C-Vine Copula Mean Reversion

Code Stratmill research code

Summary

The code implements a statistical-arbitrage strategy that uses a fitted C-vine copula to estimate conditional probabilities for a target asset from a panel of returns. It transforms each asset’s returns through fitted cumulative distribution functions, computes a conditional-probability mispricing index, subtracts its neutral midpoint, and cumulatively sums the result. Rolling means and standard deviations of that cumulative series form Bollinger bands: positions are opened when the series moves beyond a band and exited when it crosses the rolling mean. A signal-to-position table preserves, changes, or closes positions based on the prior holding.

The code also supports live position updates and translates directional signals into dollar-neutral holdings against an index fund, allocating equal dollar amounts to the target and hedge leg. It cites a published study as its basis, but the supplied material reports no backtest results. Performance therefore depends on the fitted marginals and copula, the selected assets and parameters, and execution assumptions; the code itself does not establish profitability. The excerpt is incomplete, limiting assessment of the full implementation.

Key ideas

The strategy maps asset returns to quantiles and uses a C-vine copula to estimate conditional mispricing indices.
It accumulates deviations of the mispricing index from its midpoint and applies rolling Bollinger bands.
Band breaches trigger long or short positions, while a crossing of the rolling mean signals an exit.
A configurable table maps each signal and previous position to the next position.
The code can convert signals into equal-dollar target and index-hedge holdings, but provides no performance evidence.

C-Vine Copula Mispricing Signals with Bollinger Band Entries

Summary

Key ideas

Tags