Loading data…
Price and total-return indices
Drag or pinch to zoom.
Monthly returns
Total return by calendar month; first and last months may be partial.
Calendar-year returns
Price and total return. * Partial year within the selected period.
Total return and drawdown
The two charts zoom together. Drawdown: decline from the previous peak within the period.
Volatility
EWMA (λ = 0.94) and GARCH(1,1)-t conditional volatility, daily × √252.
Rolling 3-year Sharpe ratio
Monthly excess returns over the 13-week US T-bill.
Largest drawdowns in the period
Distribution of daily returns
Histogram against a normal density with the same mean and standard deviation.
Statistics and one-day risk
Daily returns in the selected period and currency.
Value at Risk: out-of-sample backtest
One-day 1% VaR forecast by a GARCH(1,1)-t model re-estimated every year; red points are breaches. USD, log returns.
Full backtest period. Kupiec: unconditional coverage; Christoffersen: independence of breaches; CC: both. p-values below 0.05 reject the model.
Rolling first-order autocorrelation
Daily log returns, trailing 3-year windows, with the 95% band of a heteroskedasticity-robust test.
Rolling variance ratios
Lo–MacKinlay VR(5) and VR(20), trailing 3-year windows. 1 = random walk.
Full-sample tests
p-values adjusted for multiple testing with the Holm procedure.
Moving-average timing rules
VT in USD and in CHF
Total return rebased to 100 at the start of the period, and the USD/CHF exchange rate (CHF per USD).
Rolling correlation with USD/CHF
Weekly returns, trailing 52 weeks. Negative: the dollar tends to rise when VT falls.
Currency contribution
Selected period, total return.
Before taxes. Dividends from VT are subject to a 15% US withholding tax for Swiss residents, usually creditable against Swiss tax.
Buy and sell simulator
What a single investment would have become, in the selected currency.
Monthly savings plan
A fixed amount invested at each month end, dividends reinvested.
The best and the worst days
Final value of 10,000 invested over the selected period when the N best and/or N worst days are removed.
Probability of loss by holding period
Full history, USD, total return.
Historical windows overlap and understate the risk at long horizons; the stationary bootstrap resamples blocks of monthly returns (mean length 12 months).
Validation against Vanguard
Data-quality checks
Run on every update.
Report and downloads
The full analysis report (LaTeX, about 16 pages) covers return and risk, GARCH estimates, the VaR backtest, market-efficiency tests with multiple-testing correction, timing rules, the probability of loss, the Swiss franc view and data validation.