VFTSE Correlation

VFTSE itself is not directly tradeable - it's an index. You can trade VFTSE futures/options through global brokers, but this requires access to Eurex or similar exchanges. For most Singapore traders, it's easier to use VFTSE as a signal and trade Singapore instruments (STI options, stocks).

VIX is often better for Singapore traders due to higher correlation. However, VFTSE provides additional information from the European session (which closes before US opens). Using multiple indices together creates a stronger signal than any single index.

It depends. For overnight signals (VFTSE spikes after Singapore close), you might position before Singapore open or at open. For signals during the Singapore day, act within hours. Signals lose predictive power over days as markets adjust.

No. Correlation varies over time and can break down during idiosyncratic events. During global risk events (like COVID crash), correlation is high. During region-specific events, correlation can break. Always check current correlation strength before trading.

The simplest approach: When VFTSE spikes more than 15% overnight, expect STI weakness. Consider buying STI puts or shorting STI at open (with stop loss). This captures the basic inverse relationship without complex structures.

Use daily returns for both. Calculate Pearson correlation over a rolling window (e.g., 60 days). In Excel: =CORREL(range1, range2). In Python: df['VFTSE_return'].rolling(60).corr(df['STI_return']). Correlation will be negative (inverse relationship).

Weight VIX more heavily (~40% vs ~15% for VFTSE). US markets lead global sentiment and have higher correlation with Singapore. VFTSE adds value for European session information but VIX is the primary global fear gauge.

Align data properly. VFTSE closes at 11:30 PM SGT; VIX at 4:15 AM SGT (next day); STI closes at 5:00 PM SGT. For daily analysis, use closes that are meaningful - e.g., previous day VFTSE close vs same day STI. For overnight signals, use most recent data.

Pair trading involves simultaneously long and short two related securities (e.g., long DBS, short OCBC). Correlation trading uses the correlation relationship as a signal for directional or volatility trades. They're related concepts but correlation trading doesn't require offsetting positions.

Collect historical VFTSE and STI data. Define your signal (e.g., VFTSE +15% = short signal). Apply signal historically. Calculate hypothetical STI returns following signal. Account for realistic execution (can't trade at exact close, etc.). Calculate win rate, avg P&L, Sharpe ratio.

Use DCC-GARCH or similar models. These estimate correlation that varies with market conditions. Simpler approach: Calculate rolling correlation with different windows (20, 40, 60 days) and use consensus. Weight recent observations more heavily. Update estimates daily.

Pure correlation trading requires correlation swaps (institutional, OTC). Alternative: Dispersion trading (index vs component vol). In practice, most correlation trades involve some directional risk. You can minimize it by hedging delta, but gamma and other risks remain.

Research shows correlation is stronger when VFTSE rises (fear) than when it falls. Adjust by: (1) Sizing bearish signals larger than bullish, (2) Using tighter stops on bullish correlation trades, (3) Weighting spike signals more than collapse signals, (4) Different holding periods for each direction.

SGX closes at 5 PM, hours before European close and US session. Overnight moves may not be reflected until next day. This creates gap risk (signal occurs, but you can't act until gap). Consider: (1) Using options for overnight exposure, (2) Placing orders before close, (3) Accepting some timing slippage.

Build a signal generation module that: (1) Calculates composite vol index score (weighted VFTSE, VIX, VHSI), (2) Applies thresholds for actionable signals, (3) Checks correlation filter (only trade if correlation is strong), (4) Generates position size based on signal strength, (5) Passes to execution engine. Backtest rigorously; out-of-sample test; paper trade before live.