University of Warwick Ec226 Econometrics 1 Test 1

UNIVERSITY OF WARWICK EC226 ECONOMETRICS 1 TEST 1: December 2009

Time allowed 50 minutes. Answer all questions. Calculators may be used. A 2-sided sheet of notes may be consulted.

1. Over the year 2008 a survey was undertaken of people who played on-line poker to find the determinants of expenditure on gambling. The following model was estimated for a sample 586 individuals responding to the survey:

ln(Gi )      agei  3 ln(Yi )   4  ln(Yi )   5 Femi   6Ci   7 Mari   i

2

(1)

where Gi age Yi Fem C Mar

- £ spent last month on all forms of gambling (excluding lottery expenditure). - Age in years. - Gross monthly income (from all sources), in £ (min=600, max=5,800) - 1 if individual is female, 0 otherwise. - 1 if individual smokes, 0 otherwise. - 1 if individual is married, 0 otherwise.

ln is the natural logarithm. The results from estimating this equation by Ordinary Least Squares (standard errors in parentheses) were:

ln(Gi )  2.17  0.032agei  0.400 ln(Yi )  0.023  ln(Yi )   0.086 Femi

2

(4.62) (0.0083) (0.008)

2

(0.103)

(0.009)

(0.0012)

 0.231Ci  0.0221Mari  ei (0.0111)

RSS = 111.26, R  0.101 , cov(b3 , b4 )  0.0009 . (i) (ii) Give a brief interpretation of the coefficient on the variable C. What do the estimated coefficients on the income variables, tell you about the responsiveness of gambling to income. Does this make any economic sense? Find the income at which gambling expenditure is a maximum.

(Continued)

(iii)

At the 5% significance level, test that the elasticity of gambling with respect to income is zero at ln(Y )  6.5 . [36 marks]

A new equation is specified as:

ln(Gi )      agei   3 ln(Yi )   4  ln(Yi )   5 Femi  6 Ci  7 Mari

2

 8 Femi  Ci   i

(2)

(iv) (v)

Interpret the coefficient 8 in equation (2). Estimating (2) by OLS yielded an estimated coefficient on Fem of -0.1422. Explain why this coefficient estimate might be different from...