Price beliefs for LPG in Ghana

How to cite

Malagutti, Flavio; Abubakari, Sulemana; Jack, Kelsey. 2021. "Price beliefs for LPG in Ghana." AEA RCT Registry. AEARCTR-0007417. https://doi.org/10.1257/rct.7417-1.0

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

Flavio Malagutti* , Sulemana Abubakari , Kelsey Jack

* Primary investigator

Registry ID
AEARCTR-0007417
DOI
https://doi.org/10.1257/rct.7417-1.0
First registered
Last updated
Status
Completed
Trial dates
2021-02-18 – 2021-03-23
Countries
Ghana ()
Registry record
http://www.socialscienceregistry.org/trials/7417

Abstract

This project unfolds in four steps. First, we will test the hypothesis that cooking with LPG (stoves and fuel) is viewed as desirable by Ghanaians but is associated with an expensive lifestyle. Second, we will compare survey respondents’ beliefs about LPG prices with true market prices in their home regions. This comparison gives us the direction and size of possible price information failures in this market. Our third step is to correlate price beliefs with individual and household observable characteristics. This will allow us to learn about key determinants of price perceptions for LPG. Fourth, we will use an experimental design to learn whether and how individuals update their price beliefs upon receiving information about LPG consumption in specific locations. This last piece will teach us about whether and how individuals learn from others in the context of cooking technologies. Results may allow us to think about cost-effective interventions aimed at addressing an information failure in the Ghanaian LPG market.

Experimental design

To understand individuals’ beliefs over LPG prices, we elicit both the cost of refilling a cylinder and the per-day cost of cooking with LPG.

Price elicitation: refill. We capture cylinder refill price beliefs by asking the following question: “What do you think is the cost of refilling an LPG cylinder at your nearby station right now?”

Answer choices allow the respondent to answer in whichever unit she is most comfortable (full cylinder, per KG, per liter). We will convert this answer to a price-per-KG variable.

We capture per-day price beliefs by asking the following question: “What is the size of LPG cylinders you are most familiar with? If I brought you a cylinder of this size today and you began cooking all of your household’s meals with it from now on, how long do you think this cylinder would last?"

This answer comes in the “number of days” that the chosen cylinder would last in the household. We will convert this answer to a days-per-Kg variable.

For a random subset of households, we will follow the price questions with an extra question to capture the respondent’s level of confidence on her answer: “How confident would you say you are that the actual (refill cost at your nearby station/number of days the cylinder would last in this household) is what you told me just now?” These answers will be in a Likert scale ranging from “not at all confident” to “very confident.”

We experimentally test whether informing individuals about LPG usage levels (or prevalence) in general populations can impact price beliefs. We randomize survey respondents to an interaction that consists of a prior elicitation followed by the provision of information on the prevalence of LPG in a pre-specified location. For one group (the Accra- information treatment), we elicit respondents’ priors on how many households (out of ten) are currently using LPG stoves as their primary cooking stoves in Accra, the capital of Ghana. Immediately after collecting the answer, the enumerator informs the respondent of the actual statistic, as reported by the Ghanaian Statistical Service. The interaction is exactly the same for our second experimental group (the Ghana-information treatment), except that, in this group, the conversation is framed around the whole country of Ghana. We ask for respondents’ priors on how many households in the country (out of ten) are currently using LPG stoves as their primary cooking stoves.

Design details

Randomization method
Randomization programmed in the data collection instrument (tablet).
Randomization unit
Household
Clusters
0
Observations
7350
Arms
2450
Primary outcomes
Beliefs about LPG refill price and days worth at the household as a cooking fuel.

Cited by

No citation data available for this registration yet.

Keywords