CSI 300 vs CSI 500: A 16-Year DCA Backtest (China A-Shares)
If you invest in China A-shares through dollar-cost averaging, the first question is usually: large cap or mid cap?
CSI 300 and CSI 500 are the two most commonly compared broad indices in mainland China. This post walks through a 16-year backtest of both, with real numbers.
Two indices, two different beasts
CSI 300 tracks the 300 largest companies by market cap on the Shanghai and Shenzhen exchanges. Banks, insurance, liquor, home appliances, oil. Mature industry leaders with stable earnings and high dividends. Low growth potential, low volatility.
CSI 500 tracks companies ranked 301 to 800 by market cap. Pharmaceuticals, electronics, chemicals, machinery. Growth companies in niche industries. High elasticity, higher volatility.
One is the bedrock of China's economy. The other is its waist. This shapes everything about their returns and risk profiles.
Backtest setup
| Item | Value |
|---|---|
| Assets | CSI 300 (000300.SH) vs CSI 500 (000905.SH) |
| Period | 2010-01-01 – 2026-08-31 |
| Frequency | Monthly |
| Amount | 1,000 CNY per contribution |
| Total contributions | 199 |
| Total invested | ¥199,000 |
| Dividends | Excluded |
| Fees | Excluded |
The 2010 start captures the 2015 crash, the 2018 bear market, the 2019–2021 structural bull, the 2022 correction, and the 2023–2026 consolidation.
You can run the same backtest yourself:
The results
| Metric | CSI 300 | CSI 500 |
|---|---|---|
| DCA return | 38.41% | 53.52% |
| Total gains | ¥76,444.62 | ¥106,497.30 |
| Total invested | ¥199,000.00 | ¥199,000.00 |
| Final value | ¥275,444.62 | ¥305,497.30 |
| Annualized return | 3.86% | 5.02% |
| Contributions | 199 | 199 |
| Shares held | 58.58 | 38.04 |
| Max drawdown | -39.35% | -46.13% |
Same monthly amount, same number of contributions, meaningfully different outcomes.
Return side:
- CSI 500's annualized return of 5.02% is 1.16 percentage points higher than CSI 300's 3.86%
- Over 16 years, that gap compounds into about ¥30,000 more in final assets
Risk side:
- CSI 500's max drawdown is 6.78 percentage points deeper than CSI 300's
- During the 2015 crash or 2018 bear market, CSI 500 investors saw their accounts drop over 46% from peak
15 points more return. 7 points more drawdown. That's the trade.
Why CSI 500 returns more but draws down deeper
This is the key to understanding the difference.
CSI 300's constituents are industry leaders — Moutai, Wuliangye, China Merchants Bank, Ping An, CATL. Stable earnings, strong cash flow, high dividends. In bear markets, low valuations and high dividends limit the downside.
CSI 500's constituents are growth companies in niche industries. Not sector leaders yet, but catching up. In bull markets, their earnings elasticity is high and their stock prices rise faster. In bear markets, their valuations contract more sharply.
High elasticity cuts both ways. CSI 500 beat CSI 300 by about 15 points over 16 years. The cost was deeper drawdowns every time the market turned.
How the two performed across market cycles
| Phase | Winner | Why |
|---|---|---|
| 2010–2014 (sideways decline) | CSI 300 | Lower valuations, smaller drawdown |
| 2015 (leverage bull + crash) | Extreme for both | CSI 500 rose more, then fell harder |
| 2016–2018 (value return + bear) | CSI 300 | Foreign capital pushed blue chips higher |
| 2019–2021 (structural bull) | CSI 500 | Growth stocks dominated |
| 2022–2026 (consolidation) | CSI 500 | Lower valuation, better rebound |
Long-term verdict: CSI 500 won during growth-stock periods but required enduring deeper drawdowns each time.
The conclusion
| If you… | Better fit |
|---|---|
| Can tolerate -45%+ account drawdowns | CSI 500 |
| Get anxious when the account drops 35% | CSI 300 |
| Chase higher long-term returns | CSI 500 |
| Prefer a smoother holding experience | CSI 300 |
| Believe in China's growth-stock story | CSI 500 |
| Prefer blue chips and dividends | CSI 300 |
| Have a 10+ year horizon | CSI 500 |
| Have a 3–5 year horizon | CSI 300 |
The two aren't mutually exclusive. A common split is 70% CSI 300 + 30% CSI 500. Keeps the stability of blue chips, captures some growth elasticity.
In one sentence: CSI 500 offers higher returns, but you pay for it with the ability to hold on through deeper drawdowns.
FAQ
Q: CSI 500 has higher long-term returns. Why not go all-in?
A: Because returns are only obvious in hindsight. Anyone buying CSI 500 in 2015 had no idea the crash was coming. If your horizon isn't long enough, or you sell during a deep drawdown, you never capture the high return.
Q: Can DCA smooth out CSI 500's volatility?
A: Partially. CSI 500's max drawdown was -46.13% vs CSI 300's -39.35%. DCA lowers average cost, but the account still shows a loss during bear markets.
Q: Which is better for beginners?
A: CSI 300. The biggest risk for a beginner isn't picking the wrong index — it's quitting during the first major drawdown. CSI 300's lower volatility and higher dividends make it more forgiving.
Q: Can I DCA into both?
A: Yes. A common split is 70% CSI 300 + 30% CSI 500. The right ratio depends on how much account volatility you can accept.
Q: Can I run the backtest myself?
A: Yes. Use the pre-filled parameters: Run the CSI 300 backtest, Run the CSI 500 backtest. All parameters are pre-filled.
Data for reference only. Not investment advice.
Top comments (1)
The two return rows are measured differently, and that changes how to read the gap. I rebuilt both from your inputs: 199 monthly ¥1,000 contributions ending at ¥275,444.62 solve to about 3.8% a year, and ¥305,497.30 to about 5.0%, so 3.86% and 5.02% look like money-weighted IRRs. The 38.41% and 53.52% are final value over total invested, and the average yuan was invested for roughly half of the 16.6 years. 38.41% is therefore not a 16-year CAGR (that would be about 2.0% a year), and "15 points more return" and "1.16 points a year" are different measures of the gap.
Two things I'd check before reading it as large cap versus mid cap. Dividends are excluded, which costs CSI 300 more because its yield is higher. And it is one start date: re-running from every month-start between 2010 and 2018 and plotting the spread of IRR differences would show whether 1.16 points survives a different entry. Is the -39.35% / -46.13% drawdown on the index or on the DCA account value (which has fresh cash going in)?