Answer to Question #303142 in Macroeconomics for Abichu

Question

22. Suppose you have a data set on wealth (w), and access to credit (credit) for 1000 farmers randomly selected from the Gamo-Gofa Zone over the period 2015 to 2020. The data includes individuals who did not participate in the credit scheme.

(a) Suppose you want to evaluate how wealth changed over time as a result of the credit facility. How do you specify your model?

Indicate what impacts the parameters in your model measure.

[Hint: Recall the importance of time dummies].

(b) Suppose you have a reason to believe that level of education (E), gender of the head of the household (gender ), land size (land), distance from urban centers (distance), are other major variables that are believed to affect wealth status. Can you use fixed effects method in estimating your model in (a) by including theses variables? Why or why not? What other means can you craft to apply fixed effects method on the model?

(c) Given the information in (b), under what condition a random effects method could be an option?

Answer 1

a)

The credit facility facilitates collateral that may be sold or even substituted without altering the terms of the original contract. This facility may apply to different projects in business. since repaying period for the loan is flexible, there will be an increase in wealth created. Time dummy will also allow control of time-specified effects.

b)

YES, I will use time-fixed effects since they allow controlling of underlying observable and unobservable systematic differences between observed time units. This will help me in controlling larger categories by controlling everything by ensuring that is constant within that particular category.

c)

One of the conditions for using the fixed-effect model is when dealing with differing within and between effects. THE result is that between effects that are associated at the higher level it cannot be estimated.

Another condition is when dealing with random effects without within and between and separation. One of the commonly used models uses the random effects framework though it does not estimate separate relationships at each of the two levels.