China Allocation · Income Vector中国仓位 · 红利向量
A diversified income portfolio for a cyclical market为周期性市场打造的多元化收益组合
As of截至 2026-07-17 · onshore RMB ETFs在岸人民币 ETF
Diversification — not momentum — is the edge in China在中国，优势来自分散 —— 而非动量
China is cyclical and mean-reverting, and its cash leg is broken (10Y government bonds ~1.7%, money funds ~0.8%), so the US playbook — time the index versus T-bills — does not transfer. What works here is a static, diversified income core: a dividend / cashflow ETF blended with GOLD and short BONDS, the two liquid onshore assets uncorrelated to A-shares. That blend roughly doubles the Sharpe ratio and cuts the drawdown of buy-and-hold by more than half.中国市场具周期性且均值回归，现金腿收益极低（10年期国债约 1.7%，货币基金约 0.8%），因此美式打法 —— 在指数与短债间择时 —— 并不适用。真正有效的是静态、分散的收益核心：红利/现金流 ETF 搭配黄金与短久期国债（这两类在岸资产与 A 股不相关）。该组合大致使夏普翻倍，并将买入持有的回撤削减一半以上。
Recommended allocation建议仓位 · Balanced均衡
Dividend (income core)红利（收入核心）35%
Gold黄金30%
Treasury bonds国债35%
Dividend (income core)红利（收入核心）Gold黄金Treasury bonds国债
Rebalanced quarterly back to target. Three risk variants below; pick by how much drawdown you can hold.按季再平衡至目标权重。下方提供三种风险档；按你能承受的回撤选择。
What this blend has delivered该组合的历史表现 (2013-07-19→)
0.92
Sharpe
CSI 300沪深300 0.44
-17.2%
Max drawdown最大回撤
CSI 300沪深300 -45.5%
7.53%
CAGR
CSI 300沪深300 7.33%
Bootstrap 95% CI on Sharpe夏普自助法95%置信区间: 0.39 – 1.5 · vol波动 8.6% · Sortino索提诺 1.14
Versus CSI 300 buy & hold over the same window. Short China history → judged on the independent bear-episode count (below), not a bare ratio.与同期沪深300买入持有相比。中国历史较短 → 以下方独立熊市事件数为判据，而非单一比率。
Where each sleeve is now各组成部分现状
Dividend (income core)红利（收入核心）
3.135 -0.35%
-7.3% off high距高点 · ✕ 200d
CSI 300 (growth kicker)沪深300（增长）
4.589 -3.45%
-9.9% off high距高点 · ✕ 200d
Gold黄金
8.311 -0.62%
-30.2% off high距高点 · ✕ 200d
Treasury bonds国债
140.812 +0.02%
-0.3% off high距高点 · ✓ 200d
Growth of ¥1 and drawdown · vs CSI 300¥1 增长与回撤 · 对比沪深300
Growth of ¥1 (log)¥1 增长（对数）
Blue = the diversified blend, amber = the income core alone, grey = CSI 300 buy & hold.蓝 = 多元组合，琥珀 = 仅收益核心，灰 = 沪深300买入持有。
Underwater (drawdown)水下回撤
The blend (blue) barely dips where the index (grey) loses ~45%. Gold and bonds absorb the equity drawdown.在指数（灰）回撤约45%处，组合（蓝）几乎不动。黄金与国债吸收了股票回撤。
The three variants · pick your drawdown tolerance三种风险档 · 按回撤承受力选择
| Variant档位 | Mix配比 | CAGR | Sharpe | Max DD最大回撤 | Vol波动 |
|---|
| Conservative保守 |
|
6.5% |
1.03 |
-11.7% |
6.6% |
| Balanced均衡 ◀ |
|
7.53% |
0.92 |
-17.2% |
8.6% |
| Growth进取 |
|
7.91% |
0.81 |
-22.0% |
10.4% |
Conservative保守 — Bond-heavy — shallowest drawdown, steadiest line.偏债 —— 回撤最浅、曲线最稳。Balanced均衡 — The risk-parity-style default — best Sharpe per unit of drawdown.类风险平价的默认 —— 单位回撤的夏普最佳。Growth进取 — Adds a CSI 300 sleeve — higher return, still ~quarter the index drawdown.加入沪深300 —— 更高收益，回撤仍约为指数的四分之一。
We tested a momentum overlay — here is why it is OFF我们测试了动量叠加 —— 这是它被关闭的原因
You asked for a momentum / trend overlay to cut drawdown. We built and backtested one — a slow (200-day, monthly) trend filter that rotates the equity sleeves into bonds during downtrends. In China's mean-reverting market it WHIPSAWS: it does not improve the Sharpe ratio or reduce drawdown versus simply holding the diversified blend. So the strategy is static by design. The comparison is shown in full — no cherry-picking.你希望加入动量/趋势叠加以削减回撤。我们做了并回测了一个 —— 慢速（200日、按月）趋势过滤，在下行趋势中将股票腿轮动至国债。在均值回归的中国市场里它会反复打脸：相比直接持有多元组合，既不改善夏普也不减少回撤。因此策略刻意保持静态。完整对比如下 —— 不做挑选。
| Approach方法 | Sharpe | Max drawdown最大回撤 | CAGR |
|---|
| Static blend (what we ship)静态组合（采用） | 0.92 | -17.2% | 7.53% |
| + 200-day trend overlay+ 200日趋势叠加 | 0.91 | -16.6% | 6.62% |
Timing LOWERS the Sharpe ratio and does not reduce drawdown — diversification, not momentum, is the edge in China.择时降低夏普且不减回撤 —— 在中国，优势来自分散而非动量。
The instruments · liquid onshore ETFs所用工具 · 流动性强的在岸 ETF
| Role角色 | ETFETF | Tracks跟踪 |
|---|
| Income core收益核心 | 510880 | SSE Dividend (高股息红利)上证红利（高股息） |
| Growth kicker增长 | 510300 | CSI 300 large-caps沪深300大盘 |
| Crisis diversifier危机分散 | 518880 | Gold黄金 |
| Low-vol ballast低波压舱 | 511010 | Treasury bonds (5Y)国债（5年） |
Drop-in income-core alternatives可替换的收益核心
510880 SSE Dividend上证红利 (2008→)512890 CSI Dividend Low-Vol中证红利低波 (2018→)159201 Free Cash Flow ('cash-cow')自由现金流（现金牛） (2025→)
The low-vol dividend (512890) cuts drawdown a touch more; the free-cash-flow ETF (159201, the "cash-cow") is the newest income factor but has under 1.5 years of live history, so it is shown as an option, not in the backtest.红利低波（512890）回撤更小一些；自由现金流 ETF（159201，"现金牛"）是最新的收益因子，但实盘历史不足1.5年，故仅作为可选项，未纳入回测。
Independent drawdown episodes · the true sample size独立回撤事件 · 真实样本量
| Peak峰值 | Trough谷底 | Blend drawdown组合回撤 |
|---|
| 2015-06-12 | 2016-01-28 | -17.2% |
| 2026-03-03 | 2026-06-30 | -12.5% (ongoing进行中) |
≥10% peak-to-recovery episodes for the blend. China history is short, so the effective sample is small — read the Sharpe with its bootstrap CI, not as a point estimate.组合的 ≥10% 峰值至修复事件。中国历史较短，有效样本量小 —— 应结合自助法置信区间读取夏普，而非点估计。