No stamp duty = Daily trading is viable!
ETFs have NO stamp duty in Hong Kong! This completely changes the calculus for mean reversion trading.
Stock: 0.2% round-trip cost = 50% annual cost at daily trading
ETF: 0.0% round-trip cost = 0% annual cost at daily trading
This means daily mean reversion is mathematically viable on ETFs, while it remains impossible on individual HK stocks.
| Symbol | Name | Type | Data Since | Bars |
|---|---|---|---|---|
| 2800 | Tracker Fund (HSI) | ETF | 2000 | 6,524 |
| 2823 | iShares MSCI China | ETF | 2005 | 5,219 |
| 2827 | ChinaAMC CSI 300 | ETF | 2007 | 4,668 |
| 3188 | ChinaAMC CSI 300 Alt | ETF | 2014 | 2,997 |
Same mean reversion strategy: Buy when >2% below 5-day high, sell when >2% above 5-day low
| Asset | Type | Trades | Gross P&L | Stamp Duty | Net P&L |
|---|---|---|---|---|---|
| Tracker Fund (HSI) | ETF | 1,525 | +43.8% | 0% | +43.8% |
| ChinaAMC CSI 300 | ETF | 1,498 | +315.2% | 0% | +315.2% |
| Tencent | Stock | 372 | +2.1% | +74.4% | -72.3% |
| CKH Holdings | Stock | 1,033 | -29.5% | +206.6% | -236.1% |
| ETF | Total Return | Trades | Win Rate | Profit Factor | Best Params |
|---|---|---|---|---|---|
| Tracker Fund (HSI) ETF | +57.0% | 1,538 | 39% | 1.08 | 60-day lookback, 0.5σ threshold |
| ChinaAMC CSI 300 ETF | +48.8% | 1,017 | 17% | 1.10 | 10-day lookback, 1.0σ threshold |
Without stamp duty, you can trade daily without eroding returns. 250 trades/year costs 0%, not 50%.
The Hang Seng and CSI300 tend to revert to their means more reliably than individual stocks. Index components are constantly rebalanced.
Daily signals catch more mean reversion opportunities. The 60-day lookback on HSI caught 1,538 trades vs fewer at longer intervals.
Win rate of 39% can still be profitable if wins are larger than losses. Profit factor of 1.08 means for every $1 lost, you make $1.08.
Use ETFs for mean reversion, not stocks.
The stamp duty makes stock mean reversion unviable at daily/weekly frequencies.
← Back to Stamp Duty Analysis | Main Page
Analysis using EODHD data | QuestDB | Python