Distillers and Chemists: Strategic Capture of the Bottled-in-Bond Act of 1897

Capture without foreclosure: how a consumer-protection law manufactured a premium segment and handed the rents to incumbents

Regulation
Public Choice
Economic History
Working Paper
A public-choice reinterpretation of America’s first quality-certification law. Using a newly digitized state-by-fiscal-year panel (FY1877–1920) built from the Commissioner of Internal Revenue’s annual reports, the paper shows the Act did not foreclose the rival rectifier trade — instead it manufactured a certified premium segment that grew ~24% per year, with 92% of the rents flowing to incumbent Kentucky and Pennsylvania distillers.
Modified

August 3, 2026

Job market paper · sole-authored

Overview

The Bottled-in-Bond Act of 1897 is usually remembered as one of America’s first consumer-protection laws — a federal guarantee, in an age of rampant adulteration, that what the label promised was what the bottle held. This paper reads the Act through public choice instead: as a government-administered solution to a credence-goods problem the straight distillers of Kentucky and Pennsylvania could not solve privately. Their brands and marks were freely counterfeited down a distribution chain they could not police; a Treasury strip stamp, backed by federal criminal law, made their quality signal excludable in a way no private mark ever had.

The reframing yields a sharp, testable distinction. If the Act worked by raising rivals’ costs, certified output should have grown at the expense of the rectifiers (blenders of cheaper neutral spirit) and a scarcity rent should have opened in relative prices. If it worked by certifying quality, a new premium tier should have grown from a near-zero base without the rectifying trade falling. The two mechanisms make opposite predictions — and the data can tell them apart.

The evidence comes from a newly digitized state-by-fiscal-year panel built from the Annual Reports of the Commissioner of Internal Revenue, FY1877–1920 — whiskey deposited into bond, tax-paid withdrawals, spirits rectified, distilleries operated, grain consumed, and (after 1897) withdrawals under the bottled-in-bond stamp — validated against the printed totals through 459 cross-year chain assertions and 59 zero-tolerance national control totals.

Abstract

“The Bottled-in-Bond Act of 1897 allowed whiskey to be sold under a federal strip stamp certifying it as the unblended product of a single distillery and season, aged four years in a government-supervised bonded warehouse. Conventionally read as consumer protection, the Act is reinterpreted here, through public choice, as a government-administered credence-goods certification. Using a newly digitized state-by-fiscal-year panel built from the reports of the Commissioner of Internal Revenue, I estimate a common-treatment-date event study and difference-in-differences and find that the Act did not foreclose the rectifier trade: rectified output grew at least as fast as straight whiskey after 1897, and the straight-minus-rectified gap, whose pre-trends pass on every window, narrows rather than widens. Instead it created a distinct certified premium segment that expanded roughly twenty-four percent per year from a near-zero base, with its rents concentrated overwhelmingly among incumbent Kentucky and Pennsylvania producers. Capture operated not by raising rivals’ costs but through the geographic distribution of a manufactured premium segment.”

A preview of the findings

1. No foreclosure. Indexed to 1896, rectified output reaches 186 by 1905 against straight whiskey’s 181 — the trade the bill’s proponents complained of grew at least as fast as the trade the Act favored. The within-state difference-in-differences on the straight–rectified gap runs the wrong way for foreclosure — the gap narrows by about 23% (−0.258, se 0.148) on the full one-definition FY1887–1908 panel, whose pre-trends pass on every window — and at ninety-five percent one-sided the design rules out any post-Act widening at all.

2. A manufactured premium segment. The genuine expansion is the certified sub-segment itself: bottled-in-bond withdrawals grew from 0.54 million gallons in FY1898 to 16.51 million in FY1917 — a thirty-one-fold increase — at roughly 24% per year relative to rectified output. Within adopting states the elasticity of non-bonded on bonded volume is −0.185 (se 0.073) — an order of magnitude from the −1 of wholesale relabeling, with at most partial cannibalization at the benchmark rate. A hedonic model of hand-collected price quotes attaches a ~10% premium to the bonded label, conditional on age and channel, while the certifiable–rectified price gap does not widen — the price pattern a quality-signaling mechanism predicts and a scarcity-rent mechanism does not.

3. Rents captured by design. Kentucky and Pennsylvania together supplied 92.2% of all bonded withdrawals over 1898–1904, with the Louisville collection district — home market of the Act’s chief proponent, E. H. Taylor Jr. — alone taking nearly a third of the national total. And the concentration was targeting by design: over FY1898–1908, Kentucky’s share of bonded withdrawals equals its share of aged stocks in bond (0.99×) — the statute’s four-year bonding requirement attached the subsidy to the aging inventory the incumbents already held, while Pennsylvania’s 1.38× marks the discretionary margin above the designed transfer. The demand for certified whiskey was actively cultivated by Harvey Wiley’s federal purity chemists, whose authority the distillers enrolled to teach buyers to distrust uncertified spirit. The mechanism is credence-segment creation; the distribution is geographic rent capture.

The upshot: capture need not look like foreclosure. A regulation can expand a market, resolve a real information asymmetry, and still function as capture through who collects the rents it manufactures.

Bonded whiskey withdrawals by state, 1898–1920

Hover any line to read the exact figure — the certification was overwhelmingly a Kentucky (and Pennsylvania) phenomenon.

Selected figures

Line chart of national bottled-in-bond withdrawals rising from 0.54 to 16.51 million gallons, 1898–1917, with Kentucky and Pennsylvania supplying most of the volume

Bottled-in-bond withdrawals and their geographic concentration, FY1898–1920. National bonded withdrawals rise from 0.54 million gallons in FY1898 to 16.51 million in FY1917; Kentucky and Pennsylvania account for the bulk of the volume throughout.

Chart comparing the small but fast-growing bonded sub-segment with the much larger straight and rectified output series

The certified sub-segment against straight and rectified output: a small tier expanding on top of an existing trade, not a redistribution of it. (The bonded share’s 1908 spike is a denominator artifact — that year’s output collapsed while withdrawals draw on accumulated stock.)

Event-study coefficients and indexed output series showing no post-Act divergence in favor of straight whiskey

The quantity horse-race on the FY1887–1908 one-definition panel: the event-study path (left) shows no post-Act widening of the straight–rectified gap, and rectified output (right) grew at least as fast as straight whiskey — the pattern that rejects the foreclosure reading.

Key estimate

Hedonic price model (log price per proof gallon; rectified spirits the omitted category; channel, state, and year fixed effects):

Variable Coefficient Std. error
Bottled-in-bond +0.096* 0.058
Age (per year) +0.074*** 0.009

N = 187; R² = 0.66. *p < 0.10, ***p < 0.01. Stable across real-price and high-confidence-quote specifications.

JEL codes: N41 · L51 · D72 · K23

Keywords: bottled-in-bond, regulatory capture, credence goods, certification, whiskey

Suggested citation: Smith, Jacob R. (2026). “Distillers and Chemists: Strategic Capture of the Bottled-in-Bond Act of 1897.” Working paper, Middle Tennessee State University.

Note

This page is a preview. The full draft — including the common-treatment-date event study, the relabeling test, the legislative and archival record on intent (the committee report, the Aldrich “or owner” amendment, the organized rectifier remonstrances), and the complete robustness battery — is available on request at jrs2ge@mtmail.mtsu.edu.