By Accelerator Lab, UNDP Bangladesh

Context

Disasters such as tropical storms, cyclones and monsoon floods are recurring events in Bangladesh. Every year different parts of Bangladesh face different types of disasters. Exposure to disaster causes considerable damage to assets and livelihood opportunities of affected households. In response, households adopt private defensive measures investing their private funds, which acts as an additional financial burden. There is very limited information available about expenditure on disaster risk reduction and recovery in the country. In this context, to understand these expenditures and help minimize these costs, the Accelerator Lab of UNDP Bangladesh commissioned a disaster and climate change finance household expenditure survey to assess the percentage of climate-related expenditure as a share of household income and expenditure in climate-vulnerable regions of Bangladesh. This investigation primarily focused on the following issues:

i)               disaster and climate adaptation expenditures by rural Bangladeshi households;

ii)             the socioeconomic factors influencing disaster and climate adaptation expenditures;

iii)           welfare effects of disaster and climate adaptation expenditures, and

iv)            differences in disaster and climate adaptation expenditures, and their welfare effects, by gender of the household head.

This research closely complements the “Strengthening Governance of Climate Change Finance to Enhance Gender Equality” project of UNDP Bangladesh to provide technical support by developing a survey of detailed questions for a sample of households identified by UNDP to work out the percentage of climate expenditure as a share of household income and expenditure.

The disaster and climate change finance household expenditure survey covered a total of 3,095 households from the rural areas of 10 selected districts in Bangladesh. Households have different levels of exposure to a variety of disasters. Out of 3,095 households, 42.6% and 40.7% reported exposure to floods and storms, respectively. In addition, 82.6% reported being affected by at least one slow-onset disaster.

Survey design

The UK-based International Institute for Environment and Development (IIED), in consultation with UNDP Bangladesh, developed the survey questionnaire. This involved identifying locations that were not affected by either of the 2020 flood or 2020 cyclone Amphan (i.e., unaffected regions) and locations affected by only flood, only cyclone and both (i.e., flood and cyclone-affected regions).  As there are different UNDP projects ongoing in all 64 districts of Bangladesh, the project beneficiaries easiest to reach and willing to respond were targeted. Surveyed households were asked to recall their income and expenditure data prior to happening of these disasters (i.e., pre-treatment data) and current data (i.e., post-treatment data).

Data collection

For developing countries like Bangladesh where insurance and many other important markets are either absent or underdeveloped, recovery from floods and storms takes a significant toll. Families are without sufficient food and other basic consumer goods and must prioritize their survival before constructing their homesteads and other productive inputs, such as livestock. It is highly likely that their pre- and post - disaster incomes will be different. Therefore, the comparison of their expenditure on disaster risk reduction activities was in terms of proportion of total expenditure, in addition to absolute expenditures. Hence, data on household income and expenditure was collected for both the pre- and post-disaster periods.

The survey questionnaire was used to collect data on the components of income and expenditure, labor supply, and other household and farm level attributes. Data were collected on indicators such as household roster information, education, labor supply, farm and agriculture status, income, expenditure, disaster exposure, disaster adaptation and finance expenditure, and public investments in disaster management in locality/union.  The preliminary questionnaire went through field tests before finalization.

For sampling, the entire study was divided in two parts in terms of type of disasters: flood and cyclone. Then 3 affected and three unaffected districts were chosen for each disaster. Allowing for 5% margin of error and at 99% confidence level, between 610 and 650 samples were collected from four strata, with 750 samples surveyed from each stratum. Within each stratum, the number of samples was allocated according to population weight and rounded up. The enumerators interviewed the head of the household. If the head was not present, the next decision-maker or the senior-most member was interviewed.   

Gender is an important aspect that requires further investigation since women are more likely to care for emergencies than men. In addition to traditional reporting of the gender of the household head, further information was collected on 1) women’s role in financial decision making 2) women’s actual role in managing the family (instead of proxy underage male head especially when the male spouse is either deceased or away). These allowed more critical investigation of whether more inclusive roles of women are crucial to households’ adaptation finance decisions.

Data Analysis and results

The collected data were analyzed with appropriate econometric analysis method. In particular, the difference-in-difference method was implemented.

The results show that in response to the damage to assets and livelihood opportunities caused by disasters, households adopt defensive measures investing their private funds. Among the 1,320 flooded households, 31.1% raised floors to save their homestead from flood waters. In addition, they also invested in protecting their agricultural lands (2.5%), household materials (7.35%), birds and animals (22.9%), and other actions (4.7%). Altogether, the study found that a total of 545 households undertakes at least one protective measure at an average cost of BDT 9,243 (roughly equivalent to US$115 at the exchange rate of US$1=BDT 80).

Among the 1,260 storm-affected households, 11.3% raised floors, whereas they also invested in protecting their agricultural lands (6.03%), household materials (6.11%), birds and animals (16%), and other actions (2.22%). The study found that a total of 344 households undertakes at least one protective measure at an average cost of BDT 10,822 (roughly equivalent to US$135 at the exchange rate of US$1=BDT 80).

Exposure to disasters was found to vary by location. Households from northern districts of Gaibandha (93%), Jamalpur (87.7%), and Kurigram (94.2%) reported very high exposure to floods. Storm exposure was mostly concentrated in the southern coastal districts: Bagerhat (84.6%), Barguna (88.3%), and Satkhira (96.3%).

Adaptation expenditure was found to vary by gender. In general, female-headed households are more likely to undertake disaster protection measures than male-headed households. However, in absolute amount, they spend less than male-headed households in those actions. In almost all cases the female-headed households are poorer, and due to their limited ability, female-headed households cannot spend at the same level as male-headed households. However, the study found that female-headed households spend a higher share of their income than male-headed households in adaptation. They spend 2%-points higher on flood adaptation, close to 31%-points higher in storm adaptation, and around 3%-points higher in adaptation to other disasters. While the gendered differences in disaster spend as share of income are very low for floods and other disasters, they are very high for storms due to similar absolute spends by both male and female-headed households while male-headed households have almost twice the incomes.

Utilizing the survey findings

This study helped to better understand the disaster and climate adaptation expenditures by rural Bangladeshi households, the socioeconomic factors influencing these expenditures, differences in expenditures and their welfare effects by gender of the household head. These findings can help the Accelerator Lab to narrow down specific district and female-headed households, most vulnerable and in risk of exposure to floods and storms, which holds promise for the lab to pilot new climate financing approaches and innovative adaptation solutions in the future.

Broadly, this information can help make disaster and climate finance decisions of the government and development partners more effective by aligning them more closely with the priorities of individuals at the frontline of responding to disaster risks and climate change. This would also ensure “money where it matters” to get more climate finance to the local level. Ultimately, these findings can help poor households reduce their disaster preparatory costs.