This preview shows page 1. Sign up to view the full content.
Unformatted text preview: ounds 0.45
24 25 26 27 28 29 30 31 32 33 34
Age (Census day) 35 36 37 38 39 40 Fig. 4. Employment rate in previous year, Quebec 1986. Table 1
Regression discontinuity estimates of the effect of higher social assistance beneﬁts on labour supply in Quebec, 1986
Speciﬁcation for age Empl. rate last year Empl. rate at Census Difference in empl. rate Weekly hours Mean of the dependent variable
0.562
Regression discontinuity estimates
Linear
À0.045***
(0.012)
Quadratic
À0.048***
(0.013)
Cubic
À0.043**
(0.018)
Linear spline
À0.047***
(0.013)
Quadratic spline
À0.038
(0.024)
Goodness of ﬁt statistic (pvalue)
Linear
0.48
Linear spline
0.47 0.618 0.056 À0.041***
(0.012)
À0.051***
(0.012)
À0.048***
(0.014)
À0.049***
(0.011)
À0.056**
(0.018) À0.029**
(0.011)
À0.031**
(0.012)
À0.030**
(0.013)
À0.032**
(0.013)
À0.035*
(0.016) À1.45**
(0.54)
À1.75**
(0.61)
À1.47*
(0.70)
À1.72***
(0.55)
À1.66
(0.94) 0.91
0.85 0.48
0.55 0.52
0.72 24.39 Note:
***
Indicate statistical signiﬁcance at the 1% level.
**
For the 5% level.
*
For the 10% level. One nice feature of the results is that the two employment rate measures yield remarkably similar
estimates that range from À0:038 to À0:056. This suggests that the RD approach is appropriate for
the models of previous year outcomes despite some of the data shortcomings discussed in Section 4.
Note also that the goodnessofﬁt tests suggest that even the simpler models (linear or linear spline) ﬁt
the data very well. Fortin – Econ 560 Lecture 1C The authors also take advantage of two records of the employment rate, one in the
week of the Census ERCia and another in the previous year ERL'ia to estimate a FDRD, under the assumption that 1 = 1 '
ERC ia ERL 'ia ( 0 0 ' ) 1 (Treat ia Treat 'ia ) ( age ) ( ia 'ia ),
that eliminates individual fixed effects. The cancellation of the lower benefit policy in 1989 makes a pre and post1989
comparison natural, and the comparison with the rest of in Canada gives the authors
the opportunity to compare the results from the regression discontinuity estimator to
results from more traditional differenceindifference estimators. They find that without a control group placed in the same labour market as the
treatment group, the differenceindifference estimates can diverge greatly from the
regression discontinuity estimates....
View
Full
Document
This document was uploaded on 02/26/2014 for the course ECON 560 at The University of British Columbia.
 Fall '13
 NicoleFortin
 Economics

Click to edit the document details