CITIC SEC: The A-share market is currently in a stage of oversold rebound, with a focus on core assets in technology. Increase allocation to energy, chemicals, and other sectors.

CITIC SEC released a research report stating that the A-share market in early August is in a stage of oversold rebound, with varieties that had previously dropped significantly showing greater elasticity. Currently, the annual yields in the non-ferrous, chemical, non-banking, and electrical new industries are still below the theoretical center. The bank has quantitatively assessed the recovery progress of popular sectors from three perspectives: holding cost, financing clearance, and crowding. From the perspective of holding costs, the pressure of unrealized losses is concentrated in technology growth and small-cap stocks, which still need further digestion; from the financing perspective, the clearance progress of the leading sectors in this round has passed the halfway point; from the crowding perspective, trading activity in the technology sector remains unrelenting. Overall, the recovery progress is relatively fast in electronics, non-ferrous, innovative drugs, and non-banking, while it is slower in chemicals, electrical new, and communications. In terms of allocation, technology positions continue to concentrate on core assets, while the non-technology segments are increasing allocations in energy and chemical, non-ferrous, innovative drugs, and top brokers. CITIC SECs main views are as follows: The market in early August is in an oversold rebound stage, with varieties that had previously seen larger declines showing greater elasticity. In general, the correlation coefficient between the price changes in July and those in the first week of August is -0.85. The deeper the previous decline in specific sectors, the stronger the rebound this week. The leading sectors such as PCB, semiconductor materials, and cables rose by 33.0%, 27.6%, and 23.4%, respectively, while these sectors faced declines of 35.4%, 43.8%, and 45.4% in July. This characteristic aligns with our previous report's judgment that "after severe declines caused by similar liquidity shocks in the past, the oversold rebound effect typically materializes within 5 to 10 trading days after the low." It is worth noting that both within the technology sector and non-technology sectors, the repair extent in early August is related to previous declines; however, technology looks at the decline in July, while non-technology focuses on the decline in Q2. Therefore, the current market recovery is still more of an oversold rebound stage and has not yet transitioned to a stage of fundamental trend pricing. Currently, the annual yields in the non-ferrous, chemical, non-banking, and electrical new industries are still below the theoretical center. The bank has calculated the differences between the theoretical expected returns and actual returns for various style, industry ETFs, and broad-based indices. The theoretical expected returns are estimated by "the yield of the CSI Dividend Index (as an opportunity cost) + the adjustment range of annual profit expectations." Actual returns reflect the yields up to August 7 of this year. Regarding styles, we analyzed ten style factors: valuation, growth, profitability, size, liquidity, momentum, leverage, dividends, volatility, and technical analysis. Among these, the factors showing annual yields significantly higher than theoretical expected yields include momentum, technical, liquidity, and size factors, while growth (mainly represented by net profit growth rate, operating revenue growth rate, operating profit growth rate, and return on equity growth rate) and leverage (mainly represented by long-term debt ratio, asset-liability ratio, cash recovery rate, and shareholder equity ratio) are significantly lower than the theoretical expected yield. In terms of industry ETFs, after adjustments in July, some popular industry or theme ETFs have seen their annual price changes fall below their theoretical expected yields, notably including communications, batteries, and grid equipment; meanwhile, the annual growth in non-ferrous, chemical, and non-banking sectors has been below expected yield levels, reflecting that market investors continue to worry about the profit sustainability of these industries and are reluctant to assign higher valuations. In terms of broad-based indices, the annual growth of large-cap blue-chip indices is relatively lower than theoretical expected yields; after experiencing adjustments in July, the ChiNext 50 has returned below its theoretical value, while the Sci-Tech 50 is one of the rare indices that has had no significant upward revisions in profit forecasts this year but still achieved substantial gains. From the perspective of holding costs, unrealized loss pressure is concentrated in technology growth and small-cap stocks, which still need further digestion. The bank has constructed a weighted chip model, focusing on chips accumulated since Q2 of this year (accounting for about 40% to 60% of the current stock), to calculate the weighted average cost to characterize the actual holding positions of participants in this market cycle. As of August 7, the overall cost deviation rate for the entire A-share market is -3.9%. The calculation method for this indicator is (latest stock price - weighted average cost) / weighted average cost, with a profit margin of 44%. The funds accumulated since late June have already experienced a round of turnover and clearance. There is a clear differentiation at the level of industries, broad-based sectors, and themes, with holding cost pressure highly concentrated in technology growth and small to mid-cap stocks, while there is virtually no unrealized loss pressure in dividend and resource directions. From an industry perspective, the deepest negative deviation rates are in defense and military industries (-12.4%), communications (-10.6%), power equipment and new energy (-10.0%), automotive (-8.9%), and light industry manufacturing (-8.7%); in contrast, the medical sector (+3.0%), non-ferrous metals (+0.7%), petroleum and petrochemicals (+0.6%), banking (+0.5%), and coal (+0.2%) have positive deviations, with the latest prices still above the average cost estimated from the chip model since Q2 2026. In terms of styles, growth (-5.6%) and volatility (-4.7%) have the deepest negative deviations, while momentum portfolios (+7.5%) still maintain a positive deviation after this round of adjustments. From the broad-based viewpoint, the ChiNext 2000 (-9.8%), ChiNext 50 (-8.2%), and ChiNext 1000 (-7.4%) exhibit significantly greater lock-in depths compared to the CSI 300 (-1.3%), with the deeper value divergence leading to higher cost pressures. In terms of popular themes, grid equipment (-12.0%), Siasun Robot & Automation (-10.4%), and batteries (-10.3%) have the deepest negative cost deviation rates. From the structure of the chips above, the technology growth sector shows a relatively deep average unrealized loss, with a high proportion of deeply unrealized loss chips, making the pressure to reduce holdings during the rebound relatively stubborn. When categorizing chips with costs above current prices into different layers of 0-15% and over 15%, the proportion of deeply unrealized loss chips (with costs more than 15% above current prices) is highest for defense and military (44%), communications (41%), and basic chemicals and machinery (both at 35%). In the broad-based sector, this proportion reaches 40% for ChiNext 2000, and in the themes, grid equipment and Siasun Robot & Automation (both at 41%) also rank in the same tier. We have also calculated the clearance turnover adequacy for each industry. Non-AI sectors, which had adjusted earlier in Q2, have completed substantial turnover, while the technology sectors that led the declines since the end of June still require further digestion in turnover. From the financing perspective, the clearance progress of the leading directions in this round has surpassed the halfway mark. The bank measures the clearance progress of net financing growth in this round using the formula (peak financing balance - latest value) / (peak - peak value in the previous year). As of August 7, the clearance progress for the directions that peaked in financing balances from late June to early July is generally low: in the industry sector, electronics (0.51), non-bank financials (0.53), construction materials (0.55), and non-ferrous metals (0.59) are all below 0.6; in the broad-based sectors, Sci-Tech 50 is only at 0.51, and CSI 300 is at 0.67, with liquidity (0.45) and momentum (0.49) being the lowest among the styles. Conversely, the 21 industries with clearance progress exceeding 1 (indicating that the entire net increase has been returned) mostly reached their financing peak from January to March. In terms of themes, half of the semiconductor equipment (0.46), Sci-Tech chips (0.48), AI pricing chain (0.52), and grid equipment (0.53) are also around halfway in their clearance progress, with securities and insurance (0.34) as the lowest across all groups; meanwhile, batteries (1.33) are the only popular theme to achieve excess clearance. From the crowding perspective, trading activity in the technology sector has not yet declined. Currently, the concentration of trading activity in the A-share market remains high, with the top 5% of stocks accounting for 51.4% of trading volume (20-day moving average), slightly down from the peak of 51.9% on July 15, and still at a high percentile of 98.3% over nearly five years. Regarding major broad-based indices, the concentration in the innovation sectors remains quite high, while the small-cap index shows a significant decrease in trading crowding. In terms of industries, there are generally three types of crowding changes in sectors that have recently gained a lot of market attention: the first type still sees trading crowding at elevated levels, indicating a strong sentiment of collective investment, as seen in electronics, communications, and construction materials, with trading crowding having only slightly decreased compared to its peaks in 2026; the second type has experienced a rapid decline in trading crowding, including sectors like basic chemicals, power equipment and new energy, non-ferrous metals, electricity and utilities, petroleum and petrochemicals; the third type consistently shows low trading crowding, with typical representatives in pharmaceuticals, non-bank financials, food and beverage, and banking. In terms of styles and themes, the dividend style currently has low trading crowding, while AI applications and non-AI popular themes have seen varying degrees of crowding declines. Recovery progress in popular directions: electronics, non-ferrous, innovative drugs, and non-banking are faster, while chemicals, electrical new, and communications are relatively slower. By comprehensively analyzing the current clearance levels of popular industry themes from the perspectives of holding costs, financing conditions, and trading crowding, the results show that although the electronics sector has high crowding, the clearance process for financing has surpassed the halfway mark, and the unrealized loss pressure is low (sectors like semiconductor equipment have already accumulated unrealized profits); non-ferrous, innovative drugs, and non-banking sectors show little unrealized loss pressure, undergo rapid financing clearance, and currently have low trading crowding; although the electrical new and chemical sectors have acceptable financing clearance progress and their trading crowding has decreased, they still face significant unrealized loss pressure (particularly in chemicals, which have a heavy burden of deeply unrealized loss chips); the communication sector faces severe unrealized loss pressure, with relatively slow clearance in financing and high trading crowding. Following a comprehensive rebound in early August, popular sectors with annual yields significantly below theoretical expected yields include non-ferrous, chemicals, non-banking, and electrical new, while communications and innovative drugs are near their theoretical expected yields, with the electronics sector (semiconductor equipment, Sci-Tech chips) having annual growth still far exceeding theoretical expected yields, possibly reflecting a narrative of long-term profitability rather than adjustments to current profit forecasts. In terms of allocation, technology holdings continue to concentrate on core assets, while the non-technology segment is reallocating towards energy and chemical, non-ferrous, innovative drugs, and leading brokers. The market has begun the process of repair; moving forward, as market liquidity and price discovery mechanisms return to normal, the allocation strategy in August should gradually shift from trading-based oversold rebounds to a balanced approach. By leveraging the rebound to reduce exposure to marginal assets, the focus should be on core competitive technology leaders and non-technology sectors with stable fundamentals, lower valuations, and favorable chip positions for a rebalancing. We maintain a mid-term judgment of three convergences: 1) within the AI industrial chain, the excess returns of upstream hardware and price-increasing varieties are expected to converge relative to downstream platforms and cloud services; 2) the valuation discount of non-AI industrial sectors relative to comparable overseas companies is expected to be repaired in stages; 3) the extreme differentiation between technology and non-technology sectors tends to converge. Within the technology sector, we suggest that the rebound in AI-related price-increasing varieties be utilized to timely adjust holdings towards core assets (such as gas turbines, wafer manufacturing platforms, and semiconductor equipment), emphasizing quantity certainty while cautiously addressing price explosiveness. For the non-technology sectors, the emphasis should be on increasing allocations to energy and chemical, non-ferrous, innovative drugs, and leading brokers with potential for overseas expansion.