2020 Portugal Stata Conference
Porto | 25 January
The 2020 Portugal Stata Conference was held on 25 January at the University of Porto, Faculty of Economics.
Proceedings
| 9:30–10:30 | Performing meta-analysis with Stata
Abstract:
Meta-analysis provides a theoretical framework to integrate and
analyze empirical evidence from multiple studies. It has been applied
to many areas of research such as econometrics, education, psychology,
and medicine. The new suite of commands meta provides
an integrated framework to address the different aspects of our meta-analysis in
a simple way. I'll discuss how to prepare and summarize our data,
how to address heterogeneity using random-effects models, extend
these models to the use of meta-regression, and use
post-estimation commands to perform statistical tests and assess
possible issues on our data.
Additional information:
Isabel Canette
StataCorp
|
| 11:00–12:15 | Session I: Empirical analysis using StataChair: Hugo Figueiredo
Everything you always wanted to know about sex discrimination Abstract: In Portugal, over the last two decades, the proportion of women among employed workers increased from 35 to 45 percent. This evolution was accompanied by a sharp fall in the gender wage gap from 32 to 20 percent. The improvement in the wage outcome of the women, however,
is fully accounted for by the catching up of their skills in comparison to males, after
two decades of human capital investments. By 2013 women already possess observable
characteristics that enhance productivity identical to their male counterparts. This means
that gender discrimination remained roughly constant over the 1991-2013 period. In this study,
we investigate the sources of the wage gender gap and conclude that sorting among firms and
job-titles can explain about two fifths of the wage gender gap. Peer effects seem to drive
wages of women down. In more than one way.
Contributors:
Paulo Guimarães
Pedro Portugal
Pedro Raposo
Overqualification and future labor market outcomes: Evidence from recent graduates in Portugal Abstract: Using a sample of higher education graduates in Portugal, this paper analyses the career dynamics of workers who entered the labor market for the first time in a job for which they were overqualified. Exploring a large matched employer-employee data set over the
2006-2012 period, we provide evidence that overqualification is a persistent phenomenon
where workers get trap. Six years after entering the labor market, 63% of the workers
that entered overqualified remain in that status. Finally, even though, unconditionally,
wages at entry are lower for overqualified workers when compared with the wages of
well-matched workers, overqualified workers that were able to switch to a well-matched
job over the analysed period exhibit a larger wage growth. Actually, taking into account
workers observed and unobserved permanent heterogeneity, the estimates reveal that five
years after entering the labor market overqualified individuals that were able to switch
to a well-matched job experience a wage growth that exceeds the wage growth of their
similar well-matched counterparts in 12 percentage points.
Contributors:
Anabela Carneiro
Isabel Araújo
Additional information:
context of rapid massification. Using an off-the-shelf non-parametric matching technique,
we disentangle two different sources of postgraduates' relative earnings: wage premiums
within occupations and the assignment to better paid and more complex occupations. We
show that both sources are relevant but the relative importance of the former has been
steadily increasing overtime. This evidence suggests that postgraduate degrees have
largely worked as a way to avoid falling down the occupational ladder or, in other words,
to hold on to a higher ground.
Contributors:
A. Almeida
M. Portela
J. Cerejeira
C. Braga
C. Sá
P. Teixeira
Hugo Figueiredo
Universidade de Aveiro
|
| 12:15–1:00 | Technical short presentations using StataChair: Miguel Portela
Python applications in Stata 16 Abstract: Stata 16 has been recently released and one of its new features is the ability to embed and execute Python code from within Stata. In this prsentation, I present two simple applications that make use of this new feature. The first one, adrcheck, finds Google Maps'
suggestions for addresses stored in a Stata variable.
The second one, menuine, interfaces with INE's (Instituto Nacional de Estadística)
API to get data from a vast list of indicators directly into Stata.
Gustavo Iglésias
Banco de Portugal
Additional information:
Miguel Portela
University of Minho
Additional information:
|
| 2:00–3:00 | Keynote speakerOccupational sorting and wage gaps of refugees Abstract: Refugee workers start low and adjust slowly to the wages of comparable natives. The innovative approach in this study using unique Swedish employer-employee data is to show that the gap is mainly caused by occupational sorting into cognitive and manual tasks.
Within occupations, it can be largely explained by differences in work experience,
while the remaining gap might be caused by discrimination. The identification
strategy relies on a control group of matched natives with the same characteristics
as the refugees, using population-level panel data for 2003–2013 to capture unobserved
heterogeneity.
Contributors:
Hans Lööf
Andreas Stephan
Klaus F. Zimmermann
Kit Baum
Boston College
|
| 3:00–4:15 | Session II: Empirical analysis using StataChair: João Cerejeira
Tell me who you consult with and I'll tell you how compliant you are: Can physicians affect patient adherence? Abstract: Relying on a large longitudinal matched physician-prescription-patient dataset for all Portuguese e-prescriptions (SPMS, 2015-2017) we've extended the paper by Koulayev et al. (2017) which analyzes the contributions of doctor-specific, patient specific, and
drug-specific factors to the adherence decision, now applied to hypoglycemic agents and
their pharmacotherapeutic groups.
Contributors:
Nuno Sousa Pereira
Ricardo Vicente
Joana Gomes da Costa
European funds and firm dynamics: Estimating spillovers from increased access Abstract: We take advantage of a quasi-natural experiment to assess the impact of European funds on firm dynamics in regions that, while not having their status changed, saw their neighbours increased access to European funds. Causality is established in a difference-in-differences
intention to treat setting, using a rich dataset that considers the universe of
Portuguese mainland municipalities from 2003 to 2010, and controlling for
socio-economic, political and demographic variables. Our findings suggest a causal
impact of between 1 and 2 percent in private sector firms´ entry and net entry r
ates, while we find no impact on firm exit rates. We consider time and space
placebos to assure the reliability of our estimates. Our findings suggest that
EU regional funds have a greater impact in times of distress, such as the world
economic crisis, as far as entry rates are concerned. The analysis of the
cross-section of firm demonstrates it is domestic owned micro firms in the primary
and tertiary sectors that are most impacted by regional funds.
Contributors:
José Tavares
João Pereira dos Santos
NOVA SBE
Where or whom to contract? An empirical study of political spillovers in public procurement Abstract: A growing body of empirical research details how public procurement processes deviate from the competitive, efficient outcomes. In particular, municipal public procurement decisions are both affected by geographical proximity and political considerations.
This presentation extends the analysis that by now has solely focused on single
municipalities and poses the hypothesis they are also affected by political proximity.
Specifically, it asks if there is a relation between the political parties in power in
Portuguese municipalities and the frequency of contracts awarded to a given firm? We
rely on a dataset ("base.gov") with information on all bids by private firms and all
contracts awarded by the 308 Portuguese municipalities in the period between 2008 and
2017. This includes three electoral cycles and more than 250,000 contracts. Our results
show that municipalities are more likely to award contracts to firms have won in other
municipalities led by the same political party. This political proximity effect is robust
to virtually all controls, including geographic proximity. This result has political and
public governance implications.
Contributors:
Pedro Camões
João Cerejeira
|
| 4:45–6:00 | Technical short presentations using StataChair: Paulo Guimarães
checkmd: Utility tool for database quality control checks Abstract: Quality control is of paramount importance in data management and sources of data contamination can be eliminated or flagged by using quality control techniques. In this presentation, we present a Stata community-contributed command
checkmd that allows to examine logical conditions in the data. The command
produces an HTML report with a description of each condition checked, including
frequency tables of consistent and inconsistent values and lists of observations
that have the most remarkable inconsistencies.
Emma Szhao
BPLIM/BdP
validarcae: Utility tool to deal with the Portuguese classification of economic activities (CAE) Abstract: The Portuguese Classification of Economic Activities establishes the common categorical system to report economic activities in Portugal and suffered several revisions over time. In this presentation, I will present validarcae, a community-contributed
Stata command that allows to validate the codes in a string or numerical variable reporting
the economic activity according to the revision specified by the user. This tool also allows
to obtain different types of aggregation for valid codes according to each revision.
Marta Silva
BPLIM/BdP
Additional information:
for fast estimation in the presence of HDFE. Since the code is built around the reghdfe package
it has similar syntax and supports many of the same functionalities. ppmlhdfe also implements
a novel and more robust approach to check for the existence of (pseudo) maximum likelihood estimates.
Contributors:
Sérgio Correia
Tom Zylkin
Paulo Guimarães
Banco de Portugal
Additional information:
|
| 6:00 |
Open panel discussion with a Stata developer
Isabel Canette
StataCorp
|
Scientific committee
|
Miguel Portela University of Minho |
Anabela Carneiro University of Porto |
João Cerejeira University of Minho |
Paulo Guimarães |
Logistics organizer
The 2020 Portugal Stata Conference is jointly organized by
Timberlake Consultants,
the distributor of Stata in Portugal, and the Faculdade de Economia da Universidade do Porto.
View the proceedings of previous Stata Conferences.
