Factor momentum and stock momentum in the UK: Spanning and its limits

We examine whether Ehsani and Linnainmaa's (2022) factor-autocorrelation account of stock momentum extends to the UK, a market in which momentum is well established but its relationship with the wider factor structure remains unsettled. Using a broad set of UK factor returns spanning nearly four decades, we find that no statistically detectable momentum alpha remains once time-series factor momentum is controlled for - in this sense, factor momentum spans the conventional momentum factor - although the confidence interval around the residual alpha still admits economically meaningful values. This interpretation is reinforced by a pooled conditional-return difference consistent with, though not by itself establishing, factor-return persistence, correlations with stock momentum that reverse with the sign of prior factor performance, and stronger momentum among the principal components that capture the greatest share of common factor variation. The full-sample result masks subperiod variation: conventional momentum retains a significant alpha before 2006 but none is detectable thereafter, although the difference between the subperiod estimates is not itself statistically significant. Factor momentum also fails to span the wider UK factor universe. Residual momentum and a small set of valuation- and financing-related factors retain significant alphas after multiple-testing correction, whereas the Amihud illiquidity factor does not. Ehsani and Linnainmaa's account therefore carries over to the UK in a specific rather than general sense: no detectable alpha remains for conventional stock momentum, while residual momentum and a small set of other factors retain significant alphas after multiple-testing correction.