In the first seven months of the year, China's economic performance has exhibited a pronounced "K-shaped divergence": external demand and technology supply chains have maintained strong resilience, while domestic demand recovery has lagged; new growth drivers such as artificial intelligence, high-end manufacturing, and new energy have accelerated, whereas traditional industries and old momentum remain in adjustment; upstream sectors have seen faster profit improvements, while downstream enterprises and households bear the brunt of insufficient demand. The economy does not lack growth momentum; rather, there is a structural misalignment between growth drivers and the recovery of demand.
From the government side, fiscal efforts have been slow to materialize—local government bond issuance reached roughly 55% of the annual plan through July, with new special bond issuance under half, meaning fiscal funds have not yet fully converted into tangible project output. From the household side, consumption is undergoing a weak recovery, with service retail sales growth significantly outpacing goods retail; the consumption structure is shifting from goods to services, yet income expectations, employment stability, and wealth effects continue to restrain spending intentions. Meanwhile, industrial enterprise profits, while growing rapidly, remain concentrated in capital- and technology-intensive sectors, failing to translate into broader employment expansion and household income gains.
Thus, the core issue for the economy is not simply "whether growth exists," but whether its benefits can be transmitted more widely. The key to breaking the impasse lies in accelerating the development of the new economy, enabling technological innovation and export growth to achieve greater industrial scale, while unblocking the chain linking corporate profits to employment, wages, and consumption—facilitating a smooth transition between old and new growth drivers and steering the economy from localized prosperity toward more balanced recovery.
Three Dimensions of Extreme Divergence in the January-July Economy
The first seven months have witnessed clear "K-shaped divergence" across sectors, industries, and demand sides: some industries grow rapidly, supporting upward momentum, while others recover slowly, creating persistent downward pressure. This divergence manifests in three main aspects: stronger external demand versus weaker domestic demand; accelerated growth in new drivers like AI versus sluggish recovery in traditional industries; and relatively active upstream sectors versus downstream pressure from weak demand and squeezed profits.
On the upside, AI, the digital economy, and high-end manufacturing continue to generate ripple effects, driving faster investment and production expansion in related fields, which in turn boosts supply chains for equipment, components, and technical services. Some high-tech and emerging industry products maintain strong overseas competitiveness, with export chains sustaining high growth rates as a key pillar of industrial production and economic growth. Technological innovation thus not only creates new growth points directly but also forms a prominent upward force through the interlinkages among investment, production, and exports.
On the downside, insufficient domestic demand remains the primary constraint on further recovery. Household consumption recovers slowly, with income and employment expectations in some sectors and groups still needing improvement, suppressing consumer willingness and capacity. Meanwhile, traditional industries face weak demand, the property sector and its related chains remain in adjustment, and downstream firms contend with order shortfalls, intensifying price competition, and narrowing profit margins. These factors combine to create a structural pattern where upstream and emerging sectors are stronger while downstream and traditional areas lag.
Overall, the economy through July does not lack growth drivers; rather, momentum is concentrated in new technologies like AI, emerging industries, and external demand, yet has not been transmitted to broader consumption and traditional industries. Transforming the localized prosperity from tech investment and exports into household income growth, consumption expansion, and overall supply chain improvement will determine whether the economy can move from K-shaped divergence toward more balanced recovery.
Government Side of Domestic Demand: Fiscal Pace to Accelerate in August-September
From the government perspective, fiscal efforts have been relatively slow through July, with lagging local bond issuance and fund utilization, insufficiently supporting domestic demand expansion and infrastructure investment. Data show that local bond issuance reached only 55% of the annual plan through July, about 8 percentage points below historical average levels; new special bond issuance is under 50%, implying that some arranged fiscal funds have not yet been converted into tangible project volume and effective investment.
The slower fiscal pace stems both from slower issuance on the funding side and insufficient project supply. New projects accounted for only 36.9% of cumulative issuance through July, below the 45.4% in the same period of 2025, limiting the availability of quality projects eligible for special bond funds. In recent years, as "correct performance views" have standardized local investment promotion and government investment behavior, and as lifelong accountability for government project decisions is implemented, local governments have become more cautious in project selection, investment decisions, and fund utilization—curtailing past models reliant on high leverage and fast investment. Stricter project review helps improve fiscal fund efficiency but also leads to insufficient project reserves in the short term, delaying bond issuance and fund disbursement.
However, July saw marginal improvements on the project side. The share of new special bonds allocated to new projects rebounded notably, interrupting the declining trend since May, suggesting that as the 15th Five-Year Plan (2026-2030) special plans gradually materialize, local major project reserves are improving, easing the project-side constraints on fiscal funds. In other words, fiscal space is not lacking; rather, the mismatch between fund supply and project supply, influenced by project preparation and cautious investment decisions, is being addressed.
Policy signals have already pointed clearly toward faster fiscal implementation. The July Politburo meeting emphasized accelerating fiscal spending and bond fund utilization, while the State Council executive meeting called for expediting "15th Five-Year Plan" major projects, focusing on the "six networks": water networks, new-type power grids, computing networks, next-generation communication networks, urban underground pipelines, and logistics networks. These areas possess fundamental and strategic importance, with large investment scale and industrial multiplier effects, positioning them as key directions for subsequent special bond deployment.
Looking ahead, with accelerated implementation of existing policies and enriched project reserves, local bond supply and fund utilization will likely accelerate in August-September to make up for earlier delays and align fiscal spending with annual targets. Policy focus is expected to remain on executing existing measures, especially "six networks" and other major projects. Meanwhile, if domestic demand recovery remains sluggish, policymakers may also promptly design and introduce more pragmatic and targeted incremental policies to expand effective investment, stabilize market expectations, and strengthen fiscal support for domestic demand.
Household Side of Domestic Demand: Weak Recovery, Commodity Consumption Peaking, Shift to Services
Current household consumption remains in a state of weak recovery, yet the consumption structure is undergoing profound changes. On June 16th, the National Bureau of Statistics began jointly publishing "total retail sales of consumer goods and services," formally shifting the traditional "social retail" indicator—previously focused on goods—to an approach that treats goods and services consumption equally. This adjustment is not merely statistical; it reflects a clear divergence in China's consumption structure: goods consumption growth is slowing while services consumption gains importance.
Under the new framework, total retail sales of consumer goods and services grew 2.7% year-on-year in the first half, with services retail up 5.3% and goods retail up only 1.1%—a 4.2 percentage point gap. In other words, services consumption growth is nearly five times that of goods, making it the more dynamic segment of household spending. Communications, dining, cultural tourism, performances, sports, housekeeping, and elderly care services all maintain rapid growth, indicating that households are not entirely lacking consumption intent; rather, amid saturating goods markets and weakened wealth effects, they are gradually shifting from "buying goods" to "buying experiences, services, and quality of life."
From an international comparison perspective, China's household consumption structure has long favored goods, with high shares for food, tobacco, alcohol, clothing, housing, and daily necessities. As basic material needs are progressively met, some goods markets show signs of supply surplus, intensified homogenized competition, and narrowing growth space, signaling stage-peaked or decelerating goods consumption. In 2025, per capita services consumption expenditure reached 46.1% of total per capita consumption, approaching 60% in first-tier cities like Beijing and Shanghai, yet still below levels in the US (~68.5%), France (~60%), Japan (~57%), and Germany (~51%). Thus, significant structural headroom remains for China's services consumption.
China's relatively low services consumption share is not purely a development-stage issue; it relates to consumption structure and institutional environment. On one hand, rigid expenditures on housing and education crowd out household consumption, while food and tobacco shares remain notably high. On the other, differing public service and social security coverage creates significant worries about retirement, healthcare, and education, strengthening precautionary saving motives. Additionally, some service industries face access restrictions, regional segmentation, and insufficient competition—with high-quality supply shortfalls coexisting with oversupply of low-end services. Low household disposable income shares and wide income gaps also limit overall services consumption expansion.
Germany's experience demonstrates that a strong manufacturing sector does not preclude high services consumption; the key lies in aligning household income levels, social security, service supply, and consumption environments. The rise of services consumption cannot be simplified as policy stimulus or traditional consumption upgrading; it reflects multiple transformations in how the consumption system operates. First, consumption motivation is shifting from "buying what's missing" to "who I want to become," attaching greater importance to emotional value, companionship experiences, health management, and self-realization—reflecting an "experiences over ownership" trend. With weakened wealth effects, households are more cautious in big-ticket goods purchases but may sustain resilience in services that deliver immediate experiences and emotional satisfaction.
Second, digital platforms are reshaping the transaction structure of services consumption. Platforms like Meituan, Taobao Flash Deals, JD.com, and Douyin have moved a large volume of offline life services online, alleviating transaction barriers that made services consumption highly decentralized and dependent on offline scenarios and word-of-mouth. As instant retail, local life services, tourism, performances, and payments integrate further, physical goods, instant fulfillment, local services, and AI assistants are forming a new consumption ecosystem. While the services retail industry has reached considerable scale, online penetration remains low outside dining, accommodation, and health sectors, leaving substantial room for future digital transformation.
Third, service supply itself is expanding and upgrading. Healthcare is extending beyond "seeing a doctor" to consumer medical services, preventive health management, and wellness care; commercial spaces are moving from traditional shopping and cinemas to children's education, VR/AR experiences, themed entertainment, and immersive consumption. Meanwhile, new supply forms—such as medical escorts, door-to-door services, companion services, minimally invasive aesthetics, and instant home delivery—are in turn creating new consumption demand. Traditional goods consumption is also accelerating its servitization: mobile phone sales are shifting from single product transactions to integrated services encompassing trade-in, data migration, installation, and after-sales support.
The policy environment is likewise driving this transformation. The July-issued "15th Five-Year Plan for Expanding Consumption" places "promoting quality and accessible services consumption" in a prominent position, proposing improvements in life services consumption and expansion of elderly care consumption, with repeated emphasis on services. The plan explicitly sets a target of approximately 60 trillion yuan in total retail sales of consumer goods by 2030, indicating that services consumption is no longer a natural byproduct of upgrading but has been incorporated into national medium- and long-term consumption and economic transformation goals. Since 2025, more than ten national-level policies and plans have been issued to promote services consumption development, covering silver economy, childcare services, cultural tourism, health consumption, and automotive consumption.
However, the short-term stimulative effect of services consumption policies may be limited; their policy nature is more about cultivation than direct stimulus. Policy support can improve consumption scenarios and expand service supply in the short run, but transaction structure reform, services online migration, and supporting mechanisms require extended timeframes; changes in consumption motivation represent even slower structural trends, typically requiring simultaneous shifts in income expectations, social security, and lifestyle before full release. Therefore, while services consumption holds long-term upward potential, it cannot fully compensate for the gap left by moderating goods consumption in the near term.
The core constraint on household services consumption remains insufficient income and expectations. Services consumption is mostly discretionary, more sensitive to employment stability, income expectations, and social security. The rising share of new employment forms—food delivery, gig work, self-media—means some workers face high income volatility, platform commission pressures, inadequate labor protection, and weak occupational stability, which both constrains consumption capacity and erodes confidence. Additionally, real estate remains a major household wealth vehicle; housing price expectations and balance sheet changes directly affect consumption decisions. Wealth effects have not materialized, and income effects remain underdelivered. Combined with inconsistent service industry standards, high consumer rights-protection costs, and premium prices for quality services, some potential demand remains unconverted.
In the second half of the year, consumption policy focus is expected to shift beyond simple subsidies or short-term stimuli toward improving consumption institutions and environments. On the front end, stabilizing employment, boosting household income, improving social security, and implementing paid leave can reduce precautionary savings and enhance consumption willingness and disposable time. On the back end, focus should be on elderly care, childcare, cultural tourism, and health sectors—increasing affordable, quality-guaranteed service supply while relaxing services access and introducing market competition to reduce costs and improve quality. Additionally, digital consumption, green consumption, first-release consumption, and experiential consumption scenarios should be cultivated, promoting "AI + consumption" and enhancing immersive, interactive experiences in scenic areas, neighborhoods, and commercial districts.
Furthermore, consumption policies must optimize the trading environment and consumption layout—improving service industry standards, strengthening consumer rights protection, facilitating payments and after-sales services, pushing services consumption resources toward counties and rural areas, narrowing urban-rural and regional consumption gaps, and expanding inbound consumption through visa waivers, tax rebates, and payment facilitation. Overall, services consumption is an important direction for China's domestic demand transformation, but its true release depends on improved income expectations, stronger social security, higher-quality service supply, and reduced institutional barriers. In the short term, consumption will continue to exhibit weak recovery; over the medium-to-long term, services consumption is expected to become the key increment driving household consumption structure upgrades and domestic demand expansion.
Industrial Enterprise Profits: K-Shaped Divergence Persists, Profits Not Converting into Consumer Demand
Industrial enterprises achieved simultaneous revenue and profit growth in the first half, but the structural characteristics of profit improvement are pronounced. Above-scale industrial enterprises saw operating revenue grow 6.5% year-on-year, up 1 percentage point from the January-May period; profits grew approximately 18.7%, modestly down 0.1 percentage points from January-May, still at elevated levels with total profits hitting a near four-year high. Thus, while the overall industrial sector's operating conditions have improved, profit growth does not represent broad-based recovery but concentrates in a few industries, with industrial profits continuing to show clear K-shaped divergence.
By industry, the upward end primarily comprises capital-intensive and technology-intensive sectors, including high-tech manufacturing, the new energy chain, and certain resource industries. These benefit from AI, digital economy, equipment renewal, and industrial upgrading trends, with faster production expansion and investment growth, higher value-added products, and stronger profitability. However, their production processes rely heavily on automation equipment, algorithms, technology, and capital input, with relatively limited direct absorption of ordinary labor. Even rapid profit growth may not translate into large-scale hiring and worker income gains.
The downward end primarily consists of traditional manufacturing, processing, and basic industries that are large employment absorbers, including automobiles, electrical equipment, ferrous metals, and agricultural food processing. Some firms face insufficient end-market demand, intensifying price competition, low capacity utilization, and squeezed profit margins. Since these industries are more closely tied to ordinary workers' employment, their profitability directly affects income expectations for a large labor population. Hence, the more profit growth concentrates in technology- and capital-intensive industries, the more pronounced the disconnect between industrial profit improvement and household income growth.
Employment data confirms this. Employment in above-scale industrial enterprises continues to decline overall; at end-May, employment stood at approximately 72.21 million workers, down 0.9% year-on-year, with absolute scale below 2024 and 2025 levels and overall growth in negative territory. This means that despite notable profit improvement, profit growth has not brought about a rebound in above-scale industrial employment. Technology-intensive industries, in particular, while raising automation and production efficiency, have relatively limited demand for additional ordinary workers, with industry employment at multi-year lows.
By employment distribution, total employment in industries with stagnant or declining profits versus others stands at roughly a 1:1.8 ratio, implying nearly two-thirds of industrial workers are employed in profit-stressed industries. Concurrently, nearly half of workers have already or are entering new employment forms—food delivery, livestreaming, instant retail, online education, freelance work. While these new forms provide some employment buffer, they commonly feature high income volatility, inadequate labor protection, and weak occupational stability, providing limited support to consumer confidence. Therefore, this round of significant industrial profit growth reflects efficiency gains and profitability expansion in capital- and technology-intensive industries rather than broad-based labor income improvement. New profits may in the near term be used more for replenishing cash flow, debt repayment, expanding equipment investment, or raising automation—not yet transmitted to households through expanded employment, higher wages, and improved benefits. Consequently, the transmission chain from "corporate profit growth to employment expansion to household income gains to consumer demand growth" remains incomplete.
In summary, record industrial profits and weak household consumption are not contradictory; they reflect structural divergence between different sectors and industries in economic operation. The current pass-through from industrial profit growth to consumption will take time; the key depends on whether future profits flow more toward labor-intensive industries, transform into stable employment and household income, and further form effective support for domestic final consumption.
Outlook for Breaking the Impasse: When Will K-Shaped Divergence Converge?
The current K-shaped divergence in economic operation is no longer a purely short-term cyclical fluctuation but the combined result of old-new growth driver transition, industrial restructuring, and income distribution changes. The upward end is driven by AI, high-end manufacturing, new energy, and export chains; the downward end is dragged by traditional industry adjustments, weak household consumption, and inefficient employment transmission. On when this divergence will converge, market institutions have not reached a fully consistent view, but mainstream perspectives generally hold that achieving smooth old-new driver transition will require substantial time.
On timing, CICC is relatively optimistic, suggesting the peak of K-shaped divergence may have passed and economic structure may improve progressively; CSC expects the convergence inflection point may emerge around 2029; Changjiang Securities believes divergence may persist until 2030-2035; Morgan Stanley expects structural divergence to continue without providing a clear convergence timeline. Overall, mainstream expectations are not that divergence will quickly vanish in the short term, but that as new economy scale expands, old economy drag diminishes, and household income transmission mechanisms improve, economic structure may gradually move toward balance over several years.
Whether K-shaped divergence ultimately converges depends on three critical conditions. First, whether exports and external demand can continue serving as a growth engine and further unblock the transmission chain from exports and production to employment and household income. Current export growth has been achieved, but the phenomenon of "high output growth without employment growth" persists. To cite one institution's estimate, 2025 export growth of approximately 5.45% coincided with above-scale industrial enterprise employment declining by about 330,000 workers, indicating external demand expansion has not yet converted into domestic employment and consumption.
Second, whether corporate profits can convert more into wage growth and household income improvement. Currently, profit growth concentrates in capital- and technology-intensive industries, with corporate earnings improvement not simultaneously generating large-scale hiring and higher labor compensation. Some institutions note that industrial enterprise profit margins have declined from 6.09% to 5.31%, implying that even with higher total profits, firms may prefer cost-cutting, automation, and capital investment to expand labor's share in primary distribution. Only when corporate profits flow more smoothly into wages, employment, and social security can they form support for household consumption.
Third, whether the AI technological revolution can move from localized high growth to scaled, nonlinear expansion. Currently, AI, new energy, and high-end manufacturing have become the main upward forces, yet their share in macroeconomic indicators like exports, investment, and inflation remains limited. In the future, only when technological innovation completes the leap from R&D breakthroughs, equipment investment, to industrial diffusion and consumer application can new drivers achieve sufficient economic scale to offset the decline of old drivers such as property, traditional manufacturing, and basic industries.
Historical experience shows that successful old-new driver transitions do not rely solely on cultivating a few high-growth industries; more crucial is establishing an institutional environment that supports driver conversion. First, allow backward capacity and inefficient enterprises to exit in an orderly manner, preventing capital, land, and labor from remaining locked in sectors lacking growth prospects. Second, build a capital system capable of supporting high-risk innovation—through venture capital, long-term capital, and multi-tiered financial markets—providing funding for new technologies from laboratory to commercialization. Third, promote free flow of talent, capital, and other production factors toward new industries, reducing industry access and regional mobility barriers. Fourth, through basic research, application scenarios, and market cultivation, help new drivers reach sufficient industrial scale to genuinely offset the drag from shrinking old drivers.
From institutional perspectives, CSC places greater emphasis on balancing technological advantages with domestic divergence, arguing policy sequencing should prioritize technological innovation first, then narrowing domestic structural divergence, and unblock profit-to-household income transmission by raising labor compensation share. Changjiang Securities argues the AI industry chain's share in exports and investment remains limited; short-term reliance is on policy-supported investment gradually ramping up—computing power, electricity grids, urban renewal—while awaiting diminishing old economy drag. Guosheng Securities emphasizes that the transmission of AI and high-end manufacturing prosperity to household income will not happen overnight; current policy focus should be on accelerating fiscal spending, special bonds, and policy financial tools, as well as forming more tangible project output from "six networks" and major projects. CICC's relatively positive view is that the most severe phase of K-shaped divergence may be passing, but its premise is that policy focus can further shift toward domestic demand when external demand pressure intensifies. Currently, real interest rates exceed natural rates, leaving limited self-repairing momentum for the credit cycle; hence, technology and advanced manufacturing may continue receiving more policy resources. Morgan Stanley emphasizes fiscal-tax policy rebalancing, arguing structural fiscal support should gradually shift from simply investing in physical hardware toward employment priority and household consumption—reducing export tax rebate dependence, supporting consumer services and labor-intensive industries, and improving the connection between economic growth and household income.
Other institutions' proposed policy directions broadly revolve around "supply and demand dual efficacy." On one hand, link fiscal interest subsidies with monetary policy tools to lower financing costs in key areas, unclog transmission bottlenecks from easy money to easy credit, and guide resources into technology, high-end manufacturing, and new quality productive forces. On the other hand, enhance traditional domestic demand recovery through improved household income, consumer confidence, and employment expectations. Zheshang Securities further notes from a technological cycle perspective that each technological revolution's introduction phase may accompany pronounced structural divergence, and whether technological dividends become inclusive ultimately depends on institutional reconstruction—social security, income distribution, and labor rights—advancing in tandem.
Therefore, the core of future policy is not simply eliminating industrial divergence but accelerating new economy development while unblocking the chain transmitting emerging industry prosperity to household income and final consumption. In the short term, policy will likely focus on new driver cultivation—technological innovation, advanced manufacturing, new infrastructure, and "six networks"—prioritizing accelerated implementation of existing measures while maintaining overall restraint. Over the medium-to-long term, greater emphasis must be placed on employment priority, wage growth, social security, services consumption, and income distribution—through institutional reform, ensuring that profits and productivity gains from technological progress benefit households more broadly.
In aggregate, the convergence of K-shaped divergence will not occur automatically simply through growth of new industries. The true breakthrough requires accomplishing three tasks simultaneously: orderly exit of backward capacity, sufficient capital, talent and market support for new drivers, and conversion of profits generated by new industries into stable employment, household income, and consumer demand. Only when the chain from "technological innovation-industrial expansion-employment growth-income improvement-consumption enhancement" progressively takes shape can the economy move from localized prosperity toward broader, more balanced recovery.