Controlling omitted variables and measurement errors by means of constrained autoregression and structural equation modeling: Theory, simulations and application tomeasuring household preference forin-house piped water in Indonesia

I aimed this thesis at shedding further light on the preferences for in-house piped water in rural and urban Indonesia. Particularly first, I introduced constrained autoregression (CAR) Structural Equation Modeling (SEM) – ASEM for controlling time-varying omitted variables and measurement errors in a hedonic price model used in revealing household preferences for in-house piped water connection. Second, by a Monte-Carlo simulation, I compared CAR to the commonly used approaches for controlling omitted variables. Third, also by a Monte-Carlo simulation, I tested sensitivity of CAR specification in response to the presence of omitted variable bias. Fourth, I applied ASEM to analyze the preference for in-house piped water in rural and urban Indonesia.
From the first objective I found that ASEM performs well in controlling both omitted variables and measurement errors and both should be controlled simultaneously as done by ASEM. From the second objective I found that CAR constrained autoregression outperforms the alternative omitted variables correction procedures in terms of bias and mean square errors. From the third objective, I found that CAR is an acceptable vehicle to control for time-varying omitted variables. From the fourth objective, I found that on average urban and rural households have the same willingness to pay for in-house piped water i.e. 34.24% of their monthly house rent. The findings support a need for further investment in in-house piped water in both areas, particularly for the households with the lowest expenditure levels.