Spring 2017

Dondena Seminar Series
Spring 2017


All seminars (and light lunches!) will begin at 12:45 p.m. and be held in Room 3-e4-sr03, unless otherwise noted.

Titles and abstracts will be posted as they become available.


Download the series poster here.


January 30
Rachel Margolis, Western University

Demographic Change and the Increasing Prevalence of Older Adults without Close Kin

Demographic changes in marriage, fertility and mortality affect the size of the pool of family members that older adults have to rely on for instrumental, emotional, and economic support. A potentially growing kinless population is of concern because the kinless may be at higher risk for use of institutional care, expensive health services, and other social needs. However, despite the importance of the kinless older population, no current estimates exist. In this article, we document the size and characteristics of the population of older adults without close kin in the contemporary US. First, we examine the size of the kinless population using different definitions of kinlessness, and examine how it is changing across birth cohorts. Next, we establish the correlates of being kinless. Last, we present forecasts about how the size of this group will increase in the next 50 years. Our results show that the kinless population is a substantial subpopulation and likely to likely to increase substantially in the next 50 years. This demographic profile shows that the rise of kinlessness is a new and important issue to be addressed

February 6
Emilio Zagheni, University of Washington

Leveraging Twitter and Facebook data to study migration

Although international migration is a major determinant of demographic change,  estimates of migration rates are often inconsistent across countries or outdated. In this talk I discuss our recent work that leverages Facebook and Twitter data to study migration. Facebook allows advertisers to target users with certain characteristics, such as age, gender, country of origin, education level, or topical interest. Before an ad is launched, Facebook's advertising platform produces an estimate of how many users match the provided criteria. We use this functionality, akin to a 'digital census' over Facebook users to test the feasibility of nowcasting stocks of migrants within and across countries. We use geo-referenced Twitter tweets for about 62,000 users, spanning the period between 2010 and 2016, to estimate a series of US internal migration flows, under varying definitions of migration, to evaluate the relationships between short-term mobility and long-term migration, and  to test the possibility of using geo-located social media data to harmonize migration statistics.


February 13
Megan Sweeney, UCLA

The Changing Context of Contraceptive Choice across the Reproductive Life Course in the USA and Britain

The association between early and unintended childbearing and women's socioeconomic standing has long interested social scientists. Recent policy discussions emphasize the potential for long-acting reversible contraceptives (LARCs) to reduce poverty and inequality. Our understanding of social-class disparities in LARC use, however, remains limited. In particular, although the context of contraceptive decision-making varies across the reproductive life course, little work directly considers how the context of contraceptive choice varies for the time before a first birth occurs (childbearing “starters”), the time between births (childbearing “spacers”), and the time after the end of intended childbearing (childbearing “limiters”). We address this omission using recent survey data from the United States (National Survey of Family Growth) and Britain (National Survey of Sexual Attitudes and Lifestyles). We consider our findings in light of ongoing policy discussions about the potential for LARCs to reduce inequality and poverty among disadvantaged women and children.


February 20
Vincent Pons, Harvard Business School

Strategic or Expressive Voters? Evidence from a RDD in French Elections

In French parliamentary and local elections, candidates ranked first and second in the first round automatically qualify for the second round, while a third qualifies only when selected by more than 12.5 percent of the registered citizens. Using a fuzzy regression discontinuity design around this threshold, we find that the third candidate attracts both “loyal” voters, who would have abstained or cast a blank or null vote if she were not present, and “switchers”, who would otherwise have voted for the top two candidates. Switchers vote for the third candidate even when she is very unlikely to win the election. This disproportionately harms the candidate ideologically closest to the third, and changes the winner in one fifth of the races, causing the defeat of the Condorcet winner. We rationalize these results by a model of strategic voting and conclude that a large fraction of voters value voting expressively over behaving strategically to ensure the victory of their second best.


February 27
Martin Dribe, Lund University

Earnings, Social Class and Mortality in the Long Run: Southern Sweden 1815-2010

All contemporary developed countries show marked health and mortality differentials by socioeconomic status, whether measured by education, social class or income. We know a great deal about the detailed patterns of these inequalities today, but much less about when the health gradient emerged. Some argue that it has always existed, others maintain that it is of a more recent origin. In this paper we study differences in male adult mortality in Sweden by socioeconomic status at the individual level over an almost 200-year period. We analyze two different dimensions of socioeconomic status – occupational class and earnings – and how they are associated with mortality in working ages and among the elderly. Our results indicate that it is not until after 1970 that consistent class and earnings gradients in mortality of a modern kind emerge. Before that people with earnings below the tax threshold and without a registered occupation had higher mortality, while there were no socioeconomic differences in the rest of the population. The earnings gradient was quite independent of social class, while occupational differences disappear when controlling for earnings. The patterns for working age men and elderly men were fairly similar.


March 13
Raymond Duch, University of Oxford

Is Cheating A National Past-Time? Experimental Evidence

Individuals in the population, controlling for its costs and benefits, systematically lie or cheat.  And there are significant economic costs associated with this cheating.  We conduct public goods experiments in order to better understand who in the population cheats: subjects earn real money, are subject to a deduction that is distributed to their group, and can lie about their earnings. Performing better on incentivized real effort tasks results in more cheating.  High performance types also give less in a conventional Dictator Game and cheat more in a classic die game in which they privately report the results of tossing a die. We conclude that there are identifiable High Performance types in the population who are predisposed to cheating.  We also examine whether cheating or corrupt behavior varies significantly across diverse national contexts (Britain, Chile, and Russia).  Cheating, not surprisingly, is pervasive but it is also the case that this predilection for cheating by High Performance types knows no national, economic or cultural boundaries. 


March 27
Giovanni Mastrobuoni, Collegio Carlo Alberto

The Effect of Neighbourhood Police Patrolling on Crime is Zero, joint with Jordi Blanes i Vidal

A number of recent economic studies have used plausibly exogenous variation to shown that more policing reduces crime. Consensus has emerged that police patrolling deters crime, with elasticities that are estimated to be close to -1/3. We use evidence from both a natural experiment and high frequency variation in police presence to show that this relation is more nuanced than previously thought, and strongly depends on the scale and salience of the intervention. The natural experiment uses 8,662 discontinuous changes in police presence in well-defined locations and times. Exploiting this variation, we present clear evidence of a precisely estimates zero effect of police patrolling on crime. The high frequency dimension of the data further allows us to credibly estimate the contemporaneous effect of police presence on crime, which we find to be large and negative. We also find that once police patrols exit an area, deterrence lasts for up to 35 additional minutes. We reconcile the two findings and provide policy prescriptions.


April 3

Marko Koethenbuerger

Is it luring profits? The case of European patent boxes

The effectiveness of European patent boxes in triggering R&D and fostering new patentable innovations is the subject of a growing debate. These regimes are considered liable of tax-favouring already successful ideas, without imposing a nexus between the final location of the intellectual property (IP) and its related innovation. This paper brings the debate forward onto the assessment of the quantitative impact of patent box regimes on profit shifting by multinational firms. Our empirical strategy builds on a difference-in-difference model comparing the pre-tax profit of European subsidiaries affiliated to corporate conglomerates that owned patents long before the introduction of IP boxes, to that of European subsidiaries affiliated to corporate conglomerates with no historical record of patent ownership. We find that European subsidiaries affiliated to foreign IP owners report, after the introduction of a local patent box, on average 2.5 to 3.9 percent higher profit compared to European subsidiaries affiliated to non-IP-owning conglomerates.


April 10
Elisabetta De Cao, University of Oxford

The Impact of Unemployment on Child Abuse and Neglect in the United States


In this paper, we show that unemployment increased the neglect and physical abuse of children in the United States during the period from 2004 to 2012. A one percentage point increase in the unemployment rate led to a 25 percent increase in neglect and a 12 percent increase in physical abuse. We identify these effects by instrumenting for the county-level unemployment rate with a predicted county-level unemployment rate, which we create by combining national level unemployment rates across industries with differences in the initial industrial structure across counties.
We  tested whether the unemployment effects can be explained by changes in alcohol abuse or divorce, but failed to find consistent evidence with those mechanisms. 
However, we find that results for neglect are consistent with the poverty mechanism. 
By exploiting variation in unemployment policies over time within states, we find that the extension of unemployment benefits mitigates the effect of unemployment on neglect.


May 8
Thomas Aronsson, Umea University



May 15
Jungho Kim, Ajou University

Parental leave and women's employment in Korea

The paper examines the impact of the duration of parental leave on women’s employment in Korea. It exploits the natural experiment of the extension of the effective parental leave from 12 months to 15 months. Since the effect is likely to vary depending on the opportunity cost, it focuses on the heterogeneous effects. The results are as follows. First, with the job-protected leave extended, female employees indeed took up the leave more often and for a longer period than before. In general, the impact on take-up is found to be smaller, but that on duration to be larger for high wage earners, which implies that there exists a fixed cost in switching between own and paid child care. Further, those in large firms tend to benefit more than those in small or medium-sized firms in terms of usage of leave. Second, the probability of returning to work did not change significantly after the extension, which suggests that the marginal extension of leave in the vicinity of one year does not lead to a negative consequence in terms of women’s career. However, the effect on return to work is found to vary substantially depending on individual wage and firm size, which implies that the distributional effect should be considered in designing the parental leave policy.


May 22
Gabriele Ballarino, Università degli Studi di Milano

Geographical mobility and status attainment in Italy in the second half of the 20th century

In this seminar, I will present our work concerning the geographical mobility of Italians during the second half of the 20th century and its relation to status attainment. We use ILFI data, including information concerning life-long geographical mobility collected at the more detailed level, the municipality. We created a new and unique dataset by exploiting this information, overlooked by previous research, and supplementing it with geographical and demographical information at the municipality level.

First, we describe the distribution of geographical mobility, both in general and according to the length of the movement. Second, we study selection into geographical mobility, estimating logit and event history models to observe what makes individuals unequal in their opportunities of geographical mobility, both in general and according to the length of the movement. Our findings confirm the main results of the literature: mobile individuals are positively selected on education, age and marriage, and negatively on occupation and children. The selection patterns, however, vary notably between different types of geographical mobility.
Third, we study the association between geographical mobility and occupational status attainment. Our results show that geographical mobility has a marginal role in status attainment, in particular with respect to social origins and own education, but it nevertheless has an impact, in particular concerning a number of “internal” mobility paths: for instance, it significantly increases the probabilities to access the service class for those who were born in the very same class. Finally, we take a causal perspective and develop an IV analysis to check whether the positive association between geographical mobility and status attainment is robust and can be thought of as a causal relation.

May 29
Ghazala Azmat, Sciences Po

Changes in Higher Education Funding and its Consequences

This paper investigates the impact of changes in the funding of higher education in the UK on students’ choices. In recent years, England and Wales have undergone drastic reforms in the way in which higher education is funded. Until the late 1990s, higher education was completely state funded but since then, a combination of tuition fees for all students and means-tested maintenance grants and loans have been introduced and changed in three major reforms of 1998, 2004 and 2012. Using detailed longitudinal micro-data that follows all students that attend state schools in England (more than 90% of all school-aged children) from lower education to higher education, we document the distributional effects of the policy change on various margins, such as, enrollment, relocation decisions, selection of institution, and program of study. For a subset of students, we can track them after completing higher education, allowing us to study the labor market effects of the change in policy.

Last updated 07 September 2018 - 16:22:03