Quantitative Enhanced Index Products See Widening Negative Excess Returns as Extreme AI Sector Rally Erodes Alpha
N.R. Finch
In May, 1,251 quant-enhanced index products across China posted an average excess return of -0.98%, with the shortfall widening further in June — an extreme AI-sector rally has broken the multi-sector rotation that multi-factor models depend on, systematically compressing alpha.
How bad is the negative excess return?
As of the week ending June 12, CSI 500 enhanced-index products — the largest category by AUM — averaged -1.46% in return, with excess return of just 0.30%.
Year-to-date average excess stands at 3.87%, but dispersion is extreme: the top 20th-percentile averages 10.71%, the bottom 20th-percentile averages -3.16%. This means → the gap between the best and worst managers is nearly 14 percentage points; picking the wrong one wipes out the entire point of going quant-enhanced.
From October 2024 through April 2025, fewer than 14% of enhanced-index products maintained positive excess for seven consecutive months. In plain terms = out of every ten products, barely one and a half could consistently beat their benchmark.
Why is alpha shrinking across the board?
The AI rally has absorbed the bulk of capital and liquidity. Roughly 5% of listed stocks now account for about 50% of total market turnover; liquidity is concentrated in a handful of mega-cap leaders, while over half of small- and mid-cap stocks see shrinking volume.
This reflects a market that no longer rotates across sectors — the exact environment multi-factor quant models are built for has disappeared, replaced by a single-track crowding of capital into AI names.
Scale growth has made the crowding worse. The number of 10-billion-yuan-plus quant hedge funds has passed 139, and total quant AUM now exceeds RMB 1.8 trillion. Most managers share the same factor set — small-cap, reversal, price-volume — and once factors crowd, alpha decays fast.
Growth and momentum winning — fundamentals losing?
Within quant equity strategies, growth and momentum styles are clearly outperforming; fundamental-value styles are under pressure. This means → models that chase hot AI stocks are making money, while models that pick stocks on valuation and financials are stalling.
Discretionary long-only funds are likewise concentrated in AI and semiconductors, but the trade is extremely crowded and highly sensitive to marginal changes in order intentions from overseas tech giants.
In plain terms = whether quant or discretionary, everyone is in the same lane — earning the same dollar and bearing the same risk in lockstep.
Why have sector ETFs gone from hedging tool to core holding?
Multiple quant hedge funds have integrated sector ETFs into their regular factor models — no longer a temporary hedge, but a medium- to long-term portfolio staple. A single top-tier fund holding 7 to 9 different sector ETFs simultaneously is now standard practice.
Nearly 200 new ETFs have launched this year; among them, over 90 have hedge-fund names among their top-ten holders, a coverage rate of 65.47% — almost double the 34% recorded in the same period last year.
STAR Market ETFs account for over 20% of hedge-fund new-ETF holdings, the only category seeing sustained inflows. Example: the ChinaAMC STAR Semiconductor ETF (588170) has returned 113.55% year-to-date, with units outstanding up 68.32%, AUM up 249%, and latest AUM at RMB 12.83 billion.
Can quant-enhanced products rebuild alpha?
Whether alpha can be rebuilt depends on whether individual firms can achieve meaningful factor differentiation — finding stock-selection rules that differ from the rest of the industry and tap independent return sources.
Current strategy convergence suggests the window for differentiation is narrowing. This means → if most managers keep using similar factors, restoring alpha only gets harder from here.
In plain terms = quant-enhanced indexing faces a binary choice: either discover new factors no one else is using, or accept that alpha will keep getting compressed.
Content is for reference only, not financial advice.