Value at Risk (VaR) — institutional risk management dla retail

JP Morgan 1994 RiskMetrics paper. Sir Dennis Weatherstone (CEO JPM) wymaga „4:15 report" — daily firm-wide risk summary 4:15 PM. Result: VaR institutional standard. Dziś każdy bank, hedge fund, prop firm używa. Retail trader też może adapt — Marek $10k account, 2% daily VaR = $200 max loss budget. Pokazujemy framework adaptacji.

Co to VaR

VaR = Value at Risk. Maksymalna potential loss at given confidence level over given time period. Standard: 1-day 95% VaR.

3 metody calculation

„Parametric (variance-covariance) = simplest, normality assumed. Historical simulation = past data, no assumption. Monte Carlo = random sampling, computational heavy. Retail uses parametric."

Parametric VaR formula

VaR = Position × σ × Z × √(time)

• Position: notional value
• σ: volatility (standard deviation returns)
• Z: confidence level Z-score (95% = 1.645, 99% = 2.326)
• √(time): square root time scaling

Example EUR/USD position $30k, σ daily 0.6%, 1-day 95%:

VaR = $30k × 0.006 × 1.645 × 1 = $296

= 95% confidence not lose > $296 daily.

Retail adaptation

Multi-position aggregation

Multiple positions = portfolio VaR. NIE simple sum (correlation matters):

• Long EUR/USD + Long GBP/USD = correlated +0.85
• Combined VaR = √(VaR1² + VaR2² + 2×ρ×VaR1×VaR2)
• Diversification benefit: -10-30% combined VaR
• Top tip: trade uncorrelated pairs (EUR/USD + USD/CAD)

Limitations

1. Tail risk: 5% events ignored (Black Swan)
2. Normality assumption: real returns fat-tailed
3. Backward-looking: past volatility, NIE future
4. Static: NIE captures regime change
5. Solution: stress tests + CVaR (Conditional VaR)

Marek implementation

Marek $10k retail account, daily VaR rule:

• Calculate daily VaR before each trade
• Max 2% = $200 daily loss budget
• If VaR exceeded by open positions → no new trades
• Result: -8% max drawdown vs -25% pre-VaR
• Sustainable scaling enabled

Wnioski

Value at Risk = institutional risk metric. JP Morgan RiskMetrics 1994. 1-day 95% standard.

3 metody: parametric (Excel), historical simulation, Monte Carlo. Retail = parametric.

Formula: VaR = Position × σ × Z × √(time). EUR/USD $30k = ~$296 daily VaR.

Retail adaptation: 2% account daily VaR = sustainable. $10k → $200 budget.

Multi-position: correlation matters. Portfolio VaR < sum (diversification benefit).

Limitations: tail risk, normality, backward-looking, static. Solution: CVaR + stress.

Marek case: -8% max DD post-VaR vs -25% pre. Sustainable scaling.

Konkluzja: VaR upgrade z prostego „1% per trade." Holistic portfolio risk management. Retail can adapt institutional approach.

Powiązane: risk management, position sizing, Sharpe ratio.

Głębsza analiza — VaR deep dive ForexMechanics.