The Causal Impact of the Judiciary on Economic Activity: Evidence from a Field Experiment in Kenya

Last registered on May 24, 2016


Trial Information

General Information

The Causal Impact of the Judiciary on Economic Activity: Evidence from a Field Experiment in Kenya
Initial registration date
May 24, 2016

Initial registration date is when the trial was registered.

It corresponds to when the registration was submitted to the Registry to be reviewed for publication.

First published
May 24, 2016, 1:20 PM EDT

First published corresponds to when the trial was first made public on the Registry after being reviewed.


Primary Investigator

McGill University

Other Primary Investigator(s)

Additional Trial Information

On going
Start date
End date
Secondary IDs
This paper presents the pre-analysis plan of a randomized intervention providing free legal representation to a treatment group of small scale farmers in a rural community of Kenya. Our hypothesis is that legal representation will increase access to the judiciary, the security of property rights, investment, agricultural production, and thus decrease poverty.
External Link(s)

Registration Citation

Chemin, Matthieu. 2016. "The Causal Impact of the Judiciary on Economic Activity: Evidence from a Field Experiment in Kenya." AEA RCT Registry. May 24.
Former Citation
Chemin, Matthieu. 2016. "The Causal Impact of the Judiciary on Economic Activity: Evidence from a Field Experiment in Kenya." AEA RCT Registry. May 24.
Experimental Details


To measure the causal impact of access to the judiciary on economic growth, a treatment group is provided with free legal representation. The hypothesis is that, with better access to the judiciary and thereby more secure property rights, the small-scale farmers of Kianyaga will increase their agricultural productivity and exit poverty.
Intervention Start Date
Intervention End Date

Primary Outcomes

Primary Outcomes (end points)
Our questionnaire includes sections on:
-Formal sector employment and income
-Casual work for others (and income)
-Land characteristics (title,…)
-Plot characteristics (property rights, informal protection)
-Land improvement
-Land inputs (manure, fertilizer, pesticide, seed, mechanical input, fencing (informal protection)
-Days worked on plot (=investment)
-Agricultural production
-Trust and safety (in homestead, of other people)
-Justice system perception (+scenario and expected costs)
- Legal knowledge
- Risks and worries
-Stress and happiness measures
-Expectation of land holding in future (security of property rights)
-Basic ses: sex, age, education
Primary Outcomes (explanation)
For each hypothesis, the following section lists the specific indicators to be analyzed. For some hypotheses, this section identifies a primary set of indicators that most directly relate to and are most likely to be impacted by the project, and also lists a more speculative secondary set that tie less directly to program objectives to be tested separately. Note that the analysis will be conducted only if there is sufficient variation in the data.

H1: Legal aid increases use of the judiciary


• How many times did you meet with a lawyer?

• Case filed in court?

• Outcome reached? (1=yes, whatever it is, 0=no -9999) (all from legal aid database)

Secondary 1: legal aid improves perception of the judiciary:

• F1043: “I am confident that the legal system will uphold my contracts and property rights in a land conflict.”

• On a scale of 1 to 5, in resolving land conflicts, do you believe your country’s court system to: f1037 Be fair and impartial f1040 Be affordable f1038 Be honest and incorrupt f1041 Be consistent f1039 Be quick (deal with cases in timely manner) f1042 Have its decisions enforced

• F1058 “In the last three years, access to the courts has increased”

• F1069-F1076: Perceptions of costs associated with the judiciary should decrease.

Secondary 2: Treated individuals should use more formal means of conflict resolution, while decreasing the use of informal (especially violent) mechanisms

• The responses for “Codes for Protecting Land” to questions f757-f762 (Have you done anything to protect your land from land grabbing and/or theft?”) should change: From 00. Nothing 01. Plant trees on boundary 02. Put up barbed wire/mesh wire 03. Put up electric fence 04. Get guard dog(s) 05. Fence 06. Hire security guard 07. Activate a security light during the night 09. Use witchcraft to protect land 10. Other (Specify): ; To 08. Use title deed

• The responses for “In the past 10 years: Did you encounter any of the following conflicts? What did you do to resolve the conflict?”. Answers to questions f986, f997, f1008, f1019, and f1030 should change as follows: From 00. Do Nothing 01. Private negotiation 02. Block leader/Elders 03. Chief/Sub-Chief 04. Mob justice 05. Hire vigilante 08. Go to witch-doctor 10. Harvest your crops earlier/faster 11. Go to influential friends 12. Other (Specify): To 06. Land tribunal 07. Go to law courts 09. Go to police

• Responses to hypothetical scenarios f1059-f1068 should change as follows: “Imagine that you have the title deed to your shamba, and then your neighbor attempts to grab the land in your shamba. What would you do to try to stop your neighbor from grabbing your land?” From 00. Do Nothing 01. Talk to the land grabbing neighbor 02. Call your friends/family to come and help you 03. Hire a witch-doctor to curse your neighbor 04. Talk to village elders 05. Talk to block leader 06. Talk to sub-chief/chief 08. Hire a justice vigilante 10. Try and incite mob justice/use violence 12. Go to a district official (but not District Commissioner) 13. Go to District Commissioner 16. Go to clan elders 17. Go to Provincial Commissioner 18. Go to a member of Parliament 20. Other (Specify): ; To 07. Go to a land tribunal 09. Go to the police 11. Go to the judiciary/courts 14. Go to land registration office 15. Contact a land surveyor 19. Contact an advocate

Secondary 3: Legal aid reduces perception of need for “side payments”

Questions f1044-f1057

Secondary 4: Legal aid improves legal knowledge

Questions F3186-F3200

H2: Legal aid increases security of property rights


• Expectations that “the size of future land holdings due to conflict and/or land distribution will decrease” should decrease for the treatment group (F3125-F3148)


• Do you have the right to sell/rent/take loans/cultivate/leave in a will/gift this plot without/with approval from your family? (F3149-F3185)

• Treated households should make greater use of formal title deeds (F738: “Into how many formal title deeds is this homestead divided?”

• F757-762: “Have you done anything to protect your land from land grabbing and/or theft?” (the answer “use title deed” should appear more in treated households)

• f773 Fencing (in ksh) should decrease

• Safety: f1088 How safe do you feel outside of your homestead during the night? f1090 How safe do you feel inside your homestead during the night? f1089 How safe do you feel outside of your homestead during the day? f1091 How safe do you feel inside your homestead during the day?

H3: Legal aid increases agricultural investment

• Variables f3160 – f3181 in the “Land Improvements” section

• Section Land inputs as well: • F766: animal manure collected • F767: manure purchased • F768: chemical fertilizer • F769: seeds • F770: pesticides • F771: mechanical inputs • F772: other non-labour farm inputs

H4: Legal aid increases access to credit

Increased security of property rights should lead to an increase in the use of property as collateral and hence and increase in access to formal lines credit.

• We would expect to see significant increases in the following formal means of financing: • F3055-3060: employer or landlord • F3069-3074: microfinance institution • F3083-3088: bank or government agency • F3111-3116: SACCO

• Conversely, legal aid should have no significant impact on informal means of lending: • F2999-3004: family member from same HH • F3013-3018: family member from different HH • F3027-3032: friend from same village • F3041-3046: ROSCA / Merry GO Round

H5: Legal aid increases agricultural production

• There should be significant increases in the selling of agricultural commodities (both at market and to brokers) in ksh as specified under the “Household Sales and Consumption” section f785-f957

• More agricultural inputs should increase the marginal product of on-farm labor. Hence, the amount of labor used on the farms of legal aid recipients should increase . This results in a greater number of working days each of the following individuals in the past month have worked have done on the farm: • F774: household head (HHH) • F775: spouse of HHH • F776: HHH’s children • F1195: HHH’s parents • F777: HHH’s in-laws • F1196: HHH’s siblings • F778: HHH’s siblings in law • F779: other family / relatives of HHH • F780: non-family hired labour • F784: friends • F781-782: Other

• Treated households should be significantly more confident that they will not experience a food shortage next year (f385-f386).

• Household identity: less casual work for others, less formal work, more work on own farm (less diversification of activities)

H6: Legal aid improves well-being, locus of control, and trust in others

• Legal aid recipients should experiences changes in measures from the PSS-10, ICD10, and WHO5 questioning corresponding to decreased feelings of depression and anxiety and increased loci of control.

• The following worries should also decrease: • F275: problems at home with relatives • F279: not enough money for food • F280: not enough money for education • F281: not enough money for living expense • F282: not enough money for medicines and medical treatment • F284: idleness of children or spouse • F286: debts owed to a family member • F287: debts owed to others

• f1087 Do you think most people would try to take advantage of you if they got a chance, or would they try to be fair?

• Response values measuring trust in the following groups of people should significantly increase (1. Distrust completely → 5. Trust completely): • F1092: nuclear family • F1093: friends • F1094: co-workers • F1095: people of another religion • F1096: rich people in community • F1097: poor people in community • F1098: political leaders • F1099: people from same clan • F1100: other clan • F1101: other ethnicities • F1102: policemen • F1103: government workers

Secondary Outcomes

Secondary Outcomes (end points)
Secondary Outcomes (explanation)

Experimental Design

Experimental Design
We recruit households involved in cases with legal merit, and willing to receive legal aid. Once a case is deemed justiciable, we randomize it with a 50/50 probability into either the treatment or control group.
Experimental Design Details
Providing free legal aid to the treatment group is the focus of the current project. A team of paralegals first identifies potential participants by visiting the universe of households door to door following existing maps, and asking if they are involved in legal conflicts and wish to receive legal aid. Individuals answering yes to both questions automatically enter the pool of participants. Additionally, each of these participants are asked if they have friends with legal conflicts. These friends are also visited to determine their eligibility in the project. We also visit church leaders, block leaders, chiefs, and village elders to provide us with a list people involved in conflicts. After receiving the contact information of these people, we then proceed to visit them at their homes and determine their eligibility (engaged in a conflict and desire to receive legal aid). Paralegals collect basic data on the case, and forwards the case to the office.

A lawyerThe correct word is advocate in Kenyan law, but we use lawyer out of simplicity. then determines the eligibility of the case together with the paralegals. Only cases that have legal merit are considered. Factors determining whether or not a case has indeed legal merit include: 1- the respondent’s understanding of the case; 2- the availability of supporting documents; 3- when the conflict arose (cases dating back to the 70s are excluded); 4- any actions the respondent has taken so far in resolving the conflict (to judge the pro activeness of people), and finally; 5- the nature of the conflict itself (Elimu does not handle criminal cases). Legal merit is judged by the paralegal in the field, and in a meeting with the lawyer and all paralegals present.

2.2 Randomization

Once a case is deemed justiciable, we randomize it with a 50/50 probability into either the treatment or control group, without any stratification, on a purely individual basis, on the computer in the office of Prof. Chemin.
Randomization Method
we randomize justiciable cases with a 50/50 probability into either the treatment or control group, without any stratification, on a purely individual basis, on the computer in the office of Prof. Chemin.
Randomization Unit
The unit of randomization is the hosuehold.
Was the treatment clustered?

Experiment Characteristics

Sample size: planned number of clusters
Sample size: planned number of observations
The planned sample size is 344.
Sample size (or number of clusters) by treatment arms
The planned sample size is 172 in the treatment group, 172 in the control group.
Minimum detectable effect size for main outcomes (accounting for sample design and clustering)
Each participant in the treatment group of 172 individuals will receive legal aid from paralegals 2 days valued at 600 Ksh/day*2 days=1,200 Ksh, and from the volunteer advocate 52 days/172 particpants= 0.3 days at 6200 Ksh/day, hence 1,875 Ksh. This corresponds to an overall aid of 3,075 Ksh. Baseline data indicates that yearly household expenses are 78048 Ksh (approximately 1 CAD/capita/day). We are thus interested in detecting an increase of at least 3,075 Ksh on household expenditure, otherwise the benefits, despite being statistically significant, would be less than the costs of the intervention. Statistical power calculations indicate that this sample size is enough to detect such an increase in household expenditures.

Institutional Review Boards (IRBs)

IRB Name
McGill University Ethics Review
IRB Approval Date
IRB Approval Number


Post Trial Information

Study Withdrawal

There is information in this trial unavailable to the public. Use the button below to request access.

Request Information


Is the intervention completed?
Data Collection Complete
Data Publication

Data Publication

Is public data available?

Program Files

Program Files
Reports, Papers & Other Materials

Relevant Paper(s)

Reports & Other Materials