1.  (a) The UK’s Chancellor of the Exchequer hypothesises that 3 out of

10 individuals in the UK are below the poverty line. In a sample      of 100 individuals, 22 of them are below the poverty line with      standard deviation 0.4.  By stating the null and the alternative      hypothesis, perform a test of the above hypothesis if the critical      value is 1.993.                                                                        [6]

The UK’s Chancellor of the Exchequer now hypothesises that      the fraction of people that is below the poverty line is the same      in the UK as in France.  In a sample of 100 individuals from      the UK, 22 of them are below the poverty line with standard      deviation 0.4.  In a sample of 80 individuals from France, 28 of     them are below the poverty line with standard deviation 0.3. By      stating the null and the alternative hypothesis, perform a test      of the above hypothesis if the critical value is 2.575.                 [6]

(b) We have UK individual data for the year 2018 of the variables hourpay which is the hourly wage in ↔ , durun which is the duration of unemployment in months, age which is the age in years and marsta which is the marital status and has 4 cate- gories (Single, Married, Divorced and Widowed). We run an OLS regression and we obtain the following regression output:

(a) Write the econometric model, the estimated model and ex- plain.

(b) Interpret the coefficients.

(c) What is approximately the average hourpay for a 30 years old married individual who was 2 years in unemployment?

2.  (a) Describe and analyse the Blinder-Oaxaca Method (show every step clearly and use graphical analysis in your answer).

(b) Suppose that hourly wages in pounds (Wage) depend only on years of education (Educ), months of experience (Exper) and months in job related training (Train).  In the below table a researcher produced the descriptive statistics of those variables (top) and the estimated coefficients from regressing Wage on Educ, Exper and Train (bottom), separately for Male and Female:

(i) Interpret all the estimated coefficients.

(ii) Find the male-female wage differential.

(iii) How much of the male-female differential is explained by differences in Educ, in Exper and in Train between women and men and how much by discrimination?

3.  (a) We have annual UK data on the total number of crimes (C)     and on the number of armed police officers (P). By running an     OLS estimation, propose a suitable model that can capture the      effect of the number of armed police officers on total crimes.     Explain why this model might suffer from an endogeneity bias.   [4]

A researcher observed that during elections (E), which are re-     peated every four years, there is a high increase in the number      of armed police officers. Explain if E is a good instrument and      describe and analyse which methodology can be used to address      the above endogeneity bias.                                                     [9]

(b) In the article “The medical care costs of obesity: An instrumen- tal variables approach” 2012 by Cawley & Meyerhoefer:

(i) analyse whether positive correlation between BMI and Total      Medical Expenditures implies causality.                                [4]

(ii) analyse whether the instrument that they use is powerful and      valid.               [4]