Does Poverty Lower Productivity?

How to cite

Oh, Suanna; Mullainathan, Sendhil; Schilbach, Frank; Kaur, Supreet. 2018. "Does Poverty Lower Productivity?." AEA RCT Registry. AEARCTR-0002181. https://doi.org/10.1257/rct.2181-1.0

Download: BibTeX · RIS — for reference managers without a browser translator. Zotero and most others pick up this page's citation metadata directly.

Suanna Oh* , Sendhil Mullainathan , Frank Schilbach , Supreet Kaur

* Primary investigator

Registry ID
AEARCTR-0002181
DOI
https://doi.org/10.1257/rct.2181-1.0
First registered
Last updated
Status
On going
Trial dates
2017-03-01 – 2020-07-31
Countries
Private
Registry record
http://www.socialscienceregistry.org/trials/2181

Abstract

Recent work posits that poverty lowers cognitive function (Mani et al. 2013), but evidence for behaviors in the field is lacking. If true, this has the potential to lower labor productivity among the poor, providing a channel through which poverty could directly reinforce itself. We design an experiment with piece rate manufacturing workers to examine a link between financial constraints and worker productivity. We induce variation in both the extent of financial constraints as well as their salience among workers.

Intervention

We recruit daily wage laborers to participate full-time in a low-skill manufacturing job, namely producing disposable plates. Workers receive wages according to a base wage rate and a piece rate, so that their total wage depends on their individual production. We record the number of plates worker produce in each hour throughout the study in order to measure their productivity. We randomly vary the schedule of the lump-sum wage payments across individuals, so that some workers receive their initial wage payments in the middle of the study, while the rest receive all payments at the end. At several points during the study, workers are asked to participate in a half-hour activity, during which they are induced to think about their financial issues. The participants first listen to a 10-minute story in a small group of people of up to six people. The story describes the experiences of an agricultural worker who has recently faced various financial problems. Then they individually engage in an informal discussion of approximately 15 - 20 minutes with a surveyor about how the story relates to their own life and financial situation. We compare the productivity of workers after the activity, comparing those who were randomly selected to engage in the activity that day against those who had a regular work day (“financial salience effect”). We test whether the financial salience effect changes depending on whether workers have already received the initial lump-sum wage payments.

Experimental design

The experiment will be implemented in sessions, with each worker receiving a 12-day-contract for full-time employment. Each session involves approximately 30 participants, who are randomized into the following treatment groups prior to the start of each session.

(1) Wave 1: Early salience - Cash
(2) Wave 2: Late salience - Cash
(3) Wave 3: No salience - Cash
(4) Wave 4: Early salience - No cash
(5) Wave 5: Late salience - No cash
(6) Wave 6: No salience – No cash

The schedule of the 12-day session is as follows:

- Day 1: Workers provide informed consent, complete a baseline survey and go through training. They receive the first day’s wage at the end of the day. If any worker drops out after Day 1, they are replaced by a new worker by Day 3. No replacements occur after Day 3.
- Day 5: The wage payment schedules are announced in the morning. That is, Wave 1-3 learn that they will receive lump-sum wage payments in the middle of the session.
- Day 6: Those in Waves 1 and 4 (Early salience) participate in the half-hour financial salience activity in the morning.
- Day 8: Those in Waves 1, 2, and 3 (Cash) receive lump-sum wage payments at the end of the day.
- Day 10: Those in Waves 2 and 5 (Late salience) participate in the half-hour financial salience activity in the morning. Everyone is also asked a few questions about their work experience on this day.
- Day 12: Workers complete an endline survey and receive final wage payments at the end of the day.

Workers in each wave are also randomized into Type A and Type B. Those in Type A have the schedule described above. For those in Type B, the salience activities and cash payments happen with a one day lag. That is, the first financial salience activity takes place on Day 7, wave payments on Day 9, and the second financial salience activity on Day 11, respectively.

The baseline survey collects basic information on demographics (age, literacy, education, family) and wealth (work, loans, dwelling and recent expenditure). The endline survey collects information on the detailed expenditures in the last few days, major sources of financial stress, life satisfaction and happiness. Output measures collected on Day 2-11 are used for analysis.

Hypotheses

Using our randomized treatments, we will conduct two main tests. First, we will test whether the productivity of the workers that have recently received lump-sum wage payments is different from the productivity of those who have not yet received payments. Second, we will test whether having recently received wage payments reduces the financial salience effect.

Design details

Randomization method
Randomization is done in the office by one of the PIs, Research Assistants, or graduate students on the computer, prior to the start of each session.
Randomization unit
Workers are randomized into different treatment waves at the individual level.
Clusters
Over 200 individuals. Since we expect the financial salience effect to be present in the lean season, where workers are most concerned about their income and expenditure, we plan to run the experimental sessions from early March to early June of 2018. We plan to recruit more than 200 workers, according to our power calculations.
Observations
Over 10,000 individual-hours (5 work hours per day * 10 measured work days per person).
Arms
(1) Wave 1: 10% (2) Wave 2: 20% (3) Wave 3: 20% (4) Wave 4: 10% (5) Wave 5: 20% (6) Wave 6: 20%
Primary outcomes
We will collect data on work performance by recording the hourly disposable plate output by each worker. To fully understand the intervention’s effects on productivity, we will also collect output quality measures in addition to the quantity measures. For all the measures, two auditors will first independently record their scores and then their supervisors will reconcile any differences.

Cited by

No citation data available for this registration yet.

Keywords