Liquidity Risk Mispricing in Recommendations

Issue: Model Recommends Illiquid Assets; Prices Assumed Frictionless but Actual Bid-Ask Spreads Destroy Returns

Frequency: Common

Symptoms

• Backtest assumes buying at close price
• Live trading must buy at ask, sell at bid
• Bid-ask spread erodes 50-100% of expected alpha
• Position sizing fails in illiquid markets (cannot execute large orders)

Root Cause Backtest data only has closing price, not bid-ask. Model doesn’t account for trading friction. Illiquid assets have high spreads; assumptions of “free” trading cost more than expected return. Model blind to liquidity premium.

Example

Scenario: Emerging market bond fund recommendation
Backtest return (assumed perfect liquidity): 6%
Live trading (with bid-ask, slippage): 2%
Issue: Bid-ask spreads in EM bonds: 0.5-2%
Impact: Model strategy works in theory, loses money in practice

Key Statistics

• Bid-ask spread: 0.1% (liquid stocks) to 2%+ (illiquid assets)
• Impact on annual return: 0.5-2% per round-trip trade

Mitigation Strategies

1. Liquidity Adjustment: Measure and deduct actual trading costs from backtest
2. Liquidity Score: Favor more liquid assets in recommendations
3. Slippage Model: Estimate bid-ask for different order sizes
4. Position Sizing: Reduce position size in illiquid assets

Metrics

• Backtest return (bid-ask adjusted)
• Actual returns achieved in live trading

Alerts

• Live return <80% of backtest → Trading friction issue

References

• Bid-Ask Spreads and Asset Returns
• Liquidity Adjusted Performance