The Other Box Score THE OTHER Box Score
Part Three -- The Record That Wasn't Kept

The Salary Ledger.

The wage gap between Negro Leagues and Major League players was not a market outcome. It was a transfer of wealth. This chapter calculates the size of that transfer.

Chapter 08 / 15
1920 -- 1948
25 Player Ledgers
5 ML Models

Have you heard what Satchel Paige was paid?

In the archives at Notre Dame, in a financial ledger book that documents the Birmingham Black Barons from 1926 through 1930, there is an entry for a nineteen-year-old rookie pitcher named Leroy Paige. His salary in 1928 was eighty dollars per month.

That same year, the average Major League pitcher earned approximately $7,200 for the season. Across a five-month Negro Leagues season, Paige earned $400. Across a six-month MLB season, his white counterpart at the same position earned eighteen times more. This was not because the white pitcher was better. By every contemporary and modern measure, the white pitcher was not better.

The ledger is real. The math is simple. The comparison has been possible since the day the entries were recorded. This chapter makes the comparison for every documented Negro Leagues player, every documented season, and calculates what the gap cost in 2024 dollars. The total is not a metaphor. It is a number.

SALARY RATIO -- 1928
18x
MLB vs Negro Leagues
$400
Paige -- 5 months
$7,200
MLB avg -- 6 months
Source: Birmingham Black Barons Ledger Book, 1926--1930. Joyce Sports Research Collection, Hesburgh Libraries, University of Notre Dame. Digitized via Marble.
FIG 01

The Single Ledger Entry

The same five months. The same position. Two formats. The column on the right has always been calculable. The data existed. Nobody made the comparison.

Birmingham Black Barons
Financial Ledger -- 1928
Player: Leroy "Satchel" Paige
Position: Pitcher
Date Credit Debit Balance
Apr 1928 $80.00 $80.00
May 1928 $80.00 $160.00
Jun 1928 $80.00 $7.50 $232.50
(salary)    (shoes)
Jul 1928 $80.00 $312.50
Aug 1928 $80.00 $5.00 $387.50
(salary)    (fine)
SEASON TOTAL: $400.00
Ledger Comparison -- 1928 Season
Player: Leroy "Satchel" Paige
Month NL Salary MLB Equiv Ratio
April $80.00 $600.00 7.5x
May $80.00 $600.00 7.5x
June $80.00 $600.00 7.5x
July $80.00 $600.00 7.5x
August $80.00 $600.00 7.5x
Season $400.00 $3,000.00 7.5x
Full Year $400.00 $7,200.00 18x
2024 $ (CPI) $7,312 $131,616 18x

The data existed. Nobody made the comparison.

FIG 02

The Aggregate Curve

Annual Negro Leagues payroll, MLB counterfactual payroll for the same players, and the gap between them. The shaded area is the documented wage theft, 1920 through 1948.

NL Payroll
MLB Counterfactual
Wage Gap
$3.7 billion
Total documented wage gap, 1920--1948, in 2024 dollars
Modeled: pending Bayesian imputation pipeline. Placeholder figures derived from documented salary ranges and MLB comparison data. Final figures will carry 90% credible intervals.
$2.1B
CPI
$3.7B
Wage Index
$8.4B
GDP Share
$13.9B
Compounded Estate
THE ESTATE

The Compounded Transfer

The wage gap was not a one-time event. Every dollar stolen continued to grow. If those wages had been paid and invested in the S&P 500, the compounded estate value for the top 50 Negro Leagues players would be worth nearly $14 billion in 2024. Not salary. Not inflation adjustment. Compounded wealth, denied across generations.

Compounded Estate Value, Top 50 NLB Players, 2024
$13.9 billion
90% credible interval: $9.0B to $20.1B
Modeled: S&P 500 total returns (Shiller series), year-by-year compounding from season of theft to 2024
CPI Wage Gap
$82.6M
Estate Multiplier
168x
Per Player Average
$277M
# Player Pos WAR CPI Gap Estate (2024$) Multiplier

Confidence: Modeled. Estate values compound the Bayesian wage gap estimate at historical S&P 500 total returns (Shiller series). The 90% credible interval propagates from the underlying wage gap model. Top 50 documented NLB players only.

FIG 03

The Individual Ledger

Career wage gap for every documented Negro Leagues player with sufficient service time. The top five are names you know. The next twenty are not. That is part of the argument.

# Player Pos WAR NL Earned (2024$) MLB Counter (2024$) Gap (2024$) Estate (2024$) Conf
FIG 04

The Team Ledger

Cumulative wage gap by franchise. The bars move the argument from individual to institutional. Every roster, every season, systematically underpaid.

FIG 05

The Sensitivity Dashboard

Every assumption is visible. Adjust any one and watch the headline figure move. The chapter does not hide behind a single number.

$3.7B
Central estimate -- adjusts with selections below
Modeled: pending Bayesian imputation pipeline
Inflator Method
MLB Comparison Floor
NL Season Length
Barnstorming Income
Compounding Rate (Estate Value)
Variance Decomposition -- What Drives the Number
38%
27%
18%
10%
7%
Inflator (38%)
MLB Floor (27%)
Compounding (18%)
Season Length (10%)
Barnstorming (7%)
THE 1924 WORLD SERIES

Same Championship. Same Year.

In 1924, the Kansas City Monarchs won the first Negro Leagues World Series. The Washington Senators won the MLB World Series. Both were champions. The winners' shares were not comparable.

Method

Five models, each documented. Every assumption labeled at the point of claim. The methodology is the chapter's contribution. The number alone is contestable. The methodology with the number behind it is not.

Model 01
Salary Imputation
Hierarchical Bayesian model with priors from Lester, Britannica, and the Black Barons ledger. Produces posterior salary distributions per player-season, conditional on team, year, role, position, and league. Every imputed salary carries a 90% credible interval.
Model 02
MLB Counterfactual
Gradient-boosted regression trained on documented MLB salaries, 1920 through 1948. Features: prior-season WAR, position, age, team payroll quartile, year and league fixed effects. Applied to Negro Leagues player-seasons using Seamheads WAR. Extends the Hayes IBW21 framework from league average to production-specific prediction.
Model 03
Wage Gap Calculation
Gap per player-season equals predicted MLB salary minus imputed NL salary. Inflated three ways (CPI, wage index, GDP share). Aggregated to player career, franchise, and league total. Monte Carlo simulation over the joint posterior produces career-level credible intervals.
Model 04
Compounded Estate Value
Year-by-year wage gap compounded at historical S&P 500 total return (Shiller monthly series) from year of theft to 2024. Produces the intergenerational wealth transfer figure. Alternative compounding rates (T-Bills, savings, none) documented in the sensitivity dashboard.
Model 05
Sensitivity Analysis
Sobol variance decomposition across five assumption dimensions. Pre-computed lookup table of headline figures for every combination. The reader can adjust assumptions and see the number move. Every methodological choice is visible and testable.
Confidence Vocabulary
Documented
Verified
Reported
Estimated
Modeled
Reconstructed
Disputed
AI-generated
Primary Data Sources
  • Birmingham Black Barons Ledger Book, 1926--1930. Joyce Sports Research Collection, Hesburgh Libraries, University of Notre Dame. Digitized via Marble.
  • Larry Lester / SABR Negro Leagues Research Committee. Documented salary research including the $75 / $175 / $375 rookie / journeyman / star ranges for the 1920s.
  • SABR Business of Baseball Research Committee historical salary data. MLB salary leaders since 1874, average and median MLB salaries.
  • Seamheads Negro Leagues Database. Gary Ashwill and Kevin Johnson, lead researchers. Per-season WAR and production data.
  • Bureau of Labor Statistics CPI historical series. Measuring Worth historical wage index. BEA GDP deflator.
  • Robert Shiller, S&P 500 monthly total return series, 1871 through 2024. Used for compounding calculations.
  • Robert Peterson, Only the Ball Was White (1970). Foundational secondary source for scattered salary references.

The WAR-plus-average-salary reparations framework was first publicly proposed by James M. Hayes in "The Rightness of Negro League Player Reparations," Institute of the Black World 21st Century, June 2020. This chapter operationalizes that framework at full scale with hierarchical Bayesian imputation, production-specific MLB counterfactuals, and confidence intervals. The debt to Hayes is foundational.

Coda

The ledger is a document. The gap is a calculation. The transfer is a fact. What you do with the number is a choice, but the number itself is no longer missing. It is here, documented, sourced, and reproducible from the data this chapter publishes under CC0.

The wages are one accounting. There is another.

The wages are one accounting. There is another. If every team had integrated in 1947, the gap would still be vast. But every team did not integrate in 1947. The last team waited twelve more years. The next chapter asks why, and which teams, and what the delay itself cost.

Next chapter -- 09
The Last Team.
The Red Sox integrated July 21, 1959. Twelve years after Robinson. Every team. Every player passed over. Every year of delay accounted for.
Cite This Chapter
Haynes, Jeremy. "The Salary Ledger." The Other Box Score, theotherboxscore.org/chapters/the-salary-ledger/, 2026. Accessed [access date].
Chicago: Haynes, Jeremy. "The Salary Ledger." The Other Box Score. 2026. https://theotherboxscore.org/chapters/the-salary-ledger/.
Data (CC0): The Other Box Score. "Negro Leagues Salary Ledger Dataset." CC0 1.0. https://github.com/other-boxscore/chapters/08-the-salary-ledger/data/.
Primary salary source: Birmingham Black Barons Ledger Book, 1926--1930. Joyce Sports Research Collection, Hesburgh Libraries, University of Notre Dame. Digitized via Marble. https://marble.nd.edu/
Foundational reparations argument: Hayes, James M. "The Rightness of Negro League Player Reparations." Institute of the Black World 21st Century, June 2020. https://ibw21.org/editors-choice/the-rightness-of-negro-league-player-reparations/