Pairs Trading (S&P 500 vs Nasdaq 100)

Yes, but costs become proportionally larger. With spread bets (minimum bet sizes), you can trade smaller. With ETFs, trading two positions increases minimum capital needed. Start with at least £5,000-10,000 to make costs manageable. Spread bets are most efficient for smaller accounts due to no commissions.

Ideally yes, to maintain the hedge from the start. In practice, a few seconds delay is fine. The risk is if one fills and market moves before the other. Use market orders for both or bracket orders if your platform supports it. Don't try to time each leg separately.

They're both major US equity indices with high correlation (0.85-0.95) but different compositions. Nasdaq is tech-heavy (~50% tech), S&P is diversified. This creates predictable divergences (tech sentiment, sector rotation) that tend to revert. Both are very liquid with low trading costs.

If you're properly hedged, minimal loss. You're long one and short the other with similar dollar values. If both fall equally, the long loses but the short gains, roughly canceling. You only profit or lose based on their RELATIVE movement (the spread), not absolute direction.

For S&P/Nasdaq, typical holding periods are 5-30 days. The spread half-life (time to revert halfway) is usually 5-15 days. If a trade hasn't worked within 30 days, something may have changed in the relationship. Use time stops to limit duration.

Simplest: Use dollar-neutral (equal £ value). Better: Regress Nasdaq on S&P using 60 days of data, use the slope (beta) as hedge ratio. Best: Use rolling regression or Kalman filter for adaptive hedge ratio. For S&P/Nasdaq, beta is typically around 1.1-1.2.

CFDs/spread bets are usually more efficient: no stamp duty, no borrowing costs for shorts, tax-free profits (spread bets), and can trade fractional amounts. ETFs are simpler to understand and hold overnight without financing (though you need to borrow for shorts). For active pairs trading, CFDs/spread bets are typically better.

Tech sentiment (Nasdaq more sensitive), sector rotation (growth vs value), earnings season (tech earnings impact Nasdaq more), interest rates (growth stocks more rate-sensitive), and market themes (like AI boosting Nasdaq). Most divergences are temporary, which is why pairs trading works.

ETFs: You receive dividends on long leg, pay equivalent on short leg (if borrowing). CFDs: Dividend adjustments applied automatically. Spread bets: Similar automatic adjustments. Net impact is usually small since both indices have similar dividend characteristics, but it's not zero.

Spread divergence beyond your stop - the relationship temporarily or permanently changes. This happened in 2020 when tech massively outperformed. Your stop at z = 3.0 might get hit, then spread continues to z = 5.0. Accept stops as part of the strategy. Relationship breakdown (permanent) is worse than temporary extension.

Monitor multiple metrics: rolling correlation (alert below 0.80), ADF test p-value (alert above 0.10), and spread half-life (alert above 30 days). If multiple metrics deteriorate, reduce size or pause. Also watch for fundamental reasons: major index changes, persistent regime shifts, policy changes affecting sectors differently.

Kalman filter is theoretically superior - smooth adaptation without window effects. But it requires proper parameter tuning and implementation. Rolling OLS is simpler, transparent, and works well. For most traders, rolling 60-day OLS updated weekly is sufficient. Use Kalman if you have quant infrastructure.

Select pairs from different sources of divergence: S&P/Nasdaq (US indices), FTSE/DAX (European), sector pairs, factor pairs. Ensure low correlation between pair P&Ls - if all pairs move together, you're not diversified. Size each pair based on its volatility and conviction. Monitor aggregate portfolio risk.

Very high - both indices are extremely liquid. Market impact becomes relevant only for institutional-size trades (millions of dollars). For retail traders, capacity is essentially unlimited. The constraint is strategy alpha decay over time as more traders exploit the relationship, not liquidity.

Options define maximum loss (premium paid) but add theta decay. If spread takes longer to converge, options lose value. They're useful for event-driven pair trades with clear time horizon or when you want defined risk. For standard mean reversion, linear instruments (stocks/CFDs) are simpler and don't have time decay.