| name | portfolio-construction |
| description | Portfolio construction guide — asset allocation, diversification, rebalancing, personal risk-tolerance assessment, the core-satellite strategy, a simplified Kelly criterion, and performance tracking. For retail investors graduating from single-stock picking to systematic portfolio management.
|
| category | finance |
| tags | ["portfolio","asset-allocation","risk-management","diversification","investing"] |
| related | ["tw-stock-fundamental","tw-etf-investing","tw-stock-options","tw-stock-tax","tw-stock-trend"] |
Portfolio Construction
~90% of a portfolio's return is determined by asset allocation, not stock picking. Decide "how much in stocks vs cash" first; pick names later.
When to Use This Skill
- You hold several stocks and ETFs but have no overall allocation
- Want to graduate from "buy one at a time" to "manage the portfolio"
- Want to set sane allocation ratios for your risk tolerance
- Want to know when to rebalance
- Want to track and improve portfolio performance
1. Self-Assessing Risk Tolerance
Before allocating, answer these:
| Question | Conservative | Moderate | Aggressive |
|---|
| Investment horizon | < 3 yrs | 3–10 yrs | > 10 yrs |
| Age | > 55 | 35–55 | < 35 |
| Income stability | Unstable | Stable | Stable + growing |
| Max acceptable drawdown | -10% | -20% | -30%+ |
| Reaction at -20% | Sell everything | Uncomfortable but hold | Buy more |
Suggested allocation
| Type | Equity (incl. ETF) | Bonds / Fixed deposit | Cash |
|---|
| Conservative | 30–40% | 40–50% | 10–20% |
| Moderate | 50–70% | 20–30% | 10–20% |
| Aggressive | 70–90% | 0–20% | 10% |
- Be honest. Overestimating your risk tolerance = panic-selling at the bottom (the worst time).
- There is no "correct" allocation — only what fits you. A portfolio you can hold through a downturn is the only good portfolio.
2. Allocation Framework
Core-satellite (recommended)
Core (60–80%)
├── Broad-market ETF (e.g. 0050 / 006208 in TW; VOO / VTI globally)
├── Dividend ETF (optional)
└── Stable assets (fixed deposit, bond ETFs)
Satellite (20–40%)
├── Individual stocks (5–8 names, fundamentals-screened)
├── Thematic ETFs (semiconductors, ESG, etc.)
└── Options strategies (optional, small allocation)
- Core: dollar-cost average, don't time. Satellite can be actively managed.
- Core's job is "don't lose"; satellite's job is "outperformance".
3. Diversification Principles
Dimensions of diversification
| Dimension | How |
|---|
| Number of names | 5–15 holdings (too few = concentration risk; too many = unmanageable) |
| Industry | No single industry above ~40% (cap tech at 50%) |
| Market cap | Mostly large cap, with some mid/small |
| Style | Mix growth + value + dividend |
| Asset class | Equity + cash + bonds (per risk tolerance) |
| Geography (optional) | TW + US ETFs to spread regional risk |
- No single name above 15% of total capital. No matter how confident.
- No single industry above 30–40%.
- Combine low-correlation assets. TSMC + MediaTek isn't diversification (highly correlated). TSMC + Chunghwa Telecom is.
4. Position Sizing
Equal weight (simplest)
Same dollar amount per name. Ten names → 10% each.
Confidence-weighted
Weight by your conviction:
| Conviction | Weight |
|---|
| High (fundamentals + chip-flow + technical aligned) | 10–15% |
| Medium (two of three aligned) | 5–10% |
| Low (single signal, or watchlist) | 2–5% |
Simplified Kelly Criterion
Optimal position % = win_rate − (1 − win_rate) / win_loss_ratio
Example: 60% win rate, average win/loss ratio 2:1
→ 0.6 − 0.4 / 2 = 0.4 = 40%
In practice, use 1/4 to 1/2 Kelly. Full Kelly has too much volatility.
- Build new positions in tranches starting at 1/3 of the target. Add the rest after the thesis confirms.
5. Rebalancing
When
| Trigger | Action |
|---|
| Calendar: quarterly or semi-annually | Check whether allocations have drifted from target |
| Drift: any asset class > 5% off target | Sell the over-allocated, buy the under-allocated |
| Event: after a sharp move | Mechanically adjust back to target |
How
Target: 0050 60%, individual stocks 30%, cash 10%
Actual: 0050 72%, individual stocks 22%, cash 6%
→ Sell some 0050 (72%→60%), add to stocks (22%→30%), refill cash (6%→10%)
- Rebalancing mechanizes "sell high, buy low". It naturally trims winners and tops up laggards.
- Don't over-rebalance. Each adjustment has cost. 2–4 times per year is enough.
6. Performance Tracking
Core metrics
| Metric | How to compute | Good benchmark |
|---|
| Annualized return | Total return (incl. dividends) / years | > broad-market index (0050) |
| Max drawdown (MDD) | Peak-to-trough loss | < 25% (moderate type) |
| Sharpe ratio | (return − risk-free rate) / volatility | > 0.5 |
How to track
- Log total portfolio value monthly. One line in Excel or Google Sheets.
- Compare against a benchmark. Your stock-picked portfolio vs. dollar-cost averaging into 0050. If you trail 0050 long-term, raise the ETF allocation.
- Full review once a year. Which positions contributed, which dragged.
7. Construction Steps
Step 1: Decide risk tolerance
→ Set the high-level equity / cash / bond split.
Step 2: Decide core / satellite split
→ Usually 70/30 or 60/40.
Step 3: Pick core
→ 1–2 broad-market ETFs, dollar-cost average.
Step 4: Pick satellites
→ 5–8 names (fundamentals filter + chip-flow timing).
Step 5: Decide position sizes
→ Equal weight or confidence-weighted.
Step 6: Set rebalancing rule
→ Quarterly + drift trigger > 5%.
Step 7: Execute and log
→ Monthly log; annual review.
Common Traps
| Trap | Counter |
|---|
| No allocation, just a pile of stocks | Decide the high-level split first, names second |
| Concentrated in one sector | Cap industry at 30–40% |
| Never rebalance | Use calendar + drift triggers |
| Tweak every week | Quarterly is enough; over-tweaking = high cost |
| No performance tracking | One line per month; compare to benchmark yearly |
| Overestimating risk tolerance | Stress-test with a real downturn (how did 2022's 20% drop feel?) |
Pre-Flight Checklist
Related Skills
