Skip to article frontmatterSkip to article content
Site not loading correctly?

This may be due to an incorrect BASE_URL configuration. See the MyST Documentation for reference.

Model Calibration

PolicyEngine

Heterogeneous Calibration Success

The two-type consumer model achieves near-perfect calibration. With observed beer prices of $12.50, the heterogeneous model predicts a profit-maximizing price of $12.51, yielding a calibration error of only $0.01. This represents a 99.5% improvement over the homogeneous model, which predicted an optimal price of $14.59 with error of $2.09. The near-exact match provides strong empirical support for the importance of heterogeneity in consumer preferences, demonstrating this captures a genuine economic mechanism rather than serving as a statistical adjustment.

Objective

Calibrate model so observed prices ($12.50 beer) are approximately profit-maximizing.

Key Challenge

With standard demand models, profit maximization suggests much lower beer prices (\$5-7).

Why? Without internalized costs, selling high volume at low margin dominates selling low volume at high margin.

Solution: Internalized Costs

Stadiums face convex costs from excessive alcohol consumption that affect their own profits:

Cintern(Q)=α(Q1000)2C_{intern}(Q) = \alpha \cdot \left(\frac{Q}{1000}\right)^2

Where α=62.3\alpha = 62.3 (calibrated via config.yaml).

Economic Rationale

These costs are negative externalities that drunk fans impose on OTHER customers:

  1. Experience degradation: Drunk fans hurt experience → lose repeat customers

  2. Brand damage: “Cheap beer stadium” reputation → lower long-run revenue

  3. Crowd management: Security incidents scale non-linearly

  4. Capacity: Service bottlenecks and operational stress

As monopolist, stadium internalizes these because they affect future profits.

Calibration Results

PriceBeers SoldInternalized CostStadium Profit
\$5117,549\$13,814,000-\$7.8M
\$875,253\$5,665,000\$0.3M
\$12.5039,556\$1,563\$2.2M
\$12.8538,021\$1,444\$4.0M (max)
\$1531,801\$1,011\$2.6M

Profit-maximizing consumer price: \$12.85 ≈ \$12.50 observed

Parameter Summary

ParameterValueSource
Capacity46,537Official Yankee Stadium capacity
Base ticket price\$70 (model)Model-predicted optimal; observed avg \$80 varies by seat location
Base beer price\$12.50Industry data (2025)
Ticket elasticity-0.625Noll (1974), Scully (1989)
Beer elasticity-0.965Stadium-adjusted from literature
Beer cost\$5.00All-in (materials + labor + overhead)
Beer excise tax\$0.074Federal + NY + NYC
Sales tax rate8.875%NYC rate
Experience cost (α)250Calibrated to observed prices
Capacity constraint50,000Operational estimate
Price sensitivity (λ)0.133Semi-log calibration

Validation

Heterogeneous model achieves near-perfect match to all empirical targets:

Calibrated parameters (from config.yaml):