VWAP / TWAP

VWAP and TWAP are both price averages over an interval. But they weight observations differently — by volume and by time, respectively — and the divergence between them is itself a powerful signal about the quality of price moves.

The execution benchmark perspective

Berkowitz, Logue & Noser (1988) established VWAP as the standard institutional execution benchmark, arguing that it represents the "average price available to the market" weighted by actual trading intensity. An institution that consistently buys above VWAP is demonstrably executing worse than the average participant — either because of adverse selection, poor timing, or market impact.

TWAP, by contrast, treats all time equally. It's the benchmark for time-sliced execution algorithms that distribute orders uniformly across a window, ignoring volume patterns. Madhavan (2002) noted that the choice of benchmark reveals the execution strategy's implicit assumption: VWAP assumes you should trade when everyone else does; TWAP assumes you should trade at a constant rate regardless.

The signal in the divergence

The gap between VWAP and TWAP encodes information about the structure of trading within the bar:

VWAP > TWAP → Volume was disproportionately concentrated at higher prices. Heavy-volume trades happened at the top of the range. This is a bullish volume-quality signal: the market didn't just drift up on thin liquidity — participants endorsed the higher prices with real size. This aligns with the Blume, Easley & O'Hara (1994) finding that volume conveys information about the precision of traders' signals.

TWAP > VWAP → Price was elevated for extended periods but on thin volume. The asset spent time at high prices, but trading activity didn't endorse those levels. This is the signature of a thin-market drift — potentially unsustainable and vulnerable to reversal once volume returns.

For performance attribution

Measuring fills against both benchmarks simultaneously reveals different aspects of execution quality:

• Fills consistently above VWAP → adverse selection. You're systematically buying at the worst prices within each bar.
• Fills between VWAP and TWAP → depends on which is higher. If VWAP > TWAP and you're between them, you're beating the volume-weighted average but still paying above the time-average — acceptable in a strong market.
• VWAP-TWAP spread as a feature → the magnitude of this divergence predicts subsequent reversal probability. Extreme divergence means the current bar's volume-weighted price is far from its time-weighted price — an unstable configuration.

References:

• Berkowitz, S.A., Logue, D.E. & Noser, E.A. (1988). "The Total Cost of Transactions on the NYSE." Journal of Finance, 43(1), 97–112.
• Madhavan, A. (2002). "VWAP Strategies." Trading, 1, 32–39.
• Blume, L., Easley, D. & O'Hara, M. (1994). "Market Statistics and Technical Analysis: The Role of Volume." Journal of Finance, 49(1), 153–181.