Game Theory Models Reference¶
This page provides a comprehensive reference for all stage games in the NHRA Game Theory toolkit. Each game captures a specific strategic tension in healthcare funding negotiations.
Overview¶
The simulation uses nine stage games that represent key decision points in NHRA negotiations. Each game is a two-player normal-form game solved for Nash equilibria.
Master Index¶
| Game | Row Actions | Column Actions | Strategic Tension |
|---|---|---|---|
| Definition | R (Realism) / E (Efficient) | R / E | What counts as "efficient" |
| Bargaining | A (Agree) / D (Defer) | A / D | CAP convergence negotiation |
| Cost Shifting | I (Invest) / S (Shift) | I / S | Upstream vs downstream investment |
| Discharge Coordination | C (Coordinate) / F (Fragment) | C / F | Hospital discharge planning |
| Governance Integration | I (Integrate) / S (Separate) | I / S | UCC and governance structures |
| Aged Care Interface | C (Coordinate) / F (Fragment) | C / F | ACAT handoff coordination |
| NDIS Interface | C (Coordinate) / F (Fragment) | C / F | NDIS transition planning |
| Coding/Audit | H (Honest) / U (Upcode) | L (Light) / T (Tight) | Activity-based funding compliance |
| Compliance | T (Tight) / L (Light) | T / L | Regulatory intensity choice |
Game Parameters¶
All games receive a GameParams dataclass with the following state variables:
@dataclass
class GameParams:
pressure: float # System pressure index (≥~0.8 typical)
efficiency_gap: float # Divergence from NEP indexation (0..~0.6)
discharge_delay: float # Excess length of stay index (1.0 = baseline)
political_salience: float # Media/political attention (0..1)
audit_pressure: float # Audit intensity (0..1)
cost_shifting_intensity: float # Fee-for-service incentives (0..1)
political_capital: float # Government negotiating capital (0..1)
Funding Negotiation Games¶
Definition Game¶
Strategic tension: What counts as "efficient"?
- Row player: Commonwealth framing / policy narrative
- Column player: State implementation reality
| Actions | Description |
|---|---|
| R (Realism) | Acknowledge cost reality; reduce efficiency gap drift |
| E (Efficient) | Strict efficient-price framing; gap drifts upward |
Payoff drivers:
realism_benefit = 0.5 + 0.8 × efficiency_gap + 0.4 × (pressure - 1.0)realism_cost = 0.25 + 0.35 × political_saliencestrict_benefit = 0.35 + 0.45 × political_saliencestrict_cost = 0.30 + 0.50 × pressure
Equilibrium characteristics: Under high pressure and efficiency gap, (R,R) becomes the welfare-dominant equilibrium. Under low pressure with high political salience, (E,E) may be preferred.
Bargaining Game¶
Strategic tension: Converge on CAP target or defer?
| Actions | Description |
|---|---|
| A (Agree) | Converge toward target nominal share |
| D (Defer) | Defer/escalate; slow movement with conflict costs |
Payoff drivers:
converge_gain = 0.45 + 0.25 × (pressure - 1.0) + 0.20 × political_capitalconflict_cost = 0.55 + 0.90 × pressurenarrative_gain = 0.25 + 0.50 × political_salience
Equilibrium characteristics: High pressure increases conflict costs, making (A,A) more attractive. Political capital enhances agreement payoffs.
Cost Shifting Game¶
Strategic tension: Invest upstream or shift costs downstream?
| Actions | Description |
|---|---|
| I (Invest) | Invest in upstream prevention/primary care |
| S (Shift) | Shift costs to downstream acute care |
Payoff drivers:
coop_gain = 0.55 + 0.45 × (1.0 - efficiency_gap)shift_gain = 0.35 + 0.75 × efficiency_gap + 1.0 × cost_shifting_intensitypr_cost = 0.65 × pressure
Equilibrium characteristics: High cost_shifting_intensity (fee-for-service incentives) shifts equilibria toward (S,S). Pooled funding (low CSI) promotes coordination.
Discharge and Interface Games¶
Discharge Coordination Game¶
Strategic tension: Coordinate discharge planning or fragment care?
| Actions | Description |
|---|---|
| C (Coordinate) | Joint discharge planning; shared responsibility |
| F (Fragment) | Fragmented care; siloed decision-making |
Payoff drivers:
benefit = 0.70 + 0.80 × excess_delay(whereexcess_delay = max(0, discharge_delay - 1.0))cost = 0.30 + 0.10 × (1.0 - min(1.0, excess_delay))pr_penalty = 0.45 × pressure
Equilibrium characteristics: High discharge delay increases coordination benefits. Under pressure, fragmentation becomes increasingly costly.
Governance Integration Game¶
Strategic tension: Integrate governance (e.g., UCCs) or maintain separation?
| Actions | Description |
|---|---|
| I (Integrate) | Unified governance structures |
| S (Separate) | Maintain separate governance |
Payoff drivers:
safety_gain = 0.55 + 0.35 × (pressure - 1.0)integration_cost = 0.20 + 0.35 × political_saliencefragmentation_risk = 0.40 + 0.60 × pressure
Equilibrium characteristics: High pressure increases both safety gains from integration and fragmentation risks, generally favouring integration.
Aged Care Interface Game¶
Strategic tension: Coordinate ACAT handoffs or fragment?
| Actions | Description |
|---|---|
| C (Coordinate) | Coordinated aged care transitions |
| F (Fragment) | Fragmented handoffs |
Payoff drivers:
coord_benefit = 0.6 + 0.4 × (discharge_delay - 1.0)frag_cost = 0.5 × pressure
Equilibrium characteristics: A symmetric coordination game where both players prefer (C,C) when discharge delays are high.
NDIS Interface Game¶
Strategic tension: Coordinate NDIS transitions or fragment?
| Actions | Description |
|---|---|
| C (Coordinate) | Coordinated NDIS transitions |
| F (Fragment) | Fragmented handoffs |
Payoff drivers:
coord_benefit = 0.5 + 0.5 × (discharge_delay - 1.0)frag_cost = 0.6 × pressure
Equilibrium characteristics: Similar structure to Aged Care, but with slightly higher fragmentation costs under pressure.
Compliance and Audit Games¶
Coding/Audit Game¶
Strategic tension: Provider coding honesty vs auditor vigilance
- Row player: Healthcare provider
- Column player: Auditor/regulator
| Row Actions | Description |
|---|---|
| H (Honest) | Accurate activity coding |
| U (Upcode) | Inflate activity codes for higher reimbursement |
| Column Actions | Description |
|---|---|
| L (Light) | Minimal audit scrutiny |
| T (Tight) | Intensive audit regime |
Payoff drivers:
upcode_gain = 0.3 + 0.7 × efficiency_gappenalty = 0.8 × audit_pressureaudit_cost = 0.2recovery = 0.4 × efficiency_gap
Equilibrium characteristics: Mixed equilibrium common. High audit pressure deters upcoding; high efficiency gap increases temptation.
Compliance Game¶
Strategic tension: Tight vs light regulatory compliance
| Actions | Description |
|---|---|
| T (Tight) | Stringent compliance; higher admin cost, less leakage |
| L (Light) | Relaxed compliance; lower cost, more leakage |
Payoff drivers:
leakage = 0.40 + 0.70 × efficiency_gapadmin = 0.18 + 0.45 × audit_pressure
Equilibrium characteristics: Trade-off between administrative burden and revenue protection. High audit pressure makes tight compliance more attractive.
Nash Equilibrium Solver¶
All games return a TwoPlayerGame object that supports:
from nhra_gt.subgames.nash import solve_all_equilibria
game = definition_game(GameParams(pressure=1.0, efficiency_gap=0.3, ...))
equilibria = solve_all_equilibria(game)
for eq in equilibria:
print(f"Row: {eq.row_strategy}, Col: {eq.col_strategy}")
print(f"Expected payoffs: {eq.row_payoff:.3f}, {eq.col_payoff:.3f}")
The solver finds all pure and mixed Nash equilibria using support enumeration.
References¶
- Source code:
src/nhra_gt/subgames/games.py - Nash solver:
src/nhra_gt/subgames/nash.py - Agent integration:
src/nhra_gt/agents/base.py