Figure 5 · Daily synthetic-control gap

The announcement day looked like any other day

Daily difference between ExxonMobil’s return and the return of its synthetic peer benchmark, across 20 trading days before the announcement through 5 days after. Day-0 (red) is well inside the normal pre-event range (shaded).

+σ −σ T0 Announcement day: −0.03% +6% +3% 0 −3% T−20 T−15 T−10 T−5 T0 T+5 Trading days relative to the announcement
−0.03%
Announcement-day gap (essentially zero)
±0.71%
Normal day-to-day variation (1σ band)
11
Pre-period days with a larger absolute gap
Sources & methodology notes
Daily synthetic-control gap (treated minus synthetic counterfactual)gapt = RExxonMobil,t − Σj ω̂j Rj,t
σpre = sd(gapt),   t ∈ pre-period (250 trading days)
Shaded band shows the ±1σpre range; 95% credible interval is ±1.96σpre.

This figure shows the daily synthetic-control gap (treated minus synthetic counterfactual) over the [−20, +5] window, computed from the S&P Capital IQ daily-adjusted-close panel using the 10-firm pre-registered frozen donor weights. The shaded band is the pre-period 1σ range, where σpre is estimated from the 211-day pre-event distribution of daily gaps. Sources below mirror the methodology already cited in Figure 3 (donor pool), Figure 4 (Bayesian posterior), and Figure 5 (three-benchmark) of this site.

  1. Alberto Abadie & Javier Gardeazabal, The Economic Costs of Conflict: A Case Study of the Basque Country, 93 Am. Econ. Rev. 113 (2003). Original synthetic-control framework. Establishes the weighted-donor counterfactual whose daily realization is plotted here against the treated unit.
  2. Alberto Abadie, Alexis Diamond & Jens Hainmueller, Synthetic Control Methods for Comparative Case Studies, 105 J. Am. Stat. Ass’n 493 (2010). Formal statistical-inference framework. The pre-period RMSPE / 1σ band visualized as the shaded region is the inference instrument from ADH’s §3 placebo-permutation procedure. doi.org/10.1198/jasa.2009.ap08746.
  3. Alberto Abadie, Using Synthetic Controls: Feasibility, Data Requirements, and Methodological Aspects, 59 J. Econ. Literature 391 (2021). Modern review; the post/pre RMSPE ratio threshold of 2.0 (the canonical “treatment detected” threshold) is from §6.2 here. Observed ratio = 0.88, well below threshold.
  4. Stephen J. Brown & Jerold B. Warner, Using Daily Stock Returns: The Case of Event Studies, 14 J. Fin. Econ. 3 (1985). Foundational empirical-event-study paper; daily-returns research design within which this gap series is taken.
  5. Shane Goodwin, Read the Fine Print: What ExxonMobil’s Proxy Actually Says About Texas Redomiciliation, Columbia Law School Blue Sky Blog (May 2026); replication kit on file with the SMU Corporate Governance Initiative. Companion paper. The article’s headline Day-0 synthetic-control gap of +0.021% (nested-ADH variant on the same CIQ panel) is within 5 basis points of the −0.026% frozen-weights value plotted here; both are inside the published ±0.5pp reproducibility tolerance. Day-0 chart values: ExxonMobil raw return −1.54%; synthetic return −1.51%; gap −0.026%.

Data attribution. Daily adjusted closing prices from S&P Capital IQ (IQ_CLOSEPRICE_ADJ feed) for 13 series (ExxonMobil, the 10 donor firms, the S&P 500 ETF benchmark, and the Brent crude oil benchmark) over 261 trading days 2025-05-16 to 2026-05-15. Frozen donor weights from Goodwin (May 2026) fn. 27, renormalized to the unit simplex per Abadie-Diamond-Hainmueller (2010) §2.2: Chevron 43.1%, Diamondback 15.2%, EOG Resources 13.9%, Schlumberger 8.1%, Williams 5.8%, Baker Hughes 4.6%, Occidental 3.5%, ConocoPhillips 2.3%, Marathon (MRO pre-registered; MPC substituted post-delisting) 1.8%, Phillips 66 1.8%. Pre-renormalization values (sum 0.968) preserved in donor_weights.json weight_pre_renorm field.

Source: Author’s calculations from S&P Capital IQ daily adjusted closing prices; 10-firm pre-registered frozen donor weights per Goodwin (May 2026) fn. 27.