Columbia MicaSviatschi

AY #________________
4883057 FEE_______________________________________ APPL #________________
APPLICATION FOR ADMISSION
__________________________________________________________________________________________
___________________________________________
Sviatschi ________________________________________
Maria _________________________________
Micaela
LAST NAME (FAMILY/SURNAME) FIRST NAME (GIVEN) MIDDLE NAME
____________________________________________________________________________________________________ ___________________
05/29/1986 Female
_____________
VARIATIONS OF YOUR NAME ON OFFICIAL RECORDS DATE OF BIRTH SEX
_______________________
Y White
_____________________________________________________________________ ______________________________________
Other Latina/o
HISPANIC RACE ETHNICITY
y
Economics
_________________________________________________ Economic Development
________________ ________________ Autumn
availableTerm 2011 ________________
No subfield choices________________ M.A./M.Phil./Ph.D.
PROGRAM OF STUDY 1 SUBFIELD 1 SUBFIELD 2 TERM DEGREE
op
_________________________________________________ ________________ ________________ NA
_______________________________________
PROGRAM OF STUDY 2 SUBFIELD 1 SUBFIELD 2 COMPARATIVE LIT. OR INSTITUTE
__________________________________________________________________________________________
___________________________________________________________________
Cerrito 1548 floor 16 "86" CITIZENSHIP: U.S. U.S. PERMANENT RESIDENT: _____________
tC
ADDRESS ALIEN NUMBER
___________________________________________________________________
Buenos Aires ✔ INTERNATIONAL: __________________________
Argentina ____________
F-1 Visa - Desired
STREET ADDRESS LINE 2 CITY COUNTRY VISA TYPE
___________________________________________________________________
1010 Argentina _____________________________________________________________________
Buenos Aires Argentina
STATE POSTAL CODE COUNTRY CITY/STATE OF BIRTH COUNTRY OF BIRTH
an
_msviatschi@gmail.com 054 1148154832
___________________________________________________________________ _____________________________________________________________________
Argentina
EMAIL ADDRESS TELEPHONE STATE/PROVINCE OF RESIDENCE COUNTRY OF RESIDENCE
__________________________________________________________________________________________
_____________ __________________________________________________ ____________________________________________________
ic
961328 University of San Andres Economics

SCHOOL CODE SCHOOL NAME MAJOR
_____________
Complete _______________________________
BA ______________
8.12 _____________________________
03/2004 12/2007 __________________
08/2008
STATUS DEGREE GPA DATES ATTENDED AWARD DATE
pl
_____________
961328 __________________________________________________
University of San Andres ____________________________________________________
Economics
Ap
_____________
Complete _______________________________
MA ______________
8.33 _____________________________
03/2008 12/2008 _________________
12/2010
_____________ __________________________________________________ ____________________________________________________

_____________ _______________________________ ______________ _____________________________ __________________

_____________ __________________________________________________ ____________________________________________________

SCHOOL CODE SCHOOL NAME DATES ATTENDED
_____________ __________________________________________________ ____________________________________________________

SCHOOL CODE SCHOOL NAME DATES ATTENDED
__________________________________________________________________________________________
_________________________________________________
Guillermo Cruces _____________________________________________ ______________________________________
Professor
NAME OF EVALUATOR INSTITUTION TITLE
Rafael Di Tella
_________________________________________________ _____________________________________________ D.Phil
______________________________________
Martin Rossi
_________________________________________________ _____________________________________________ D.Phil
______________________________________
Yuri Soares
_________________________________________________ _____________________________________________ Ph.D
______________________________________
__________________________________________________________________________________________
DEPARTMENT RECOMMENDATION AND COMMENTS INSTITUTE RECOMMENDATION AND COMMENTS
ADMIT WITH FELLOWSHIP WAIT LIST
ADMIT DEFERRED FELLOWSHIP REJECT ADMIT REJECT
ADMIT PUBLIC HEALTH FORWARD TO ANOTHER SCHOOL ADMIT WITH MONEY
ADMIT NON-DEGREE FORWARD TO ANOTHER DEPARTMENT
ADMIT MA ONLY
_____________________________________________________ _____________________________________________
DEPARTMENT SIGNATURE DATE INSTITUTE SIGNATURE DATE
__________________________________________________________________________________________
TOEFL (CBT / PBT) TOEFL (IBT) IELTS ALP/ECT
TEST DATE: TEST DATE: 04/2010 TEST DATE: TEST DATE:
LISTENING LISTENING 28 LISTENING SCORE
STRUCTURE WRITING WRITING 27 WRITING
READING READING 28 READING
ESSAY / TWE SPEAKING 22 SPEAKING
TOTAL SCORE TOTAL SCORE 105 OVERALL BAND SCORE
GRE GENERAL GRE SUBJECT GMAT

TEST DATE: 09/2010 SCORE PERCENTILE TEST DATE: TEST DATE: SCORE PERCENTILE
VERBAL 320 13 TEST TAKEN VERBAL
QUANTITATIVE 790 91 CODE QUANTITATIVE
ANALYTICAL WRITING 4 45 SCORE TOTAL
y
PERCENTILE ANALYTICAL WRITING
op
LSAT MCAT
TEST DATE: TEST DATE:
SCORE SCORE
__________________________________________________________________________________________
N
HAVE YOU PREVIOUSLY APPLIED TO GSAS? _____________________________________________________________________________________________________
tC
WERE YOU PREVIOUSLY ENROLLED AT GSAS? ___________________________________________________________________________________________________
N
ARE YOU A CURRENTLY ENROLLED AT GSAS? ___________________________________________________________________________________________________

N
HAVE YOU PREVIOUSLY APPLIED TO ANY DIVISION OF COLUMBIA UNIVERSITY? __________________________________________________________________
ARE YOU CURRENTLY APPLYING TO ANY OTHER DIVISION OF COLUMBIA UNIVERSITY? ___________________________________________________________
N
an
N
ARE YOU APPLYING TO ANY PROGRAM OFFERED BY GSAS AT THE COLLEGE OF PHYSICIANS AND SURGEONS? ______________________________________
ARE YOU APPLYING FOR A DUAL-DEGREE PROGRAM? ___________________________________________________________________________________________

N
Harvard, Berkeley, NYU, Oxford and LSE
TO WHICH SCHOOLS OTHER THAN COLUMBIA UNIVERSITY ARE YOU APPLYING TO? _______________________________________________________________
__________________________________________________________________________________________
ic
Y
DO YOU HAVE ANY FAMILY MEMBERS WHO ARE CURRENTLY ATTENDING COLUMBIA UNIVERSITY? _______________________________________________
N
pl
DO YOU HAVE ANY FAMILY MEMBERS WHO HAVE GRADUATED FROM COLUMBIA UNIVERSITY? ___________________________________________________
Y
DO YOU HAVE ANY FAMILY MEMBERS WHO ARE AFFILIATED WITH COLUMBIA UNIVERSITY? ______________________________________________________
__________________________________________________________________________________________
Ap
ARE YOU CURRENTLY PARTICIPATING IN A PROGRAM WHICH IS ELIGIBLE FOR A FEE WAIVER? _____________________________________________________
N
WHICH CONFERNECE, GRAD FAIR OR RECRUITMENT EVENT IN WHICH GSAS PARTICPATED IN DID YOU ATTEND? _____________________________________
N
IS YOUR APPLICATION BEING COMPLETED BY AN IIE/FULBRIGHT REPRESENTATIVE? _______________________________________________________________
N
ARE YOU APPLYING THROUGH AN EXTERNAL FELLOWSHIP ORGANIZATION? ______________________________________________________________________
ARE YOU CURRENTLY SERVING IN THE U.S. MILITARY AND WILL BE SUPPORTED BY A GRANT? _____________________________________________________
__________________________________________________________________________________________
HAVE YOU DISCUSSED YOUR APPLICATION WITH A MEMBER OF THE COLUMBIA FACULTY? _______________________________________________________
Moreira and Urquiola 10/2010
Leigh Linden/ Miguel Urquiola/Cristian Pop-Eleches
IS THERE A FACULTY MEMBER WITH WHOM YOU WOULD LIKE TO WORK? ________________________________________________________________________
__________________________________________________________________________________________
Spanish
FROM WHICH LANGUAGES CAN YOU TRANSLATE SCHOLARLY MATERIAL INTO ENGLISH? _________________________________________________________
IF PROFICIENCY WAS ACQUIRED BY MEANS OTHER THAN COURSEWORK, PLEASE DESCRIBE: ______________________________________________________
__________________________________________________________________________________________
N
WERE YOU EVER SUBJECT TO DISCIPLINARY SANCTIONS AS A RESULT OF PERSONAL MISCONDUCT BY ANY COLLEGE OR UNIVERSITY? _____________
N
WERE YOU EVER SUBJECT TO SANCTIONS RELATED TO ACADEMIC CONDUCT BY ANY COLLEGE OR UNIVERSITY? __________________________________
__________________________________________________________________________________________
I CERTIFY THAT THE INFORMATION GIVEN BY ME ON THIS APPLICATION (INCLUDING ALL SUPPLEMENTARY PAGES) IS COMPLETE AND
ACCURATE AND THAT ALL MATERIALSUBMITTED AS MINE IS MY OWN, ORIGINAL WORK. I UNDERSTAND THAT WILLFUL
MISREPRESENTATION OR FAILURE TO DISCLOSE COMPLETE AND ACCURATE INFORMATION MAY RESULT IN THE CANCELLATION OF MY
APPLICATION OR THE RESCINDING OF AN OFFER OF ADMISSION.
________________________________________________________________________________________________________________________________________________
SIGNATURE DATE
AY #________________
4883057 APPL #________________
EXTERNAL FUNDING
__________________________________________________________________________________________
___________________________________________
Sviatschi ________________________________________
Maria _________________________________
Micaela
LAST NAME (FAMILY/SURNAME) FIRST NAME (GIVEN) MIDDLE NAME
____________________________________________________________________________________________________ ___________________
05/29/1986 _____________
Female
VARIATIONS OF YOUR NAME ON OFFICIAL RECORDS DATE OF BIRTH SEX
y
____________________________________________________________________
Economics Autumn Term 2011
____________________________ M.A./M.Phil./Ph.D.
______________________________________
PROGRAM OF STUDY TERM DEGREE
op
__________________________________________________________________________________________
DO YOU WISH TO BE CONSIDERED FOR FELLOWSHIP AID FROM COLUMBIA? ✔ YES NO

tC
_________________________________________________________________________________________
DO YOU HOLD AN OUTSIDE AWARD? ✔ NO

YES
NAME OF AWARD ______________________________________________________________________________________________________________________________

an
________________________________ _____________________________ ______________________________ ______________________________________
TOTAL AMOUNT TUITION PER YEAR STIPEND PER YEAR TERM (NUMBER OF YEARS) OF AWARD
__________________________________________________________________ _________________________________________________________________________
AWARD CONTACT PERSON EMAIL OR TELEPHONE NUMBER OF AWARD CONTACT PERSON
ic
NAME OF AWARD ______________________________________________________________________________________________________________________________
________________________________ _____________________________ ______________________________ ______________________________________

pl
__________________________________________________________________ _________________________________________________________________________
__________________________________________________________________________________________
Ap
ARE YOU APPLYING FOR AN OUTSIDE AWARD? ✔ NO

YES
NAME OF AWARD ______________________________________________________________________________________________________________________________
________________________________ _____________________________ ______________________________ ______________________________________

__________________________________________________________________ _________________________________________________________________________
NAME OF AWARD ______________________________________________________________________________________________________________________________
________________________________ _____________________________ ______________________________ ______________________________________

__________________________________________________________________ _________________________________________________________________________
NAME OF AWARD ______________________________________________________________________________________________________________________________
________________________________ _____________________________ ______________________________ ______________________________________

__________________________________________________________________ _________________________________________________________________________
__________________________________________________________________________________________
ARE YOU APPLYING FOR A FOREIGN LANGUAGE AND AREA STUDIES FELLOWSHIP (FLAS)? ✔ NO
YES
ARE YOU APPLYING FOR THE CONSERVATION BIOLOGY SCHOLARSHIP? ✔ NO
YES
ARE YOU APPLYING FOR THE SOCIOMEDICAL SCIENCES PREDOCTORAL FELLOWSHIP IN GENDER, SEXUALITY AND HEALTH? ✔ NO
YES
Transcript 1 Maria Sviatschi
4883057
y
op
tC
an
ic
pl
Ap
4883057
y
op
tC
an
ic
pl
Ap
4883057
y
op
tC
an
ic
pl
Ap
Statement
MariaofMicaela
Academic
SviatschiPurpose
– 05/29/1986 Department of Economics, Columbia Maria Sviatschi
University
Applicant to the PhD. in Economics 4883057
Statement of Purpose
My application to the PhD. program of the Department of Economics at Columbia University reflects
my strong interest and commitment to pursue a successful academic career. Upon graduation, my aim is to
obtain employment at a research-oriented academic institution to contribute to the understanding of economic
phenomena. In particular, I would like to contribute to the design of economic policies by studying their
theoretical and empirical impact.
Over the years I discovered my passion for understanding the impact that certain economic policies
might have on societal behavior, and I found economics to be the perfect environment to achieve this
objective. My interests are multifarious but nevertheless focused on applied microeconomics and development.
y
I am especially concerned in understanding microeconomic problems that affect developing countries in the
areas of education, health, crime and justice. I am interested in not only building models to understand these
op
problems, but also corroborating their predicted effects empirically. I find the faculty of the Economics
Department at Columbia University to be an ideal community to deepen my knowledge in these as well as
other fields that might interest me in the future.
tC
My interests in economics began during high school. At that time (2001), my country was affected by
one of the worst economic crises in country’s history. Many families, including my own, saw their income
directly affected by the economic cycle. The crisis motivated my desire to understand not only the causes of
such fluctuations in economic activity but especially the role that policies might play in improving such
situations. With that in mind, after graduating from high school, I enrolled in the economics program at
Universidad de San Andrés, one of the top two schools in Argentina in this field.
an
During my undergraduate studies, my passion for the study of economics grew further. The work of
many of my professors inspired me to devote my profession to research. Courses like Economic Growth,
Political Economy, Economics of Income Distribution and Applied Economics made me realize that the field
ic
could provide me the tools to understand the causes and consequences of real-life phenomena. In my last year
of studies, I became especially interested in understanding the economic impact of legislation. In particular, I
sought to shed light on a number of important policy debates around the effectiveness of alcohol related laws
pl
on reducing car accidents. In my undergraduate thesis, I studied the impact of a law that prohibits the sale of
alcohol at night in convenience stores and gas stations on car accident mortality rates. This was the first study
to provide rigorous empirical evidence of the effect of dry laws in a developing country. For this paper, I
Ap
received two awards: the Best Undergraduate Thesis Award, given by the Universidad of San Andrés; and the
Award for Young Researchers, given by the Argentinean Association of Political Economy, which is one of the
most important prizes for young Argentinean economists. In addition, I received several grants to present this
research at various international conferences.
Consistent with my strong interest in pursuing a research-oriented academic career, I began to pursue a
Masters in Economics immediately after completing my undergraduate studies. I received an important merit-
based fellowship for my Masters studies. During this program I took Macroeconomics, Econometrics and
Microeconomic Theory courses which involved extensive use of graduate level bibliography. In the last
semester of the program I received the great opportunity to work as a Research Assistant for professors Rafael
Di Tella (Harvard Business School) and Martin Rossi (Universidad de San Andrés) in the areas most related to
my core interests in Economics. This experience further encouraged my passion for doing research by
improving my academic writing and providing training on how to organize and coordinate a research project
from the beginning.
After my graduation, I was able to join the Inter-American Development Bank as a Research Assistant
in Washington, DC. I worked on the impact evaluation of poverty-targeted programs designated to improve
access to justice in rural and poor areas in Peru and Costa Rica, providing the methodological, statistical and
econometric support to evaluate these programs. I also participated in country evaluation missions, including
conducting interviews and collecting data. During my work there, I have not only improved my analytical skills
but also learned that much of the effectiveness of public policies is limited by the absence of good research to
guide the effective use of resources.
Statement
MariaofMicaela
Academic
SviatschiPurpose
– 05/29/1986 Department of Economics, Columbia Maria Sviatschi
University
Applicant to the PhD. in Economics 4883057
Two examples of my research during my experience in an international organization include a paper

that evaluates the causal effect of the construction of courts in rural regions of Peru on conflict resolution and
perceptions regarding social mores and laws (which has an invitation for revise and resubmit at the journal
Economic Development and Cultural Change); and another that attempted to determine how court inefficiencies
might increase the economic incentives of criminal behavior. These two papers are innovative in the sense that
there is little empirical literature analyzing the effect of enhancing the access to justice on household conflict
resolution and criminal behavior. This is an unexplored topic where there are many questions to answer and I
would like to continue working on it during my PhD studies. In particular, I would like to analyze the extent to
which formal mechanisms of justice might improve the position of disadvantaged groups in rural societies such
as women or ethnic and religious minorities that are often discriminated against by informal traditional means
of justice. I had the opportunity to present much of my work on this topic in several conferences and seminars
y
such as the annual meeting of the Latin American and Caribbean Economic Association (LACEA) and the
Bolivian Conference in Development Economics in 2010.
op
Currently, I am also working on three other projects. One of these studies is on the impact of legalized
gambling on household behavior in USA for the period 1956-2007. I am especially interested in understanding
the mechanisms in which casino openings might have affected divorce, domestic violence and child abuse rates.
tC
The second topic in which I would like to continue working during my studies is on the impact of peer effects
on education and crime. For this project, I had interviewed people from the Department of Education in NYC
to get data on test scores in order to analyze the effectiveness of gifted and talented education. Since children
can enroll in this program in every grade, I would expect the program effect to be smaller for those that enter
the talented school in older years. This could be explained by the fact that since most of their peers and friends
an
are not in the talented schools, those peers might negatively influence the academic performance of students in
the talented school. Finally, I am also working on many topics related to economics of crime, including an
analysis of open regimes’ effectiveness on juvenile crime recidivism and of family visits on prisoners’
misconduct. For this last one, I will exploit the swine flu pandemic of mid-2009 as a source of exogenous
variation for prison visits, thus allowing identification of causal link between family support and inmates’
ic
behavior in prison.
In addition to my research background, I also have experience as Teaching Assistant in the most
pl
demanding courses at Universidad de San Andrés: Introduction to Economics, Economic History, Applied
Economics and Economics of Regulation. In these courses, I had an active participation in the elaboration of
the syllabus, lectures, assignments and examinations. According to the feedback of my professors and students,
Ap
my work has been exceptional. Indeed, I am currently helping two of my students in the elaboration of their
undergraduate thesis. Apart from this teaching experience at San Andres, during my year in Washington DC I
became part of Latino Student Fund, an NGO whose aim is to maximize educational opportunities for
underprivileged PreK-12 students of Hispanic descent. In this experience, I not only improved my teaching
skills but I also learned that positive change on a large scale is often limited without effective institutions and
collaborative partnerships.
Considering my set of interests and the quality of the program and faculty, Columbia University is
clearly my top choice for pursuing a PhD. in Economics. I have visited the campus twice this year, and after
talking to students and professors, including Marcelo Moreira from the Economics Department, I concluded
that Columbia University is the perfect place for me to pursue my interests. I am absolutely aware of the
enormous effort and skills required to successfully pursue a doctoral degree, and I consider myself prepared to
meet the challenge. I profoundly hope to have the opportunity to develop my potential in Columbia’s
outstanding academic environment.
Sincerely,
Maria Micaela Sviatschi

Resume Maria Sviatschi
4883057
MARÍA MICAELA SVIATSCHI
Cerrito 1548 16 “86”, Buenos Aires, Argentina
Mobile: (+54911)-41983980
msviatschi@gmail.com
_____________________________________________________________________________________
PERSONAL INFORMATION
Date of Birth: May 29, 1986
Citizenship: Argentine
y
op
EDUCATION
2008-2009: Master in Economics, Universidad de San Andrés.
2004-2007: Licentiate in Economics, Universidad de San Andrés.
Research interests: Applied Microeconomics, Health, Crime, Justice, and Development Economics.
RESEARCH PAPERS
2010: tC
“The Impact of Costa Rica’s Modernization for the Administration of Justice Program on Court’s Efficiency” jointly with
an
Yuri Soares. Forthcoming OVE working paper at IDEAS, Inter-American Development Bank, Office of Evaluation and
Oversight.
2009: “The Impact of Improving Access to Justice on Conflict Resolution” jointly with Yuri Soares, Jimena Montenegro and Raul
Andrade. Under revised and resubmitted for Economic Development and Cultural Change, Chicago Journals.
ic
2009: “Dry Law for Drunk Drivers. The Impact of Alcohol-Related Laws on Car Accident Mortality Rates.”
http://ssrn.com/abstract=130589. Under revision in Fiscal Studies.
pl
WORKING PAPERS
Ap
2010: “Does Court Efficiency have a Deterrent Effect on Crime? Evidence from Costa Rica” jointly with Yuri Soares.
Forthcoming OVE working paper at IDEAS. Inter-American Development Bank, Office of Evaluation and Oversight.
“The Impact of Legalizing Gambling on Divorce Rates in the US” jointly with Nicolas Bottan.
“Using a Regression Discontinuity Design to Measure the Benefits of Gifted and Talented Education. Evidence from
New York City” jointly with Diego Aparicio
TEACHING EXPERIENCE
2010: Applied Microeconomics with Martin Rossi. Universidad de San Andrés.
Economics of Regulation with Martin Rossi. Universidad de San Andrés.
Applied Microeconomics (applications in stata) with Martin Rossi. Ministerio de Coordinación de Desarrollo Social.
Ecuador.
2009: Latino Student Fund. Mentoring and Teaching PreK-12 students of Hispanic descent. Washington DC.
2008: Economics with Juan Carlos de Pablo. Universidad de San Andrés.
Economic History with María Inés Barbero and Patricia Saporiti. Universidad de San Andrés.
RESEARCH EXPERIENCE
2010: Research Assistant of Martin Rossi. Universidad de San Andrés. Design of Program Evaluation Methodologies for
the National Commission of Scientific and Technological Research (CONICYT). Chile.
4883057
Research Fellow at Inter-American Development Bank’s Office of Evaluation and Oversight (OVE). Buenos Aires,
Argentina. Projects: The Impact of Justice Reform in Costa Rica. Design of Program Evaluation Methodology for Justice
Modules Project Peru (Phase II). How Impact Evaluations Can Inform Policy?
Research Assistant of Rafael Di Tella. Harvard Business School. Project: Laboratory experiment to test the emergence of
self-serving biases in fairness evaluations.
2009: Research Assistant of Martin Rossi. Universidad de San Andres. Project: Impact Evaluation of the Infrastructure Services
Project in Rural Malawi.
y
Research Fellow at Inter-American Development Bank’s Office of Evaluation and Oversight (OVE). Washington DC.
op
Projects: Country Program Evaluation of Paraguay 2004-2009. The Impact of Justice Modules in Rural Areas of Peru
(Phase I).
2008: Research Assistant of Rafael Di Tella, Harvard Business School. Project: Government Advertising and Media Coverage of
tC
Corruption Scandals.
CONGRESSES AND PRESENTATIONS

2010: “The Impact of Improving Access to Justice on Conflict Resolution” presented at the 2010 Bolivian Conference on
an
Development Economics in La Paz, Bolivia.
2010: “Dry Law for Drunk Drivers. The Impact of Alcohol-Related Laws on Car Accident Mortality Rates” presented at the 2010
Bolivian Conference on Development Economics in La Paz, Bolivia.
ic
2010: “The Impact of Improving Access to Justice on Conflict Resolution” presented at XV Annual Meeting of the Latin American
and Caribbean Economic Association (LACEA) in Medellín, Colombia.
2010: “Dry Law for Drunk Drivers. The Impact of Alcohol-Related Laws on Car Accident Mortality Rates” presented at XV
pl
Annual Meeting of the Latin American and Caribbean Economic Association (LACEA) in Medellín, Colombia.
October 2010: presented “Dry Law for Drunk Drivers. The Impact of Alcohol-Related Laws on Car Accident Mortality Rates.” at
Ap
John Hopkins University. Bloomberg School of Public Health. Baltimore MD.

June 2010: presented “Dry Law for Drunk Drivers. The Impact of Alcohol-Related Laws on Car Accident Mortality Rates.” at the
Association for Social Economics. Montreal, Canada.
June 2010: presented “Does Court Efficiency have a Deterrent Effect on Crime? Evidence from Costa Rica.” Inter-American
Development Bank’s Office of Evaluation and Oversight. Washington, DC.
April 2010: presented “Dry Law for Drunk Drivers. The Impact of Alcohol-Related Laws on Car Accident Mortality Rates.” at the
Population Association of America. Dallas, Texas.
November 2009: “The Impact of Improving Access to Justice on Conflict Resolution” selected to be presented at the XLIV
Annual Meeting of the Argentinean Association of Political Economy. Mendoza, Argentina.
September 2009: presented “The Impact of Improving Access to Justice on Conflict Resolution” at Inter-American Development
Bank’s Office of Evaluation and Oversight. Washington, DC.
November 2008: presented “Dry Law for Drunk Drivers. The Impact of Alcohol-Related Laws on Car Accident Mortality Rates.”
at the XLIII Annual Meeting of the Argentinean Association of Political Economy. Córdoba, Argentina.
AWARDS AND SCHOLARSHIPS

2010: Institute for Advanced Development Studies. Special Distinction for “The Impact of Improving Access to Justice on
Conflict Resolution”
4883057
2010: Grant from conference committee to present “The Impact of Improving Access to Justice on Conflict Resolution” and “Dry
Law for Drunk Drivers. The Impact of Alcohol-Related Laws on Car Accident Mortality Rates” at the 2010 Bolivian Conference
on Development Economics in La Paz, Bolivia.
2010: Grant from conference committee for best student papers to present “Dry Law for Drunk Drivers. The Impact of Alcohol-
Related Laws on Car Accident Mortality Rates” at the XV Annual Meeting of the Latin American and Caribbean Economic
Association (LACEA).
2010: Grant from the Inter-American Development Bank to present “Dry Law for Drunk Drivers. The Impact of Alcohol-Related
y
Laws on Car Accident Mortality Rates” at the XV Annual Meeting of the Latin American and Caribbean Economic Association
op
(LACEA).
2008: Best Undergraduate Thesis “Dry Law for Drunk Drivers. The Impact of Alcohol-Related Laws on Car Accident Mortality
Rates” Universidad de San Andrés.
tC
2008: Argentinean Association of Political Economy: Young Researcher Award for “Dry Law for Drunk Drivers. The Impact of
Alcohol-Related Laws on Car Accident Mortality Rates.”
2008: 75 percent scholarship. M. S in Economics, Universidad de San Andrés.
2004: 75 percent scholarship. B. S in Economics, Universidad de San Andrés.
an
OTHER TRAINING
2010: “Applied Microeconomics”. David Card (University of California, Berkeley). Inter-American Development Bank.
ic
2010: “Lectures on Evaluation Methods”. Guido Imbens (Harvard University). Impact Evaluation Network of the Latin
American and Caribbean Economic Association (LACEA).
pl
ADDITIONAL SKILLS
Statistical software: Stata, Eviews.
Ap
Languages: Spanish (native), English (advanced), Latin (written) and Portuguese (basic).
Writing Sample Maria Sviatschi
4883057
Dry Law for Drunk Drivers:

The Impact of Alcohol-Related Laws on Car Accident Mortality Rates
y
op
María Micaela Sviatschi*
Universidad de San Andrés & Inter-American Development Bank’s Office of Evaluation and Oversight
October 2010
tC
an
Abstract
This paper studies the effect of a law that prohibits the sale of alcohol at night for convenience
ic
stores and gas stations (usually called dry law) on car accident mortality rates. The analysis exploits the
pl
fact that each province decided to pass the law at different points in time. This observed variation in the
implementation of the law across time and space provides a potential instrument to identify the causal
Ap
effect of alcohol restriction in car accident mortality rates using the panel nature of the data. After
controlling for province and year fixed effects, I found that the dry law is associated with reduction of 14
percent in car accident fatalities. This result is especially important in the case of Argentina where 8,000
people die annually in car accidents and where 37 percent of these deaths are caused by alcohol abuse.
* María Micaela Sviatschi, Teacher Assistant at Universidad de San Andres, Vito Dumas 284, Victoria, Provincia de Buenos
Aires, Argentina, Research Fellow at Inter-American Development Bank‘s Office of Evaluation and Oversight
msviatschi@iadb.org. I would like to thank Martín Rossi, for his guidance and valuable comments. I am also grateful to
Ignacio Caronia, Mateo Caronia, Andrés Maggi, Ricardo Nicolás Peréz Truglia and Alexander Triantaphillys for their helpful
discussions and cooperation.
1
4883057
I. Introduction
In the last years there has been an important concern in terms of public policy about reducing the
risks associated with car accidents. According to the World Health Organization approximately 1,200,000
people die annually in the world in car accidents, 400,000 of whom are young people under the age of 25.
Road traffic crashes rank as the 11th leading cause of death and account for 2.1 percent of all deaths
y
globally. Motor-vehicle accidents are responsible for more annual deaths than AIDS, tuberculosis, and
op
malaria. The most important feature is that 90 percent of these deaths have occurred in developing
countries (see Table 1).1
One of the main causes of car accidents fatalities is alcohol abuse when driving. A report of the
tC
World Medical Association shows that in many countries where alcohol consumption is part of daily life,
driving under the effects of alcohol is the cause of almost half of deaths and serious injuries in car
accidents. In many high-income countries about 20 percent of fatally injured drivers have an excess of
an
alcohol in their blood. In low-income countries the number is even worse, as alcohol is present in
between 33 percent and 69 percent of fatally injured drivers. In the case of Argentina, where nearly 8,000
people die per year, 37 percent of motor-vehicle accidents are attributable to excessive alcohol
ic
consumption and these alcohol-related accidents are the main cause of young people deaths, leaving
pl
behind tumors and heart diseases.2

In addition to human losses there is an enormous economic cost, estimated as 1 percent of the
Ap
National Gross Product (NPG) in low-income countries, 1.3 percent in those with middle income, and 2
percent in countries of high income.3
Many policies have been implemented in order to restrict alcohol consumption and reduce
alcohol-related accidents such as higher beer taxes, driving laws, a minimum legal drinking age, and
limited stores hours. There is strong evidence that alcohol consumption is sensitive to some of these
policies. Using U.S. state level data, Cook and Tauchen (1982) find that excise taxes significantly reduce
alcohol consumption and heavy drinking. Saffer and Chaloupka (1989), using a cross sectional time series
data set for forty-eight U.S. states, verify that both alcohol taxes and minimum drinking-age laws reduce
state level motor vehicle fatalities which are strongly correlated with drinking. In contrast, there is less
agreement regarding the impact of other alcohol control regulations, such as limitations on selling hours.
1 World Bank & World Health Organization (2004). ―World Report on road traffic injury prevention.‖
2 Data provided by Luchemos por la Vida (Let's fight for life). This is a non-profit organization whose only purpose is to help prevent traffic
accidents in Argentina.
3 Report of the World Bank (January of 2000).
2
4883057
In this study I test the effectiveness of a law that deals with the prohibition of alcohol selling
during night-time hours at convenience stores and gas stations in Argentina. This law is often called dry
law when it is applied in countries where alcohol is accepted as an allowed drug. 4 The purpose of this
law is to reduce street alcohol consumption and therefore alcohol-related car accidents and delinquency
under the effects of alcohol. In this paper, I study whether the dry law has a causal effect on motor
y
vehicle fatalities. It is especially important to study the potential effect in the case of Argentina which
op
counts among the countries with the highest car mortality rates (see Table 2).
The expected effect of this law on car accident fatalities is not evident. One might expect a
decrease in the number of car accidents fatalities by decreasing the availability of alcohol during the hours
tC
in which alcohol fatalities are substantially higher.5 However, if people simply find an alternative location
to purchase alcohol the impact of the law might be ameliorated. Furthermore, the existence of black
markets could eliminate the positive effect of the law.
an
This paper attempts to find the causal effect of the law, using detailed data on alcohol laws in
every province of Argentina between 1990 and 2005. Two important contributions are made.
First, to the best of my knowledge this paper is the initial study done for developing countries.
ic
There are many empirical studies for the United States that provide some evidence of the effect of dry
pl
laws in alcohol-related accidents but there are no studies for low and middle income countries, which
account for 90 percent of car accident mortality rates. Nevertheless, these U.S. studies give an idea of the
Ap
expected effect, suggesting that alcohol control regulations reduce fatalities. Brown, Jewell and Richer
(1996), using a two-stage estimation and county-level data on the state of Texas find that county-level
alcohol prohibition decreases alcohol-related motor vehicle accidents and fatalities. Winn and Giacopassi
(1993) report that Kentucky counties that prohibit alcohol sales have significantly lowered alcohol-related
motor vehicle accident rates. In the same line, Chaloupka et al. (1993) find a negative impact on state-
level car mortality rates, and McCarthy (1999) concludes that bans on common site sale of alcohol and
gasoline increases the number of alcohol-related fatalities outside metropolitan areas. The intuition
behind this result is that citizens may react to the ban by traveling to other counties to purchase and
consume alcohol, leading to the possibility of increased time on the road for drunk drivers and thus more
accidents.6
4 Similar laws which restrict alcohol consumption are applied in Chile, Ecuador, the Dominican Republic, Spain, United States, Colombia
and Costa Rica.
5 Dee (1999) suggests that alcohol involvement in fatal accidents is higher at night-time than at day-time. In the same line Grabowski et al.
(2001) report that teen traffic fatalities are concentrated during night-time, between these hours teen driver death rates are nearly three times
greater per trip than during the hours from 6 a.m. to 10 p.m.
6 This may not be the case of Argentina where the restriction is applied only to convenience stores and gas stations.
3
4883057
Second, while the existing literature focused on cross-state comparisons of alcohol policies and
car accident fatalities without controlling for unobserved variables that are correlated with cross-state
variations in alcohol policies and thus producing biased estimates, in this study I provide evidence
controlling for province time-invariant unobserved heterogeneity and common shock for all provinces.
To do so I take advantage of the fact that in Argentina there is no national dry law and that each
y
province decides to pass or not to pass the law. In addition, each province decided to set the law at
op
different moments of time. Since there is variation across time and space, using panel data, I will try to
identify the causal effect of alcohol restriction on motor. It is important to note that none of the
previous studies use matching methods and very few of them incorporate time-fixed effects and state-
tC
fixed effects. Furthermore, many previous studies have ignored that differences in car mortality rates are
influenced by economic conditions, suffering from omitted variables bias. In attempt to address this
issue I include controls for unemployment, public expenditure, and gross domestic product (GDP). 7 In
an
the case of Argentina it is not necessary to control for other programs or laws which may be
simultaneously operating to reduce drunk driving because most of them are all national, such as the
National Traffic Law 24,449 and the Law 24,788, and therefore any effect arising from these laws would
ic
be captured by the time effect.8 A major methodological concern is that the choice to pass the law may
pl
not be orthogonal to unobservable factors that also affect car mortality. Using panel data and matching
techniques I will try to consider the most comparable provinces in order to estimate a causal effect.
Ap
The main finding of this paper is that alcohol restriction on sales in convenience stores and gas
stations is actually associated with a reduction of 14 percent in the number of car accident fatalities. This
finding does not change as a result of different specifications, suggesting the existence of a causal effect.
The validity of the causal interpretation of the estimates is tested using other causes of mortality. It is
found that while the imposition of the law is correlated with car mortality rates, it is uncorrelated with
deaths from causes unrelated to alcohol consumption.
The rest of the paper proceeds as follows: Section II describes the data and Section III presents
the specification strategy. Finally, Section IV concludes.
7 Ruhm (1995) have pointed that only a few studies (Wagenaar and Maybee, 1986; Saffer and Chaloupka, 1989; Chaloupka et al., 1993;
Evans, 1993) have accounted for differences in unemployment rates.
8 The Law 24,788 was designed in 1997 to reduce alcohol consumption. One of it main regulation was to prohibit alcohol selling to eighteen
year-old people.
4
4883057
II. Data
The data is a panel of observations of 23 provinces of Argentina over the period 1990-2005.
Three sets of data are used: province-level data on unintentional deaths in road accidents; information on
the dates in which the law was passed; and province-level information on gross domestic product,
unemployment rates, expenditure rates, citizen involvement, the number of registered vehicles, and
y
weather conditions.
op
As there is no measure of alcohol-related car fatalities, the study uses car fatalities. This first
dataset comes from two sources: Argentina‘s national office on crime statistics (Registro Nacional de
Reincidencia); and the National Crime Information System of the Ministry for Justice (Sistema Nacional de
tC
Información Criminal del Ministerio de la Justicia). Information on car accident mortality from 1990 to 1997
was provided by the Registro Nacional de Reincidencia while the information from 1999 to 2005 was
provided by the National Crime Information System.9 These two sources compile the information
an
registered by security forces, provincial police, federal police, gendarmerie, and naval prefecture based on
the monthly number of criminal reports. The number of fatalities in road accidents registered by the
police and law enforcement agencies does not include unreported casualties or those reported directly to
ic
some judicial court (Judge, Office of the Public Prosecutor, Chamber).10

pl
It is important to note that the implementation of the dry law does not affect the procedure by
which the observed outcome (the number of car accident fatalities) is measured and therefore there is no
Ap
potential bias in the estimated effect coming from this source.11 Car accident mortality rates per 100,000
inhabitants are presented in Table 3. A total of 336 observations are available for estimation and
variability exists across provinces and time in car mortality rates.
Table 4 presents average car mortality rates before and after the implementation law. In most
provinces where the dry law was implemented, car mortality rates decreased.
The information about the dates in which the law was passed was obtained from the legislatures
of each province. In every province the law is very similar. In general it prohibits the sale of alcohol in
the entire territory between 11:00 pm and 8:00 am, with the exception of bars, dance clubs, and
restaurants.12 The treatment variable is a dummy indicating if the province has passed the law or not.
Table 5 describes the changes in the status of the alcohol law between 1995 and 2005. Over the analyzed
9 There is no available data for the provinces in the year 1998.

10 Although the data was provided by two different offices, they came from the same Ministry and thus, there was no methodological change
11 This type of bias might appear, for example, if a policy consisting of changes in blood alcohol content requirements is accompanied by
increased efforts by police officers to find drunk drivers.

12 In the case of Buenos Aires, Santa Fe, and Chubut, there is a more restricted law which prohibits alcohol sale in convenience stores and gas
stations the whole day. No separation into two different treatments was done due to lack of statistical power.
5
4883057
period, most of the provinces have passed the law in different moments. This variation in time and in
space is used to identify the causal effect in car accident deaths.
The third dataset contains province-level information on control variables. The unemployment
rate, the gross domestic product, and the alcohol expenditure rate were obtained from the National
Institute of Statistics and Census (Instituto Nacional de Estadística y Censos) and the Bureau of Federal
y
Investments (Consejo Federal de Inversiones). For the population and the gross domestic product variables,
op
projections are made for every year based on information in 1990, 1991, 1993 and 2001. Data
corresponding to public expenditure and percentage of paved roads was collected from the Ministry of
the Economy.13 I use the precipitation rate as a proxy for weather conditions. This variable was
tC
obtained from the National Meteorological Service (Servicio Nacional Meteorológico).
registered vehicles was provided by the National Registry of Motor-Vehicle Property (Dirección Nacional
del Registro Automotor). Citizen involvement was calculated as the ratio of voters that actually vote in
The number of
an
presidential elections and was obtained from the Ministerio del Interior de la República Argentina.14 Finally, the
number of doctors and life expectancy was obtained from the Ministry for Health.
Table 6 compares some characteristics related to car mortality across all provinces. In all cases I
ic
cannot reject the hypothesis of equality between the control and the treated group in the pre-intervention
pl
period.15 In particular, the treated and the control group showed similar levels of car mortality rates
before the implementation of the law.
Ap
These results can be interpreted as evidence that the pre-intervention characteristics of the
treatment and control group are similar and suggests that the control group maybe a good counterfactual
of what would have happened to the treated group in the absence of the intervention. It is shown not to
be the case that the provinces where this law was implemented were those with higher rates of car
accidents and thus where citizens pressured for this type of law.
III. The Effect of the Dry Law on Car Accident Mortality

A. Empirical Strategy
The purpose of this paper is to identify the average effect of the dry law on car mortality rates,
comparing car accident fatalities of provinces that passed the law to provinces that did not pass the law.
13 There is no data available of public expenditure for the year 1991.

14 For the variable citizen involvement, I have data only for the years 1989, 1995, 1999 and 2003. For the missing years, I suppose the same
citizen involvement (for example if the citizen involvement was about an 87 percent in 1990 for Buenos Aires, I take the same percentage
for the following years 1991,1992,1993,and 1994).
15 I can only reject the hypothesis null of equality between the control and the treated group for the number of registered cars. Nevertheless
this might not be a concern because I will include this variable in my regression model.
6
4883057
As there is no experimental design, it cannot be asserted that the decision to implement the legislation
was random in each province. Different particular characteristics in the provinces may have affected the
decision of whether or not to implement the law. It could be the case that some provinces—for example
those whose governors are more concerned about alcohol consumption—have implemented this type of
laws earlier than other provinces. In this sense, correlation between the dry law and certain factors that
y
may influence fatalities may lead to a biased estimate. Many variables that potentially confound the
op
identification are those that vary across provinces but are fixed over time, such as the number of
highways, customs, and attitudes of each society towards alcohol consumption. 16 In order to control
these time-invariant unobserved variables that may affect the outcome and the imposition of the law, I
tC
use panel data and estimate a difference-in-differences model. This model compares the evolution in
outcomes in the treatment group before and after the intervention to the evolution in outcomes in the
control group. The evolution in the control group is an estimate of what would have happened to the
an
treatment group with no intervention.
The difference-in-differences estimator includes year fixed effects that control for any common
shocks for all provinces and province fixed effects that control for time-invariant influences. Formally,
ic
the difference-in-differences model can be represented by the following equation:

pl
CarMortalityit= DryLawit + Xit + i + t + it (1)

Ap
where CarMortalityit is the car mortality rate in province i in year t (car fatalities of province i in year t per
100,000 inhabitants); DryLawit is a dummy variable that takes the value of one if the province i has a dry
law in year t and zero otherwise; Xit is a vector of control variables that varies across provinces and time;
t is a time effect common to all provinces in period t; and i is a province fixed effect.
The inclusion of province and time fixed effects is crucial for the identification of the parameters
of interest as it guarantees that there will not be a bias coming from differences in car mortality rates
across provinces that are time-invariant and differences in car mortality rates across time that are
common to all provinces.
The Xit vector includes variables like regional gross domestic product, unemployment rate, public
spending, citizen involvement, number of registered vehicles and precipitation.
GDP and the unemployment rate are used as proxies for macroeconomic conditions. GDP may
be positively correlated with the implementation of the law in two ways. First, provinces with higher
16 While many highways were constructed in Buenos Aires in the 1990‘s, there is no variability across other provinces.
7
4883057
gross domestic product may tend to have better institutions and hence more laws. Second, provinces
with higher GDP can have more alcohol related accidents since alcohol consumption increases in better
economic times, therefore leading to more car accident fatalities and thus a higher likelihood of passing
the law. The possible level and direction of correlation between unemployment rates and car accident
mortality rates is unclear. On the one hand, the two might be negatively correlated since total alcohol
y
consumption and the proportion of drinking that occurs in bars and restaurants is likely to fall during
op
recessions.17 On the other hand, the unemployment rate may be positively correlated with car accident
mortality rates since unemployed workers could drink more due to depression.
The public spending variable gives an idea of the expected probability of detection of drunk
tC
drivers and it controls for road conditions through expenditure on urban improvement. It is important
to note that I did not include whether or not the roads are paved because these variables are time
invariant and therefore it is captured by the province fixed effect.18
an
The number of registered vehicles captures the traffic density; the expected sign of this variable is
not evident. On the one side, traffic density is positively correlated with car accident frequency while on
the other side it may decrease fatal accidents as driver speed is lower when traffic density is higher.
ic
The vector of control variables also includes a proxy for citizen involvement represented by voter
pl
participation. The inclusion of this variable is important for two reasons. First, as the vote in Argentina
is mandatory, it represents citizens‘ willingness to obey laws. In provinces where citizen involvement is
Ap
higher, laws are more likely to be enforced and people might be more respectful of laws. Furthermore, in
these provinces citizens may conduct themselves more responsibly and partake less frequently in drunk
driving. Second, citizen involvement may be correlated with the number of laws passed in each province.
Finally, as the correlation between the alcohol law and car accident deaths may be confounded
with a weather effect, I include the amount of precipitation. Provinces with rainy weather and thus a
higher frequency of car accidents might be more likely to pass the law.
The error  it is a province time-varying error which is generally assumed to be independent
across time and space; however, as the analysis uses panel data, the errors could be correlated across time
in the same province. In the case of a positive correlation, the standard errors could be computed
smaller and the null hypothesis could be over rejected. To avoid potential biases in their estimation,
standard errors are clustered at the province level, allowing an arbitrary covariance structure within
17 Ruhm (1995) has shown that individuals may shift drinking away from bars where alcohol is relatively expensive during downturns.
18 See Statistical yearbooks of the National Institute of Statistics and Census (INDEC 1990-2005).
8
4883057
provinces over time.19 It is important to note that if the province errors are highly correlated, clustering
standard errors may reduce the statistical power of the estimation.20
The most important identifying assumption of this model is that the change in mortality in the
control group is an unbiased estimator of what would have happened to the treated group if there had
been no intervention. In order to test this assumption following Galiani (2005), I will test if the trends in
y
the pre-treatment period are similar for the treated and the control groups. If in the pre-intervention
op
period the trends are not different then it is possible to assume that in the absence of the treatment the
trends would remain the same in the post-intervention period. Formally, this assumption is tested
estimating a modified version of equation (1).21 In this model, the null hypothesis that the pre-
tC
intervention year dummies are the same for the eventually treated provinces and the control provinces
cannot be rejected, giving validity to the difference-in-differences approach.22
an
B. Results
Column 1 in table 7 presents the results for a model including only the dry law dummy, the fixed
effects for each province, and the year dummies. Provinces that have passed the alcohol-related law
ic
experience a statistically significant reduction in car mortality rate of about 1.57, which amounts to an
pl
11.4 percent reduction of the baseline rate. That is, we passed from an average of 14 annually deaths per
100,000 inhabitants to 12 deaths with the imposition of the law.
Ap
Since the differences-in-differences model does not take into account the characteristics that vary
across time and across provinces and that might be correlated with car mortality and the implementation
of the law, column 2 includes a set of socioeconomic characteristics such as the unemployment rate,
public spending, and GDP. The variable public spending includes road infrastructure investment and
police expenditure. It could happen that provinces which passed the law were also spending more
budgets on road improvements, so the effect of the dry law could be coming from public expenditure.
Nevertheless, the estimated impact of the law is unchanged. It is important to note that the population is
not included as a control variable since all the variables are calculated on a per capita basis. The only
19 See Bertrand et al. (2004).

20 This could be a problem in the case of not finding any effect of the law. Unfortunately, in the case of not finding an effect, nothing can
be done to solve this problem because the number of provinces cannot be increased, which is one possible solution to the power problem.
One possible solution to the power problem is to increase the sample size, including more groups. In this study, each province conforms to
a group (so there are 23 groups). Therefore, the sample size cannot be increased because the analysis uses all the provinces and there are no
more groups (provinces) to add.
21
As in Galiani et al. (2005) the estimation only considers the observations of the pre-intervention period. The control group has 16
observations and the treated group only consists of pre-treatment observations. Therefore, a dummy is generated, which takes the value of
1 if the province will be eventually treated and 0 if the province will never be treated. Finally, an interaction is made between this dummy
and the year dummy; and the model (1) is estimated without the dry law dummy.
22 See results in the appendix (table A1).
9
4883057
significant control variable is unemployment and it coefficient suggests that during recessions alcohol
consumption decrease and thus car accidents. 23
In the fourth column, I add controls for citizen involvement, number of registered vehicles and
precipitation. Little change in the policy‘s effect is observed (16 percent with five percent of statistical
significance).24
y
Overall, these results suggest that the introduction of a law that restricts the sale of alcohol for
op
convenience stores and gas stations generates a significant decline in car mortality rates. 25 The rest of the
paper will focus on the specification of the model.
C. Matching Techniques
tC
The difference-in-differences approach does not take into account the possibility that some
provinces may not serve as good controls. In other words, this approach may be comparing provinces
an
which are not comparable. In this section, I use matching methods to attempt to solve this problem.
Matching methods eliminate the aforementioned bias by comparing treated provinces with
control provinces that have similar pre-characteristics. If selection is based on observable characteristics,
ic
this method estimates treatment effects consistently in a non-experimental context. The assumption
pl
behind matching methods is that conditional on some observable variables, the counterfactual outcome
distribution of the treated provinces is the same as the observed outcome distribution of the units in the
Ap
control group.
First, I estimate a propensity score for each province using a logit model, which estimates the
probability of being treated conditional on a set of pre-intervention characteristics.26 In this case, I
condition the probability of passing the law on GDP, the public expenditure rate, the unemployment
rate, the citizen involvement rate, the car mortality rate, the amount of precipitation, the number of
registered vehicles, the proportion of doctors, life expectancy, the alcohol expenditure rate, and the
percentage of paved roads.
Second, in order to make the working sample more comparable, I restrict the sample to districts
with probabilities that lie within the common support, that is, the area were the propensity score for the
23 Chaloupka et al. (1993) has also found a similar result.

24 For robustness purpose, all the models were also estimated without the control citizen involvement and results did not change.
25 The results are unaffected when standard errors are computed by clustering at the provincial level.
26
Rosembaum and Rubin (1983) show that if it is valid to conduct matching on X´s, it is equally valid to do so for the propensity score.
The advantage of using the propensity score is that it has one dimension. The propensity score is P (T=1/X) which, of course, has one
dimension.
10
4883057
treated group is similar to that of the control group.27 All control provinces whose propensity scores are
less than the minimum propensity score of the treated provinces and all treated provinces whose
propensity scores are higher than the maximum propensity score of the control provinces are excluded.
Therefore, with these observations, I estimate a difference-in-differences model on the common
support.28
y
Finally, I use the kernel density weighting procedure29 to obtain a fixed effect matching estimator
op
similar to the difference-in-differences matching estimator.30 This procedure estimates the average
treatment effect on the treated after using propensity score matching to minimize the differences
between the treated and control group. This estimator combines the advantages of using matching
tC
methods with the benefits of estimation using the difference-in-differences model.
The results of these models are reported in Table 7. The results do not change when the
observations are restricted to the common support. Column 9 presents the results of the kernel
an
matching model in the common support. A reduction of about 14 percent is found and is significant at
the 5 percent level.31
ic
D. No Mean Reversion
pl
One possible concern is that those provinces that decided to pass the law may have experienced a
negative shock (higher fatalities) in car accident mortality rates before the treatment, relative to those
Ap
provinces not treated. That is, maybe the treated provinces decided to pass the law because the year
before they experienced a bad outcome as a result of an unlucky event or circumstance. Since they are
unlikely, on average, to experience a bad outcome again the following years, their car accident mortality
rates will tend to fall, ―they will revert towards the mean‖ even in the absence of the law. If this is the
case the provinces treated would have less number of car accident fatalities even in the absence of the
law. In order to address this concern, I estimate the same model with the treatment status minus two
years. If there is a surge before the treatment, the coefficient associated should be positive and
32
significant. Fortunately, the results of this regression suggest no evidence of such surge. In the
27 All the specifications reported satisfy the balancing condition.

28
The region of common support is [.14341644, 1] and the balancing property is satisfied.
29 The kernel matching procedure gives a positive weighting to all provinces in the control group but the weighting is inversely proportional
to the distance in propensity scores.

30 See Heckman et al. (1997).
31 For robustness purpose, all the models were also estimated in logs with similar results. All regressions mention and not show are
available.
32 Regressions are available upon request. I also exploit the timing of the imposition of the law to construct a placebo treatment. I estimate
whether the introduction of the law affects car accident mortality rates between 1993 and 1995. The imposition of the law should not affect
the rates between years in which the law did not exist. I find no evidence of a significant correlation in this time period. The results suggest
11
4883057
appendix, I present the trends for each province that passed the law, and it can be seen for most of the
provinces that there is no increase in the car mortality rates just before the implementation of the law. In
the case of Buenos Aires there is a surge two years before the treatment, so I repeat the previous analysis
while excluding this province from the sample and I found that the results do not change.
y
E. Impact of the Dry Law in Other Causes of Mortality
op
Despite of the robustness of the estimates to different estimation methods, one could still argue
that when the dry law was passed other unobserved changes correlated with mortality in general may
have an effect in provinces where the law was implemented. For example, it could be the case that
tC
provinces that pass the law are more active in terms of public policy. Thus, the governments who pass
the law may be more concerned about mortality in general and not only on car accident mortality. In this
case, there may be unobserved variables which are correlated with the passing of the law. To address this
an
issue, a difference-in-differences model is estimated for other types of mortality, such as infections,
tumors, and cardiovascular diseases.33 If the estimation presented in the previous section is correct, the
dry law should not operate affecting mortality rates which are not correlated with alcohol abuse.
ic
As shown in Table 8, the null hypothesis cannot be rejected. It is not statistically significant,
pl
which suggest that the dry law has no influence on other sources of mortality and that it has only affected
alcohol-related casualties such as car accident mortality rates. This result provides evidence that the law
Ap
has only affected car accident mortality rates through a reduction of alcohol consumption when driving.
IV. Conclusions
This paper shows that alcohol-related laws in Argentina have reduced car accident casualties. It
provides evidence that decreasing the availability of alcohol during the hours in which fatalities are
substantially higher is a good policy that reduces car accident casualties. Using a combination of
methods, I found that provinces which passed the law experienced, on average, a reduction of about 14
percent in car accident mortality rates. This study suggests one way to reduce alcohol consumption and
hence car accidents fatalities in the context of developing countries, where the number of car accident
deaths have been substantially increasing in the last years. While the previous literature has focused on
that the imposition of the law did not affect the number of fatalities between 1993 and 1995. Nevertheless as the sample is reduced, these
results could be due to the lost of power.
33The impact of the dry law is only estimated for these particular alternative causes of mortality because there was no data available on other
causes.
12
4883057
the effect of dry laws in developed countries, this paper provides evidence for a country which belongs to
the group of developing countries, which together account for 90 percent of car accident mortality rates.
Many factors suggest that the relation between the dry law and car accident fatalities might be
causal. First, the treatment and the control group exhibited similar time trends in the pre-intervention
period, validating the difference-in-differences identification strategy. Second, the model includes
y
provinces and time fixed effect and the conclusions are robust to the inclusion of variables that may
op
affect drunk driving behavior and may also influence car accidents casualties. It also remains robust after
controlling for heterogeneous provinces, redefining the control and treatment groups according to each
province‘s propensity score. Furthermore, the results remain unchanged when the control group is re-
tC
weighted with the kernel density procedure. Third, it demonstrates that the dry law only affected car
mortality rates and not other types of deaths where alcohol is not involved.
The results shed light on a number of important policy debates. This result is very important in
an
Argentina where the dry law has been criticized frequently, mainly by consumers and retailers who have
argued that the dry law restricts personal liberties and generates losses as it limits the sale of alcohol to
fewer hours of the day. Moreover, a commonly held opinion was that this type of law could not be
ic
enforceable and that it was only a strategy of politicians to show the society a commitment to alcohol
pl
concerns. This work, however, shows that the law was enforceable and with significant results in car
accidents fatalities. As Hobbes said ―man is a wolf to man,‖ the society needs this kind of laws to protect
Ap
people from others and from themselves. In this sense, more attention ought to be paid to policies
regarding alcohol consumption because they may have highly beneficial effects.
13
4883057
Tables
Table 1. Motor-Vehicle Fatalities in the World

Number of motor-
vehicle fatalities Rate per 100,000 habitants Proportion of the total
Countries with low and middle
income 1,065,988 20.2 90
Countries with high income 117,504 12.6 10
Total 1,183,492 19.0 100
Source: World Health Organization (2002).
y
op
tC
an
ic
pl
Ap
14
4883057
Table 2. Car Mortality Rates (per 1,000,000 vehicles)

China 2,033 Denmark 267
Rumania 1,414 Luxembourg 258
Argentina 1,310 Austria 257
Poland 890 Belgium 249
Turkey 752 Switzerland 212
Greece 650 United States 209
Slovak Republic 590 Holland 206
Bulgaria 519 Germany 188
Hungary 511 Italy 186
Portugal 466 Canada 180
y
Rep. Checa 444 Finlandia 179
Ireland 345 Australia 164
op
Israel 332 Great Britain 160
France 329 Japan 145
Spain 302 Sweden 132
Source: Luchemos por la Vida.
tC
an
ic
pl
Ap
15
4883057
Table 3. Car Accident Mortality Rates per 100,000 inhabitants

Mean Std. Dev. Min Max
Buenos Aires 12.23 5.02 6.63 22.75
Ciudad de Buenos Aires 7.79 6.34 2.94 27.90
Catamarca 14.49 5.97 6.20 29.36
Córdoba 10.49 2.02 5.71 14.60
Corrientes 13.42 4.34 9.59 26.59
Chaco 10.50 1.69 7.13 12.68
Chubut 7.95 6.78 1.42 29.13
y
Entre Rios 11.96 3.36 7.47 18.43
Formosa 7.54 2.56 2.70 12.50
op
Jujuy 13.05 5.25 6.55 21.70
La Pampa 19.90 9.40 9.95 41.13
La Rioja 16.08 5.15 1.05 2.51
Mendoza 14.29 2.06 9.80 16.56
Misiones
Río Negro
tC 13.02
11.08
2.48
3.15
8.08
6.05
17.17
16.71
Salta 10.45 2.96 5.21 17.86
San Juan 17.86 5.37 12.16 33.64
San Luis 15.75 6.24 10.00 36.65
an
Santa Cruz 15.44 5.32 5.10 25.53
Santa Fe 14.03 2.61 11.18 18.52
Santiago del Estero 11.31 3.51 4.23 17.97
Tierra del Fuego 6.58 2.62 2.76 10.76
ic
Tucumán 11.35 1.99 8.60 14.87

Note: Each mean was calculated taking into account the whole period (1990-
2005).
pl
Ap
16
4883057
Table 4. Before and After Comparisons of Car Accident

Mortality Rates
Pre-Treatment Post-Treatment
‗‗Dry Law‘‘ Jurisdictions period period
Buenos Aires 16.13 9.64
Ciudad de Buenos Aires 8.34 4.18
Córdoba 10.18 12.48
Corrientes 15.51 11.04
Chaco 9.8 11.29
Chubut 9.33 2.4
La Pampa 25.54 16.15
La Rioja 17.68 11.69
Misiones 12.88 13.92
Río Negro 11.36 9.26
y
San Juan 18.94 13.9
Santa Cruz 14.99 18.39
op
Santa Fe 14.31 12.22
Tierra del Fuego 7.79 4.97
Tucumán 12.1 9.29
All 13.66 10.72
tC
an
ic
pl
Ap
17
4883057
Table 5. Jurisdictions with Alcohol Restrictions

‗‗Dry Law‘‘ Jurisdictions Date Department
Buenos Aires 1996 All
Ciudad de Buenos Aires 2004 All
Córdoba 2004 Ciudad de Córdoba
Corrientes 1999 Ciudad de Corrientes
Chaco 1999 All
Chubut 2003 All
La Pampa 1996 All
La Rioja 2001 All
Misiones 2004 Posadas
Río Negro 2004 All
San Juan 2003 All
y
Santa Cruz 2004 All
Santa Fé 2004 Ciudad de Santa Fé, Rosario
Tierra del Fuego 2000 All
op
Tucumán 2002 All
tC
an
ic
pl
Ap
18
4883057
Table 6. Pre-Treatment Characteristics

Mean for the Mean for the
Control Group Treated Group Difference
Car mortality rate
(per 100,000 inhabitants) 13.375 13.133 .241
(1.426) (1.323) (2.096)
Population 668395 1771208 -1102813
(140325.5) (807348.6) (1122217)
Gross domestic product
(per capita) 7364.986 6191.364 1173.622
(3611.714) (1072.775) (2980.195)
Unemployment rate .055 .070 -.001
(.009) (.006) (.011)
y
Public Expenditure
(per capita) 746.493 867.566 -121.072
op
(82.146) (156.474) (224.448)
Citizen involvement 82.125 83.06667 -.941
(1.652) (1.232) (2.075)
Number of registered cars
(per 100,000 inhabitants) 13690.63 20739.27 -7048.642
Precipitation
tC (2381.666)
783.25
(117.602)
(2069.292)
834.733
(130.224)
(3332.612)**
-51.483
(198.928)
Percentage of paved road 73.375 75.866 -2.491
(7.485) (5.991) (9.870)
Life expectance 70.281 70.984 -.703
an
(.509) (.287) (.539)
Alcohol risk consumption rate 1.046 1.2 -.153
(.078) (.068) (.109)
Number of doctors
(per 100,000 inhabitants) 237.600 310.929 -73.329
ic
(19.940) (57.810) (81.295)

Note: Standard errors are in parentheses.
pl
Ap
19
o
4883057
tC
Table 7. Impact of the Dry Law on Car Accident Mortality Rates
KERNEL
an
FULL SAMPLE USING OBSERVATIONS ON COMMON SUPPORT MATCHINGa
(1) (2) (3) (4) (5) (6) (7) (8) (9)
DryLaw(=1) -1.571208 -2.062532 -2.201858 -1.902302 -1.599648 -2.077977 -2.330818 -1.986002 -1.832
(.8852299)* ( .9312431)** (.9412235)** (.9564898)** (.8994801)* (.9422438 )** (.9555845)** (.9739893)** (0.740)**
[ .8614502]* [.9036023]** [.9430633]** [.9665846]** [ .894417]* [.079158]** [.9871087]** [1.00801]**
{.9618942} {.6616625}*** {.7176608}*** { .7882377}** {.96953} {.6657079}*** {.7353158}*** {.8072872}**
% in mortality rate
GDP per capita
ic -11.4 -15
-.0001796
(.0002242)
[.000245]
-16
-.0002149
(.0002269)
[.0002433]
-13.8
-.0001186
(.0002361)
[.0002575]
-11.6 -15.1
-.0001924
(.0002242)
[ .0000201]
-16.9
-.00025
(.0002271)
[.0002442]
-14.4
-.0001489
(.0002365)
[.0002563]
-13.3
pl
{.0004567} {.000435} {.0004311} { .000461} {.0004278} {.0004198}
Unemployment rate -30.72676 -30.31064 -31.96489 -27.31604 -25.81227 -27.91007
( 12.24095)** (12.24698)** (12.41652)** (12.68169)** (12.69179)** (12.84406)**
[12.23258]** [12.14504]** [12.37045]*** [1.120588]** [12.47874]** [12.63554]**
Ap
{12.83959}** {13.04981}** {12.62346}** {13.67664}* {13.60569}* {13.45073}**

Public Expenditure rate -.0003859 -.0005679 -.0003878 -.0003666 -.0006457 -.0004637
(.0007709) (.0007914) (.0008079) (.0007731) (.0007943) (.0008101)
[.0007667] [.0008349] [.0008281] [.0000668 ] [.0008555] [.0008454]
{.000805} {.0008836} {.0008374} {.0008344} {.0009161} {.0008573}
Citizen Involvement .1634018 .1432987 .2437653 .2160122
(.1608231) (.1624297) (.1661662) (.1675703)
[.1650702] [.1664365] [.1656662] [.1646803]
{.1878166} {.1996894} {.1741906} {.180947}
Number of registered vehicles -1526576 -1513959
-107662 -1074202
[14.60204] [14.61882]
{10.59999} {10.19451}
Precipitation .0011331 .001282
(.0009906) (.0010119)
[.0010849] [.0011099]
{.001523} {.0015558}
R2 0.12 0.22 0.26 0.20 0.11 0.21 0.26 0.22
Observations 336 282 282 282 322 271 271 271
Note: Each column reports the estimated coefficients of a regression model in which the dependent variable is the car mortality rate, whose mean is 13.73 during 1990-1995. Standard errors are in parentheses. Robust standard
errors are in brackets. Standard errors clustered at the province level are in braces. All the regressions include year and provinces fixed effects. a- Standard errors for the kernel matching estimate are bootstrapped standard errors.
* Statistically different from zero at the .1 level of significance.
** Statistically different from zero at the .05 level of significance.
*** Statistically different from zero at the .01 level of significance.
20
4883057
Table 8. Impact of the Dry Law in other Causes of Mortality
1990-1995 Mean Estimated Impact %in Mortality

Mortality Rate Coefficients Rate
Mortality in General 6.65942 0.0863804 12.9
(.1033147)
[.1132787]
y
{.1491665}
Infectious Diseases 64.68841 1.547875 2.3
op
(8.63611)
[4.600039]
{6.041139}
Tumors 128.8043 7.654269 5.9
tC
(20.54315)
[8.092138]
{7.306383}
Cardiovascular diseases 236.1159 2.279264 0.96
(36.26529)
[15.30938]
an
{16.17147}
Note: Each cell reports the estimated coefficient from a different difference in-differences
regression. Standard errors are in parentheses. Robust standard errors are in brackets.
Standard errors clustered at the province level are in braces. All the regressions include year
and province fixed effects.
ic
pl
Ap
21
4883057
Appendix
Table 4. Test of Trends
Inter1990 .1379642
(.2310363)
Inter1991 -.1886652
(.2322094)
y
Inter1992 .2108176
(.2310363)
op
Inter1993 -.0683793
(.2310363)
Inter1994 .0337442
(.2326525)
tC
Inter1995 -.01365
(.2310363)
Inter1996 .2234399
(.2452311)
Inter1997 .1522996
an
(.2452311)
Inter 1999 .2697993
(.2366806 )
Inter2001 -.0120813
(.2420228 )
ic
Inter2002 -.0711589
(.2459588)
pl
Inter2003 .0436593
(.2620646)
Year fixed effect Yes
Ap
Province fixed effect Yes

R 271
Observations 0.10
Note: standard errors are in parenthesis
25
Car Accident Mortality Rates
20
15
10
0
90
91
92
93
94
95
96
97
99
00
01
02
03
04
05
19
19
19
19
19
19
19
19
19
20
20
20
20
20
20
Buenos Aires
22
Writing Sample
Car Accident Mortality Rates Car Accident Mortality Rates Car Accident Mortality Rates
Ap
10
15
20
25
30
0
5
10
12
14
16
0
2
4
6
8
10
15
20
25
30
0
5
19 19 19
90 90 90
19 19 19
91 91 91
19 19 19
pl
92 92 92
19 19 19
19
93
19
93
ic 19
93
94 94 94
19 19 19
95 95 95
an
19 19 19
96 96 96
19 19 19
97 97 97
23
19
98
tC
19 19 19
99 99 99
Capital Federal
Cordoba
Corrientes
20 20 20
00 00 00
20 20 20
01 01 01
op
20 20 20
02 02 02
20
20
20
03
20
20
03
20
y
03
04
04 04
20 20 20
05 05 05
4883057
Maria Sviatschi
Writing Sample
Car Accident Mortality Rates Car Accident Mortality Rates Car Accident Mortality Rates
Ap
10
15
20
25
30
35
40
45
0
5
10
15
20
25
30
35
0
5
10
12
14
0
2
4
6
8
19 19 19
90 90 90
19 19 19
91 91 91
19 19 19
pl
92 92 92
19 19 19
93
19
93
ic 19
93
19
94 94 94
19 19 19
95 95 95
an
19 19 19
96 96 96
19 19 19
97 97 97
24
tC
19 19 19
Chaco
99 99
Chubut
99
La Pampa
20 20 20
00 00 00
20 20 20
01 01 01
op
20 20 20
02 02 02
y
20 20 20
03 03 03
20 20 20
04 04 04
20 20 20
05 05 05
4883057
Maria Sviatschi
4883057
20

18
16
14
12
10
y
8
6
op
4
2
0
90
91
92
93
94
95
96
97
99
00
01
02
03
04
05
19
19
19
19
19
19
19
19
19
20
20
20
20
20
20
18
tC Misiones
an
16
14
12
10
ic
8
6
pl
4
2
0
Ap
1990 1991 1992 1993 1994 1995 1996 1997 1999 2000 2001 2002 2003 2004 2005
Rio Negro
30
25
20
15
10
0
1990 1991 1992 1993 1994 1995 1996 1997 1999 2000 2001 2002 2003 2004 2005
Santa Cruz
25
4883057
40
35

30
25
20
y
15
op
10
5
0
1990 1991 1992 1993 1994 1995 1996 1997 1999 2000 2001 2002 2003 2004 2005
20
tC San Juan
an
18
16
14
12
ic
10
8
6
pl
4
2
Ap
0
1990 1991 1992 1993 1994 1995 1996 1997 1999 2000 2001 2002 2003 2004 2005
Santa Fe
12
10
0
1990 1991 1992 1993 1994 1995 1996 1997 1999 2000 2001 2002 2003 2004 2005
Tierra del Fuego
26
4883057
16
14
Car Accident Mortality Rates 12
10
y
8
6
op
4
2
0
tC
1990 1991 1992 1993 1994 1995 1996 1997 1999 2000 2001 2002 2003 2004 2005
Tucuman
an
ic
pl
Ap
27
4883057
References
Bertrand, Marianne, Esther Duflo & Sendhil Mullainathan (2004).―How Much Should We Trust.
Differences-in-Differences Estimates?‖ National Bureau of Economic Research 119, No.1: 249-275.
Brown, Robert, Todd Jewell & Jerrell Richer (1996). ―Endogenous Alcohol Prohibition and Drunk
Driving.‖ Southern Economic Journal 62, No.4: 1043-1053.
y
Chaloupka, Frank J., Henry Saffer & Michael. Grossman (1993). ―Alcohol-Control Policies and Motor
op
Vehicle Fatalities.‖ Journal of Legal Studies 22, No.1: 161-186.
Cook, Philip J., & Michael J. Moore (1994). ―This Tax is For You: The Case of Higher Beer Taxes.‖
National Tax Journal 47, No.3: 559-573.
tC
Cook, Phillip J., & George Tauchen (1982). ―The Effect of Liquor Taxes on Heavy Drinking.‖ The Bell
Journal of Economics 13, No. 2: 379-390.
Cook, Phillip J., & George Tauchen (1984). ―The Effect of Minimum Drinking Age Legislation on
Youthful Auto Fatalities, 1970-77.‖ Journal of Legal Studies, Vol. 13, No. 1: 169-90.
an
Dee, Thomas. S. & William Evans (2001). ―Behavioral Policies and Teen Traffic Safety.‖ The American
Economic Review 91, No.2: 91-96.
Evans, William N., Neville, D. & John D. Graham. (1991). ―General Deterrence of Drunk-Driving:
ic
Evaluation of Recent American Policies.‖ Risk Analysis 11, No.2: 279-289.

Galiani, Sebastian, Gertler, Paul & Ernesto Schargrodsky (2005). ―Water for Life: The Impact of the
pl
Privatization of Water Services on Child Mortality.‖ Journal of Political Economy 113, No.1: 83-120.
Grabowski, David C., & Michael A. Morrisey (2001). ―The effect of State Regulations on Motor Vehicle
Ap
Fatalities for Younger and Older Drivers: A Review and Analysis.‖ The Milbank Quaterly 79, No.4: 517-545.
Heckman, James J., Hidehiko Ichimura, and Petra Todd. 1997. ―Matching as an Econometric Evaluation
Estimator: Evidence from Evaluating a Job Training Programme.‖ Rev. Econ. Studies 64: 605–54.
McCarthy, Patrick S. (1999). ―Public Policy and Highway Safety: A City-Wide Perspective.‖ Regional Science
and Urban Economics 29: 231-244.
Rosenbaum, Paul R., & Donald B. Rubin (1983). ―The Central Role of the Propensity Score in
Observational Studies for Causal Effects.‖ Biometrika 70: 41–55.
Ruhm, Christofer J. (1995). ―Alcohol Policies and Highway Vehicle Fatalities.‖ National Bureau of Economic
Research, No. 5195.
Toomey, Traci L. & Alexander C. Wagenaar (1999). ―Policy Options for Prevention: The Case of
Alcohol.‖ Journal of Public Health Policy 20, No. 2: 192-213.
World Bank & World Health Organization (2004). ―World Report on road traffic injury prevention.‖
Winn, Russell G. & David Giacopassi. (1993). ―Effects of County-Level Alcohol Prohibition on Motor
Vehicle Accidents.‖ Social Science Quarterly 74, No.4: 783-792.
28

Columbia MicaSviatschi

Uploaded by

Document Information

Original Description:

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Columbia MicaSviatschi

Uploaded by

Copyright:

Available Formats

AY #________________

4883057 FEE_______________________________________ APPL #________________

APPLICATION FOR ADMISSION

961328 University of San Andres Economics

_____________ __________________________________________________ ____________________________________________________

_____________ _______________________________ ______________ _____________________________ __________________

_____________ __________________________________________________ ____________________________________________________

_____________ __________________________________________________ ____________________________________________________

GRE GENERAL GRE SUBJECT GMAT

ARE YOU A CURRENTLY ENROLLED AT GSAS? ___________________________________________________________________________________________________

HAVE YOU PREVIOUSLY APPLIED TO ANY DIVISION OF COLUMBIA UNIVERSITY? __________________________________________________________________

ARE YOU APPLYING FOR A DUAL-DEGREE PROGRAM? ___________________________________________________________________________________________

DO YOU WISH TO BE CONSIDERED FOR FELLOWSHIP AID FROM COLUMBIA? ✔ YES  NO

DO YOU HOLD AN OUTSIDE AWARD? ✔ NO

NAME OF AWARD ______________________________________________________________________________________________________________________________

NAME OF AWARD ______________________________________________________________________________________________________________________________

________________________________ _____________________________ ______________________________ ______________________________________

ARE YOU APPLYING FOR AN OUTSIDE AWARD? ✔ NO

NAME OF AWARD ______________________________________________________________________________________________________________________________

________________________________ _____________________________ ______________________________ ______________________________________

NAME OF AWARD ______________________________________________________________________________________________________________________________

________________________________ _____________________________ ______________________________ ______________________________________

NAME OF AWARD ______________________________________________________________________________________________________________________________

________________________________ _____________________________ ______________________________ ______________________________________

Two examples of my research during my experience in an international organization include a paper

Maria Micaela Sviatschi

CONGRESSES AND PRESENTATIONS

John Hopkins University. Bloomberg School of Public Health. Baltimore MD.

AWARDS AND SCHOLARSHIPS

Dry Law for Drunk Drivers:

behind tumors and heart diseases.2

some judicial court (Judge, Office of the Public Prosecutor, Chamber).10

9 There is no available data for the provinces in the year 1998.

increased efforts by police officers to find drunk drivers.

III. The Effect of the Dry Law on Car Accident Mortality

13 There is no data available of public expenditure for the year 1991.

the difference-in-differences model can be represented by the following equation:

CarMortalityit= DryLawit + Xit + i + t + it (1)

19 See Bertrand et al. (2004).

23 Chaloupka et al. (1993) has also found a similar result.

27 All the specifications reported satisfy the balancing condition.

to the distance in propensity scores.

Table 1. Motor-Vehicle Fatalities in the World

Table 2. Car Mortality Rates (per 1,000,000 vehicles)

Table 3. Car Accident Mortality Rates per 100,000 inhabitants

Tucumán 11.35 1.99 8.60 14.87

Table 4. Before and After Comparisons of Car Accident

Table 5. Jurisdictions with Alcohol Restrictions

Table 6. Pre-Treatment Characteristics

(19.940) (57.810) (81.295)

{12.83959}** {13.04981}** {12.62346}** {13.67664}* {13.60569}* {13.45073}**

Table 8. Impact of the Dry Law in other Causes of Mortality

1990-1995 Mean Estimated Impact %in Mortality

Province fixed effect Yes

Car Accident Mortality Rates

Car Accident Mortality Rates

Tierra del Fuego

Evaluation of Recent American Policies.‖ Risk Analysis 11, No.2: 279-289.

You might also like

4883057 FEE_______________________ APPL #

_ ______________

_ ___ _ ____

_ ______________

_ ______________

DO YOU WISH TO BE CONSIDERED FOR FELLOWSHIP AID FROM COLUMBIA? ✔ YES NO

__________________ _ ______________________

__________________ _ ______________________

__________________ _ ______________________

__________________ _ ______________________

{12.83959} {13.04981} {12.62346}** {13.67664}* {13.60569}* {13.45073}**