Overindebtedness and debt relief: Evidence from a randomized experiment at the Swedish Enforcement Authority

The intervention consists of one attempt at contacting the treated individual by case handlers at the Swedish Enforcement Authority (SEA). The aim of the contact is to provide information about debt relief and to incentivize the individual to apply. The main priority is to make contact by phone. If this is not possible, the individual is sent a letter and (if possible) a text message.

In Sweden, debt relief is a five-year program during which the debtor pays as much as possible of their debts conditional on upkeeping a minimum standard of living. Afterwards they are relieved of all debt, except for mortgage debt corresponding to the value of the property and debt pertaining to alimony. Compared to salary garnishment, the maximum amount that the debtors may keep for monthly subsistence during debt enforcement is the same. However, the debt relief program offers two months every year without payment of debts, while salary garnishment gives no such breaks. This implies that not only may regular debt enforcement last for much longer than the five years of debt relief, it is also more severe.

About 400,000 individuals are currently subject to debt enforcement by the SEA. According to SEA estimations, as many as 70,000 of these individuals may qualify for debt relief. The SEA has the ambition to incentivize these individuals to apply for debt relief. As a result, the authority has conducted a series of information interventions during 2019-2021. The strategy for these interventions was to select overindebted individuals from the debt enforcement register based on predetermined selection criteria that identified these individuals as likely to qualify for debt relief. The interventions consisted of a combination of attempts at contacting the debtor both in writing and by phone with the aim to provide information about debt relief and to incentivize the debtor to seek additional advice and to actually apply for debt relief. The case handlers at Kronofogden followed a protocol that consisted of three distinct steps: preemptive contact in writing, contact by phone, and follow-up contact by phone. These interventions and preliminary results are described in a separate document that is attached to this registration.

The current trial employs the same strategy while slightly altering the protocol for the intervention. The SEA aims to test a more focused intervention that includes only one attempt at making contact with the treated individual according to the following protocol:
1. If there is no telephone number: send information letter.
2. If there is one or more telephone numbers: make one attempt at making contact by phone. If there is no answer, send information letter and text message with information.
3. If there is contact by phone: Provide oral information about debt relief and how to apply.

The case handlers manually record all steps of the intervention protocol for each treated individual. They also record the total time they allocate to working with the information intervention. The identities of the case handlers are known.

2022-04-29

2022-05-31

APPLY: the individual applies for debt relief.
GRANTED: the individual is granted debt relief.

APPLY will be measured six months after the end date of the intervention. I.e., APPLY is an indicator for whether the individual applied for debt relief somewhere between 2022-04-29 to 2022-11-30.

GRANTED is an indicator taking the value of one if APPLY=1 and the application was approved and zero otherwise. That is, an application that is handed in between 2022-04-29 and 2022-11-30, but where a decision is made after 2022-11-30 will also be included.

Aside from the primary outcomes, we will also study other aspects, including

TERMINATED: the individual has been granted debt relief but has been terminated from the program during its five-year duration.
CONCLUDED: the indiviudal has been granted debt relief and concluded the five-year program.
EMPLOYED: the individual is employed.
INCOME: the individual´s annual income.
DEBT: the individual´s total debt subject to enforcement at the SEA.
MEDICINE: the individual is prescribed medicines that indicate various forms of somatic and mental illness.
HEALTHCARE: the individual has been subject to out- or inpatient care for conditions related to stress, mental illness, or substance abuse.

We have not completely decided on the secondary outcomes to study as it will depend on the strength of the primary outcomes (i.e., we can think of it as a "first stage". That is: If there is a strong effect of the experiment on the probability of being granted debt relief, then it is more likely that we can study labor market and health outcomes. The plan will be updated to specify more on the secondary outcomes.

Overindebted individuals subject to debt enforcement at the Swedish Enforcement Authority are included in the trial based upon pre-determined inclusion criteria that identify them as likely to qualify for debt relief and as having had recent contact with the authority. The individuals are randomized to treatment (n=2,000) and control (n=1,647) groups in five stratas corresponding to geographically organized debt enforcement divisions at the SEA.

The SEA intends to provide the information treatment to 2,000 individuals. This number is selected based upon considerations of available staff resources.

The population of individuals subject to debt enforcement at the SEA was 369,038 at the point in time for the application of the inclusion criteria (April 2022). In the SEA organization, these individuals belong to five geographically organized divisions. This means that each division handles debt enforcement of a specific share of the total number of debtors according to the following table:

TABLE 1: Share debtors of population by debt enforcement division at the SEA
Division, Share
1, 28 %
2, 12 %
3, 23 %
4, 19 %
5, 17 %

Overindebted individuals that may qualify for debt relief are selected from the population based upon the following inclusions criteria:
1. Individual with Swedish personal identification number
2. Not deceased
3. Not personal bankruptcy (swe. konkurs)
4. Not prohibition to conduct business (swe. näringsförbud)
5. Not subject to intermittent sequestration (swe. kvarstad)
6. Not handled by special debt enforcement section
7. Private debt > SEK 90,000
8. Share of private debt pertaining to damages pertaining to criminal offence < 30 per cent
9. Not applied for debt relief during 2021-2022
10. Not included in the information intervention conducted in 2021
11. The debtor either has no other recorded properties by the SEA than bank account, equity fund, listed shares, or car or other vehicle with a value that is too low to motivate attachment.
12. The debtor is not subject to salary garnishment.
13. The SEA has previously investigated the debtor with the aim to conduct debt enforcement but found that there is not enough attachable properties to pay the debts on at least one occasion (swe. utredningsrapport).
14. There is recorded contact with the debtor since July 2020.

The application of the inclusion criteria on the total population of individuals subject to debt enforcement identifies 3,647 individuals. 2,000 individuals are randomized to treatment in five stratas corresponding to the debt enforcement divisions at the SEA. The remaining individuals are assigned to the control group. Each division is assigned a specific number of individuals in the treatment group based upon their shares of the total population according to Table 1. This means that each division handles only debtors that belong to their share of the total population. The debtors are assigned to the divisions based upon their residency, i.e., where they live.

However, inclusion criterium no. 14 is based on a specific registration in the SEA database regarding contact with the individual subject to debt enforcement. The application of this registration in day-to-day debt enforcement operations is unevenly distributed across the five divisions according to Table 2.

TABLE 2: Share debtors with registered contact by debt enforcement division at the SEA

Division, Share debtors with registered contact
1, 16 %
2, 14 %
3, 17 %
4, 33 %
5, 31 %

Table 2 shows that the share of individuals with the contact registration in the database is approximately 30 per cent for two of the debt enforcement divisions, while it is approximately 15 per cent for the remaining three. This is explained by biased implementation of the case handling protocol for this specific registration across the divisions. Accordingly, the stratas in the control groups corresponding to the two divisions with greater propensity to register contact with the individual subject to debt enforcement, are larger compared to the other three divisions as shown in Table 3.

TABLE 3: Number of debtors that fullfill inclusion criteria and numbers of debtors in the treatment and control groups by debt enforcement division at the SEA
Division, Debtors that fullfill all inclusion criteria, Debtors in treatment group, Debtors in control group
1, 711, 568, 143
2, 258, 248, 10
3, 467, 464, 3
4, 1,234, 377, 857
5, 977, 343, 634
Total, 3,647, 2,000, 1,647

Randomization done in Excel

Individual

No

0

3,647 individuals

2,000 individuals in treatment group and 1,647 individuals in control group

For APPLY. If the baseline application rate (in the control group) during six months is 10 percent, then there is 80 percent power for rejecting the sharp null when the treatment effect is a little more than 3.5 percentage points (i.e., an application rate of more than 13.5 percent in the treatment group. See attached analysis plan.