# delimit ;
* Estimation code for exercise 11: estimating the potential demand for ecolabeled oranges.;
use "C:\Users\AutoLogon\Desktop\apple.dta", clear;
log using "C:\Users\AutoLogon\Desktop\exercise11.log", replace;
* First, we verify the number of people without any demand for ecolabeled oranges at the suggested prices;
count if ecolbs==0;
* We observe that 248 out of the 660 participants declared zero demand for ecolabeled oranges. A high proportion of our sample is truncated at the value of zero (approximately one third of the observations). Since we have such a high proportion, the tobit model should be preferred to an OLS estimation;
* When estimating a tobit model, we need to declare the lower bound of our dependent variable (in our case, the value zero). It is always mandatory to declare the value of our corner solution.;
tobit ecolbs ecoprc regprc faminc hhsize numlt5 num5_17 num18_64 numgt64,ll(0);
* First, we observe that the 95% confidence interval for sigma (=var(e.ecolbs)) does not include zero. This means that the coefficient is statistically significant. This finding validates the choice of the tobit model. If the confidence interval includes zero, we cannot reject the hypothesis that var(e.ecolbs)=0 and OLS estimation would be appropriate.
The price of regular and ecolabeled apples have the expected signs. Marginal increases in the price of ecolabeled apples decrease the demand. On the other hand, an increase in the price of conventional apples will increase the consumption of organic apples, since the latter became relatively more cheap.
We also note that household income is significant at a 10% level: families with higher income tend to consume higher ecolabeled orange quantities. However, household size and household age structure do not seem to influence on the decision to consume ecolabeled oranges.
In terms of marginal effects interpretation, we can only interpret the sign of the coefficients. Positive coefficients are associated to an increase in the demand. However the coefficients do not provide the magnitude of the marginal effect.;
log close;