Last week, A-shares experienced a modest rebound with increased volume, leading to a marginal recovery in risk appetite, with oversold bounces serving as the primary trading theme. Small-cap stocks demonstrated significantly greater flexibility than their large-cap counterparts.
Looking ahead, we believe there is still room for oversold rebounds to develop, with a particular focus on the ChiNext Index. The pace of this recovery will likely be tested during the intensive mid-year report period in mid-to-late August. In terms of positioning, we recommend a balanced allocation anchored to mid-year reports.
Oversold rebounds in AI hardware remain accessible for participation, but the lack of clearing in public fund positions may limit upside potential, requiring control over position exposure. Within this sector, communication equipment is the preferred choice. Among other oversold categories, we suggest focusing on energy metals and minor metals.
For sectors that held up relatively well in July, we continue to see reasonable value in non-bank financials and CXO leaders. From a medium-term perspective, cost-effective overseas expansion chains and essential consumer goods are entering a window for increased allocation. The defensive dividend base portfolio should be maintained, with a structural emphasis on banks and transportation.
Risk Appetite Marginal Recovery, Oversold Bounce in Progress
Last week, A-shares rebounded with moderate volume expansion. First, panic sentiment has somewhat eased. The implied volatility of CSI 300 and ChiNext Index ETFs declined by 14% and 11%, respectively. The average daily number of stocks hitting the upper limit in August increased compared to July, while those hitting the lower limit decreased. Second, oversold bounces are the main trading theme. In terms of style, large-cap stocks began recovering earlier, but small-cap stocks showed greater elasticity. The CSI 2000 has recovered approximately 50% of its July decline, while the CSI 300 has recovered about 30%. In terms of sectors, the Spearman correlation coefficient between August and July performance reached -0.8, with electronics, communications, and building materials recovering around 30% of their July losses. Third, from a positioning perspective, the height of the oversold rebound may be constrained. While trading-oriented funds have been gradually reducing positions since July, active public funds have not significantly reduced their holdings in AI hardware, as indicated by position estimates and net value trends. If net values recover to near cost levels, these funds may gradually reduce positions due to factors like redemptions, limiting the rebound's potential.
Historical Precedent Suggests ChiNext Oversold Bounce Has Room to Develop
Since 2010, the ChiNext Index has experienced 11 instances of sharp declines exceeding 20% (excluding the current one). In terms of magnitude, a conservative estimate suggests further room for a rebound. The lower quartile of the first rebound amplitude after each sharp decline bottom was 21%. The current rebound has reached 10%, implying potential for around another 10%. Even based on historical lower limits, there is approximately 4% of additional space. In terms of time, the lower quartile of the first rebound's duration for these events was 26 trading days (at least 12), whereas the current rebound has only lasted 6 trading days. Notably, the slope of this rebound is relatively steep, with an 8.4% gain in the first 5 days after bottoming (median 5.5%). Historically, there have been 3 instances where the index fell below its previous low within 120 trading days after an oversold bounce. Common characteristics of these cases included a relatively short first rebound duration, with a median amplitude of 20.5% and a median duration of 14 trading days. The verification window for the current cycle might be around the end of August, with the quality of mid-year reports being a key factor.
Mid-Year Report Season Approaching, Screening High-Cost-Effectiveness Sectors
As the intensive mid-year report disclosure period approaches, we observe that investors are responding more positively to sectors that exceed expectations in their mid-year reports, such as CXO, indicating improved effectiveness of fundamental-based pricing. Anchoring to mid-year reports, we screen for high-value sectors by considering the retracement magnitude from the early June low, the upward revision in 2026E earnings expectations, and the current valuation level. One area is select AI hardware. We previously noted that during the July decline, capital and sentiment factors led to an overshoot in valuations and relative performance. By sub-sector, communication equipment is the top choice, followed by semiconductor equipment (valuations are not cheap) and storage (valuations offer value, but the pace of price increases is slowing, and rising supply expectations limit the room for recovery). Beyond AI hardware, we recommend focusing on energy metals and minor metals.
Allocation Strategy: Balanced Positioning Anchored to Mid-Year Reports, Controlled Participation in AI Hardware Oversold Bounce
Maintain a balanced allocation, prioritizing position discipline over directional choices. While there is still room for an oversold bounce, the steep slope and uncleared public fund positions suggest anchoring allocation to mid-year report value while retaining a defensive dividend base. After the valuation overshoot in July, AI hardware still has room for recovery, with communication equipment as the top pick. Beyond AI hardware, other high-value oversold sectors include energy metals and minor metals. Among sectors that held up well in July, non-bank financials and CXO leaders offer reasonable value. From a medium-term perspective, cost-effective overseas expansion chains and essential consumer goods are entering a window for increased allocation. The defensive dividend base portfolio should be maintained, with a structural focus on banks and transportation.
Risk Factors
External risks exceeding expectations; domestic fundamental recovery falling short of expectations.