No announcement yet.
  • Filter
  • Time
  • Show
Clear All
new posts

  • Analyzing the properties of OLS and IV estimators when an important variable "ability" is omitted

    Plssss, can somebody help me with my university assignment of Econometrics Course and help me with creation of do.file with appropriate codes . I will be forever gratefulllll))))))

    We are interested in estimating the impact of elite college attendance on the expected wage income. Abstracting from other factors, suppose we have omitted an important variable ability (this may capture both physical and intellectual characteristics of a person). For this exercise, rather than using actual data, simulate data for 900 respondents. For this analyze I need to proceed with this steps.

    • Generate variable ability as follows: ability = 0 (poor) for the first 300 respondents, ability = 1 (moderate) for the observations in 301-600, and ability = 2 (talented) in observations 601-900.
    • Create an indicator (dummy) variable for those people who live near an elite college. This will be used as an instrument. [Hint: You can use the following command gen nearcol = mod(n, 2)].
    • Generate a college attendance dummy variable, such that the person attends a college if he/she is talented or has a moderate ability and lives near an elite college.
    • Generate a white noise variable , following a normal distribution with a standard deviation of 0.01 and a mean of zero, εi ∼ N(0, 0.1).
    • Generate an income variable such that the true return to ability is 1 and the return to college is 2. Add the white noise to the constructed income variable so that income has a small random component [Hint: Incomei = 1 ∗ Abilityi + 2 ∗ AttendCollegei + εi ].
    • Regress income on college attendance and ability and verify that OLS uncovers the data generating process you created.
    • Now suppose ability is unobservable to the econometrician but remains an important determinant of income (still has a return of 1). Regress income on college attendance and state your conclusions.
    • Still assuming that ability is unobservable to the econometrician, use ivreg2 command to obtain an unbiased estimate of the impact of college attendance on income using the nearcol dummy as an instrumental variable.

  • #2
    Posted more appropriately in the General Forum, with a reply, at