Handbook of Social Economics SET: 1A, 1B
1600
Handbook of Social Economics SET: 1A, 1B
1600eBook
Related collections and offers
Overview
How can economists define and measure social preferences and interactions?
Through the use of new economic data and tools, our contributors survey an array of social interactions and decisions that typify homo economicus. Identifying economic strains in activities such as learning, group formation, discrimination, and the creation of peer dynamics, they demonstrate how they tease out social preferences from the influences of culture, familial beliefs, religion, and other forces.
- Advances our understanding about quantifying social interactions and the effects of culture
- Summarizes research on theoretical and applied economic analyses of social preferences
- Explores the recent willingness among economists to consider new arguments in the utility function
Product Details
| ISBN-13: | 9780444537140 |
|---|---|
| Publisher: | Elsevier Science |
| Publication date: | 11/10/2010 |
| Series: | ISSN , #1 |
| Sold by: | Barnes & Noble |
| Format: | eBook |
| Pages: | 1600 |
| File size: | 17 MB |
| Note: | This product may take a few minutes to download. |
About the Author
Read an Excerpt
HANDBOOK OF SOCIAL ECONOMICS VOLUME 1A
ELSEVIER
Copyright © 2011 Elsevier B.V.All right reserved.
ISBN: 978-0-444-53714-0
Chapter One
Nature and Nurture Effects On Children's Outcomes: What Have We Learned From Studies of Twins And Adoptees?Bruce Sacerdote Dartmouth College and NBER
Contents
I. Introduction and Overview 2
II. The Behavioral Genetics Model 5
III. Canonical Results from the Behavioral Genetics Literature 8
IV. Critiques and Challenges to Interpretation of the Behavioral Genetics Results on IQ and Schooling 14
V. Treatment Effects and Regression Coefficients 17
VI. Results from Economics on Adoptees 19
VII. Putting It All Together: What Does It Mean? 25
References 26
Further Readings 29
Abstract
There is a rich history of using data from twins and from adoptees to control for genetic influences and thereby examine the impact of environment on children's outcomes. The behavioral genetics model is the workhorse of this literature and for a variety of outcomes including IQ scores and personality measures behavioral geneticists find that the bulk of the variance that can be explained is correlated with genetic influences. However, finding that variation in test scores has a large genetic component is quite different than asking whether test scores can be improved by interventions and changes in policy or whether such interventions pass a cost benefit test. Economists have recently begun asking how the intergenerational transmission of educational attainment, income and health vary when a child is being raised by adoptive rather than biological parents. Results suggest that both the biological and the nurturing parents contribute a great deal to the transmission of income and education to their children JEL Codes: I0, J0, J24
Keywords
twins
adoption
intergenerational transmission
nurture
educational attainment
I. INTRODUCTION AND OVERVIEW
A fundamental question in social science has long been the degree to which children's outcomes are influenced by genes, environment, and the interaction of the two. One sensible way to attempt to separate out the effects of genes and environment is to examine data on twins or adoptees since we may be able to make plausible assumptions about the genetic relationships between identical versus fraternal twins, or between parents and their adoptive and non-adoptive children.
I begin this chapter by reviewing the methods used by psychologists and behavioral geneticists to identify the effects of nature and nurture, and I summarize some of the key results from this large literature. I discuss the assumptions underlying the behavioral genetics model and explain some of the challenges to interpreting the results. I use these issues of interpretation to motivate why economists and sociologists have used a different approach to measuring the impact of environment on children's outcomes. And I discuss the results from the recent literature in economics on environmental versus genetic determinants of children's education, income, and health. Finally, I try to bring the results from both literatures together to address the issues of what we do know, what we don't know and whether this work has implications for social policy or other research on children's outcomes.
Behavioral geneticists have estimated the "heritability" of everything from IQ to "shrewdness" to alcoholism. Their most frequently cited result is that genetic factors explain about 50 to 60% of the variation in adult IQ, while family environment explains little of the variation in adult IQ. This finding is incredibly robust (see Devlin, et al. [1994]). But researchers' interpretation of the finding varies. Harris [1998] uses the finding of almost no effect from family environment as a key piece of evidence for her thesis that parents do not have a direct effect on their children's outcomes. Both Hernnstein and Murray [1994] and Jensen [1972] interpreted the lack of measured effects from family environment to mean that policies aimed at improving the home and school environment of children are likely to have small impacts on outcomes.
Jencks et al. [1972], Jencks [1980], and Goldberger [1979] provide a series of reasons why such strong interpretations may be unwarranted. First, understanding the determinants of IQ is different than understanding the determinants of educational attainment, income, and health. Second, the assumptions of the behavioral genetics model may be tilted towards overstating the importance of genes in explaining variation in outcomes. Positive correlation between family environment and genes raises the heritability estimate. Third, family environment is likely endogenous and may depend heavily on genes (Jencks [1980], Scarr and McCartney [1983], Dickens and Flynn [2001]). This endogeneity makes any simple nature nurture breakdown difficult to interpret.
Fourth, noting that variation in a given outcome for some population has a large genetic component is different from saying that the outcome is predetermined or cannot be changed by interventions. Genetic effects can be muted just as environmental effects can be. To take Goldberger's example, a finding that most of the variation in eyesight is due to genes does not imply that we should stop prescribing eyeglasses for people. The use of eyeglasses may add enormous utility for people (and offer an excellent return on investment), regardless of what fraction of eyesight is measured as being environmental.
In other words, knowing what fraction of the existing variance is environmental does not tell us whether a given environmental intervention is doomed to failure or success. Imagine a state with uniformly mediocre schools. Perhaps in that population, school quality doesn't explain any of the variation in student outcomes. But there may be great benefits from introducing a new school with motivated peers, high financial resources and high teacher quality. It is critical to bear in mind that the variance breakdown only deals with variation in the sample. Mechanically, expanding a sample to encompass a broader range of environment (e.g., considering children in both Africa and the US as opposed to the US alone) increases the variation in inputs and outcomes and likely the proportion of the variation in outcomes that is due to environment.
What then do we learn from behavioral geneticists' estimates of the relative contribution of genes, family environment, and non-shared environment? We are getting a breakdown of the variance of the outcome in the current population, assuming a particular structural model. In the case of adoption studies, heritability is a measure of how much more biological siblings resemble each other relative to adoptive siblings. Similarly, in the case of twin studies, heritability is a measure of how much more outcomes for identical twins are correlated relative to outcomes for fraternal twins or other siblings. See the next section for the algebra. If heritability estimates were labeled as the additional correlation in outcomes that is associated with being identical rather than fraternal twins, there might be less misinterpretation of these numbers.
Such a variance breakdown may be worth something to social scientists as an estimate as to whether genetic variation is particularly important in determining an outcome. Even if the functional form of the behavioral genetics model is simplified, the model may still deliver useful relative rankings of how much variation in genes contributes to variation of different outcomes (e.g., height versus age at first marriage.)
Economists and sociologists have suggested several ways to reframe the question so as to use adoption data to estimate some of the causal impacts from family environment without having to know the true model by which outcomes are determined and without having to deliver a complete nature, nurture breakdown. This line of research consists of regressing child outcomes on parental characteristics, i.e., using the more standard approach within economics. For example, Plug and Vijverberg [2003], and Sacerdote [2007] regress adoptee's years of schooling on the mother's years of schooling, family income, and family size. The advantage of using regression is that it tells us which specific parental inputs are most correlated with child outcomes and the slope of the relationships.
Certainly one cannot necessarily take these regressions coefficients as causal due to measurement error, endogenous relationships among variables, and unobservables. But these regressions provide a starting point for understanding which parental inputs matter and how much they matter even in the absence of a genetic connection between parents and children. We can then compare the observed coefficients on parental inputs that we find for adoptees to those that use other sources of variation in family characteristics. For example, Sacerdote [2007] finds little evidence for a direct effect of parental income on adoptees' income and education. This finding is generally consistent with the work of Mayer [1997] and Blau [1999]. And one can compare the effects of family size found in adoption studies to those found by Black Devereux and Salvanes [2005b] and Angrist, Lavy, and Schlosser [2005] who use the birth of identical twins and sex preferences as an exogenous shock to family size.
One can also generate separate transmission coefficients for adoptees and nonadoptees by regressing the child's outcome on that of the parent. See Björklund Lindahl and Plug [2006] and Björklund Jäntti and Solon [2007] for transmission coefficients of income (education) from parents to adoptees and non-adoptees. This enables one to see how transmission varies when there is and is not a genetic link to the parents. This work also has the advantage of providing comparability between existing estimates of transmission coefficients from parents to children such as those in Solon [1999], Zimmerman [1992], and Mazumder [2005].
Several bottom lines emerge from my summary of the nature and nurture literature. First, economists who are not already familiar with the literature are generally surprised by how much genes seem to matter, or more precisely stated, how much less adoptees resemble their adoptive parents and siblings than do non-adoptees. Second, the estimated effects of family environment on adoptee outcomes are still large in some studies and leave scope for children's outcomes to be affected by changes in family, neighborhood, or school environment. And the importance of family environment can rise significantly when the model is made more flexible. Third, the precise breakdowns of variance provided by behavioral genetics are subject to a number of important issues of interpretation.
Ultimately, the evidence is consistent with the widely held view that both nature and nurture matter a great deal in determining children's outcomes. Parental characteristics matter even in the absence of any genetic connection to their children. A more deeply informed view will also recognize that certain measured parental effects or transmission coefficients from parents to children drop significantly, when one considers adoptees rather than children raised by their biological parents. However, that fact does not negate any of the findings of researchers who measure directly the causal effects of changing school, neighborhood, and family environment on outcomes.
II. THE BEHAVIORAL GENETICS MODEL
In the simplest version of the model, child outcomes (Y) are produced by a linear and additive combination of genetic inputs (G), shared (common) family environment (F) and unexplained factors, which the BG literature often calls non-shared or separate environment, (S). This implies that child's educational attainment can be expressed as follows:
Child's years of education (Y) = G + F + S. (1)
The key assumption's here are that nature (G) and shared family environment (F) enter linearly and additively. Separate environment (S) is by definition the residual term and is uncorrelated with G and F. In the simple version of the model, one further assumes that G and F are not correlated for a given child. On the surface, this seems like a strange assumption and one that could perhaps be defended for some subsets of adoptees but not for children being raised by their biological parents. At a deeper level, behavioral geneticists often take F to represent that portion of family environment that is not correlated with genes, and they assume that G represents both the effects of gene and gene-environment correlation. The correlation between G and F can be modeled explicitly. If F itself is endogenous, modeling becomes very difficult. With these caveats in mind, one can already see that the BG breakdown into genes versus family environment is not necessarily an easily interpreted decomposition.
(Continues...)
Excerpted from HANDBOOK OF SOCIAL ECONOMICS VOLUME 1A Copyright © 2011 by Elsevier B.V.. Excerpted by permission of ELSEVIER. All rights reserved. No part of this excerpt may be reproduced or reprinted without permission in writing from the publisher.
Excerpts are provided by Dial-A-Book Inc. solely for the personal use of visitors to this web site.
Table of Contents
Social Preferences- Cultural Transmission and Socialization. (A. Bisin and T. Verdier)
- Social Construction of Preferences (J. Benhabib and A. Bisin)
- Preferences for Status (R. Frank and Ori Heffetz)
- Evolutionary Selection (A. Robson and L. Samuelson)
- Nature and Nurture (B. Sacerdote)
- Social Beliefs (A. Alesina and P. Giuliano)
- Does Culture Matter? (R. Fernandez)
- Social Capital (L. Guiso, P. Sapienza, L. Zingales)
Empirical Analysis of Social Interactions
- Identification of Social Interactions (L. Blume, W. Brock, S. Durlauf, Y. Ioannides)
- Neighborhood Effects and Housing (Y. Ioannides)
- Peer and Neighborhood Effects in Education (D. Epple and R. Romano)
- Labor Markets and Referrals (G. Topa)
- Risk Sharing Among and Between Households (M. Fafchamps)
- Credit and Labor Networks in Development (K. Munshi)
- Econometric Methods for the Analysis of Assignment Problems in the Presence of Complimentary and Social Spillovers (B. Graham)
Social Actions
- Norms, Customs, and Conventions (P. Young and M. Burke)
- Social Norms and Social Assets (A. Postlewaite)
- Local Interactions (O. Ozgur)
- Group Formation and Local Interactions (S. Durlauf)
- An Overview of Social Networks and their Analysis (M. Jackson)
- Formation of Networks and Coalitions (F. Bloch and B. Dutta)
- Diffusion, Strategic Interaction, Social Structure (M. Jackson and L. Yariv)
- Herding (C. Chamley)
- Learning in Social Networks (S. Goyal)
- Experiments in Social Learning (S. Kariv)
- Matching, Allocation, and Exchange of Discrete Resources (T. Sonmez&U. Unver)
- Discrimination (G. Loury)
- The Importance of Segregation, Discrimination, Peer Dynamics, and Identity in Explaining Trends (R. Fryer)
- Theories of Statistical Discrimination and Affirmative Action: A Survey (H. Fang and A. Moro)
What People are Saying About This
Offers high-level perspectives on the economics of social choice and the effects of social influences on economic decisions