Impact of Children on Parents' Well-being Through Causal Effect Estimation
Explore the complexities of estimating the causal effect of children on parents' well-being using instrumental variable approaches. Delve into different research questions, methods, strengths, and weaknesses in this intriguing research domain. Learn about the challenges researchers face in handling data and interpreting results.
Download Presentation
Please find below an Image/Link to download the presentation.
The content on the website is provided AS IS for your information and personal use only. It may not be sold, licensed, or shared on other websites without obtaining consent from the author. If you encounter any issues during the download, it is possible that the publisher has removed the file from their server.
You are allowed to download the files provided on this website for personal or commercial use, subject to the condition that they are used lawfully. All files are the property of their respective owners.
The content on the website is provided AS IS for your information and personal use only. It may not be sold, licensed, or shared on other websites without obtaining consent from the author.
E N D
Presentation Transcript
Estimating the Causal Effect of Children on Parents Well-being Strengths and Weaknesses of an Instrumental Variable Approach Gerrit Bauer Thorsten Kneip Dominik Steinbei er RC-Seminar, Venice, November 22, 2016
Motivation The belief that parenthood makes people happier seems to be rather pervasive across the world (Hansen, 2012, p33) [ ] adding virtually nothing to life, apart from perpetual difficulty and worry (Donath, 2015, p359) As well as providing reasons to expect the more satisfied to have higher fertility, the literature also provides other reasons which might lead one to expect lower fertility among the more satisfied (Parr, 2010, p639) [ ] there is no reason to expect that parents will have better or worse lives than nonparents (Deaton & Stone, 2014, p1328) One of the central questions in the analysis of subjective well-being is whether people adapt to conditions. If this is the case, then [ ] conditions or circumstances may not matter, in the long run at least. (Clark et al., 2008, p222) 2
Different Research Questions & Methods Cross sectional Twin-FE FE IV Myrskyl & Margolis (2014) Parenthood Deaton & Stone (2014) Kohler et al. (2005) ? Margolis & Myrskyl (2011) Myrskyl & Margolis (2014) Higher parities Kohler et al. (2005) ? # of children Stutzer & Frey (2006) Kohler et al. (2005) Stutzer & Frey (2006) ? (In-)direct effects Alesina et al. (2004) ? Pollmann-Schult (2014) ? Myrskyl & Margolis (2014) Dynamic effects Frijters et al. (2011) ? ? Reverse effect Aassve et al. (2016) ? ? ? 3
Better Methods or Data? Unfortunately, even the most advanced statistical approaches that have been used in this research fail to handle all [ ] problems, so reported results should be interpreted very cautiously (Kravdal, 2014, p263) 4
Different Research Questions & Methods Cross sectional Twin-FE FE IV Myrskyl & Margolis (2014) Parenthood Deaton & Stone (2014) Kohler et al. (2005) ? Margolis & Myrskyl (2011) Myrskyl & Margolis (2014) Higher parities Kohler et al. (2005) ? # of children Stutzer & Frey (2006) Kohler et al. (2005) Stutzer & Frey (2006) ? (In-)direct effects Alesina et al. (2004) ? Pollmann-Schult (2014) ? Myrskyl & Margolis (2014) Dynamic effects Frijters et al. (2011) ? ? Reverse effect Aassve et al. (2016) ? ? ? 5
Distinguishing Causal Structures #kids happy #kids happy #kids happy #kids happy #kids happy 6
Distinguishing Causal Structures #kids happy #kids happy IV IV #kids happy IV #kids happy IV #kids happy IV 7
Distinguishing Causal Structures #kids happy #kids happy IV IV X #kids happy IV #kids happy IV X #kids happy IV X 8
Distinguishing Causal Structures #kids happy #kids happy twins twins X #kids happy twins #kids happy twins X #kids happy twins X 9
Assuming Conditional Exogeneity happy #kids twins age infertility child mortality education + FE 11
Estimated Effects Men Women -1 -.75 -.5 -.25 0 .25 .5 .75 1 1.25 1.5 -1 -.75 -.5 -.25 0 .25 .5 .75 1 1.25 1.5 Source: pairfam 7.0 W1-W7; own calculations POLS POLS + controls FE FE + controls 3,495 obs. in 11,823 obs. years; 321 twin years IV IV + controls FE-IV FE-IV + controls 12
Mediation or Confoundedness? Men Women 0 .25 .5 .75 1 1.25 -.75 -.5 -.25 0 .25 .5 basic IV + health + hh income + marrdur + income + cohabdur + workload + reldur + housework + relsat + pregnant + socsat + sleep + sexsat 14
Lessons Learned IV can help identify total causal effects in the presence of confounders reverse causality mediatiors correlated to the error of the treatment 17
Lessons Learned IV non-IV 18
Lessons Learned IV can help identify total causal effects in the presence of confounders reverse causality mediators correlated to the errors of treatment or oucome In case of doubt: Do not control variables that might be mediators or confounders of the treatment and outcome IV hardly relaxes the assumptions for identifying direct & indirect causal effects (unless the mediator is also instrumented) 19
Lessons Learned IV non-IV 20
THANK YOU 21
Example labour orientation income happy #kids twins family orientation 22
Costs of Children Women Men 1 1 Effect on Happiness Effect on Happiness 0,5 0,5 0 0 -0,5 -0,5 -12 -6 Months before and after birth 0 6 12 18 24 30 36+ -12 -6 Months before and after birth 0 6 12 18 24 30 36+ Note: 4025 Respondents in 14424 Observations Note: 4318 Respondents in 15301 Observations total effect direct effect 23
Discussion of findings According to the IV estimator, more children lead to an increase in happiness for men, but not women. OLS and FE estimates are insignificant for both men and women. Results substantially consistent over models for women, but not for men Possible reasons: OLS/FE: Ommitted variable bias OLS/FE: Over-control IV: Exclusion restriction does not hold Comparability of estimates across different model types? (N)ATE vs. ATT vs. LATE vs. LATT FE Models exploit different sample: those begetting children/twins (i.e. have young children) Result transferable to transition into parenthood? What if effects were only transitory? 24
