The Limits of Mathematics and NP Estimation in ... - Chichilnisky

More documents

Recommendations

Info

The Impact of Government-SponsoredTraining Programs on the Labor Market Transitions of Disadvantaged MenThe Impact of Government-Sponsored Training Programs on the Labor Market Transitions of Disadvantaged Men 2975The last section of the panel reports the implicit correlations between the origin and thedestination states. Note that the correlation matrix need not be symmetric nor does thediagonal need be equal to unity. On the other hand, the restrictions that were imposed toachieve identification of the loading parameters imply that the first row of the matrix is equalto the first row of the matrix of the middle section.For the sake of brevity we will focus our attention on the most interesting correlations. Theestimates suggest that those who are likely to have short welfare training spells are also lesslikely to transit through welfare or JRP and more likely to enter employment (row 2). Similarly,row 3 indicates that individuals who are likely to have short JRP spells are less likely to returnto either welfare or welfare training in the future, and much more likely to enter employment.Finally, those who have short UI training spells (row 5) have higher transitions rates intowelfare and welfare training, and lower transitions rates into employment.These correlations suggest there is considerable selectivity into the training programs.Furthermore, they show that those who are selected into welfare and JRP training programsappear to be different from those who participate in UI training programs. As a matter of fact,all the correlation coefficients of the last section of the panel pertaining to UI training havethe opposite sign to those of welfare and JRP training. Consider, for example, those who haveunexpectedly long UI training spells and those who have unexpectedly short welfare trainingor JRP spells. According to the last section of the panel, all these individuals are more likelyto move into employment upon exiting their respective spells than average. Yet, the middlesection indicates that only those on welfare training or JRP are likely to have long employmentspells. Those who were on UI training are more likely to have short employment spells.That those who are likely to have short JRP or welfare training spells are more likely toexperiment long employment spells may be somewhat surprising. In fact, when studying theimpact of the Youth Training Scheme in the UK, Mealli et al. (1996) and Mealli & Pudney (2003)conjectured that early program termination may result from more intensive search stemmingfrom better than average motivation. Hence, early termination may be associated with ahigher probability of transition into employment and longer employment spells. Conversely,if failure to complete the full term is a consequence of low ability and motivation, it may beassociated with poorer employment outcomes. Although we have no information regardingprogram completion, our results are consistent with the first possibility, whereas those ofMealli et al. (1996) were consistent with the second possibility.5. ConclusionThe analysis has focused on an examination of the impact of government-sponsored trainingprograms aimed at disadvantaged male youths on their labour market transitions. We haveelected to concentrate our attention on this group since they have fared relatively poorly onthe labour market over the past decade in Canada by all accounts. The richness of the dataat our disposal has allowed us to recreate very detailed individual histories over a relativelylong period. As many as seven distinct states on the labour market could be identified in thedata.This study has applied a continuous time duration model to estimate the density of durationtimes in these seven states, controlling for the endogeneity of an individual’s training status.Most previous studies have used survey or administrative data that were less amenableto the kind of analysis performed in this chapter. Depending on the nature of the data,
76 Advances in Econometrics - Theory and Applications30 Will-be-set-by-IN-TECHcomplex adjustments to the model were often required to account for potential problemsrelated to stock sampling and initial conditions. Fortunately, we were able to avoid thesedifficulties by recreating each individual’s history as early as age 16, the legal school-leavingage in Canada. Consequently, the initial state can be safely considered exogenous, and thesubsequent duration times void of any form of bias.There is no consensus in the literature concerning the appropriate treatment of unobservedheterogeneity in multi-states multi-episodes duration models. When few states areconsidered, two-factor loading models with a finite set of points of support have becomerelatively standard. When the analysis focuses on more states, factor loading models require alarge number of parameters to be flexible or become relatively restrictive if a parsimoniousspecification is used. In this chapter we have chosen to investigate the sensitivity of theparameter estimates by comparing a typical non-parametric specification and a series ofparametric two-factor loading models. These models implicitly assume that the intensityof transitions are related to the state of destination. We have also estimated a parametricthree-factor loading model. The novelty of this specification lies in the fact that the intensitiesof transitions are related to both to the state of destination and the state of origin.The estimation of the model yields a number of interesting results. As found in previousstudies, unobserved heterogeneity appears to play an important role in determining whoselects or gets selected in training programs. On the other hand, the slope and baseline hazardparameter estimates are not very sensitive to the choice of a particular distribution functionfor the unobserved heterogeneity variables. The two-factor loading models, either parametricor non-parametric, yield essentially the same results as the three-factor loading model. Theseshow that the duration times in any of the seven states considered are sensitive to variationsin program parameters such as welfare benefits, policy variables such as the minimum wagerate, and in the economic environment as proxied by the unemployment rate. Nearly all theparameter estimates have the expected sign when statistically significant.The results pertaining to the impact of the welfare training programs and JRP are similarto those found earlier by Gritz (1993), Bonnal et al. (1997) and Fortin et al. (1999a). Inessence, young, poorly educated males who participate in these programs do not fair as wellon the labour market compared to non-participants, even after controlling for unobservedheterogeneity. On the other hand, participation in training programs while on unemploymentinsurance provides them some benefits in the form of increased transitions into employment.6. ReferencesBeaudry, P. & Green, D. A. (2000). Cohort patterns in canadian earnings: assessing the role ofskill premia in inequality trends, Canadian Journal of Economics 33: 907–936.Blasco, S., Crépon, B. & Kamionka, T. (2009). Evaluation of the Impact of Training on IndividualLabor Market Transitions, Mimeo, Crest, Paris.Bonnal, L., Fougère, D. & Sérandon, A. (1997). Evaluating the impact of french employmentpolicies on individual labour market histories, The Review of Economics Studies64(4): 683–718.Brouillette, D. & Lacroix, G. (2010). Heterogeneous treatment and self-selection in a wagesubsidy experiment, Journal of Public Economics 94(7-8): 479 – 492.Brouillette, D. & Lacroix, G. (2011). Assessing the impact of a wage subsidy for single parentson social assistance in canada, Canadian Journal of Economics, Forthcoming.
Page 2 and 3:
ADVANCES INECONOMETRICS - THEORYAND
Page 4:
free online editions of InTechBooks
Page 8 and 9:
PrefaceEconometrics is a (sub)disci
Page 10:
PrefaceIXLastly, I would like to th
Page 14:
10The Limits of Mathematicsand NP E
Page 17 and 18:
6 Advances in Econometrics - Theory
Page 19 and 20:
8 Advances in Econometrics - Theory
Page 21 and 22:
10 Advances in Econometrics - Theor
Page 23 and 24:
Page 25 and 26:
Page 27:
Page 32 and 33:
Instrument Generating Functionand A
Page 34 and 35:
Instrument Generating Functionand A
Page 36 and 37: Instrument Generating Functionand A
Page 38 and 39: Instrument Generating Functionand A
Page 41 and 42: 30Advances in Econometrics - Theory
Page 44: Recent Developments in SeasonalVola
Page 47 and 48: 36 Advances in Econometrics - Theor
Page 58 and 59: The Impact of Government-Sponsored
Page 60 and 61: The Impact of Government-SponsoredT
Page 90 and 91: Returns to Education and Experience
Page 108 and 109: Using the SUR Model of Tourism Dema
show all

The Limits of Mathematics and NP Estimation in ... - Chichilnisky

Create successful ePaper yourself

Delete template?

Save as template?