Improving state effectiveness in environmental risk mitigation: Experimental evidence from Bangladesh

In Bangladesh, the DPHE (Department of Public Health Engineering) is the governmental body in charge of allocating and installing arsenic-safe public wells. Each fiscal year, villagers submit applications to the local DPHE office asking for the installation of DPHE wells in their neighborhood.

Within each upazila (an administrative unit, of which there about 500, comprising on average 70,000 households), one upazila-level DPHE engineer advises the upazila's chief bureaucrat (called "UNO") on which of these applications should be approved. Historically, the upazila-wide budget for DPHE wells has been evenly distributed across all of its unions (a sub-upazila administrative unit of 5,000 HHs on average). A union on average consists of 13 villages.

On top of such advisory roles, the upazila-level engineer presides over the installation and maintenance of DPHE wells. The engineer is aided by local subordinates, called "mechanics".

Despite their important role in providing arsenic-safe water, DPHE wells have been subject to suboptimal allocation–due to the lack of local knowledge on arsenic contamination status and political pressure. The intervention aims to overcome these challenges by providing the upazila-level engineers access to a website ("Arsenic Information Dashboard") that visualizes and summarizes the arsenic contamination status within their upazila. The information is provided at the sub-union level. The website also allows the users to digitize and store the list of applications that they received for this fiscal year, and to simulate the public health benefits of approving a set of applications over another. Information on the website is delivered via official mail from the central DPHE office, and field enumerators provide training sessions to both the DPHE engineer and the UNO’s subordinates.

The intervention and the baseline survey will be followed by midline and endline surveys; in the former, we will collect information on the applications received for the DPHE wells from the villagers. At the endline phase, enumerators will visit a randomly selected union within an upazila to ensure the veracity of the administrative data and to conduct simple household surveys for households surrounding the wells.

2025-09-07

2026-12-31

Arsenic exposure reduction of well allocation.

To compute the Arsenic Exposure Reduction of a well allocation we look at how wells are allocated across villages within a union. We will use information on the list of applications for wells in each union to calculate the minimum and maximum number of arsenic-drinking families that could be affected by an allocation. The maximum and minimum impacts are B_{max} and B_{min}, and then an allocation that reaches B_k households is scored as $(B_k - B_{min})/(B_{max} - B_{min})$.

Note that the calculation of arsenic exposure at the union level presumes that the number of wells allocated to each union will remain fixed. I.e. that the UNOs and DPHE engineers do not move wells across unions: their allocation decisions are all taken within unions. In all past years, each union has been allocated the same number of wells, and this is a strong equity norm that we do not expect will be violated during our study. In case we do observe deviations from such rules, we will assess the robustness of our results to dropping such upazilas.

DPHE Engineers
-DPHE engineers’ understanding of local arsenic contamination
-Time spent on activities unrelated to well installation
DPHE engineers’ self-reported importance assigned to arsenic

UNOs
-UNOs’ self-reported importance assigned to DPHE engineers’ advice
-Shift of union-level well budgets

Well/villager experience characteristics
-Perceptions of the DPHE well installation process
-Accessibility of DPHE wells

DPHE engineers’ understanding of local arsenic contamination will be measured by asking engineers to score the mauzas in their upazilas from 0 to 100, with 0 meaning “all wells are safe in arsenic” and 100 “all wells are unsafe in arsenic.” This will then be compared with the true % of arsenic contamination.
Time spent on activities unrelated to well installation will be measured by asking the engineers, in the baseline and endline surveys, how much time is spent per week on the maintenance of DPHE wells–which is another major activity that the DPHE engineer office attends to.
DPHE Engineers’ self-reported importance assigned to arsenic and UNOs’ self-reported importance assigned to DPHE engineers’ advice are respectively measured by asking the engineers and UNOs to score the importance of the individual categories/criteria (e.g., “opinions of local politicians,” “number of families drinking from the well”) in their decisions of installing the wells. Their scores are subject to a budget constraint of 100, to assure comparability across scores.
Shift of union-level budgets is measured by observing whether all unions within the upazilas received identical numbers of well installations or not.
Perceptions of the DPHE well installation process and Accessibility of DPHE wells will be evaluated using household surveys, where we will solicit households’ satisfaction with the well installation process and will evaluate the public’s capacity to access the DPHE wells–and check for evidence of illicit privatization (e.g., linkage of the well to a submersible pump for private shower facilities).

We randomly assign 142 upazilas to either treatment or control, where the randomization is stratified by the arsenic exposure reduction of past public well allocations, by the severity of arsenic contamination, and by division (details below).

The 142 upazilas belong to 5 divisions. Except for division (which is predetermined for all upazilas), all data that are required for stratification come from ARRP (Arsenic Risk Reduction Program), which tested 6.3 million drinking-water wells across the country for arsenic. The data included the type of the well (public vs. private), along with the arsenic testing results from field test kits. A random subset of the well water samples were analyzed in DPHE laboratories, the results of which were used to recalibrate the test kit results.

Using the ARRP test results and their arsenic test results, we compute for each upazila the % of wells unsafe in arsenic, from Bangladesh national standard levels (50 ppb).

Finally, we compute the arsenic exposure reduction of past well allocation in each union using the same formula we describe in the explanation for its use as a primary outcome. We then average this union-level measure across unions in an upazila to obtain the upazila’s arsenic exposure reduction of past well allocations.

We split the sample of upazilas above vs below the median of the arsenic exposure reduction of their past well allocations, and above vs below the median of their % arsenic contamination. Combined with the 5 divisions, this leads to 2x2x5 = 20 strata. Removing 2 empty strata and pooling a singleton stratum with other upazilas in the same division results in 16 strata. Our randomization is conducted within these strata.

Not available

Randomization done in office by a computer using R, using set.seed for reproducibility

Upazila

Yes

142 upazilas

2,288 unions / 2,500 DPHE wells / 5,000 households

70 treated upazilas, which contain 1,184 unions, 1,232 wells surveyed / 2,464 households surveyed
72 control upazilas, which contain 1,104 unions, 1,268 wells surveyed / 2,536 households surveyed