Nudging Good Politicians: How Conditional Incentives Impact Political Selection

Workshop: The Foundational Training for Aspiring Young Politicians workshop will be held in Sorsogon City, the namesake capital of the Province of Sorsogon. It will be an all-expense-paid 3-day workshop. The workshop will be held in 4 batches. In 2 of these batches, the workshop will have a conditional incentive, while in the other 2 batches the workshop will have an unconditional incentive.

The workshop will be conducted by a hired consultant, Alvin Samson. The PI will also be present at all workshops to oversee the whole event. Alvin Samson will be assisted by the study team members that will serve as workshop facilitators. Subjects selected to attend the workshop will be asked to bring along with them signed assent to participate in the workshop and the parent's permission letter to participate in the workshop with their parent's signature. To ensure that the workshop runs smoothly and safely, we will ask the subject and the parent to read and agree to the house rules of the workshop.

Workshop Content: The workshop is entitled "LEAD2SERVE SUMMIT: Foundational Training for Aspiring Young Politicians," which I am developing in close consultation with Alvin Samson, the main workshop facilitator. The Lead2Serve Summit provides aspiring young leaders (ages 15 to 17) a shared platform to interact with each other, to be grounded on leadership principles that found application both in the corporate and public sectors, and to evolve a plan of action that they can readily implement and deploy in their immediate community in the context of their prospective roles as elected youth council members.

The workshop is a combination of plenary sessions, individual activity and small group discussions, and structured learning exercises. (Please see attached workshop framework document in Docs & Materials tab). Workshop outputs are in the form of worksheets to be accomplished by participants in every individual activity and small group discussion.

Scoring: Participants will receive scores for their participation in each of the plenary sessions and for each of the worksheets that they accomplish and submit.
Scoring is simple. A participant will either receive a score of 1, 2, or 3 (3 being the highest score) for each of the following: Participation in Plenary 1, Participation in Plenary 2, Participation in Plenary 3, Worksheet 1, Worksheet 2, Worksheet 3, Worksheet 4

Participation in plenary sessions will be scored by the participant's Small Group Facilitator. Scores obtained in the three plenary sessions will then be averaged out. This average score from participation will carry a weight of 20% of the overall score. Worksheets will be scored anonymously. Every time worksheets are submitted, they will be randomly redistributed equally among the Small Group Facilitators. Worksheets will be redistributed for scoring twice, so that each worksheet will receive two scores from two different, randomly assigned Small Group Facilitators. The two scores will be averaged out.

Average score for worksheet 1 will carry a weight of 20% of the overall score.
Average score for worksheet 2 will carry a weight of 20% of the overall score.
Average score for worksheet 3 will carry a weight of 40% of the overall score.

Hence each participant will receive an overall score that would range between 1 and 3. (Please see attached workshop framework document in Docs & Materials tab).

Incentives: The workshop content will be the same across the 4 batches. This is because maintaining symmetry across the treatment groups is crucial for the science of the study. The only difference between workshops with conditional incentive and workshops with unconditional incentive is as follows:

In workshops with conditional incentive (Group B), participant output from and performance in the different workshop sessions will be assigned scores, although participants will not know during the workshop that performance is actually being monitored and scored to prevent Hawthorne effects (i.e., reactivity of respondents to knowing that they are being scored). A predetermined overall cutoff score will be set before any of the workshops. Participants that make the cutoff score will be
given an incentive at the end of the workshop during the closing ceremony.

The incentive is a certificate of merit, and 5 pieces of campaign posters that the Angara Centre will donate to the respondent if he/she subsequently files for a certificate of candidacy. In workshops with unconditional incentive (Group C), participant output from and performance in the different workshop sessions will also be assigned scores, although participants will not know during the workshop that performance is actually being monitored and scored. However, these scores will have no bearing, since at the end of the workshop during the closing ceremony, each and all participants will be given the same incentive as described in the preceding paragraph. That is, everyone will receive a certificate of merit and--should they choose to subsequently file an official certificate of candidacy--5 pieces of campaign posters each from the Angara Centre. Hence, the two types of workshops are symmetric in all ways but one. The difference between the two sets of workshops is in the manner by which the incentive is earned by the participants. Whereas in Group B, earning the incentive is conditional on making the score cutoff, in Group C the incentive is unconditional and awarded to everyone in the group.

At the end of every workshop, all respondents receiving the incentive will be awarded the certificate of merit. They will also be asked to inform the study team by text message or email if they subsequently decide to file their official certificate of candidacies so that the study team can prepare the campaign posters to be donated to them in time for the campaign period. Respondents will also be informed at the end of every workshop that to verify their candidacy, the study team will collect copies of COCs, which is publicly available from COMELEC.

The predetermined cutoff score / rule known only to the principal investigator is as follows:

For any given batch of the workshop, if the number of participants receiving a score of 2.25 or higher is less than 15, then participant scores will be rank-ordered
and the top 15 highest scoring participants will receive the incentive. If the number of participants receiving a score of 2.25 or higher is more than 30, then participant scores will be rank-ordered and the top 30 highest scoring participants will receive the incentive. If the number of participants receiving a score of 2.25 or higher is between 15 and 30, then all of them will receive the incentive. In short, for any given workshop batch, the minimum possible number of participants receiving the incentive is 15, and the maximum number is 30. If the workshop is one with conditional incentive, then the cutoff score / rule will matter. If, on the other hand, the workshop is one with unconditional incentive, then the cutoff score has no bearing since everybody in that workshop will receive the incentive.

2013-09-20

2013-10-13

The main outcomes of interest are: (1) subjects' subsequent decision to run or not run for office, (2) conditional on filing an official certificate of candidacy, their vote-share margin of victory, (3) a binary variable for winning the elective seat, and (4) conditional on winning an elective seat, subsequent performance in office measured in several ways including session attendance, legislations passed, projects delivered, funds raised, among others.

EDIT 08/05/14:
In October 2013, in the middle of the treatment interventions, Pres.Benigno Aquino III signed the law that postpones the SK ( Youth Council) Elections from October 28, 2013 to a date between October 28, 2014 February 28, 2015 (please see supporting document file, "Law Postponing SK Elections.pdf") Along with this law is the Implementing Rules and Regulations (please see supporting document file, "Implementing Rules of RA10632.pdf") that details the creation of a Task Force on Youth Development in the interim, when the SK offices are held vacant.

These changes imply that I cannot measure the original outcome of interest until after the new scheduled election date. In the interim (Agust 2014), I will conduct a follow-up survey that will measure the following:
(1) On a scale of 0-10 (10 being most likely), respondent's likelihood of running for youth council post
(2) On a scale of 0-10 (10 being most interested), respondent's interest in joining the Task Force on Youth Development
(3) An indicator variable for whether they got nominated to be part of the Task Force on Youth Development
(4) An indicator variable for whether they actually got designated as member of the Task Force on Youth Development.
And other measures that get at their political participation and interest in public service since they were exposed to the treatment interventions.

The design of the field experiment is as follows. Before a scheduled election, call for applications will be sent out to citizens who are contemplating on running for an elective office. The call for application is for an all-expense-paid workshop that would equip participants with knowledge and leadership skills useful in campaigning and in policymaking. There will be limited slots and so interested participants will be informed that selection is not guaranteed but that everybody will have an equal chance of being selected.

There will be a pre-workshop session in which selected applicants will be asked to take a baseline survey and undergo a series of tests. The tests, in particular, will measure "quality" such as aptitude (measured by a test of memory for digit span), personality trait (measured by Big Five Inventory), motivational profile (measured by Perry's 1996 scale of Public Service Motivation), and aspiration (measured by Kasser and Ryan's 1993 aspiration index). Additionally, I will measure the ratio of the
length of the index and ring fingers (also called the "digit ratio") as proxy for biologically determined factors of altruism and prosocial behavior.

After the pre-workshop session, respondents will then be randomly assigned into three experimental groups: a Control Group (A), a Workshop with Conditional Incentive Treatment Group (B), and a Workshop with Unconditional Incentive Treatment Group (C). Participants in the Control Group (A) will not be invited to the workshop and will not receive any further intervention. They will act as the counterfactual, the "business-as-usual" group.

Those assigned to the treatment groups B and C, on the other hand, will be invited to attend a workshop. However, none of them will be informed of any incentives at the time of the invitation, to prevent differential take-up across these two treatment arms. In the workshop, participants will attend different sessions in which their output and performance will be monitored and scored. If the screening theory is correct, then "high quality" participants (e.g. those who have high aptitude, are more altruistic, or have high public service motivation, etc.) should be delivering quality outputs and performing better than their counterparts.

If a participant is assigned in Treatment Group B (workshop with conditional incentive), then he/she will receive an incentive at the end of the workshop only if he/she makes the predetermined cutoff score. On the other hand, if he/she is assigned in Treatment Group C (workshop with unconditional incentive), then he/she will receive an incentive regardless of his/her score.

The incentive is a certificate of merit to be awarded during the closing ceremony of the workshop, and, should the incentive recipients decide to subsequently run for an elective office, the sponsoring partner organization shall donate 5 pieces of campaign posters to them. In other words, the incentive is designed to lower the cost of campaigning and to increase the psychic benefit of running for office.

I am interested in measuring the effect of the treatments on the composition of those who choose to stand for office. If the screening theory is correct, I only expect treatment group B to work in selectively nudging high quality citizens to run for office, where quality is measured by their baseline characteristics (i.e. memory for digit span, public sector motivation, digit ratio, aspiration, and personality). Because the incentive in treatment group B is only given conditional on making the workshop performance scores cutoff, and because ‘high quality’ respondents are more likely to make the cutoff (given that effort is less costly to them), then the mechanism works to selectively incentivize high quality citizens to run for office. In effect, this causes a higher proportion of high quality citizens running for office in this group relative to the other experimental groups, A and C.

Practical Considerations: A foremost practical concern of my research project is how to identify the population of interest – these aspiring political candidates. It is virtually impossible to find a group of political amateurs who would be willing to participate in a large-N experimental study in which some of them are randomly
assigned into being exposed to a workshop described above. I look to a particular and germane political setting to implement this proposed field experiment my
country, the Philippines. The Philippines is by far the only country in the world that popularly elects youth representatives to its smallest political unit—the barangay (village). Each of the 42,000+ barangays in the country is mandated by law (under the Local Government Code of 1991) to establish a Sangguniang Kabataan (SK), a governing body comprised of youth aged 15 to 17 years old. Crucially for my proposed research project, the SK is a low-stakes public office (both in terms of
remuneration and in scope of responsibilities) and it often serves as a jump-off point for a political career for aspiring politicians. Therefore, their group offers an apt opportunity to implement my proposed policy intervention.

Those selected to attend the workshop will be randomly assigned into a workshop batch (since there will be several batches of workshop) and within a batch, randomly assigned into small groups to randomly generate networks among subjects. This is so that one can test whether social networks explain political selection (e.g. a subjects likelihood of standing for office depends on another subject's (within the small group) decision to stand for office.)

Randomization done in office by a computer.

Individual.

No

570 individuals

570 individuals

190 individuals control, 190 individuals workshop with conditional incentives, 190 individuals workshop with unconditional incentives

This paper evaluates a policy intervention designed to attract good political candidates – competent and honest ones – to public service. Inspired by the idea that schooling can act as a screening mechanism, and that non-monetary status awards can be a cost-effective tool to incentivize individuals, we evaluate whether a leadership training workshop with performance-based awards can screen and incentivize good people to serve in public office. In the context of a randomized field experiment among aspirants for the village youth councils in the Philippines, we find that this policy intervention is effective in terms of attracting individuals with above-median measures of public service motivation, intellectual ability, integrity, and aspiration.

http://www.nicoravanilla.com/uploads/2/4/1/1/24114923/ravanilla_-_nudging_good_politicians.pdf