Synthetic control in practice
The synthetic control method appears on 29 recent papers, mostly in economics, and its use is falling (4 in the last quarter against about 8 expected), while difference-in-differences is the most-used method on this site. Do published synthetic control studies report what readers need to trust them: pre-period fit, donor pools and placebo tests?
What prompted this#
Since October 2025, "synthetic control method" appears on 29 papers (24 in economics journals, 5 in finance), and "synthetic difference-in-differences" on 2. Only 4 were published in July–September 2026 against about 8 expected. On the equation ranking, difference-in-differences is mentioned far more often, often in settings with a single treated unit where a synthetic control would fit better.
Questions#
- In recent synthetic control studies, how many report the donor pool, the pre-treatment fit, the weights and placebo-based inference?
- How sensitive are their estimates to the choice of donors and of the pre-treatment period, when re-run with the authors' data?
- In difference-in-differences papers with one treated region or country, would a synthetic control give a different answer?
Data that could help#
Replication packages of published synthetic control and single-unit difference-in-differences studies.
Next step#
Code the reporting practices of every synthetic control paper in the archive since 2025 and re-run a sample with alternative donor pools.