uu.seUppsala University Publications
Change search
CiteExportLink to record
Permanent link

Direct link
Cite
Citation style
  • apa
  • ieee
  • modern-language-association
  • vancouver
  • Other style
More styles
Language
  • de-DE
  • en-GB
  • en-US
  • fi-FI
  • nn-NO
  • nn-NB
  • sv-SE
  • Other locale
More languages
Output format
  • html
  • text
  • asciidoc
  • rtf
Comparing Discrete Time Multi-state Models Using Dissimilarities
Bocconi University, Department of Decision Sciences.
Bocconi University, Department of Social and Political Sciences.
Uppsala University, Disciplinary Domain of Humanities and Social Sciences, Faculty of Social Sciences, Department of Economics. Uppsala University, Units outside the University, Office of Labour Market Policy Evaluation.
(English)Manuscript (preprint) (Other academic)
Abstract [en]

In this paper we propose different methods for comparing the predictive performance of non-nested event history models. These models are generally used to study transitions across states and over time in a variety of settings, both in biomedical studies and in the social sciences. Our goal is to compare the ability of competing models to generate trajectories that are similar to those observed in the sample at hand. To this aim, we first introduce alternative distance-based criteria to compare pairwise dissimilarities computed between observed and model-generated sequences. Next, we estimate two alternative discrete time semi-Markov models using data on family formation and childbearing decisions from the Dutch Fertility and Family Surveys study. Finally, we use the estimated models to simulate event histories and to illustrate the proposed comparison criteria.

Keywords [en]
multi-state models, dissimilarity, micro-simulation, sequence analysis, survival analysis
National Category
Economics
Identifiers
URN: urn:nbn:se:uu:diva-382739OAI: oai:DiVA.org:uu-382739DiVA, id: diva2:1312956
Available from: 2019-05-01 Created: 2019-05-01 Last updated: 2019-05-03
In thesis
1. Essays on Event History Analysis and the Effects of Social Programs on Individuals and Firms
Open this publication in new window or tab >>Essays on Event History Analysis and the Effects of Social Programs on Individuals and Firms
2019 (English)Doctoral thesis, monograph (Other academic)
Abstract [en]

Essay I: This paper studies threat effects of unemployment insurance (UI) benefit sanctions on job exit rates. Using a difference-in-differences design, I exploit two reforms of the Swedish UI system that made monitoring and sanctions considerably stricter at different points in time for different jobseeker groups. I find that men and long-term unemployed respond to the stricter UI rules by finding jobs faster. I also estimate the effect of receiving a sanction on the job exit rates, and find significant sanction imposition effects. However, a decomposition exercise shows that these effects explain very little of the overall reform effects, which instead are driven the threat of sanction imposition.

Essay II (with Gerard J. van den Berg and Johan Vikström): We use an Empirical Monte Carlo design and rich administrative data to generate realistic placebo treatment durations. First, we highlight important confounders to be controlled for when estimating selection models. Next, we omit some of the covariates used to simulate placebo treatments, and we estimate Timing-of-Events models. The model is generally able to adjust for a large share of the resulting unobserved heterogeneity. However, we find that specifying too many or too few support points to approximate the unobserved heterogeneity distribution leads to large bias. Information criteria that penalize parameter abundance can help selecting the appropriate number of support points.

Essay III (with Oskar Nordström Skans and Johan Vikström): We study how targeted wage subsidies affect the performance of the recruiting firms. Using Swedish linked employer-employee data from 1998–2008, we show that the firms hiring through subsidies substantially outperform other recruiting firms, despite identical pre-treatment performance levels and trends in a wide set of key dimensions. The pattern is less clear from 2007 onwards, after a reform removed the involvement of caseworkers from the subsidy approval process. Our results suggest that targeted employment subsidies can have large positive effects on outcomes of the hiring firms, at least if the policy environment allows for pre-screening by caseworkers.

Essay IV (with Raffaella Piccarreta and Marco Bonetti): We propose different methods for comparing the ability of competing non-nested event history models to generate trajectories that are similar to the observed ones. We first introduce alternative criteria to compare pairwise dissimilarities between observed and simulated sequences. Next, we estimate two alternative multi-state models using data on family formation and childbearing decisions from the Dutch Fertility and Family Survey. We use the estimated models to simulate event histories and to illustrate the proposed comparison criteria.

Place, publisher, year, edition, pages
Uppsala: Department of Economics, 2019. p. 152
Series
Economic studies, ISSN 0283-7668 ; 180
Keywords
Labor Economics, Unemployment Insurance, Job Search, Monitoring and Sanctions, Policy Evaluation, Dynamic Treatment Evaluation, Duration Analysis, Firm performance, Employer-employee Match, Monte Carlo simulation
National Category
Economics
Research subject
Economics
Identifiers
urn:nbn:se:uu:diva-382742 (URN)978-91-506-2769-5 (ISBN)
Public defence
2019-08-30, Hörsal 2, Ekonomikum, Kyrkogårdsgatan 10, Uppsala, 10:15 (English)
Opponent
Supervisors
Available from: 2019-06-25 Created: 2019-05-03 Last updated: 2019-06-25

Open Access in DiVA

No full text in DiVA

Authority records BETA

Lombardi, Stefano

Search in DiVA

By author/editor
Lombardi, Stefano
By organisation
Department of EconomicsOffice of Labour Market Policy Evaluation
Economics

Search outside of DiVA

GoogleGoogle Scholar

urn-nbn

Altmetric score

urn-nbn
Total: 18 hits
CiteExportLink to record
Permanent link

Direct link
Cite
Citation style
  • apa
  • ieee
  • modern-language-association
  • vancouver
  • Other style
More styles
Language
  • de-DE
  • en-GB
  • en-US
  • fi-FI
  • nn-NO
  • nn-NB
  • sv-SE
  • Other locale
More languages
Output format
  • html
  • text
  • asciidoc
  • rtf