scholarly journals Decoupling synthetic control methods to ensure stability, accuracy and meaningfulness

SERIEs ◽  
2021 ◽  
Author(s):  
Daniel Albalate ◽  
Germà Bel ◽  
Ferran A. Mazaira-Font
Keyword(s):  
Optimization Problem ◽  
Control Method ◽  
Economic Effect ◽  
Causal Effects ◽  
Synthetic Control ◽  
Control Methods ◽  
Synthetic Control Method ◽  
Synthetic Control Methods ◽  
Quasi Experimental ◽  
Accuracy And Stability

AbstractThe synthetic control method (SCM) is widely used to evaluate causal effects under quasi-experimental designs. However, SCM suffers from weaknesses that compromise its accuracy, stability and meaningfulness, due to the nested optimization problem of covariate relevance and counterfactual weights. We propose a decoupling of both problems. We evaluate the economic effect of government formation deadlock in Spain-2016 and find that SCM method overestimates the effect by 0.23 pp. Furthermore, we replicate two studies and compare results from standard and decoupled SCM. Decoupled SCM offers higher accuracy and stability, while ensuring the economic meaningfulness of covariates used in building the counterfactual.

scholarly journals Synthetic control method with convex hull restrictions: A Bayesian maximum a posteriori approach

2021 ◽  
Author(s):  
Gyuhyeong Goh ◽  
Jisang Yu
Keyword(s):  
Convex Hull ◽  
Control Method ◽  
Maximum A Posteriori ◽  
Synthetic Control ◽  
Control Methods ◽  
A Posteriori ◽  
Bayesian Approaches ◽  
Synthetic Control Method ◽  
Statistical Inferences ◽  
Synthetic Control Methods

Abstract Synthetic control methods have gained popularity among causal studies with observational data, particularly when estimating the impacts of the interventions implemented to a small number of large units. The synthetic control methods face two major challenges: a) estimating weights for each donor to create a synthetic control and b) providing statistical inferences. To overcome these challenges, we propose a Bayesian framework that implements the synthetic control method with the parallelly shiftable convex hull and provides a Bayesian inference, which is from the duality between a penalized least squares and a Bayesian Maximum A Posteriori (MAP) approaches. Our approach differs from the recent Bayesian approaches, which allow violating the convex hull restriction and face the potential extrapolation bias. Simulation results indicate that the proposed method leads to smaller biases compared to alternatives. We revisit Abadie and Gardeazabal (2003) by applying our proposed method.

scholarly journals The Costs of Economic Nationalism: Evidence from the Brexit Experiment*

2019 ◽  
Vol 129 (623) ◽  
pp. 2722-2744 ◽  
Author(s):  
Benjamin Born ◽  
Gernot J Müller ◽  
Moritz Schularick ◽  
Petr Sedláček
Keyword(s):  
Synthetic Control ◽  
Control Methods ◽  
Economic Nationalism ◽  
Output Loss ◽  
Structural Time Series ◽  
Time Series Estimation ◽  
Series Estimation ◽  
Synthetic Control Methods ◽  
Quasi Experimental ◽  
Unexpected Outcome

Abstract Economic nationalism is on the rise, but at what cost? We study this question using the unexpected outcome of the Brexit referendum vote as a natural macroeconomic experiment. Employing synthetic control methods, we first show that the Brexit vote has caused a UK output loss of 1.7% to 2.5% by year-end 2018. An expectations-augmented VAR suggests that these costs are, to a large extent, driven by a downward revision of growth expectations in response to the vote. Linking quasi-experimental identification to structural time-series estimation allows us not only to quantify the aggregate costs but also to understand the channels through which expected economic disintegration impacts the macroeconomy.

scholarly journals Joint culpability: The effects of medical marijuana laws on crime

2020 ◽  
Author(s):  
Yu-Wei Chu ◽  
W Townsend
Keyword(s):  
Regression Analysis ◽  
Fixed Effects ◽  
Control Method ◽  
National Level ◽  
Medical Marijuana ◽  
Property Crime ◽  
Causal Effects ◽  
Synthetic Control ◽  
Synthetic Control Method ◽  
Medical Marijuana Laws

© 2018 Elsevier B.V. Most U.S. states have passed medical marijuana laws. In this paper, we study the effects of these laws on violent and property crime. We first estimate models that control for city fixed effects and flexible city-specific time trends. To supplement this regression analysis, we use the synthetic control method which can relax the parallel trend assumption and better account for heterogeneous policy effects. Both the regression analysis and the synthetic control method suggest no causal effects of medical marijuana laws on violent or property crime at the national level. We also find no strong effects within individual states, except for in California where the medical marijuana law reduced both violent and property crime by 20%.

scholarly journals Joint culpability: The effects of medical marijuana laws on crime

2020 ◽  
Author(s):  
Yu-Wei Chu ◽  
W Townsend
Keyword(s):  
Regression Analysis ◽  
Fixed Effects ◽  
Control Method ◽  
National Level ◽  
Medical Marijuana ◽  
Property Crime ◽  
Causal Effects ◽  
Synthetic Control ◽  
Synthetic Control Method ◽  
Medical Marijuana Laws

© 2018 Elsevier B.V. Most U.S. states have passed medical marijuana laws. In this paper, we study the effects of these laws on violent and property crime. We first estimate models that control for city fixed effects and flexible city-specific time trends. To supplement this regression analysis, we use the synthetic control method which can relax the parallel trend assumption and better account for heterogeneous policy effects. Both the regression analysis and the synthetic control method suggest no causal effects of medical marijuana laws on violent or property crime at the national level. We also find no strong effects within individual states, except for in California where the medical marijuana law reduced both violent and property crime by 20%.

scholarly journals Synthetic Difference-in-Differences

2021 ◽  
Vol 111 (12) ◽  
pp. 4088-4118
Author(s):  
Dmitry Arkhangelsky ◽  
Susan Athey ◽  
David A. Hirshberg ◽  
Guido W. Imbens ◽  
Stefan Wager
Keyword(s):  
Asymptotic Behavior ◽  
Panel Data ◽  
Time Factors ◽  
Causal Effects ◽  
Synthetic Control ◽  
Control Methods ◽  
Difference In Differences ◽  
Latent Time ◽  
Synthetic Control Methods ◽  
Consistency And Asymptotic Normality

We present a new estimator for causal effects with panel data that builds on insights behind the widely used difference-in-differences and synthetic control methods. Relative to these methods we find, both theoretically and empirically, that this “synthetic difference-in-differences” estimator has desirable robustness properties, and that it performs well in settings where the conventional estimators are commonly used in practice. We study the asymptotic behavior of the estimator when the systematic part of the outcome model includes latent unit factors interacted with latent time factors, and we present conditions for consistency and asymptotic normality. (JEL C23, H25, H71, I18, L66)

Using Synthetic Controls to Evaluate the Effect of Unique Interventions: The Case of Say Yes to Education

2017 ◽  
Vol 41 (6) ◽  
pp. 593-619 ◽  
Author(s):  
Robert Bifulco ◽  
Ross Rubenstein ◽  
Hosung Sohn
Keyword(s):  
New York ◽  
School District ◽  
Graduation Rates ◽  
Control Method ◽  
New York State ◽  
Synthetic Control ◽  
Control Methods ◽  
Public And Private ◽  
Synthetic Control Methods ◽  
Synthetic Controls

Background: “Place-based” scholarships seek to improve student outcomes in urban school districts and promote urban revitalization in economically challenged cities. Say Yes to Education is a unique district-wide school reform effort adopted in Syracuse, NY, in 2008. It includes full-tuition scholarships for public and private universities, coupled with extensive wraparound support services in schools. Objectives: This study uses synthetic control methods to evaluate the effect of Say Yes on district enrollment and graduation rates. It also introduces the synthetic control method and provides guidance for its use in evaluating single-site interventions. Method: Combining school district-level data from the National Center for Education Statistics’ Common Core of Data and New York State School Report Cards, this article uses synthetic control methods to construct a synthetic comparison district to estimate counterfactual enrollment and graduation trends for Syracuse. Results: We find that Say Yes to Education was associated with enrollment increases in the Syracuse City School District, a district that had previously experienced decades of sustained enrollment declines. We do not find consistent evidence of changes in graduation rates following adoption of the program. Conclusions: Graduation rate analyses demonstrate that estimates of treatment effects can be sensitive to choices that the researcher has to make in applying synthetic control methods, particularly when pretreatment outcome measures appear to have considerable amounts of noise.

Sign in / Sign up

Export Citation Format

Share Document