Flood exposure and pro-social behaviour: Evidence from Vietnam

Last registered on August 18, 2022

Pre-Trial

Trial Information

General Information

Title
Flood exposure and pro-social behaviour: Evidence from Vietnam
RCT ID
AEARCTR-0009939
Initial registration date
August 18, 2022

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
August 18, 2022, 3:25 PM EDT

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

Locations

Region

Primary Investigator

Affiliation
University of Economics Ho Chi Minh City (UEH) and International Institute of Social Studies (ISS), Erasmus University Rotterdam

Other Primary Investigator(s)

PI Affiliation
University of Economics Ho Chi Minh City (UEH)
PI Affiliation
International Institute of Social Studies (ISS), Erasmus University Rotterdam

Additional Trial Information

Status
In development
Start date
2022-08-23
End date
2022-11-30
Secondary IDs
Prior work
This trial does not extend or rely on any prior RCTs.
Abstract
In the study, we examine the impact of flood experience on individuals' altruism and trust levels in incentivized lab-in-the field games, controlling for their risk and time preferences. We divide subjects into three groups: (i) non-flood-prone households; (ii) flood-prone households who have suffered no flood event since 2019; and (iii) flood-prone households who have suffered at least one flood event since 2019. These subjects will play a dictator game and a trust game to explore how flood exposure affects their giving and trust in others.
External Link(s)

Registration Citation

Citation
Duong, Tien Ha My, Khanh Nam Pham and Matthias Rieger . 2022. "Flood exposure and pro-social behaviour: Evidence from Vietnam." AEA RCT Registry. August 18. https://doi.org/10.1257/rct.9939-1.0
Sponsors & Partners

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

Request Information
Experimental Details

Interventions

Intervention(s)
There is no intervention as such. Households participating in the experiment come from different flood-prone areas. Differences in pro-social behaviors are explored across these three groups. We will also consider whether donations to an anonymous poor family and to an anonymous poor family from a flood prone area are different.
Intervention Start Date
2022-08-23
Intervention End Date
2022-11-30

Primary Outcomes

Primary Outcomes (end points)
Subjects from different groups will join the dictator game. We record the amount that participants donate to an anonymous poor family in the community, as well as an anonymous poor family from a flood-prone area.
Subjects will also play the trust game. We aim to explore the level of trust and trustworthiness in others between flood-prone and non-flood-prone households.
Primary Outcomes (explanation)
We explore the impact of flood exposure on altruism by comparing the difference in the amounts given in the dictator game among people from group 1, group 2, and group 3. Players will play two rounds. Each player will have 50,000VND in each round, and they can decide the amount they want to donate to an anonymous poor household (in round 1), and the amount they want to donate to an anonymous poor flood-prone household (in round 2). The donation amount in each round will range from 0 to 50,000VND (with 1,000VND steps).
Findings on the impact of natural disasters on individuals' altruism or generosity are mixed. For example, Rao et al. (2011) investigate how the 2008 Chinese earthquake affects prosocial behaviour by surveying people in non-devastated and devastated areas. They indicate that the degree of prosociality is positively related to the residential devastation level. Likewise, Chantarat et al. (2015) suggest that people who have experienced the 2011 flood in Cambodia have higher altruism. By contrast, Becchetti et al. (2017) examine the long-term effect of the 2004-tsunami on the generosity of individuals in Sri Lanka. They claim that the tsunami can reduce the generosity of people. Based on previous studies, our hypothesis is that the degree of altruism has a statistically significant (but ambiguous) difference between non-flooded and flooded households, and the effects of flood exposure on altruism will be more pronounced among group 3 due to salience of recent floods. The direction of the effect depends on underlying mechanisms that we will test as specified using secondary outcomes.
Regarding the amount given to an anonymous poor household and an anonymous poor flood-prone household, Chantarat et al. (2015) indicate that the mean of giving amount in the dictator game to the former is lower than the mean of giving amount to the latter. Therefore, our other hypothesis is that the amount given to an anonymous poor flood-prone household is higher than that given to an anonymous poor household.
We also examine the effect of flood exposure on trust and trustworthiness, which will be measured through the amount of investment and the amount of return in the trust game. We expect that to some extent, people may have to work together to deal with negative weather shocks. This in turn can promote trust and trustworthiness in others. Therefore, our hypothesis is that flood exposure is positively associated with trust and trustworthiness.

Secondary Outcomes

Secondary Outcomes (end points)
We aim to uncover the underlying mechanisms through which flood exposure can impact individuals’ altruism and trust. Particularly, the mechanisms by which flood experience influences altruism can operate through psychological and economic mechanisms, as well as indirect reciprocity. Meanwhile, flood experience can affect trust through social networks.
Secondary Outcomes (explanation)
Mechanism for altruism
Psychological mechanism
Natural disasters exposure can negatively affect individual wellbeing (Vastfjall, Peters, and Slovic, 2008). An individual's unhappiness is in turn related to the willingness to reduce the payoffs of others below his or her own (Charness & Grosskopf, 2001). In this way, negative emotions after experiencing a disaster can reduce individuals' generosity.
Based on previous studies, we construct the following hypothesis: Flooding experience causes individuals to feel negative, this negative emotion in turn reduces selflessness towards others. In other words, on average, group 3's dictator game donation is the lowest due to recent flooding's salience, followed by group 2's generosity and then group 1's.
Economic mechanism
Floods can cause people loss of income and property. In developing countries, where insurance programs are not well developed, people may have to spend their own resources on disaster recovery, such as repairing their infrastructure or rebuy the necessary items. Even after recovering from a disaster, individuals might increase their savings to cope with future risks (Roson, Calzadilla, & Francesco, 2007). Hence, a natural disaster can reduce generosity through decreasing people's resources, such as savings, and increasing precautionary savings for future shocks (Becchetti, 2017).
Based on previous studies, we suggest that flood exposure causes people to use their savings or borrow money/credit to recover from floods. They also tend to prepare precautionary savings to cope with future natural shocks. These actions reduce the generosity of individuals vulnerable to flooding. In other words, the generosity of group 1 is lower than that of group 2, while the generosity of group 3 is the lowest due to the salience of recent floods.
Indirect Reciprocity
Floods can cause great damage to people. However, if they receive help or aid to recover from a flood, this can trigger indirect reciprocity. That is, people may not forget to help when they are in trouble, and therefore behave more generously towards others (Becchetti, 2017).
Hence, our hypothesis is that people who receive help (or receive help above the average) will be more generous than people who do not receive help (or receive help below the average).
Mechanism for trust
The mechanisms by which flood experience can influence trust may be through social networks. Individuals can improve their protection capacity against disaster risks by strengthening relationships with community members who are also vulnerable to disasters. For example, natural negative shocks can spur people to cooperate and work together to rebuild infrastructure, borrow money to cope with damage due to disasters, and ask for other support. These activities will reinforce social networks and trust among community members (Dang and Dang, 2021).
Hence, we suggest the following hypothesis: Social networks can be a channel that links flood exposure and trust level.

Experimental Design

Experimental Design
The fieldwork includes three parts: (i) a round to elicit time preferences, (ii) a round to elicit risk preferences, (iii) two rounds to elicit altruism (dictator game), and (iv) two rounds to elicit trust (trust game).
Non-flood-prone households, flood-prone households who have suffered no flood event since 2019, and flood-prone households who have suffered at least one flood event since 2019 will join the experiment.
Experimental Design Details
Participants will play four games: risk preference game, time preference game, dictator and trust game. Details of the games are described below.
Risk preference
Subjects will be given an amount of money and they can choose between two options:
Option 1: They can keep all the money and not invest, or
Option 2: They can invest specific amount of money.
If subjects choose option 2, we will toss a coin. If it is heads, subjects’ investment will be tripled; otherwise, they receive no money back from the investment. The variable will be used as control in robustness checks, as well as dependent variable to explore group differences if any.
Time preference
They will choose whether they want to receive the money early (with less money) or later (with more money). The variable will be used as control in robustness checks, as well as dependent variable, as well as dependent variable to explore group differences if any.
Dictator game
There will be two rounds. At the beginning, players will perform a task to earn 100,000VND. This 100,000VND is divided into 50,000VND for round 1 and 50,000VND for round 2. They will then decide the amount they want to donate to others. In each round, the lowest donation amount is 0VND and the highest is 50,000VND (in 1,000VND steps). Particularly, in round 1, they will decide how much they want to donate to an anonymous poor family in the same village. In round 2, they will decide how much they want to donate to an anonymous poor family who is prone to floods.
Trust game
In this game, two subjects, namely a player 1 and a player 2, participate in a two-stage game. At the beginning, subjects are randomly chosen to be player 1 or player 2.
Both player 1 and player 2 will receive 40,000 VND to start. Player 1 will decide the amount (between 0 and 40,000VND) he wants to keep and the amount he sends to player 2. The amount sent to player 2 will be tripled. For example, if player 1 selects to keep 30,000VND and send 10,000VND, then player 2 will receive 30,000VND (=10,000VND*3). In this case, player 2 has a total of 70,000VND (30,000VND plus the original 40,000VND). Player 2 now decides how much of this 70,000VND to send back to player 1 and how much to keep.
If people are more risk-averse and more patient, they may increase their subjective assessment of the likelihood of future flooding in particular and other disasters in general. This in turn can reduce their generosity in response to potential negative events.
Survey
After the experiments, subjects will be asked a series of questions. These will be used to explore mechanisms relating to wellbeing, stress and networks. SES-related questions will be used as potential controls.
Randomization Method
There is no randomization. Experimental groups are allocated based on flood experience and risk as described earlier.
Randomization Unit
There is no randomization.
Was the treatment clustered?
No

Experiment Characteristics

Sample size: planned number of clusters
We will survey four villages in Dak Lak that include households from comparison groups (group 1, group 2, and group 3). We will then select all the households that meet the above criteria in these villages. That is, the study will consist of 3 groups and at least 145 people in each group.
Sample size: planned number of observations
Each group includes at least 145 subjects, with a total of at least 435 subjects. However, we aim to recruit up to 200 subjects per group, depending on field availability. The experimental protocols and procedures are pre-tested in up to two villages outside of the study area. These data will not enter the analysis.
Sample size (or number of clusters) by treatment arms
Participants in 3 groups will join the following games:
- Risk game (1 round);
- Time preference measure (1 round);
- Dictator game (2 rounds);
- Trust game (2 rounds). In this game, half of the participants are player 1, the other half are player 2.
Minimum detectable effect size for main outcomes (accounting for sample design and clustering)
We aim to discover the impact of natural disaster experience on altruism with a power of 80%. We build on Becchetti et al. (2017) who examine the effect of a tsunami on giving in Sri Lanka. They show that the mean givings in the dictator game of the non-damaged group and the damaged group are 0.372 and 0.31, respectively. The corresponding standard deviations are 0.205 and 0.168, respectively. Therefore, we should include at least 145 subjects for each group with power of 80%. Given prior knowledge/data of our study area, we aim to recruit up to 200 subjects per group.
Supporting Documents and Materials

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

Request Information
IRB

Institutional Review Boards (IRBs)

IRB Name
International Institute of Social Studies
IRB Approval Date
2022-07-07
IRB Approval Number
Ethics 2022-07

Post-Trial

Post Trial Information

Study Withdrawal

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

Request Information

Intervention

Is the intervention completed?
No
Data Collection Complete
Data Publication

Data Publication

Is public data available?
No

Program Files

Program Files
Reports, Papers & Other Materials

Relevant Paper(s)

Reports & Other Materials