 |
 |
 |
 |
 |
|||
 |
|||
 |
|||
 |
 |
 |
 |
 |
 |
 |
 |
 |
 |
 |
 |
 |
 |
 |
 |
 |
 |
 |
 |
 |
Methodological Deficiencies: Misspecification or Misdirection? Professors Barron and Staten contend that a 'credit card debt crisis' among college students is "exaggerated" (p. 25) even though their own data clearly show that college students are significantly more likely than "older adults" to be "delinquent (12.1% versus 8.1%) and "seriously delinquent" (3.1% versus 1.1%) in paying their credit card accounts. The key question is whether this conclusion is based on objective empirical "facts" or, instead, is a statistical artifact of their research design. More specifically, did the authors purposively select a research methodology that appears to be "value neutral" to the uninformed reader whereas, in actuality, they systematically manipulated the analytical focus in order to statistically attenuate patterns of student debt accumulation? The most salient methodological deficiencies of their research design are briefly reviewed below. First, the analytical focus on "universal" bank credit card accounts deflects attention from the appropriate unit of analysis. That is, individual college students in their institutional context with specified controls for variations in cost of living (geographic region, urban/suburban) and educational expenses (public/private). The authors' deftly attempt to justify their research design by basing their decision on the U.S. General Accounting Office's (GAO) 2001 report on student credit card use. Although Barron and Staten imply that an "objective" analysis of student/nonstudent credit card accounts was a key objective of the GAO study upon its inception, this belies the reality of the initial GAO research agenda When time and resource constraints led the GAO to abandon both the national student survey and credit card account projects, Dr. Staten explained in his 2002 U.S. Senate testimony that a grant from "credit card companies" financed the completion of the authors' study (Staten, 2003, p. 26). It was subsequently released by the Credit Research Center (Staten and Barron, 2002) and then revised for the JSFA. Interestingly, the GAO's 2001 College Student and Credit Cards report complains about the lack of cooperation of the credit card industry in providing requested information. According to Harvard Law Professor Elizabeth Warren, this response is consistent with the credit card industry's strategy of rewarding "friendly" researchers with access to proprietary data while denying access to objective or "unfriendly" scholars (Warren, 2002). Second, the exclusive focus on credit card accounts diverts attention from the complex dynamics of student credit card debt and, ultimately, the estimation of total credit card-related indebtedness. In particular, it fails to address the institutional and social context that explains the relatively low delinquency rate of student card accounts. These include access to federal and private student loans, consumer loans from college credit unions or banks, household/ student college savings (especially freshman year), informal family loans, and even other credit cards. Indeed, one of the most popular debt survival strategies is to allocate student loan monies for full or partial payment of outstanding credit card balances. For instance, in our 2002 survey of a mid-sized public university (Manning et al, 2002), more than two-thirds (68.3%) of students with student loans reported using some of these funds for payment of their credit card bills (Manning, 2003, p. 53). Hence, students often rotate their "revolving" card debt into college loans which highlights the problematic classification of student debt into mutually independent and discrete categories. Third, Barron and Staten fail to track the use of student credit cards beginning with the first credit card through the end of college. Indeed, the overwhelming proportion of students with bank credit cards have received their first card by the fall of their freshman year (Jamba-Joyner, 2000; Student Monitor, 2001; Manning et al, 2002; Nellie Mae, 2002; Ohio State University, 2002; Tan, 2003). This means that the typical college student has more time to accumulate consumer debt--over their entire academic career—than earlier student cohorts who typically obtained their first credit card in their junior or senior year. It also means that educational interruptions due to consumer debt are more likely to produce career threatening consequences if they occur early in a student's academic career. As a result, it is necessary to examine the borrowing patterns of individual students throughout their undergraduate years in order to understand the cumulative social and economic consequences of credit card debt. Clearly, empirical models of college credit card "usage" must include all credit card accounts of each student; note, Barron and Staten do not include retail credit cards in their analysis nor do they examine card usage patterns of students with multiple accounts. This requires a longitudinal analysis that examines total credit cards and aggregate revolving debt by class standing (freshmen through senior years). For example, student credit card debt is strongly associated with length of time of card accounts as well as number of credit card accounts. This is confirmed by all extant survey studies (Institute for Higher Education Policy, 1998; Jamba-Joyner, 2000; Pinto et al, 2001; Student Monitor, 2001; Manning et al, 2002; Nellie Mae, 2002; Norvilitis and Maria, 2002; Ohio State University, 2002; Norvilitis et al, 2003; Tan, 2003, Gnizak et al, 2004; Mattson et al, 2004). Yet, the Barron and Staten study does not statistically control card usage patterns by class standing. In fact, the median age of "student" accounts at the end of the 12 month observation period of the CRD Pooled Sample is 21.9 years old—a common age at graduation (pp. 11-12). This suggests that the sample virtually excludes traditional college freshmen, includes few college sophomores, overrepresents juniors and seniors and, significantly, features a large number of students whose age suggests that they have either graduated or dropped out of college. Note, the authors do not contend that their sample is representative of students' use of credit cards but rather is representative of "active credit card accounts" based on their sample selection and statistical "weighting" protocols (p. 9-10). This is not a statistical nuance since the data may exclude over one-half of the population universe of undergraduate students (freshmen and sophomores). Furthermore, the misclassification of many "student accounts," that may actually be used by recent college graduates, would substantially reduce the actual student delinquency rate by increasing the size of the subgroup and the proportion of wage-earning accountholders. It could also explain an interesting anomaly in the account usage patterns--the large number of inactive accounts. Indeed, many students retain their first credit card after graduation as a collegiate memento but may use it infrequently since it does not offer competitive interest rates or a sufficient line of credit. The overrepresentation of juniors and seniors in the CRC sample is especially significant since these student subgroups are most likely to have multiple credit card accounts. For instance, in our 2002 study of a public university in Virginia, only 13.7 percent of all freshmen reported two credit cards and only 5.2 percent reported 3 or more. This climbs to 29.8 percent of all seniors with two credit cards and 27.3 percent with 3 or more cards. When this distribution of credit card accounts is calculated among only student cardholders, which would be comparable to Barron and Staten's analysis of their "student accounts," the proportion of seniors with 2 credit cards rises to 33.8 percent and those with 3 or more is 31.0 percent (Manning, 2003, p. 51). In Mattson et al's study (2004) of a Wisconsin university, the proportion of students with two or more credit cards jumped from 11.2 percent of all freshmen (42.8% possessed at least one credit card) to 53.0 percent of all seniors (88.2% percent possessed at least one). Similarly, the statewide study of student credit card usage in Oklahoma reported that 33 percent of all students had 4 or more bank credit cards (43.8% of student cardholders) plus 2.0 retail cards, 1.4 gasoline cards, and 1.2 revolving lines of bank credit (Tan, 2003, p. 12). Although Barron and Staten challenge the national representativeness of the 2001 Nellie Mae survey, since its sampling framework is restricted to student loan borrowers, these findings are clustered at the high end of expected number of credit card accounts. For example, the average number of credit card accounts rises from 2.5 during freshman year including 26 percent with 4 or more to 6.1 in senior year including 66 percent with 4 or more (Nellie Mae, 2002, p. 3). Unquestionably, students tend to accumulate more credit cards with higher outstanding balances as they approach the collegiate finish line of graduation. Regardless of the cost of living and educational expenses, students tend to double their outstanding credit card debt between their freshman and senior years. In Gnizak et al's 2002 study of a Western Kansas university, average credit card debt climbed from $174 among freshmen to $1,859 among seniors. Mattson's 2002 study in Wisconsin reports the percentage of students with monthly balances of more than $1,000 rose from 4.3 percent among freshmen with a high of $8,115 to 18.5 percent among seniors with a high of $13,760. In the 2002-03 Oklahoma survey, average credit card debt jumped from $2,077 among freshmen to $3,559 among seniors. In the 2001 Nellie Mae survey, average credit card debt doubled from $1,533 among freshmen (12% with balances over $3,000) to $3,262 among seniors (40% with balances over $3,000). Furthermore, the U.S. Department of Education's 1999-2000 National Postsecondary Student Aid Study found that 41 percent of graduating seniors had a credit card balance that averaged $3,071. Among student loan borrowers, 48 percent of graduating seniors had a credit card balance that averaged $3,176 (King and Bannon, 2002). Not incidentally, these credit card debt statistics tend to be underestimated by respondents and do not include past credit card debts that were paid with student loans, family loans, or other bank consolidation loans. Consequently, individual student account data cannot possibly capture the larger dynamics of college consumer debt patterns. The most appropriate methodology is a cohort analysis that tracks student credit card usage and other consumer borrowing patterns from freshmen orientation through graduation (Manning 1999; Manning and Smith, 2005). It also has the advantage of recording student attrition "events" which can then be classified by social, economic, or academic factors.
|
|||
