Transaction Cost Analysis: Why Your Backtest Lies About What You Will Pay
Transaction costs are the gap between backtest performance and live performance that most retail algorithmic traders systematically underestimate. The commissions you see on your broker's fee schedule are only the beginning. The real costs — spread, slippage, market impact — can easily double or triple your assumed transaction costs. Understanding and modeling these accurately is one of the most important steps in realistic backtesting.
The Components of Transaction Cost
Explicit commissions. The fees you pay directly to your broker. In the "commission-free" era for US equities, these are often zero for stocks but still significant for options ($0.65-$1.00 per contract typically), futures, and most international markets.
Exchange and regulatory fees. SEC fees (for sells), FINRA Trading Activity Fees, options exchange fees. Individually small but accumulate. Typically a few cents per share or per contract.
Bid-ask spread. The difference between the highest buyer price and lowest seller price. For liquid stocks like SPY, this is a penny or less. For less liquid stocks or options, this can be 0.1-5% of position value per round trip.
Slippage. The difference between the expected execution price (typically the mid-market quote) and the actual execution price. Depends on order type, timing, and position size relative to market liquidity.
Market impact. The amount your order moves the market against you. Negligible for small retail positions in liquid names; substantial for larger positions or less liquid instruments.
Opportunity cost of waiting. For orders that don't fill immediately (limit orders), the market may move before your order fills. The "price" of this is harder to quantify but real.
Why Backtests Typically Underestimate
Most backtesting frameworks apply transaction costs as a simple percentage or dollar amount per trade. This handles explicit commissions but typically fails to capture:
- Bid-ask spreads that vary by instrument and time
- Slippage that increases with position size
- Market impact that compounds with order volume
- Opportunity cost of partial fills or missed fills
For high-frequency strategies, these hidden costs can be larger than the explicit commissions. For low-frequency strategies on liquid instruments, they may be modest. The specific impact depends heavily on strategy characteristics.
Realistic Cost Assumptions
For retail-scale strategies in 2026:
Liquid US equities (SPY, QQQ, major names): $0 commission, 0.5-1 cent per share slippage, 1-2 cent per share spread. For a $400 SPY position, round-trip cost is approximately 2-4 cents per share, or 0.005-0.01% of position value.
Less liquid US equities (small cap, less-traded names): $0 commission, 1-5 cent per share slippage, 2-10 cent per share spread. Round-trip cost can be 0.05-0.5% of position value.
US equity options: $0.65 per contract commission, slippage of 1-5 cents per contract on liquid options (SPY, QQQ, major names), 5-25 cents per contract on less liquid options. Spreads can be substantial — 5-25% of option price on less liquid names.
Futures: commission varies by broker and contract, typically $0.50-$2.50 per contract per side. Spreads on liquid futures (ES, NQ, CL) are tight; less liquid futures can have wider spreads.
International markets: generally wider spreads and higher commissions than US markets. Add 50-100% to typical US assumptions.
Cryptocurrency: varies dramatically by exchange and pair. Spreads on major pairs (BTC/USD, ETH/USD) on major exchanges are tight. Smaller altcoins can have very wide spreads.
The Specific Impact on Different Strategies
Low-frequency strategies (monthly or less rebalancing): transaction costs typically modest. 1-5% annual drag on returns from costs is common.
Medium-frequency strategies (weekly rebalancing): costs become meaningful. 5-15% annual drag common.
High-frequency strategies (daily or more): costs often dominate. Can easily be 20-50% of gross returns.
Options strategies: more expensive than equity strategies due to wider spreads and per-contract commissions. Even monthly rebalancing of options positions can incur substantial costs.
The practical implication: strategies that look profitable in backtests with low cost assumptions may be unprofitable in live trading. Testing with realistic (even conservative) cost assumptions reveals which strategies have true edge.
Testing Methodology
For strategies you develop:
- Model commissions explicitly. Use your actual broker's fee schedule, not assumed zero.
- Add bid-ask spread cost. For each trade, assume you buy at ask and sell at bid, not mid-market.
- Add slippage estimate. Model additional cost based on strategy characteristics. For retail-size liquid-stock strategies, 0.1% per round trip is a reasonable starting assumption.
- Vary cost assumptions. Test the strategy with 2x your base cost assumption. If it is still profitable, the strategy likely has robust edge. If not, the strategy depends on low costs that may not be achievable in live trading.
- Monitor live transaction costs. Compare actual live costs to your backtest assumptions. Adjust future backtests based on observed reality.
The Order Type Decision
Order type affects actual transaction costs substantially:
Market orders. Fast execution, but you take whatever price is available. Costs you the spread plus any slippage from market impact.
Limit orders at mid-market. Attempts to capture the spread. May not fill, or may fill only partially.
Limit orders at bid (buy) or ask (sell). Pays the spread but typically fills immediately.
Stop-loss orders. Trigger market orders at the stop price. Can experience significant slippage during fast market moves.
For most retail algorithmic strategies, limit orders at the bid/ask for immediate fill with explicit spread cost is simpler to model and typically acceptable. More sophisticated execution requires more sophisticated modeling.
The Realistic Framework
The transaction cost assumptions in most backtests are too optimistic. The practical approach:
Conservative baseline: assume commissions per your broker's schedule, plus bid-ask spread (buy at ask, sell at bid), plus 0.1% additional slippage per round trip for liquid US equities. More for options, less liquid instruments, international markets.
Stress test: run the strategy with 2x your baseline cost assumption. Verify it is still profitable.
Live validation: compare actual costs to assumptions over the first few months of live trading. Update cost models based on reality.
Strategies that survive stress testing with 2x conservative costs are substantially more likely to be profitable in live trading than strategies that only work with optimistic cost assumptions. The gap between backtest profitability and live profitability is often entirely in transaction costs.
The traders who model transaction costs realistically have substantially better live trading outcomes than those who use default assumptions. The additional work is modest; the insight gained is substantial.
