Do Institutions Cause Prosperity? The AJR (2001) IV Datasets

Downloads

Each dataset is available as a labeled Stata .dta and its source file.

⇩ Download all data (ZIP)stata_codebook.do

Run stata_codebook.do in Stata once to attach long-form per-variable notes to the .dta files.

Load directly in code

Every file loads straight from GitHub (raw URLs). Swap the file name to load any dataset.

Stata

* Stata 14+ : `use` reads an https URL directly
global BASE "https://raw.githubusercontent.com/cmg777/starter-academic-v501/master/content/post/stata_iv/data/"
use "${BASE}maketable1.dta", clear
describe
notes

Python

!pip install -q pyreadstat
import pandas as pd
BASE = "https://raw.githubusercontent.com/cmg777/starter-academic-v501/master/content/post/stata_iv/data/"
df = pd.read_stata(BASE + "maketable1.dta")

# load every dataset at once
files = ["maketable1", "maketable2", "maketable3", "maketable4", "maketable5", "maketable6", "maketable7", "maketable8"]
data = {f: pd.read_stata(BASE + f + ".dta") for f in files}

# pyreadstat (richest metadata) reads LOCAL files -> download first
import pyreadstat, urllib.request
urllib.request.urlretrieve(BASE + "maketable1.dta", "maketable1.dta")
df, meta = pyreadstat.read_dta("maketable1.dta")

Copy and paste this snippet in Google Colab app. https://colab.research.google.com/notebooks/empty.ipynb

R

# R : haven::read_dta auto-downloads an https URL
library(haven)
BASE <- "https://raw.githubusercontent.com/cmg777/starter-academic-v501/master/content/post/stata_iv/data/"
df <- read_dta(paste0(BASE, "maketable1.dta"))

Overview & sources

Companion data for a hands-on Stata tutorial that replicates Acemoglu, Johnson and Robinson (2001), The Colonial Origins of Comparative Development. The study instruments modern property-rights institutions (avexpr) with European settler mortality during colonization (logem4) to recover the causal effect of institutions on income (logpgp95) across ex-colonies. The naive OLS slope is 0.522; the two-stage least-squares (2SLS) estimate is 0.944 — about 81% larger — implying that measurement error dominates OLS bias and that institutional reform is roughly twice as valuable as naive regressions suggest. The eight datasets here are AJR's original replication archive: one per table of the paper, each a country cross-section that progressively narrows from the full ~163-country world to the 64-country base sample of ex-colonies with valid settler-mortality data.

Data sources

Cite this data

Please cite this dataset as follows.

APA

Mendez, C. (2026). Do Institutions Cause Prosperity? An IV Tutorial in Stata (AJR 2001 replication data) [Data set]. https://carlos-mendez.org/post/stata_iv/

Acemoglu, D., Johnson, S., & Robinson, J. A. (2001). The Colonial Origins of Comparative Development: An Empirical Investigation. American Economic Review, 91(5), 1369–1401. Albouy, D. Y. (2012). The Colonial Origins of Comparative Development: An Investigation of the Settler Mortality Data. American Economic Review, 102(6), 3059–3076. Imbens, G. W., & Angrist, J. D. (1994). Identification and Estimation of Local Average Treatment Effects. Econometrica, 62(2), 467–475.

BibTeX

@misc{mendez2026stataiv,
  author       = {Mendez, Carlos},
  title        = {Do Institutions Cause Prosperity? An IV Tutorial in Stata (AJR 2001 replication data)},
  year         = {2026},
  howpublished = {\url{https://carlos-mendez.org/post/stata_iv/}},
  note         = {Data set}
}

@article{ajr2001colonial,
  author  = {Acemoglu, Daron and Johnson, Simon and Robinson, James A.},
  title   = {The Colonial Origins of Comparative Development: An Empirical Investigation},
  journal = {American Economic Review},
  volume  = {91}, number = {5}, pages = {1369--1401}, year = {2001}
}
@article{albouy2012colonial,
  author  = {Albouy, David Y.},
  title   = {The Colonial Origins of Comparative Development: An Investigation of the Settler Mortality Data},
  journal = {American Economic Review},
  volume  = {102}, number = {6}, pages = {3059--3076}, year = {2012}
}
@article{imbens1994late,
  author  = {Imbens, Guido W. and Angrist, Joshua D.},
  title   = {Identification and Estimation of Local Average Treatment Effects},
  journal = {Econometrica},
  volume  = {62}, number = {2}, pages = {467--475}, year = {1994}
}

Variable explorer search & filter all 55 variables

Type to filter by name or label, or use the chips to filter by type. Each row shows a mini distribution. Click a header to sort.

Cross-file variable index

Which file each variable appears in (● = present).

Construction & formulas

The tutorial estimates the causal effect of institutions on income via two-stage least squares (2SLS). The structural model has an endogenous regressor X = avexpr correlated with the error, an outcome Y = logpgp95, and an instrument Z = logem4:

• Structural equation: Y_i = α + β·X_i + U_i, with Cov(X_i, U_i) ≠ 0 (this endogeneity is what biases OLS).
• First stage (relevance): X_i = π_0 + π_1·Z_i + v_i — settler mortality must move institutions (here π_1 = −0.607, Kleibergen–Paap F = 16.32).
• Reduced form: Y_i = γ_0 + γ_1·Z_i + e_i — the total effect of the instrument on the outcome (here γ_1 ≈ −0.573).
• 2SLS estimator (just-identified): β̂_2SLS = Cov(Y,Z) / Cov(X,Z) = γ̂_RF / π̂_FS = −0.573 / −0.607 = 0.944.
• Exclusion restriction: Z affects Y only through X — settler mortality circa 1700 influences 1995 GDP only by shaping inherited institutions. This is untestable in a just-identified model.
• Overidentification (Hansen J): with more instruments than endogenous regressors, the J-test checks whether the instruments agree on one causal effect; non-rejection is consistent with (but does not prove) joint exogeneity.

The 2SLS estimate identifies a Local Average Treatment Effect (LATE) for "complier" countries whose institutions would respond to a change in settler mortality (Imbens & Angrist 1994); under constant effects, LATE = ATE.

The datasets

Switch datasets with the tabs. Each shows the full variable dictionary plus a sortable statistics table with mini distributions and data coverage.

Expand all datasets expand to search (Ctrl/⌘+F) or print across all datasets

Known limitations & caveats

• Cross-section, not panel. Every file is a country cross-section keyed on shortnam; there is no time dimension. Most variables are measured at a single point (e.g. GDP in 1995, religion shares in 1980).
• baseco coding. The base-sample flag baseco takes the value 1 for the 64 base-sample ex-colonies and is missing (not 0) for other countries. Restrict with keep if baseco==1.
• Imputed mortality (Albouy 2012). Roughly 36% of AJR's settler-mortality observations are imputed or shared across countries, which weakens the exclusion restriction and gives Hansen J low power against shared imputation bias.
• Weak instruments in robustness specs. Several specifications (e.g. Table 6 resource/geography columns, Table 7 health-channel overidentified columns) drop the first-stage F below 5 — read their confidence intervals via the weak-IV-robust Anderson–Rubin test, not their point estimates.
• LATE, not ATE. The headline 0.944 is the effect for colonization-margin compliers, not a population-average treatment effect.
• Missing values. Coverage varies by variable (institutions, mortality, and many controls are missing for a subset of countries); the labeled .dta preserves AJR's original missing-data pattern byte-for-byte.