Cash Drag and Idle Balance Accounting

A long-short strategy that's only 70% invested on average, with the rest sitting in cash waiting for signals or margin buffer, doesn't earn the same return as a fully invested version of the same strategy — the idle 30% is a drag unless it's earning something. This gap between a strategy's "on paper" return and what it would actually deliver with real idle cash is called cash drag, and backtests routinely get it wrong in both directions.

The two common mistakes are opposite: some backtests assume idle cash earns nothing, silently understating real-world returns because in practice idle cash usually earns at least the risk-free rate in a margin account. Others assume idle cash earns the risk-free rate for free with no friction, which overstates returns for strategies that in reality hold cash at a broker paying below-market rates, or that need a cash buffer for margin calls that isn't actually free to hold. Neither assumption is automatically correct — it depends on the exact broker and account structure being modeled.

The fix is to model the idle balance explicitly: track average uninvested cash as its own line item, credit it with whatever rate the real account would actually pay, and separate that carry return from the strategy's trading return in the performance report. This also exposes an easy lever some strategies rely on to look better than they are — a "long-short equity" strategy might owe half its Sharpe ratio to cash carry rather than stock-picking skill.