans, occupation, employment status, type of college, level of degree, and time in repayment do not vary significantly by party. Most notably, given previous findings in the literature, borrower race does not produce dramatically different effects on evaluations of deservingness for Democrats and Republicans.
Direct effect of attributes (AMCE) on borrower preference by party ID and race
Note: Treatment Effects from Conjoint Experiment. Points are coefficients (AMCEs) from OLS regression (n=3,776 for Democrats, 1,360 for Republicans, 4,296 for white, 880 for black) and bars represent 95% confidence intervals.
The intersecting partisan and racial identity hypotheses suggest that Republicans might view racial minorities as relatively less deserving of debt relief compared to Democrats, owing in large part to 1) the convergence of white identity and Republican partisanship described previously and 2) the degree to which Democrats rely on voters of color as a key constituency. These results offer only very weak evidence of that. Relative to white borrowers, Democrats evaluate borrowers of color as more deserving of government debt relief, with Black and Latinx borrowers garnering the biggest positive effect. For Republicans, borrowers of color produce only marginally significant positive shifts in deservingness evaluations relative to white borrowers.Footnote 22
There is, however, one notable difference between partisans: the importance of outstanding debt on evaluations. Democrats view those with more debt as more deserving of government assistance, while for Republicans the amount of debt does not consistently matter. Given that this serves as the most explicit measure of need, it may be the case that Republicans focus more on reciprocity and control when evaluating deservingness, while Democrats also consider need. This fits with deservingness studies in other areas of social policy, where Republicans emphasize personal responsibility more and judge the unemployed more harshly (Gollust and Lynch Reference Gollust and Lynch2011), and it would be consistent with our hypotheses.
When it comes to the race of the participant—perhaps the clearest measure of in versus out group evaluation—we once again see both similarity and difference. As figure 3 illustrates, for both white and Black participants, occupation and time in repayment influence perceptions of deservingness, while the level of education and repayment status do not. Interestingly, however, for Black participants, neither the type of college attended nor the amount of debt outstanding consistently influence evaluations as they do for white participants. And being unemployed without taking action to find a job—one of the strongest negative predictors for most groups, does not diminish evaluations for Black participants (although that may be an artefact of sample size). Finally, and perhaps most interestingly, identity vis-à-vis race does not fully conform to our expectations. While Black borrowers do get the biggest boost in deservingness among Black participants, consistent with in-group preference, white participants deem out-group racial members as more deserving of debt forgiveness.Footnote 23
Deservingness Frames in Political Rhetoric
Based on the previous results, we might expect that rhetoric priming groups who do or do not deserve debt forgiveness based on their perceived control over debt accumulation, their need for assistance, their perceived reciprocity, and their racial identity will influence people’s support for student debt relief policy. What happens when politicians draw on these attributes in broader deservingness-based arguments to promote or oppose plans to forgive student loan debt? In reality, elites invoke combinations of borrower attributes rather than singling out one particular characteristic. Thus, the following results from the framing experiment explore how people respond to some of the most common deservingness-based arguments for and against government student loan debt forgiveness (described in detail in table 2).Footnote 24 The results for the following analysis are presented in table 4.
Effect of deservingness frames on support for debt forgiveness
Note: 1=strongly oppose, 5=strongly support
*p<.05, +p<.1; For the racialized need, competitive, and endorsement frames, significance indicators are listed relative to the control/relevant baseline treatment.
As evidenced by the examples from President Biden and Senator Cotton in the introduction, need and control arguments are some of the most prevalent types of elite messages both for and against debt forgiveness. Do they resonate with people? The need treatment employs a common pro-forgiveness message, explaining that student debt relief would help those who are struggling economically through no fault of their own.Footnote 25 As the results in table 4 show, however, this frame does not produce a shift in participants’ stated support for debt relief. This is especially notable for Democrats, who were most responsive to need in the conjoint. A slightly different result emerges with the control frame, which suggests that wealthy professionals will benefit most from student debt relief as a way to oppose the policy. While this negative frame does not produce a shift in the full sample, nor does it move Democrats, it does lead to a marginally significant decrease of four-tenths of a point in reported support among Republicans. On one hand, this partisan result is surprising given that the concept of control resonated with both Democrats and Republicans when it came to evaluations of individual borrowers, suggesting we might see uni-directional movement from both groups in response to the frame. However, it is consistent with the notion that reciprocity is especially critical to Republican evaluations of deservingness. Notably, when the two arguments are placed in a competitive frame, there is no effect on people’s support for debt relief.
Another increasingly common argument in support of student debt relief primes racial identity alongside need, noting that debt forgiveness would be especially meaningful for borrowers of color. Footnote 26 The racialized need treatment combines the broader positive need-based argument with a racial deservingness appeal. Recall that racial minorities, and especially Black and Latinx borrowers, were deemed more deserving of debt relief when compared to white borrowers by participants in the conjoint. Yet priming those racialized identities in deservingness frames to support debt relief does not produce the same result. Democrats are no more likely to support federal efforts to forgive student loan debt after receiving this argument, and Republicans are about one-third of a point less supportive relative to the control—a statistically significant result. This is a striking outcome that is at odds with the role race plays in individual-level evaluations of deservingness, particularly for Republicans. It is, however, consistent with our intersecting partisan and race hypotheses.
Finally, in an effort to explore how partisan identity might influence people’s responses to deservingness arguments, both the control and racialized need treatments are replicated in the context of Democratic and Republican endorsement experiments respectively, drawing on actual quotes from party leaders. Democrats do not move in either direction in response to either elite frame. Republicans, however, continue to be moved by the endorsements: relative to the control, Republicans are statistically less supportive of debt forgiveness when receiving both the Democratic endorsement in favor of loan forgiveness (which primes need in both a race-neutral and a racialized context) and the Republican endorsement in opposition to forgiveness (which primes lack of need and personal responsibility for debt). Interestingly, however, the endorsements do not produce statistically distinct results from the baseline control and racialized need treatments, suggesting the partisan endorsement is not doing as much work to move opinion among Republicans as are the deservingness arguments.
Discussion and Conclusion
At the time of this writing, Biden’s student debt relief proposal is on hold while the Supreme Court decides its fate, with a decision expected in June of 2023. Pending that decision, federal student loan payments are set to resume from their pandemic pause in August 2023. Thus, student debt relief remains a salient policy issue for millions of American borrowers. Our study offers some of the first systematic evidence to help understand what shapes support for federal student loan debt relief. Our results suggest several relevant takeaways for both theory and practice: First, and perhaps most importantly, support for student debt relief at the critical juncture in which the survey was fielded in late spring 2022 was broad and was particularly strong for Democratic-leaning groups. While people may be able to distinguish between factors that make borrowers more or less deserving of debt relief, the median support for student debt forgiveness in our survey was very positive. So too was support for a plan to forgive student loan debt—and among Democrats, that support appears resistant to arguments that prime deservingness to oppose debt relief. This suggests that support for President Biden’s proposed student debt relief plan is robust, especially among key Democratic constituencies.
The second major takeaway is that evaluations of deservingness in the context of student loan debt replicate many, but not all, of the dynamics of similar evaluations in the social policy context. Consistent with existing literature, we find evidence to suggest that certain characteristics of borrowers that indicate need (e.g., occupation and amount of outstanding debt), control (e.g., type of institution attended), and reciprocity (e.g., employment status, time in repayment) can condition people’s evaluations of whether borrowers deserve loan forgiveness. We can see many of the elements coming to fruition in Biden’s debt relief plan. For example, placing an income cap on eligible borrowers and offering an additional $10,000 in debt relief to Pell Grant recipients are consistent with an emphasis on need as a qualification for federal assistance. Similarly, revisions to the income-based repayment program that reduce the debt burden for those who have been repaying consistently reflect the idea of reciprocity—awarding those who have “paid their dues.”
There were also small partisan differences in which attribute was most heavily weighted, with Republicans placing more consistent emphasis on reciprocity and control relative to need. We can observe elements of this division playing out in real time: The White House continues to frame Biden’s debt relief plan in terms of its disproportionate effect on households making less than $75,000 annually (White House 2022), while Republicans continue to frame their opposition in terms of both (lack of) reciprocity and control over one’s situation (see, for example, public statements made by prominent Republicans including Mitt Romney, Marjorie Taylor Greene, and others).
The final major takeaway is that, while deservingness attributes play a central role in how people evaluate individuals as worthy of government support—potentially shaping baseline predispositions to debt relief—those results may not translate to broader deservingness arguments for or against student debt forgiveness in a clear manner. This is most obvious in the case of race. While a person being identified as Black or Latinx increased how individual borrowers were rated as deserving of federal debt relief, priming racial group identity in messaging actually undermined support for debt relief among Republicans (while not moving Democratic voters). A similar disconnect occurred for the control and need mechanisms. This finding suggests that—at least for some subset of borrowers—evaluations of individual versus group-based deservingness may not be parallel and are worthy of additional research in future. It is also important to note that the lack of positive movement by Democratic participants in response to deservingness arguments at the policy rather than borrower level may be the result of a ceiling effect. Democratic support for student debt forgiveness is exceptionally high, leaving little room for positive movement.
There are, of course, limitations to the current study that offer opportunities for future work. In addition to those queries already mentioned, we might want to know more about whether specific policy details influence people’s evaluations of borrower deservingness for debt relief. Larger subgroup samples might also allow us to better capture variations in what drives perceptions of deservingness across groups. And, of course, we might be interested to know whether these attitudes change in response to Biden’s concrete plan for debt relief, especially as it becomes subject to greater polarization in elite discourse. Ultimately, however, our findings offer the first systematic evidence for how people’s perceptions of deservingness unfold for this newly salient policy area. The results suggest that, for Democratic elected officials, student debt relief may be both good policy and good politics.
Supplementary Materials
A. Survey Administration and Sample
B. Additional Details of Conjoint Design and Analysis
C. Additional Details of Open-Ended Coding
D. Tabular Results for Figures
E. Supplemental Analysis of Student Borrowers
F. References
To view supplementary material for this article, please visit http://doi.org/10.1017/S1537592723001457.
Acknowledgments
The authors thank Abigail Williamson, Delphia Shanks, Sarah Maxey, panel participants at the 2022 Annual American Political Science Association and Midwest Political Science Association meetings, and the anonymous reviewers for their generous feedback on this article. They also thank Trinity College and Duke University for providing faculty research funds for the project.
Open access
