The economics and public policy of student loans in Korea

Transcription

1 The economics and public policy of student loans in Korea Bruce Chapman Dung Doan New approaches to tax and welfare in Australia and Korea workshop 9 November 2017

2 Outline The economics and public policy of student loans Why are student loans from government necessary? What are TBRLs and what are their implications? What are ICLs and what are their implications? International experience of repayment burdens: TBRLs vs. ICLs International experience of defaults: TBRLs vs. ICLs Understanding repayment burdens of student loans in Korea Repayment burden of TBRL

3 The economics and public policy of student loans

4 Why are student loans needed from the government? The investment is risky: non-graduation, poor jobs early Poor outcomes for students/graduates can lead to default With no saleable collateral banks are unprotected Two types of loans: Time-Based Repayment Loans (US, Canada, Colombia, China, Japan) Income-Contingent Loans (Australia, New Zealand, England, Hungary)

5 What are TBRLs and What are their implications? Major point relates to collection: TBRLs have repayment obligation fixed over time Thus TBRLs can mean: Repayment difficulties: consumption hardship, repayment anxiety Defaults (credit reputation loss) Loss of government revenue High collection inefficiencies

6

7 What are ICLs and what are their implications? Major point relates to collection: ICLs have repayment obligation which depend on income, not time Repayment only if a debtor s income exceeds a certain level Thus no repayment when unemployed, in low paying job Therefore: No repayment hardships No default Further: Highly efficient if collected by employers

8 TBRLs: International experience of repayment burdens The critical role of repayment burdens in TBRLs: Repayment burden i = Loan repayment i Income i

9 Per Cent Repayment burdens for bottom 15% of graduate incomes in countries with TBRLs ? 20 0 Colombia China indonesia Ireland USA Malaysia Vietnam Japan

10 ICLs: International experience of repayment burdens (max percent) Australia: 8 NZ: 10 England: 9 Hungary: 6 Thus, RB risks for borrowers a zero concern for ICLs

11 International experience of defaults (%) TBRLs ICLs US: Colombia: Thailand: Malaysia: Australia: 0 England: 0 New Zealand: 0 Canada: 15 Default risks for borrowers is a major concern Default risks for borrowers a zero concern

12 Images of TBRLs and ICLs:

13 TBRLs

14 ICLs

15 An ideal student loan system? ICL are much preferred to TBRL (RB problems) for debtors ICL are preferred to TBRL for government (likely higher revenue) Employer with-holding is very efficient and properly targeted Universality is critical to maximise government revenue Mandatory ICL minimises adverse selection

16 Understanding repayment burdens of student loans in Korea

17 Preliminary steps to understand Korea s student loans Consequences of high repayment burden To debtors: Consumption hardship Repayment anxiety Default and credit reputation damage Discourage fertility decision To government: High administrative cost to collect repayment Loss of revenue due to defaults Repayment burden i = Loan repayment i Income i Loan design: loan size Interest rates duration of repayment deferment period Life-time income of university graduates

18 Korea s student loans TBRL ICL Avg. loan size (2016) 33 mil KRW 22 mil KRW Interest rate 2.25% (fixed) 2.25% (floating) Duration of repayment Max 10 years of grace period (interest payment only) + Max 10 years of repayment (interest and principle) Until pay off the debt or retire Repayment amount Depends on interest rate, duration, and loan size 0 if annual income below mil KRW 20% of annual income above mil KRW

19 Estimate life-time income profile To understand RB at each point in life of a debtor: need to estimate income by age of university graduates Conventional OLS regression is not helpful because it only predicts the mean income among graduates at a certain age Low-income graduates are most prone to repayment hardship Income distribution at each age is more helpful to understand RB Solution: calculate after-tax income quantiles conditional on age by gender, then smooth them by age polynomials (quintic)

20 Life-time income of Korean graduates KLIPS 2016 data, representative of urban population (83% of total population) Consider only university graduates aged (2,382 obs) Male Female After-tax annual income ('0000 KWW) Age in survey year Age in survey year Male Female

21 Time-based loan, without grace period Heavy repayment burden (and high default risks) for almost all young graduates below 26 years old, except top 20% of males Impossible for bottom 20% of females to repay Female - no grace period Male - no grace period 100% 100% 80% 80% 60% 60% 40% 40% 20% 20% 0% % Q1 Q2 Q3 Q4 Q1 Q2 Q3 Q4

22 Time-based loan, with 10-year grace period Does grace period help ease the burden? not really! Even interest payment causes hardship to B40% of males and B20% of females Extremely high default risk for those below 27 years old. After the grace period, RB still exceeds 40% for the poorest 40% of females Given relatively low reported delinquency rate, how did they pay? Female - 10 yrs grace period Male - 10 yrs grace period 100% 100% 80% 80% 60% 60% 40% 40% 20% 20% 0% % Q1 Q2 Q3 Q4 Q1 Q2 Q3 Q4

23 % of graduates below 30 yrs old Who are zero-income young graduates? Large portion of graduates below 30 years old reported zero income in the last calendar year Most are single, living with at least another adult who earns income Potential financial support from family to repay student loans 100% 80% 60% 64.6% 53.0% % 40 20% 35.5% 47.0% 20 0% Female Zero income Male Positive income 0 Live with other incomeearning adult(s) Zero income Single Positive income

24 Time-based loan: repayment burden revisited Assume household income sources are pooled and shared among household members Low-income debtors might receive additional funding to repay their debts RB with respect to income per capita is much lower for all quantiles But still approx. 40% for debtors living in the poorest 20% of households Per capita income - 10 yrs grace Per capita income - no grace 50% 60% 40% 30% 20% 10% 50% 40% 30% 20% 10% 0% % Q1 Q2 Q3 Q4 Q1 Q2 Q3 Q4

25 Time-based loan: repayment burden revisited Lower RBs when accounting for economies of scale Adult equivalence scale: first adult=1, each other adult=0.7, each child under 14=0.5 Important to account for family support to debtors and intra-household reallocation of resources RB with respect to adult-equivalence income is sensitive to equivalence scale Future area of research: RB w.r.t consumption Per adult equivalence income - 10 yrs grace Per adult equivalence income - no grace 40% 50% 30% 20% 10% 40% 30% 20% 10% 0% % Q1 Q2 Q3 Q4 Q1 Q2 Q3 Q4

26 Appendix

27 Policy implications for Korea s student loan system The current TBRL creates heavy repayment burden to low-income debtors, most of whom are below 30 years old ICL is currently limited only to students from low-income families, who are more likely to earn less than those from richer families in the future Co-existence of both ICL and TBRL causes adverse selection Eligible students who expect low future income are more likely to chose ICL

28 Outstanding amount ('mil KRW) Outstanding amount ('mil KRW) ICL: low RB, but substantial subsidy Bottom 20% of male and bottom 80% of female debtors will never pay off 40 ICL outstanding amount - Male 80 ICL outstanding amount - Female Age Q1 Q2 Q3 Q Age Q1 Q2 Q3 Q4

29 Average student loan size Actual loan supply 2016 Average loan per student per semester (mil KRW) Average loan after 4-yr degree (mil KRW) ICL DL Amount ('00 mil KW) 11,983 No. of students 470,375 Amount ('00 mil KW) 7,145 No. of students 186, Source: Assume fixed interested rate of 2.5% - same as the actual interest rate on the time-based loan as of 2016 (second semester)