The SDID estimate of the average effect of Proposition 99 on California per-capita cigarette sales over the 12 post-treatment years is -15.6 packs.headline
Consistent
paper -15.6observed -15.6038
Paper -15.6 packs per capitaObserved -15.6038 packs per capitaAccepted if within ±0.1
Estimate -15.6 -19.6 -27.3 -20.2 -11.1
Table 1, p. 8 · quote verified on p. 8 · equivalence check · observed/published 1 · registered at event 9, before the first recorded result
Method, tolerance rationale and evidence
Metric. SDID point estimate (synthdid_estimate on the California Prop 99 panel)
Why this tolerance. The estimator is a deterministic optimization on the same public data; the paper rounds to 0.1, so one rounding unit allows for rounding plus minor solver differences.
Event 45. Observed -15.6038, delta -0.003828, consistent. Evidence events [42] (value matched by the legacy rule; not bound to an excerpt). Table 1 code extracted verbatim from vignettes/paper-results.Rmd at synthdid commit 2408a924 (R 4.0.5, MCPanel 6b2706fd, glmnet 4.1-1), run once; value printed at full precision by the table1 job.
The placebo-method standard error of the SDID estimate is 8.4.headline
Consistent
paper 8.4observed 8.36799
Paper 8.4 packs per capitaObserved 8.36799 packs per capitaAccepted if within ±0.84
Standard error (8.4) (9.9) (17.7) (11.5) (9.5)
Table 1, p. 8 (placebo method, Section 5, Algorithm 4) · quote verified on p. 8 · equivalence check · observed/published 0.9962 · registered at event 10, before the first recorded result
Method, tolerance rationale and evidence
Metric. sqrt of the placebo variance estimate for SDID
Why this tolerance. Placebo SEs are Monte Carlo estimates from random reassignment; with 200 replications the relative sampling error of an SD estimate is about 5%, so a 10% band (about two sampling SDs) is used.
Event 50. Observed 8.36799, delta -0.03201, consistent. Evidence events [42] (value matched by the legacy rule; not bound to an excerpt). Table 1 code extracted verbatim from vignettes/paper-results.Rmd at synthdid commit 2408a924 (R 4.0.5, MCPanel 6b2706fd, glmnet 4.1-1), run once; value printed at full precision by the table1 job. Placebo standard error with set.seed(12345) before each estimator, 200 replications, as in the vignette.
The synthetic control estimate is -19.6.headline
Consistent
paper -19.6observed -19.6197
Paper -19.6 packs per capitaObserved -19.6197 packs per capitaAccepted if within ±0.1
Estimate -15.6 -19.6 -27.3 -20.2 -11.1
Table 1, p. 8 · quote verified on p. 8 · equivalence check · observed/published 1.001 · registered at event 11, before the first recorded result
Method, tolerance rationale and evidence
Metric. SC point estimate (sc_estimate)
Why this tolerance. Deterministic optimization; one rounding unit.
Event 46. Observed -19.6197, delta -0.01966, consistent. Evidence events [42] (value matched by the legacy rule; not bound to an excerpt). Table 1 code extracted verbatim from vignettes/paper-results.Rmd at synthdid commit 2408a924 (R 4.0.5, MCPanel 6b2706fd, glmnet 4.1-1), run once; value printed at full precision by the table1 job.
The difference-in-differences estimate is -27.3.headline
Consistent
paper -27.3observed -27.3491
Paper -27.3 packs per capitaObserved -27.3491 packs per capitaAccepted if within ±0.1
Estimate -15.6 -19.6 -27.3 -20.2 -11.1
Table 1, p. 8 · quote verified on p. 8 · equivalence check · observed/published 1.002 · registered at event 12, before the first recorded result
Method, tolerance rationale and evidence
Metric. DID point estimate (did_estimate)
Why this tolerance. Closed-form estimator on fixed data; one rounding unit.
Event 47. Observed -27.3491, delta -0.04911, consistent. Evidence events [42] (value matched by the legacy rule; not bound to an excerpt). Table 1 code extracted verbatim from vignettes/paper-results.Rmd at synthdid commit 2408a924 (R 4.0.5, MCPanel 6b2706fd, glmnet 4.1-1), run once; value printed at full precision by the table1 job.
The matrix completion estimate is -20.2.
Consistent
paper -20.2observed -20.2181
Paper -20.2 packs per capitaObserved -20.2181 packs per capitaAccepted if within ±0.5
Estimate -15.6 -19.6 -27.3 -20.2 -11.1
Table 1, p. 8 · quote verified on p. 8 · equivalence check · observed/published 1.001 · registered at event 13, before the first recorded result
Method, tolerance rationale and evidence
Metric. MC point estimate (MCPanel nuclear-norm estimator as used in the authors code)
Why this tolerance. MC chooses its regularization by cross-validation over randomly formed folds, so results can move slightly between runs and package versions.
Event 48. Observed -20.2181, delta -0.01814, consistent. Evidence events [42] (value matched by the legacy rule; not bound to an excerpt). Table 1 code extracted verbatim from vignettes/paper-results.Rmd at synthdid commit 2408a924 (R 4.0.5, MCPanel 6b2706fd, glmnet 4.1-1), run once; value printed at full precision by the table1 job.
The synthetic control with intercept (DIFP) estimate is -11.1.
Consistent
paper -11.1observed -11.1046
Paper -11.1 packs per capitaObserved -11.1046 packs per capitaAccepted if within ±0.1
Estimate -15.6 -19.6 -27.3 -20.2 -11.1
Table 1, p. 8 · quote verified on p. 8 · equivalence check · observed/published 1 · registered at event 14, before the first recorded result
Method, tolerance rationale and evidence
Metric. DIFP point estimate
Why this tolerance. Deterministic optimization; one rounding unit.
Event 49. Observed -11.1046, delta -0.004645, consistent. Evidence events [42] (value matched by the legacy rule; not bound to an excerpt). Table 1 code extracted verbatim from vignettes/paper-results.Rmd at synthdid commit 2408a924 (R 4.0.5, MCPanel 6b2706fd, glmnet 4.1-1), run once; value printed at full precision by the table1 job.
The placebo-method standard error of the SC estimate is 9.9.
Consistent
paper 9.9observed 9.91743
Paper 9.9 packs per capitaObserved 9.91743 packs per capitaAccepted if within ±0.99
Standard error (8.4) (9.9) (17.7) (11.5) (9.5)
Table 1, p. 8 · quote verified on p. 8 · equivalence check · observed/published 1.002 · registered at event 15, before the first recorded result
Method, tolerance rationale and evidence
Metric. sqrt of the placebo variance estimate for SC
Why this tolerance. Monte Carlo placebo estimate; 10% band, as for the SDID standard error.
Event 51. Observed 9.91743, delta +0.01743, consistent. Evidence events [42] (value matched by the legacy rule; not bound to an excerpt). Table 1 code extracted verbatim from vignettes/paper-results.Rmd at synthdid commit 2408a924 (R 4.0.5, MCPanel 6b2706fd, glmnet 4.1-1), run once; value printed at full precision by the table1 job. Placebo standard error with set.seed(12345) before each estimator, 200 replications, as in the vignette.
The placebo-method standard error of the DID estimate is 17.7.
Consistent
paper 17.7observed 17.7403
Paper 17.7 packs per capitaObserved 17.7403 packs per capitaAccepted if within ±1.77
Standard error (8.4) (9.9) (17.7) (11.5) (9.5)
Table 1, p. 8 · quote verified on p. 8 · equivalence check · observed/published 1.002 · registered at event 16, before the first recorded result
Method, tolerance rationale and evidence
Metric. sqrt of the placebo variance estimate for DID
Why this tolerance. Monte Carlo placebo estimate; 10% band, as for the SDID standard error.
Event 52. Observed 17.7403, delta +0.04027, consistent. Evidence events [42] (value matched by the legacy rule; not bound to an excerpt). Table 1 code extracted verbatim from vignettes/paper-results.Rmd at synthdid commit 2408a924 (R 4.0.5, MCPanel 6b2706fd, glmnet 4.1-1), run once; value printed at full precision by the table1 job. Placebo standard error with set.seed(12345) before each estimator, 200 replications, as in the vignette.
The placebo-method standard error of the MC estimate is 11.5.
Consistent
paper 11.5observed 11.5099
Paper 11.5 packs per capitaObserved 11.5099 packs per capitaAccepted if within ±1.15
Standard error (8.4) (9.9) (17.7) (11.5) (9.5)
Table 1, p. 8 · quote verified on p. 8 · equivalence check · observed/published 1.001 · registered at event 17, before the first recorded result
Method, tolerance rationale and evidence
Metric. sqrt of the placebo variance estimate for MC
Why this tolerance. Monte Carlo placebo estimate; 10% band, as for the SDID standard error.
Event 53. Observed 11.5099, delta +0.009864, consistent. Evidence events [42] (value matched by the legacy rule; not bound to an excerpt). Table 1 code extracted verbatim from vignettes/paper-results.Rmd at synthdid commit 2408a924 (R 4.0.5, MCPanel 6b2706fd, glmnet 4.1-1), run once; value printed at full precision by the table1 job. Placebo standard error with set.seed(12345) before each estimator, 200 replications, as in the vignette.
The placebo-method standard error of the DIFP estimate is 9.5.
Consistent
paper 9.5observed 9.45566
Paper 9.5 packs per capitaObserved 9.45566 packs per capitaAccepted if within ±0.95
Standard error (8.4) (9.9) (17.7) (11.5) (9.5)
Table 1, p. 8 · quote verified on p. 8 · equivalence check · observed/published 0.9953 · registered at event 18, before the first recorded result
Method, tolerance rationale and evidence
Metric. sqrt of the placebo variance estimate for DIFP
Why this tolerance. Monte Carlo placebo estimate; 10% band, as for the SDID standard error.
Event 54. Observed 9.45566, delta -0.04434, consistent. Evidence events [42] (value matched by the legacy rule; not bound to an excerpt). Table 1 code extracted verbatim from vignettes/paper-results.Rmd at synthdid commit 2408a924 (R 4.0.5, MCPanel 6b2706fd, glmnet 4.1-1), run once; value printed at full precision by the table1 job. Placebo standard error with set.seed(12345) before each estimator, 200 replications, as in the vignette.