Capital Allocator

There's no one-size-fits-all answer - it depends on your risk tolerance, time horizon, and goals. A common starting framework: Age-based rule (100 - age = equity %). For a 30-year-old with long horizon: 70% equity, 20% debt, 10% gold is reasonable. For someone near retirement: 40% equity, 50% debt, 10% gold. Start conservative if unsure and adjust as you gain experience and clarity about your risk tolerance.

For most investors, annual rebalancing is sufficient and cost-effective. More frequent rebalancing (quarterly) can help in volatile markets but increases transaction costs and tax events. Consider threshold-based rebalancing: rebalance only when an asset class drifts more than 5% from target. This balances discipline with cost efficiency.

Yes - international diversification can reduce portfolio volatility, and it matters more in Singapore because the STI is concentrated in banks and REITs. Consider 20-40% international exposure through SGX-listed global/regional ETFs or an overseas broker. Singapore has an open capital account, so there is no outbound-investment limit (unlike India's LRS) - you can invest abroad freely. Two things to watch: currency risk (you are taking SGD-vs-foreign exposure) and foreign-dividend withholding tax, which you can manage through fund domicile (e.g. Irish UCITS ETFs).

Start simple. With limited capital, focus on 2-3 asset classes. A basic allocation: 70% equity (via one diversified fund such as an STI or global equity ETF), 20% debt (one Singapore bond fund, or Singapore Savings Bonds), 10% gold (a gold ETF). As capital grows, add diversification within each class. The key is starting with a proper allocation mindset, even if simplified.

For most investors, mutual funds (especially index funds) are better for core equity allocation: instant diversification, professional management, easier to manage. Direct stocks require more time, knowledge, and ability to stomach individual stock volatility. A hybrid approach works well: core allocation in index funds, smaller portion in direct stocks for learning and potential alpha.

Set clear rules: Define allowable deviation bands from strategic allocation (e.g., ±10%). Have specific triggers for tactical shifts (valuation thresholds, trend signals). Limit frequency of tactical changes (no more than quarterly). Require documented rationale for each shift. Set time horizons and exit triggers for tactical positions. This prevents emotional over-trading while allowing disciplined tactical adjustments.

Consider: Risk-adjusted performance (Sharpe ratio), correlation with other strategies (low correlation is valuable), capacity constraints, and your confidence in the strategy. Risk parity across strategies (equal risk contribution) is a good default. Use fractional Kelly for sizing based on edge and variance. Start conservative with new strategies and increase allocation as live performance validates backtests.

All-time highs are normal in growing markets - they do not predict corrections. Maintain your strategic allocation unless you have strong valuation-based views. If very concerned: a modest tactical underweight (reduce equity by 5-10%, increase cash). Avoid market timing based on price levels alone. Keep up regular savings-plan (dollar-cost averaging) investments - time in the market beats timing the market. And review your risk tolerance: if all-time highs make you nervous, perhaps your strategic allocation is too aggressive.

Pure risk parity typically requires leverage (a high bond allocation). For an unleveraged implementation in Singapore, use inverse-volatility weighting (allocate inversely proportional to each asset's volatility). Example: equity vol 18%, debt vol 5%, gold vol 15%. Weights: 1/18 : 1/5 : 1/15 normalised = roughly 14% : 50% : 36%. This achieves more balanced risk without leverage, though returns may be lower than a traditional 60/40.

If you own a home it is already a significant allocation to real estate (often 50%+ of net worth) - factor that in, and you may need less property exposure in your investment portfolio. For investment property or REITs, treat it as a separate asset class with a 10-15% allocation. Note that Singapore has one of Asia's deepest S-REIT markets (e.g. CapitaLand Integrated Commercial Trust, Mapletree and Ascendas REITs), giving liquid, tax-exempt-distribution real-estate exposure without the hassles of direct ownership.

Simplified implementation: (1) Use market-cap weights as the baseline (the STI weight for equity, with proportional debt/gold). (2) Calculate implied returns using a risk premium (equity ~5% over the risk-free rate, debt ~1%, gold ~2%). (3) Express views as simple adjustments, e.g. 'I expect equity to return 2% less than equilibrium.' (4) Blend: adjusted return = equilibrium + (view x confidence). (5) Reallocate based on the adjusted returns. Python libraries such as PyPortfolioOpt make a formal Black-Litterman implementation accessible.

Correlations impact both risk and optimal allocation. High correlation strategies should share risk budget, not each get full allocation. Approach: (1) Calculate rolling correlations (60-day). (2) If strategies A and B correlate >0.7, treat as partially same strategy. (3) Combined allocation = sqrt(A² + B² + 2ρAB) should equal intended risk budget. (4) Monitor for correlation regime changes. (5) In stress, assume correlations spike - stress test with higher correlations. (6) Prefer adding low-correlation strategies to existing portfolio.

Alternative data (satellite imagery, social sentiment, payment flows) can inform tactical allocation: (1) Establish predictive relationship - does data actually forecast returns? (2) Determine latency - how fresh is the signal? (3) Build into allocation framework as additional view (Black-Litterman style) rather than replacing fundamental allocation. (4) Confidence weight based on historical accuracy. (5) Avoid overfitting - out-of-sample validation essential. (6) Start with small tactical tilts before increasing based on live performance.

Regime-based allocation: (1) Define regimes - typically 4: growth, recession, inflation, deflation. Or simpler: risk-on, risk-off. (2) Identify regime signals - economic indicators (PMI, yield curve), market signals (trend, volatility). (3) Map optimal allocation per regime - backtest historical performance. (4) Build regime probability model (HMM, rules-based, ML). (5) Either hard switch (100% regime A) or probability-weighted (60% regime A, 40% regime B allocations blended). (6) Transaction cost awareness - don't switch too frequently. (7) Continuous validation - regimes and relationships can change.

Capacity constraints are critical for realistic allocation: (1) Estimate capacity per strategy (based on average daily volume, market impact). (2) Add capacity as constraint in optimization: allocation ≤ capacity. (3) Account for capacity dynamically - more AUM reduces available capacity. (4) For multi-manager allocation, consider each manager's capacity. (5) Build capacity buffer - don't allocate to full capacity. (6) Monitor for capacity pressure - declining performance as AUM grows. (7) In optimization, use diminishing returns assumption: first dollar to strategy earns full edge, marginal dollar earns less.