Hello, dear friend, you can consult us at any time if you have any questions, add WeChat: daixieit

This tutorial is a replication of Abadie et al. (2015) [https://onlinelibrary.wiley.com/doi/full/ 10.1111/ajps.12116.]  If you intend to use Synthetic Control for research, I recommend you read this paper. The section on falsifications/ Placebo Studies is excellent (and not just in the context of Synthetic Control).

The background to the study is at the conclusion of WW2 Germany was eventually divided up into two parts:  East and West.1   After years of being governed as separate countries, they were reunified in 1990.

The authors wish to estimate the effect of the 1990 German reunification on per capita GDP in West Germany.  Cunningham (2021, chapter 10) notes ’countries were simply too dissimilar from any one country to make a compelling comparison group, so they used synthetic control to create a composite comparison group based on optimally chosen countries.’

Abadie et al. (2015) describe their methodology as:

We use annual country-level panel data for the period 1960–2003. The German reunifi- cation occurred in 1990, giving a preintervention period of 30 years. Our sample period ends in 2003 because a roughly decade long period after the reunification seems like a reasonable limit on the span of plausible prediction. Recall that the synthetic West Ger- many is constructed as a weighted average of potential control countries in the donor pool. Our donor pool includes a sample of 16 OECD member countries: Australia,Aus- tria, Belgium, Denmark, France, Greece, Italy, Japan, the Netherlands, New Zealand, Norway, Portugal, Spain, Switzerland, the United Kingdom, and the United States.14

We provide a list of all variables used in the analysis in the data appendix, along with data sources. The outcome variable is the real per capita GDP in country j at time t. GDP is Purchasing Power Parity (PPP)-adjusted and measured in 2002 U.S. dollars (USD, hereafter). For the pre-reunification characteristics..., we rely on a standard set of eco- nomic growth predictors: per capita GDP, inflation rate, industry share of value added, investment rate, schooling, and a measure of trade openness.

In their notes to table 2, they say "GDP per capita, inflation rate, trade openness, and industry share are averaged for the 1981–90 period.  Investment rate and schooling are averaged for the 1980–85 period"

In the ’Tutorial week 3: Synthetic control’ folder, there is a text file called ’Germany_example_base.txt’ . Copy the code in that file into R, and see if you can replicate the Abadie et al.   results based

on the above information.  Reference the smoking example [https://colab.research.google.com/ drive/1VVyLxKHsszsVNaaT4OIiZmH2wOq6Lmyh#scrollTo=bPm-3Wvkoh7I] to figure out what you meed to add the code.

In terms of the column names of the cross_cty dataset:

• GDP: Real GDP per captia;

• infrate: inflation rate

• trade: trade openness

• schooling: Percentage of secondary school attained in the total population aged 25 and older.

• industry: industry share of value added

• invest80: Ratio of real domestic investment (private plus public) to real GDP. The data are reported in five-year averages.  So this is the five year average 1976-80 (ignore invest60 and invest70).