The Department of Energy is interested in methods of reducing household energy c

Question

The Department of Energy is interested in methods of reducing household energy consumption and in particular electricity consumption. One demand-side management scheme that is currently being considered for adoption involves energy audits. This involves an expert visiting the home and advising on potential energy saving initiatives. The crucial policy question is: How effective are audit programs in promoting energy savings? The Department has collected data for a sample of 296 households in order to conduct this analysis. The available variables are: people number of people in household rooms number of rooms in dwelling where household lives income household income exist'000 audit dummy =1 if household was audited, =0 otherwise gas dummy =1 if household has gas, =0 otherwise audit.gas interaction of audit and gas elec annual electricity expenditure exist lelec log (elec) Table 1 provides Stata output including summary statistics and results from two regression models. The first regression simply provides a convenient way to estimate the difference in electricity expenditure for those who were audited and those who were not. This model can be represented as: (1) elec = beta_0 + beta_1 audit + u The second model is given by: (2) log (elec) = beta_0 + beta_1 audit + beta_2 gas + beta_3 audit.gas + beta_4 people + beta_5 rooms + beta_6 income + u (i) The results from model (1) indicate a very large impact of energy audits on electricity expenditure. Why might you have concerns about the reliability of the estimated impact obtained from this regression? (ii) In what ways does model (2) address the concerns you expressed in (i) (iii) Explain how the impact of an energy audit is being modelled in equation (2). (iv) Use the results from model (2) to give a full description of the impact of energy audits on electricity expenditure. In particular comment on: a. whether the impact is large in an economic sense: b. whether the estimates of the impact are statistically significant: and c. how the estimates of the impact from model (2) compare with those provided by model (1).

Explanation / Answer

Answers:

i) The model is not reliable, because the regression model is build uing just one variable, which is again a binary variable, which only says that the audit was done or not. But we are not abto get a true picture, as to how much difference other variables are making, i.e. number of rooms, gas or no gas, income etc. Also, the elec variables must be having huge variance, which again might be inflating the results and estimations.

ii) Model 2 is better in the sense that it takes care of huge variance in the target variable by taking a log transformation of that. Also, it's aking other variables into consideration which makes it more relaible as we can now have the understanding which variable is actually more influential to the target variable, by seeing the variable coeffecients.

iii) Here the impact of audit is being seen by taking it as a binary variable. If we are supposed to get a negative coeffectint, i.e. Beta 1 has to be negative to show that audit has reduced the electricity bill. Also, we must get a that variable as significant by doing certain tests. Higher the absolute value of the coeffecient compare to other explanatory variables, higher is the effext of the audit.

Note: NO results have been provided in the question, and hence question (iv) can't be answered.

Related Questions

Navigate

• Browse (All)
• Browse T
• Subjects

---

---

Integrity-first tutoring: explanations and feedback only — we do not complete graded work. Learn more.