A lucrative trading pattern that generated roughly 3.7 percent returns per year by holding U.S. equity futures for a single hour in the middle of the night has vanished. In a new analysis from the Federal Reserve Bank of New York, researchers found that the "overnight drift"—a return spike between 2:00 and 3:00 a.m. Eastern time when European markets open—has averaged close to zero since 2021 after persisting for more than two decades. The report, published July 1, 2026, by Nina Boyarchenko, Lars C. Larsen, and Paul Whelan, revisits their 2021 discovery with five additional years of data and pinpoints what killed the pattern.

From 1998 through 2020, the 2:00–3:00 a.m. window was responsible for more than 60 percent of the S&P 500 E-mini futures contract's 5.9 percent annualized close-to-close return. But between January 2021 and December 2025, that same hour went flat. The same pattern holds in the E-mini Nasdaq-100 and E-mini Dow Jones contracts. The dispersion of end-of-day order imbalances—a measure of how wildly buy and sell pressure tilts at the close—has collapsed. The standard deviation of end-of-day relative signed volume (RSV) fell from 6.5 percent in the original sample to 2.9 percent in the post-publication period, a compression of more than half. Meanwhile, volatility barely budged: the VIX fell only modestly from a mean of 20.4 to 19.4, with median values of 18.6 and 18.2. Overnight liquidity also remained stable, with the overnight share of total E-mini volume edging up only slightly from 15 percent to 16 percent.

The authors argue that the overnight drift was compensation earned by liquidity providers who absorbed lopsided order flow at the U.S. close and carried risky inventory overnight. When institutional flows tilt heavily to one side in the final hour of trading, liquidity providers step in as buyers of last resort and demand a discount at the close. As overseas buyers arrive a few hours later—most notably at the Frankfurt and London open—prices rebound, and the liquidity providers earn back their discount. The report notes that in June 2022, NightShares launched two ETFs—NSPY and NIWM—designed to capture overnight equity returns, citing the inventory-risk mechanism, but both funds were closed fourteen months later, consistent with the weakened pattern.

The report attributes the fade to one primary cause: extreme closing-imbalance days have become markedly rarer. The predictive relationship between closing imbalances and overnight returns has broken down—in 2007–20, large negative closing imbalances were followed by positive overnight returns, but in 2021–25, the spread is much narrower. In ongoing work, the researchers find that limit orders posted at the close have become smaller in the post-publication sample, consistent with algorithmic liquidity providers slicing flow more finely and transmitting less residual inventory onto end-of-day counterparties. The framework yields a falsifiable prediction: if order-imbalance dispersion widens back toward its pre-2020 range, the overnight drift should reappear in the same 2:00–3:00 window. For now, the 3 a.m. edge that once minted returns for two decades has gone dark—and it won't come back until the wild swings at the closing bell do too.