Investment Banks as Corporate Monitors in the Early Twentieth Century United States

Carola Frydman; Eric Hilt (2017) · American Economic Review

Published

October 10, 2026

import sys
from pathlib import Path

ROOT = next(
    p for p in [Path.cwd(), *Path.cwd().parents]
    if (p / "scripts/tranche50_report.py").exists()
)
sys.path.insert(0, str(ROOT / "scripts"))
from tranche50_report import report
report("aeraej-13", tranche="tranche150")

Carola Frydman; Eric Hilt (2017). American Economic Review. Figure 3: high-underwriting railroad group, Tobin’s Q event study.

Collection index · Results JSON · QMD · Reproduction code

Source specification

The source Figure 3 uses the authors’ binary high-underwriting group, not Table 4’s continuous concentration measure. The template averages the post-1920 path and uses its observed minus-one baseline; the source event-study reference is 1905. Assets and surplus controls are retained.

The original event-study design independently reproduces all 26 coefficients on 1,084 observations; maximum coefficient difference 1.8e-13. The original reference year is 1905; the comparison uses 1919.

Source coefficient Template on source sample Source N Template source-sample N
0.012992 -0.036083 1084 1084

The source and comparison coefficients use different specifications or aggregation targets, as stated above; they are not an exact-replication comparison.

Role Column
Outcome q
Treatment policy
Unit RRid
Time year
Cluster RRid

2 retained control columns. lLassets; surplus3

Estimates below divide the outcome by its source-sample standard deviation, 0.16998. Static TWFE is the treatment coefficient. Dynamic TWFE, BJS and Sun–Abraham average supported post-treatment cells by observation count. CS-DID additionally reports an equal-event-time average.

Analysis sample and treatment timing

Observations Units Treated Control units Cohorts Clusters
1081 58 15 43 1 58
Sample restriction Count
additional rows removed 558

Adoption is the first observed treated period. Initially treated units and units without an observed minus-one baseline are excluded. Post-adoption untreated rows are excluded for switching policies. The plot shows at most 100 seeded-sample units; estimation uses the full eligible sample. Empty cells remain missing.

y__d: 58 of 58 units shown (seed 20261008). All eligible units enter estimation. Purple: untreated; yellow: treated; white: unobserved.

Estimates

Estimator Estimate (SD) SE 95% lower 95% upper Target Status
Static TWFE -0.21228 0.20076 -0.6143 0.18974 Coefficient estimated
Dynamic TWFE -0.12406 0.10404 -0.3324 0.08427 Treated cells estimated
BJS / did2s -0.25588 0.19384 -0.64404 0.13227 Treated cells estimated
Sun–Abraham -0.12406 0.10404 -0.3324 0.08427 Treated cells estimated
CS-DID (equal event-time) -0.19564 0.11462 -0.42029 0.029012 Equal post-event times estimated
CS-DID (treated cells) -0.18629 0.11164 -0.4051 0.032513 Treated cells estimated
SDID nan nan nan nan Treated cells unsupported
Matrix completion nan nan nan nan Treated cells unsupported
Horizontal ridge nan nan nan nan Treated cells unsupported
Multisynth (cohort-averaged) nan nan nan nan Treated cells unsupported
Multisynth (individual) nan nan nan nan Treated cells unsupported

Intervals use the configured clusters and the estimator’s available inference. Hollow points have no computed interval; no new refit-bootstrap batch is run.

Estimator Reason
SDID sdid: this adapter does not implement covariate adjustment; controls retained.
Matrix completion matrix-completion: this adapter does not implement covariate adjustment; controls retained.
Horizontal ridge horizontal-ridge: this adapter does not implement covariate adjustment; controls retained.
Multisynth (cohort-averaged) multisynth-cohort: this adapter does not implement covariate adjustment; controls retained.
Multisynth (individual) multisynth-individual: this adapter does not implement covariate adjustment; controls retained.

Event studies

Cohort homogeneity

The imported TestingInEventStudies deviations design tests positive-event-time cohort deviations jointly. This is the homoskedastic nested-model F-test; the estimator confidence intervals above use clustered inference.

**y__d: homogeneity test not identified.** Only one treated cohort remains on the common supported sample.

Fitting settings and checks

Specification Estimator Settings Pooling
y__d CS-DID {“est_method”: “dr”, “covariate_basis”: “cell-specific rank-reduced column space”, “control_group”: “notyettreated”, “xformla”: “y ~ 1 + x0 + x1”, “panel”: true, “effective_panel”: false, “allow_unbalanced_panel”: true, “weights_name”: “weight”, “fix_weights”: null, “base_period”: “universal”, “clustervar”: “cluster”, “aggregate”: “dynamic”, “faster_mode”: true} {}
y__d SDID {} {}
y__d Matrix completion {} {}
y__d Horizontal ridge {} {}
y__d Multisynth (cohort-averaged) {} {}
y__d Multisynth (individual) {} {} 
{
  "did2s_vs_direct_imputation_max_post_difference": 1.1295831396085365e-06,
  "package_aggregate_vs_estimable_cell_weights": null
}