Modeling Commercial Demand for Water: Exploring Alternative Prices, Instrumental Variables, and Heterogeneity

Evaluating the Sensitivity of Residential Water Demand Estimation to Model Specification and Instrument Choices

“…Out of 31 different sets of instruments that are used in Equation ( 5), we find the sole use of days of service as a weak instrument with a Kleibergen-Paap F stat of 9.53, below the critical value (see Table 3). Though Flyr et al (2019) discussed it as a valid instrument for addressing exogenous variation across both households and time, we find that the sole use of days of service is insufficient without the inclusion of an additional price metrics or other instrument. Consistent with this result, using only days of service results in a severe loss of precision.…”

“…To perform this test, we use one‐period‐lagged fixed charge ( FC it‐1 ), price for block one ( block1 it‐1 ), successive difference between block groups ( blockdiff1 it‐1 and blockdiff2 it‐1 ), and days of service ( DOS it‐1 ) as the set of instruments, which are commonly used in literature. While most of the water demand literature rely on the theoretical explanation to justify the validity of the instruments, few—Flyr et al (), for instance—have tested the validity of instrument choice empirically; accordingly, we follow Flyr et al () (see Text S3 for details). Rejecting the null hypothesis of the Davidson‐Mackinnon Test, that price variables (marginal and average) are exogenous, confirms the presence of endogeneity and requires the use of price instruments to get unbiased estimates of coefficients.…”

“…While different studies continue to use average price, marginal price or some combination of both as the relevant price, Shin (1985) proposed an econometric test to identify the price signal to which households respond. Numerous studies have since used this test, reaffirming average price as the appropriate price signal in residential and commercial water demand (Flyr et al, 2019;Nieswiadomy & Molina, 1991). Most recently, Ito (2014) employed the Encompassing Test, as suggested by Davidson and MacKinnon (1993), to choose among alternative prices (marginal price, expected marginal, or average price), though in electricity pricing, and also found consumers respond to the average price.…”

“…Wichman et al (2016) followed Olmstead (2009) and many others and use basic service fees and marginal prices (full block rate structure, i.e., price for each block group), since these instruments are uncorrelated with water consumption but highly correlated with average price (Nieswiadomy & Molina, 1989;Pérez-Urdiales et al, 2016;Price et al, 2014;Reynaud et al, 2005). Most recently, the number of billing days was introduced as the possible instrument to allow for variation across households, within a time period, whereas simply using block rates to instrument does not capture interhousehold variation because price structures are usually identical across a service area (Flyr et al, 2019).…”

Exploring Water Consumption in Dhaka City Using Instrumental Variables Regression Approaches

“…using different statistical estimation techniques such as, ordinary least squares (OLS), generalized method of moments (GMM) and quantile regression (Deyà-Tortella et al 2019;Dikgang et al 2019;Kumar and Ramachandran 2019;Binet et al 2014;Cardoso 2013;Wentz and Gober 2007;Worthington and Hoffman 2008). The three most central aspects of empirical works on water consumption are the identification of water demand covariates, nature of water consumption data (i.e., aggregate or household) and the magnitudes of estimated price and income elasticities (Flyr et al 2019;Oliveira et al 2017;Sebri 2016;Gardner 2010;Bartczak et al 2009;Espey et al 1997;Worthington and Hoffman 2008;Ahmad et al 2016;Nauges and Thomas 2000;Martínez-Espiñeira 2002;Strand and Walker 2005). Despite different aspects of empirical evidence on water consumption in cities, the common procedure involves identification of most important determinants from a pool of covariates and estimates their separate effects using standard statistical estimation techniques mentioned above.…”