Column 2 addresses the role that debt attitudes play in accounting for student loan take-up

For an average individual, a one-unit change in debt attitude increases the probability of taking out a student loan by 0.8 percentage point in model 2, when all other variables are included. The effect decreases when adding debt avoidance mechanisms, which is probably due to debt averse students using these mechanisms to shun loans.

Finally, column 3 adds both living at home and working during term-time to the model, behaviours that could be aimed at reducing or completely avoiding debt. These two variables are the outcome of e time as the decision on student loans. They can be inputs or outputs of the decision-making process. They are not separate exogenous factors. Nevertheless, model 3 is https://getbadcreditloan.com/payday-loans-mi/benton-harbor/ informative about whether these are negatively associated with loan take-up, and hence whether these can be seen as debt avoidance mechanisms. Students who always live at home while studying have probabilities of taking out student loans that are 11.5 percentage points lower than those of their peers who never lived at home.

Our estimates show that living at home is indeed negatively associated with loan take-up, but working during term-time is not

small business personal loans

Using a bivariate probit regression, the second analysis (shown in Table 3) estimates two probit models simultaneously to analyse the take-up of tuition fee loans and maintenance loans. This estimation procedure allows for the possibility that unobserved factors might affect the take-up of both types of loan. Allowing the residuals to be correlated can lead to a statistically more efficient estimation. We fit this model under the hypothesis that the decisions to take out tuition fee and maintenance loans are taken simultaneously by the student. This hypothesis is confirmed by the significant correlation of the errors, as shown by the athrho (the Fischer z transformation of the correlation) in Table 3. Models identical to that in Table 2 are evaluated simultaneously for tuition fee loans (Panel A) and maintenance loans (Panel B). The same variables are included in all equations to assess whether they have different effects depending on the type of loan. Estimates are reported for model 2, except when discussing debt avoidance mechanisms.

This is a substantial effect size, revealing living with parents as an important mechanism to avoid student loans

A slightly different picture emerges when it comes to deciding to borrow for tuition fees or for maintenance. Students whose family owns their home outright, who live in less-deprived areas and whose parents earn more are less likely to borrow money for both purposes. In both cases, family socio-economic background does not play a role. While family’s highest educational level is unrelated to borrowing for tuition fees, it is linked to maintenance loans except when debt avoidance mechanisms are added. This supports our former assumption of greater geographical education mobility among the children of more highly educated parents and their need to borrow to afford to live away from home.

The gender differences observed in the probit model hold for both types of loans, although effect sizes are larger for tuition fee loans. Ethnicity, however, does not play a role in the probability of taking out tuition fee loans, except for students of Indian origin. Indian students are the only ethnic group differing from White students when it comes to both types of loans, Footnote 7 although this effect disappears for tuition loans and diminishes for maintenance loans when adding debt avoidance mechanisms. This last result is probably explained by the higher propensity of Indian students to live at home and therefore not to need maintenance loans. Similarly, religion is a factor for both types of loans, with Muslim students less likely to borrow.