We propose a detailed analysis of heterogeneity in households’ responses to water conservation programs (price increase, voluntary and mandatory conservation) during periods of water shortage. Using a unique dataset covering water consumption of all residential users in Los Angeles (California) during the drought (1988-1992), we show that households generally were responsive to the conservation measures but that the magnitude of households’ responses varies depending on the instrument and on households’ characteristics, in particular the size of their lot. Price elasticity is estimated between –0.29 and –0.47 in the high season (June-October), and between 0 and –0.19 in the low season (November-May). Results suggest that the voluntary conservation program [resp. mandatory conservation program] induced a reduction in water use which varies from 1 to 13% [resp. 21% to 29%] depending on the season and the size of the lot. The achieved reduction in consumption is however very similar across households. These data also allow us to compare the effectiveness of price and nonprice policies in terms of water savings. Finally, welfare calculations suggest that households with the smallest lot sizes (and lowest income) suffered the greatest loss during the implementation of the water conservation programs.

A major environmental tragedy of modern times is the widespread arsenic contamination of shallow drinking water wells in Bangladesh. High levels of arsenic present in many wells went unrecognized for years. Now large numbers of people show a range of symptoms associated with chronic arsenic exposure. Most of the economics literature follows an epidemiological approach effectively monetizing a dose response relation. We take a different approach, given widespread exposure, and examine impacts on household labor supply. We find significant effects broadly consistent with available epidemiological information in terms of the percent of the population impacted and which demographic groups are most impacted. The nature of the arsenic contamination provides a high quality statistical instrument that identifies a labor supply reduction of over 8%. Particular attention is paid to large substitution effects involving within household labor supply as this is the primary means of insurance among poor households in developing countries.