Columbia University Arsenic Mitigation Project


Funding Agency : 

The Columbia University, New York, USA

Period : 

1st April 2007 - 30 September 2011



  • The general objectives of this project are to determine what allocation of decision-making authority maximizes for social welfare in a problem of providing local public goods; and 
  • Collect original data generated by an organizational experiment to test and revise the theory and distill policy implications. In the empirical context, the organizational change, it was taken under an observation that due to a natural disaster, the natural occurrence of high concentrations of arsenic increased in groundwater of Bangladesh. 


Main Objectives of the research: 

Theory: Determine what allocation of decision-making authority maximizes social welfare in a problem of providing a local public good. It was focused on: 
* Compare the outcomes when a central organization such as a NGO, government or private firm, makes all decisions and the community, which will be using public goods, makes all decisions in a dynamic model of interaction between individual agents. 
* Determine what allocation of decision-making authority maximizes for social welfare. 
* Determine how the welfare-maximizing allocation of decision-making authority and outcomes in the case of individual and average attributes of social networks. 

Empirical: Collect original data generated by an organizational experiment to test and revise the theory and distill policy implications. It will:

* Conduct an organizational experiment.

* Determine which intervention yields the best average outcome. 

* Use the control group to study communities in which collective action occurs in the absence of any intervention


Expected Results :

The project will develop an understanding of the effect, which dynamic interactions between individuals have on the outcomes achieved by different organizational structures. Little formal analysis of the effect of social dynamics exists in the literature on decentralization. The understanding of the relationship between social interaction and organizational structure and the resulting categorization of conditions under which delegation of decisions improve outcomes will promote research into the complex topics of decentralization, optimal organization and social networks in many disciplines such as economics, political science, sociology, epidemiology, computer science. 

The project requires an approach to modeling which integrates skills from several disciplines. The model will include elements from economic such as agents who are rational and use their information to increase their utilities. We will also use mathematical tools from statistics and probability such as modeling and analysis of stochastic processes and Monte Carlo techniques, to model more realistic and complex behavior than is standard in economics. We will assume that agents have less information, that behavior has a stochastic element, i.e. people do not always choose the action which is best for them and that amount of information and rules of interaction depend on an agent’s place in a social network. The research will result in contributions to the analysis of stochastic processes. 

Our study will collect original data to test and refine the theory. We will contribute a unique data set documenting an organizational experiment, which will allow us to separate the effects of organization from those of other factors which influence the outcomes. The second strength of the data is that, it will consist of household information, enabling us to study the distribution of outcomes. The data analysis will use econometric tools as well as the statisticians’ expertise and potential contributions to design of experiments and statistical analysis for network data. 


Major Activities: (Software/Hardware)

Safe water options (DTW, AIRP, DSP,STW) - 280 

  • Upazila coverage - 3 
  • Union coverage- 33 
  • Village coverage- 114 
  • HHs Coverage- 5363