ECONOMETRICS

  Two data files to use on STATA and produce a pdf translation of the do-file commands with the results as well as the graphs. Answer all questions. INSTRUCTIONS TO CANDIDATES: 1. The coursework has two questions. You must answer both questions. 2. The deadline for submission of the coursework is Friday 13th April at 10:00 am. Work should be submitted on KEATS. 3. The file that you upload on KEATS should contain two parts: • Short written answers to the questions • The Stata output in pdf format You can merge two pdf files using Acrobat Professional or an online pdf merger. 4. To avoid collusion, each student is given a unique version of the datasets. This means that you should answer the questions with the datasets that have been provided to you. If your answers or the Stata output file are based on the datasets given to another student, you may face an allegation of collusion. Question 1 (50 marks) Download the dataset ajr.dta from the course website. The data set contains per capita income in 1995 as well as a number of other variables for 62 non-European countries. The data have been collected by Acemoglu, Johnson, and Robinson (AJR). AJR believe that rich countries are rich primarily because they have institutions which are more conducive to growth. Institutions refers to a wide set of political and economic arrangements, including democracy versus autocratic rule, the security of property rights, the enforcement of law and contracts, the efficiency of the bureaucracy versus corruption, etc. In this question we want to assess the particular hypothesis tested by AJR that the protection of property rights is conducive to growth, and hence should be correlated with the level of contemporary per capita income. The data set contains a variable protection, indicating the protection of property rights (with larger values indicating more protection). The log of GDP per capita is called loggdp. a) Run an OLS regression of loggdp on protection. Comment on your result. b) Why might you be worried about interpreting the effect of property rights or expropriation risk on GDP per capita causally? Explain the different sources of bias and in which direction the bias would go. c) AJR suggest using the mortality of European colonial settlers as an instrument for property rights protection. Their argument is that European colonisers set up different institutions in various countries depending on whether they decided to settle there (as in the USA, Argentina, or Australia) or whether they decided simply to exploit the natural resources of the colony (as in many African countries). Some colonies had conditions more conducive to settlement than others. These conditions are measured by the variable logmort0, the log of European settler mortality (measured mostly in the 1800s). 1. What conditions need to be satisfied for settler mortality to be a valid instrument for property rights? Discuss whether these conditions are likely to hold in this context. 2. Estimate the first stage equation and explain what you find. 3. Estimate the reduced form and explain what you find. 4. Run the IV regression of loggdp on protection using logmort0 as instrument. Comment on your result and compare them to your OLS results in part 1. Are the differences you find explained by the biases you discussed in part b)? 5. Construct the IV estimate from your results in parts 2. and 3. d) A critic of these regressions is worried that the current level of GDP is correlated with the disease environment in a country, which in turn will be correlated with European settler mortality in the 1800s. 1. If the critic is right, what is the consequence of this for the IV results you obtained in part c)? 2. The critic therefore proposes to include another regressor, the current incidence of malaria in the country. Explain why this is a potential solution to the problem identified by the critic. 3. Repeat the first stage including malaria as a regressor. Compare your result to that from part c) and comment on your findings. 4. Repeat your OLS and IV regressions of loggdp on protection including malaria as a regressor. Compare your results to those from parts a) and c) and comment on your findings. 5. Could the inclusion of malaria as a regressor generate selection bias? Explain. References: Acemoglu, Daron, Johnson, Simon and Robinson, James A. (2001), “The Colonial Origins of Comparative Development: An Empirical Investigation”, The American Economic Review, Vol. 91, No. 5, pp. 1369-1401 Question 2 (50 marks) Download the dataset minwage.dta from the course website. It contains data collected by David Card and Alan Krueger on fast food restaurants in New Jersey (NJ) and Pennsylvania (PA) during two interview waves in March (wave 1) and November/December (wave 2) of 1992. On April 1, 1992 New Jersey raised its minimum wage from $4.25 to $5.05. The minimum in Pennsylvania remained at the federal level of $4.25. Card and Krueger use these data to analyse the impact of the minimum wage increase in New Jersey on employment in the fast food industry. a) We start by focusing on average employment. 1. Calculate the average full time equivalent employment (fte) separately for restaurants in NJ and in PA, for each interview wave. 2. Calculate the difference in the average employment between the second and first interviews for each state. 3. Now calculate the difference between NJ and PA of the time differences just obtained. 4. What is the interpretation of such a difference-in-differences estimate of the employment effect? Under what condition does this provide a valid estimate of the of the minimum wage increase on employment in the fast food industry? b) Repeat the analysis in part a) for the wage (wage_st). c) Instead of using averages across restaurants, difference-in-difference estimates can also be calculated from a regression using data on individual restaurants: ??,?,?=?+???????,?+??????+?(??????,?×?????)+??,?,? Where ??,?,? is employment or wages in restaurant i in state s and period t, ??????,?is an indicator for the treatment area (NJ), ????? is an indicator for the treatment period (Nov/Dec) and ??????,?×????? is the interaction of these two dummies. 1. What are the regression DD estimates on employment and wages using this regression? How do they compare to the results you found in a) and b)? What can you conclude about the effect of the increase in the minimum wage on employment and wages? 2. The regression allows you to control for other factors. Repeat the regressions, including a dummy variable for whether the restaurant is company owned (co_owned) as compared to franchised) and three dummy variables for three of the four chains in the dataset (Burger King, KFC, Roy Rogers, and Wendy’s; you will have to construct the dummies from the variable chain or use i.chain). Do your results change when you enter restaurant-specific covariates? Would you have expected the results to change? Explain why or why not. d) An alternative to comparing NJ and PA restaurants is comparing restaurants within NJ which have high and low wages before the minimum wage increase. The minimum wage should only be binding for restaurants with low starting wages. 1. Would you expect the DD assumption to be satisfied more easily for the within NJ comparison than for the NJ - PA comparison? 2. Restrict your sample to restaurants in NJ. Construct a variable for those restaurants paying starting wages of less than $5.00 before the minimum wage increase. Use the regression to obtain a DD estimate of the employment and wage effects of the minimum wage increase. What is the relative impact of the minimum wage on employment and wages within NJ? How do your within NJ estimates compare to the ones obtained in part c) for the NJ - PA comparison? e) You can do the same analysis as in part d) for PA. 1. Restrict your sample to restaurants in PA. Construct a variable for those restaurants paying starting wages of less than $5.00 in PA in the initial period. Use the regression to obtain a DD estimate of the employment and wage effects in PA. How do your results differ from those just for NJ obtained in part d)? 2. Why is this a check on how well the DD methodology is doing at uncovering the effects of the minimum wage increase? What do you conclude? References: Card, David and Krueger, Alan B. (1994), “Minimum Wages and Employment: a Case Study of the Fast-Food Industry in New Jersey and Pennsylvania”, The American Economic Review, Vol. 84, No. 4, pp. 772-793.