For a taxable account, every realised gain triggers a tax. Short-term gains (held < 1 year in the US) are taxed at ordinary income rates; long-term gains (held ≥ 1 year) at preferential rates. A strategy with monthly rebalancing realises almost all gains as short-term — substantially reducing the after-tax CAGR. The Tax Drag sub-pill computes the expected drag using a configurable ST/LT rate split.
The decomposition
For each rebalance period, FM103 separates the realised gains into:
- Short-term: gains on positions held less than 1 year
- Long-term: gains on positions held 1 year or more
For a quarterly-rebalanced strategy, virtually all gains are short-term unless a position carries through several periods. The default rate assumption: 37% ST, 20% LT (US federal top brackets; ignores state tax for portability).
The drag formula
Drag as a percentage of starting capital:
And after-tax CAGR:
How big can tax drag be?
| Strategy | Typical ST/LT split | Tax drag on gross CAGR |
|---|---|---|
| Buy-and-hold annual rebalance | 10% ST / 90% LT | ~0.5–1.0% |
| Quarterly factor rebalance | 75% ST / 25% LT | ~2–4% |
| Monthly active strategy | 95% ST / 5% LT | ~3–5% |
| Weekly momentum | 100% ST / 0% LT | ~4–6% |
Numbers assume a base gross CAGR of ~10–15%. The pattern: every step up in rebalance frequency adds roughly 1% to annual tax drag.
Wash-sale, harvesting, and other adjustments
FM103's tax model is a simplification. Real-world taxable-account considerations the model does NOT capture:
- Wash-sale rule. A loss realised within 30 days of repurchasing the same security is disallowed. Factor strategies with mechanical rebalancing routinely trigger this; the platform doesn't track it.
- Tax-loss harvesting. Discretionary realisation of losses to offset gains. Done well, this reduces drag by 1–2% annually (Berkin & Ye 2003).
- Specific lot identification. Selling the highest-cost-basis lot first minimises realised gain. Default behavior on most brokers is FIFO; the platform assumes FIFO.
- State-level taxation. Varies; the model ignores state to keep results portable.
When tax drag doesn't apply
- Tax-deferred accounts (IRA, 401k): No annual tax on realised gains; drag is zero. The sub-pill should be ignored for these.
- Tax-exempt accounts: Same.
- Institutional accounts: Pass-through entities may avoid entity-level tax but face investor-level tax.
Reading the verdict
FM103's default warning: "Tax drag > 30% of gross return." Translation: more than 30 cents of every dollar of gross excess return goes to tax. At that level, the strategy is taxable-account-impaired — either restructure to favour LT realisations (longer holding periods), or deploy in a tax-advantaged account.
Strategies that minimise tax drag
- Annual rebalance instead of quarterly.
- Use ETFs as the rebalanced layer instead of individual stocks (ETF tax efficiency via in-kind redemption).
- Hold the strategy in an IRA or 401k.
- Build a tax-loss-harvesting overlay (outside platform scope).
Further Reading
Textbook references
- Siegel, J. J. (2014). Stocks for the Long Run (5th ed.). McGraw-Hill.
Related QuanterLab articles
Try it in QuanterLab
Tax drag rarely matters in IRA / 401k contexts. For taxable accounts, prefer annual rebalance over quarterly; the tax-rate split shifts from ~75/25 ST/LT to ~10/90, saving 2–3% per year of after-tax CAGR.