ECON

12345678910111213141516171819202122
Across
  1. 3. Type of regression where two variables look related only because they share a trend
  2. 4. An extreme data point that may be dropped if it is the result of a mistake
  3. 7. Errors in different periods being related is called ___ correlation
  4. 9. In DiD, the treatment and control groups must follow ___ trends without the policy
  5. 10. What you do to a model by adding the variable t
  6. 13. Unobserved ___: the time-invariant characteristic of each entity, treated as an omitted variable
  7. 15. The estimator that subtracts each entity’s own average over time
  8. 17. The group not affected by the policy in a DiD
  9. 20. The “P” in LRP
  10. 21. Type of effect that captures the unobserved, time-invariant characteristics of each entity
  11. 22. Cross-sections from different periods, with different individuals each time
Down
  1. 1. In an FDL model, the ___ propensity is the immediate change in y from a one-unit rise in z
  2. 2. Type of propensity (written as one word) that sums all the lag coefficients
  3. 5. OLS estimators have this property if TS 1-3 hold
  4. 6. A series that tends to grow or shrink over time
  5. 7. A time series is viewed as one realisation of this kind of random process
  6. 8. Acronym for what OLS is when TS 1-5 hold
  7. 10. Subtracting one period from the next to remove the time-invariant unobserved effect
  8. 11. The group affected by the policy in a DiD
  9. 12. A past value of z included in a finite distributed model
  10. 14. TS 6 assumes the errors follow this distribution
  11. 16. In DiD, the term made by multiplying the period dummy by the treatment dummy
  12. 18. The error term is uncorrelated with the explanatory variables (strict = all periods)
  13. 19. A 0/1 variable, used to control for each time period in pooled cross-sections
  14. 20. Data where the same individuals are followed over several periods