As China plays a growing role in global portfolios, investors increasingly require timely, granular data to understand how sentiment is shifting across different access channels and share classes.
For nearly three decades EPFR has been tracking flows into mutual funds dedicated to Chinese equity. For over two decades it has tracked the allocations of diversified Equity and Bond Funds to mainland Chinese assets. Since 2020, the company has provided, through its China Share Class Allocations data set, a tool for navigating the nine major Chinese share classes that portfolio managers use to balance their exposure to China.
The latest evolution of EPFR’s China-focused datasets is China Share Class Flows, which provides a credible estimate of the money that mutual funds and ETFs are either committing to or pulling out of the major Chinese share classes. It is based on the marriage of as-reported allocations data, with a monthly frequency, and daily fund flow data.
As clients work with this relatively new dataset, a recurring question that has surfaced – which this Quant’s Corner will address – is: Does China Stock Flows correlate with the portfolio capital flows carried by China’s Stock Connect?
For overseas investors, another long march
China’s 40-year journey from the world’s ninth largest economy to its current position has included a growing, and complex, relationship with global capital markets. As China became the ‘engine of global growth’ following the great financial crisis, Chinese companies increasingly sought access to foreign investors. Those investors were equally anxious to get direct access to China’s equity markets.
In the early stages of this shift, many large, multinational Chinese firms chose to list their shares on overseas exchanges rather than solely in domestic markets. The upshot was a proliferation of share classes that ranged from the Hong Kong-listed H-shares and Red Chips to N-shares traded in the foreign market.
In response, EPFR introduced its China Share Class Allocation product in 2020 which aggregates monthly fund-level equity exposures based on the specific exchanges where Chinese stocks are listed. This provides, on a T+26 frequency, a precise and granular view of global investor sentiment and toward Chinese equities.
Chart 1 – China Share Class Allocation Coverage – Total AuM (US$ billions), Active vs Passive
Meanwhile, those investors were stepping up their utilization of the Shanghai-Hong Kong Stock Connect channel. Over two-thirds of the Chinese equity owned by foreign investors is now linked to this channel.
On the back of these developments, investors started looking for higher frequency China-focused data. They also sought greater insights into the composition, pace and shifts in the capital flows carried by the Stock Connect channel – especially the Northbound flows through Hong Kong into China.
Flows with foreign and Chinese characteristics
While the overall numbers for Northbound flows capture all investment activity routed through the Stock Connect program, including trades from retail investors, hedge funds, proprietary desks, and institutional allocators, EPFR’s China Share Class Flow offers a more targeted perspective. It focuses exclusively on flows driven by the ETF and mutual fund universe.
Those fund flows, though only a subset of the overall flow, are generally viewed as particularly stable, transparent and reflective of persistent investor sentiment. This makes China Share Class Flow particularly valuable for identifying short to medium term positioning trends, rather than high frequency trading noise that may dominate the broader Northbound channel.
Chart 2 compares flows into Chinese share classes for foreign domiciled funds with Northbound Flow data, which represents direct foreign participation in mainland markets via the Stock Connect program.
Chart 2 – China Share Class Flows (foreign domiciled funds) vs Northbound Flow
As you can see from Chart 2, until mid-2022, aggregate China Share Class Flows tracked the total Northbound Flow closely, suggesting alignment between investor sentiment and broader market behavior. Since then, however, a clear divergence has emerged. While the Northbound Flow trend has remained steady, EPFR-tracked fund flows have become more volatile and, at times, negative. This shift may indicate rising caution among longer-term asset managers, driven by macroeconomic uncertainty, regulatory developments, or shifting geopolitical dynamics.
The relatively consistent and steady trajectory of Northbound Flows over time, with very infrequent daily outflows, suggests a sustained interest in mainland Chinese equities via the Stock Connect program. It may also reflect domestic policy-driven buying behavior aimed at maintaining a stable flow of capital into onshore markets.
Connecting the diǎn
To account for the potential bias introduced by these two trends – Northbound Flows reflecting sustained foreign interest and domestic policy-driven buying behavior – we first standardize the two time series to enable a more balanced, meaningful analysis. Specifically, we express China Share Class Flows as a percentage of total fund AUM, and Northbound Flows as a percentage of the domestic market capitalization (Chart 3A).
This normalization adjusts for scale differences between the fund universe and the overall market, allowing for a clearer and more direct comparison.
Chart 3A – China Share Class Flows (% of Fund AuM) vs Northbound Flow (% of Market Cap)
Additionally, the detrended series in Chart 3B removes long-term allocation bias and enhances visibility into short-term, tactical shifts in investor behavior. This adjusted view helps isolate flow movements driven by event-specific catalysts such as U.S.-China geopolitical tensions, rebalancing related to changes in MSCI index inclusion, or monetary policy easing by the People’s Bank of China.
To achieve this, a simple yet effective detrending technique was applied: the historical average of each series was subtracted from its corresponding original values. This normalization enables clearer detection of deviations from baseline sentiment and allows for a more meaningful comparison of directional shifts across time.
Chart 3B – China Share Class Flows (% of Fund AuM) vs Northbound Flow (% of Market Cap), detrend
The results are both visually and statistically significant. The detrended China Share Class Flows exhibit a 50.78% correlation with Northbound Flows, indicating a strong directional alignment between the two series. Furthermore, the regression analysis confirms a statistically meaningful relationship, with a highly significant coefficient and an R-squared value of 25.79%, suggesting that a substantial portion of the variation in Northbound can be explained by movements in fund flows activity.
Chart 4A – Scatter Plot: China Share Class Flows (% of Fund AuM) vs Northbound Flow (% of Market Cap)
Chart 4B – Regression: China Share Class Flows (% of Fund AuM) vs Northbound Flow (% of Market Cap)
A timely look under the hood
Overall, the comparison between China Share Class Flows and Northbound Flows offers valuable insights into how two key investor channels respond to developments in China’s markets. While Northbound Flow reflects broad market participation, including potentially short-term or policy-driven activity, EPFR’s fund-based flow data captures the sentiment-driven positioning of global investors and asset managers on a daily basis.
The divergence between the two flows observed post-2022 underscores the importance of looking beyond headline flow figures to understand underlying investor behavior. By normalizing and detrending the data, we enhance the clarity of short-term tactical shifts, revealing statistically robust patterns that support more informed portfolio decisions.
As global investors continue to navigate an increasingly complex Chinese investment landscape, tools like EPFR’s China Share Class Flow product will play a crucial role in decoding real-time sentiment and positioning signals.
Did you find this useful? Get our EPFR Insights delivered to your inbox.
