Education, Employment and Earnings of Secondary School-Leavers ...
Education, Employment and Earnings of Secondary School-Leavers ... Education, Employment and Earnings of Secondary School-Leavers ...
eplaced by the assumption that y * i | xi ~ N(the log likelihood function as follows: 17x ' i ,σ 2 ). This then allows specification ofJL = j=0yi= jlog e [Φ(aj− x i' σ) – Φ(aj - 1− x i' σ)] [2]where the first summation operator sums across individuals within the given jcategory and log e (·) denotes the natural logarithmic operator. The maximumlikelihood procedure involves the estimation of the β parameter vector and theancillary standard error parameter σ. Given that the introduction of the knownthresholds fixes the scale of the dependent variable, the estimated coefficients areamenable to a direct and more intuitive OLS-type interpretation. The estimatescontained in the β parameter vector are interpretable on the assumption that we haveactually observed the y * ioutcome for each of the i individuals in the sample.It is important to evaluate our empirical models in regard to certain key econometricassumptions. The adequacy of the estimated models is assessed using the efficientscore tests suggested by Chesher and Irish (1987). These tests require computation ofthe models’ pseudo-residuals. In general terms, the pseudo-residual for the i thindividual is defined for the interval regression model as:u i =aj- 1−'φ(xi) −φσaj−'σ[Φ(xi) −Φσ((aj− x'i)σaj- 1−x'iσ)][3]where φ(·) denotes probability density function operator for the standard normal. 1817 The only difference between this log-likelihood function and that of the ordered probit is that in ourcase the threshold values are known and σ is estimated and not constrained to unity. The exactknowledge of the thresholds allows identification of the scale of the dependent variable and estimationof σ.18 The pseudo-residuals can be obtained by differentiating the log-likelihood expression in [2] withrespect to the constant term implicit in the x vector.13
- Page 1 and 2: Education, Employment and Earnings
- Page 3 and 4: The structure of the paper can now
- Page 5 and 6: Despite large absolute increases in
- Page 7 and 8: nine studies reviewed for sub-Sahar
- Page 9 and 10: It should be noted, however, that e
- Page 11 and 12: history, current activity and incom
- Page 13: In order to understand how the mode
- Page 17 and 18: where I( ∧ ) is the information
- Page 19 and 20: 5. Empirical ResultsThe approach ad
- Page 21 and 22: preferred specification. This varia
- Page 23 and 24: mechanisms that encourage an inter-
- Page 25 and 26: 5.3 The Estimated School EffectsOne
- Page 27 and 28: 5.4 Estimated Private Rates of Retu
- Page 29 and 30: obtained using the tracer study dat
- Page 31 and 32: elates to the role of father’s ed
- Page 33 and 34: ReferencesAl-Samarrai, S. and Benne
- Page 35 and 36: Ramsey, J.B. (1969) Tests for speci
- Page 37 and 38: Table 4: Interval Regression Estima
- Page 39 and 40: Table 5: Interval Regression Estima
- Page 41 and 42: Notes to table 6:(a) See expression
- Page 43: Table A2: Summary StatisticsVariabl
eplaced by the assumption that y * i | xi ~ N(the log likelihood function as follows: 17x ' i ,σ 2 ). This then allows specification <strong>of</strong>JL = j=0yi= jlog e [Φ(aj− x i' σ) – Φ(aj - 1− x i' σ)] [2]where the first summation operator sums across individuals within the given jcategory <strong>and</strong> log e (·) denotes the natural logarithmic operator. The maximumlikelihood procedure involves the estimation <strong>of</strong> the β parameter vector <strong>and</strong> theancillary st<strong>and</strong>ard error parameter σ. Given that the introduction <strong>of</strong> the knownthresholds fixes the scale <strong>of</strong> the dependent variable, the estimated coefficients areamenable to a direct <strong>and</strong> more intuitive OLS-type interpretation. The estimatescontained in the β parameter vector are interpretable on the assumption that we haveactually observed the y * ioutcome for each <strong>of</strong> the i individuals in the sample.It is important to evaluate our empirical models in regard to certain key econometricassumptions. The adequacy <strong>of</strong> the estimated models is assessed using the efficientscore tests suggested by Chesher <strong>and</strong> Irish (1987). These tests require computation <strong>of</strong>the models’ pseudo-residuals. In general terms, the pseudo-residual for the i thindividual is defined for the interval regression model as:u i =aj- 1−'φ(xi) −φσaj−'σ[Φ(xi) −Φσ((aj− x'i)σaj- 1−x'iσ)][3]where φ(·) denotes probability density function operator for the st<strong>and</strong>ard normal. 1817 The only difference between this log-likelihood function <strong>and</strong> that <strong>of</strong> the ordered probit is that in ourcase the threshold values are known <strong>and</strong> σ is estimated <strong>and</strong> not constrained to unity. The exactknowledge <strong>of</strong> the thresholds allows identification <strong>of</strong> the scale <strong>of</strong> the dependent variable <strong>and</strong> estimation<strong>of</strong> σ.18 The pseudo-residuals can be obtained by differentiating the log-likelihood expression in [2] withrespect to the constant term implicit in the x vector.13