Yu Yongding Defends Investment-Led Growth, Dismisses Consumption-Driven Model
Influential Chinese policy advisor questions effectiveness of consumer subsidies, pushes for renewed infrastructure stimulus
As the 14th five-year plan is closing to its end this year, the discussion on the upcoming 15th five-year plan is gradually taking center stage on the agenda. Jun.15, the CPC flagship journal of Qiushi求是, published a compilation of Xi's remarks on long-term planning. The same issue featured an article by the NDRC Party Group on formulating the 15th Five-Year Plan outline, signaling the top-level focus on charting China's next phase of development. (
has the English translation.)In late May, Yu Yongding, a senior policy advisor, delivered a speech at the China Macroeconomic Forum Seminar (Session 105), defending investment as a primary driver of Chinese economic growth while questioning the long-term effectiveness of current consumption-boosting measures such as vouchers. He appears to be defending the investment-led development model that has powered China's growth for decades—a model now being challenged by advocates of demand-driven growth.
Yu is the former president of the China Society of World Economics and former director of the Institute of World Economics and Politics at the Chinese Academy of Social Sciences. He also served on the Monetary Policy Committee of the People's Bank of China from 2004 to 2006. Given his position and long-standing influence in policy circles, I believe his speech gives us a glimpse into the current policy debate on the direction of the upcoming 15th Five-Year Plan.
Yu Yongding argues that from a long-term economic growth perspective, there is no such thing as "consumption-driven" growth. All else being equal, increased consumption inevitably leads to reduced savings and investment, which in turn slows economic growth. The notion of "consumption-driven growth" only makes sense in the context of short-term macroeconomic management when addressing insufficient aggregate demand.
He contends that China's consumption rate is not as low as some claim. In terms of physical goods ownership, China actually exceeds the United States. The consumption gap between China and the US primarily lies in service consumption, largely due to price differentials and statistical methodology issues. China's real problem is not low consumption rates but excessive income inequality, with the Gini coefficient remaining elevated at 0.46-0.47.
Yu believes that direct consumption stimulus measures like vouchers and tax cuts have limited effectiveness and tend to fall into circular reasoning: "to increase consumption, we need to increase income; to increase income, we need to increase consumption." The true "prime mover" should be government-led infrastructure investment. He recommends launching major infrastructure projects during the 15th Five-Year Plan period such as advancing the Western Development Strategy and building economic corridors to Central Asia as flagship projects for this new investment push.
（In the origional article published by CMF, His proposals span traditional infrastructure that boosts permanent income, social infrastructure like elderly care facilities and childcare centers, targeted subsidies for EVs and appliances, support for service industries where supply is constrained, expanded social benefits for vulnerable groups, tax reforms to reduce inequality, financial market stabilization, and fostering innovation. But I chose to translate the more readable version he published on Guancha Net）
It’s a long article, so I’ll attach the full translation below.
Investment-Driven or Consumption-Driven?
1. "Consumption-Driven" Growth Doesn't Exist
The concept of "consumption-driven" growth is most frequently discussed by American economists like Krugman and Summers. Krugman claims that China's long-standing "investment-driven" growth model has brought the Chinese economy to the brink of financial crisis, while Summers argues that China's "investment-driven" growth model has reached its end.
When discussing the choice between "investment-driven" and "consumption-driven" growth, we must first clarify whether we're discussing long-term economic growth or short-term macroeconomic management. When Western academics discuss China's "investment-driven" model, they're actually referring to China's economic growth strategy and paradigm. How to achieve China's long-term stable growth (development) is fundamentally different from how to achieve a specific growth target in a given year, such as achieving around 5% GDP growth in 2025. The former focuses on medium and long-term economic growth, studying supply-side issues; the latter emphasizes short-term adjustment, studying demand-side issues.
As a matter of long-term economic growth, "consumption-driven" growth simply doesn't exist as an economic growth model.
Taking Marxist reproduction theory as an example, when all savings are converted into investment, the higher the savings rate, the faster the growth of total social product. Conversely, the higher the "consumption rate," the lower the growth rate of total social product.
Taking Lenin's reproduction theory as an example, at age 23, Lenin engaged in a polemic with the Narodniks in an article titled "On the Question of Markets," constructing what was essentially a growth model using difference equations. Lenin used this model to prove that despite Russia's poverty and insufficient consumption—with the second department (producing consumer goods) unable to develop—the first department (producing means of production) could grow independently of the second department's slow growth due to the continuous increase in organic composition. Therefore, Russia would inevitably enter capitalist society, making socialist revolution inevitable.
Capital accumulation driving economic growth is not only a core tenet of traditional Marxist political economy but also true in neoclassical growth theory. For example, in the earliest modern economic growth model—the Harrod-Domar model—"economic growth rate = savings rate/capital-output ratio." Since savings are assumed to be completely converted into investment, savings rate = investment rate. This model shows that economic growth is determined by the investment rate and investment efficiency, with increased consumption meaning reduced savings and investment. All else being equal, increased consumption will lead to decreased economic growth.
The economic growth model derived from the standard production function (constant returns to scale Cobb-Douglas production function) shows that given capital and labor output elasticities, economic growth is a function of investment, labor input, and technological progress. Obviously, for policymakers, investment, labor input, and technological progress are all very important, and the proportional relationship among the three must be properly managed.
Currently, the Chinese government strongly emphasizes industrial and product innovation, placing technological progress's role in driving economic growth in a prominent position. In this context, one could say China faces a choice between "investment-driven" and "technology-driven" growth, but technological progress doesn't fall from the sky—it cannot be separated from various forms of investment. Therefore, one could also say China faces a choice between "general investment-driven" and "innovation-focused investment-driven" growth.
The production function can be expanded to include factors like institutional change and human capital. Improvements in human capital are related to education, health, and other factors. If consumption has a pulling effect on economic growth, this effect should be realized through human capital improvement. However, the transmission mechanism of consumption→human capital improvement→economic growth requires specific quantitative research; conclusions cannot be drawn simply. Moreover, consumption that can improve human capital is of specific types, and the provision of such consumption is inseparable from traditional investment.
In summary, for policymakers, economic growth's dependence on investment, labor input, and technological progress varies in different periods due to different conditions. Growth models in different periods can be labeled as "investment-driven," "labor-driven," or "technology-driven." But regardless, there is no such thing as a "consumption-driven" economic growth model or strategy.
Before the 1997 Asian financial crisis, Krugman criticized East Asian economic growth for relying solely on capital and labor inputs rather than technological progress. Whether his claim was correct aside, he did not advocate that East Asian countries should implement a "consumption-driven" growth strategy at the time.
An economic activity may simultaneously have both investment and consumption characteristics, and a product may serve as both an investment good and a consumer good. For example, mountain climbing and fitness expenses can be viewed as either pure consumption expenditure or human capital investment. A car can be either a consumer good or an investment good (depending on its use). Undeniably, certain types of consumption, as human capital investment, have a driving effect on economic growth. How significant this effect is requires empirical research to determine. However, describing consumption as a "driver" of economic growth without strict qualification and advocating that China's growth model should shift from "investment-driven" to "consumption-driven" is completely wrong theoretically, with predictable consequences.
Denying the existence of a "consumption-driven" growth model doesn't mean more investment is always better. The relationship between investment and economic growth is not linear. Under given conditions, there exists a theoretical optimal investment rate. Using the Harrod-Domar model as an example, an excessively high investment rate (savings rate) may lead to an increase in the capital-output ratio, so excessively raising the investment rate may lead to decreased rather than increased economic growth.
Of course, rising capital-output ratios aren't necessarily caused by excessive investment rates. Due to certain factors causing the capital-output ratio to rise, the government may have to further increase the investment rate to maintain target economic growth. Before reform and opening up, China's one-sided emphasis on raising savings and investment rates resulted in decreased economic efficiency and rising capital-output ratios—a lesson that should not be forgotten.
2. The Relationship Between Consumption and Investment
There exists a trade-off between current consumption and investment, but in the long term, the relationship between consumption and investment is complementary and mutually reinforcing. The ultimate purpose of production is consumption. Given income, allocating more resources to investment means less consumption, and vice versa. But from a long-term perspective, the relationship between investment and consumption is essentially a choice between "consuming now or consuming less now but more in the future."
How to make trade-offs between consumption and investment has always been an important research topic in Marxist political economy. The Feldman model proposed by Soviet economist Feldman in 1936 was built on Marx's production theory. The core problem this model aimed to solve was whether to "consume more now" or "consume less now to achieve more consumption in the future."
Feldman and others believed that the planner's most important task was to correctly decide how to allocate investment between the two major departments. If more investment goes to producing capital goods and less to producing consumer goods currently, current consumer goods output will be smaller. Due to the higher growth rate of means of production used for producing consumer goods, over time, consumer goods production will grow rapidly and at some point exceed the consumer goods provided on the initial growth path, achieving a "bitter first, sweet later" result.
The trade-off between consumption and investment is also an important research area in neoclassical growth theory. Solow's classic growth model assumes that given the production function, labor force growth rate, and capital depreciation rate, different economic growth paths defined by different savings rates all have certain points defined by per capita capital and per capita income, where there exists a steady state with zero growth rates for per capita output, capital, and consumption.
Beyond theory, Western economists have also conducted empirical research on the relationship between economic growth and consumption. For example, Barro explicitly states in his famous textbook "Economic Growth" that cross-country empirical evidence shows consumption expenditure has a negative impact on economic growth. He also notes that total investment as a share of GDP has a positive relationship with initial human capital embodied in education and health.
Recent research by John Ross and others on extensive empirical data proves that there is an extremely high correlation between GDP growth and consumption growth. The share of net fixed capital formation in GDP is highly correlated with consumption growth, and the larger the economic scale, the stronger the correlation. These empirical research results also show that the relationship between consumption and investment is not an opposing either/or relationship, but rather a relationship of "more consumption now and less in the future, or less consumption now and more in the future."
Regarding China's national conditions, citizens generally value retirement security and tend to leave savings for future generations, resulting in relatively low consumption propensity and relatively high investment rates. There is no inherent "good" or "bad" to this preference. How to balance between "present" and "future" is a public choice issue. Generally, as per capita income levels rise, the investment rate (capital formation/GDP) tends to gradually decline. However, it's difficult for the government to determine the optimal investment rate level, and even more difficult to directly decide the entire economy's investment rate. What the government can do is adhere to market economy principles while using macroeconomic policy tools to implement counter-cyclical adjustment, smooth economic fluctuations, and stabilize consumer and investor confidence.
The notion of "consumption driving the economy" is only meaningful in situations of insufficient effective demand. Due to insufficient effective demand, economic growth is constrained by the demand side, and we can analyze the reasons or possibilities for GDP growth in a given year based on the national income identity. According to the expenditure method, GDP = C + I + X - M, where C, I, and X-M represent final consumption, capital formation, and net exports respectively; government expenditure G is decomposed into other parts of GDP.
In Western economic growth theory, the implicit assumption made to simplify analysis is "supply creates demand." In short-term macro analysis, because supply cannot be generated immediately, supply is assumed to be given, so "demand creates growth." China's current situation is insufficient effective demand, so we can temporarily set aside supply constraints and structural issues to focus on the main contradiction: insufficient aggregate demand.
The purpose of short-term macroeconomic regulation is to make the GDP growth rate determined by the theoretical model equal to potential GDP growth. If the economic growth rate determined by consumption demand, investment demand, and net exports is lower than potential economic growth, the government should adopt expansionary fiscal and monetary policies to increase the growth rates of consumption, investment, and net exports, making actual economic growth equal to potential economic growth.
The statement "consumption drives economic growth" is only correct in the sense that increasing consumption demand can compensate for insufficient effective demand, making actual economic growth equal to or closer to potential economic growth. But if aggregate demand exceeds potential growth, it will trigger inflation. In this case, the government needs to suppress consumption demand or other aggregate demand to control inflation around the predetermined inflation target.
3. Is China's Consumption Rate Too Low? Actually, Goods Consumption Rate May Exceed the US
Many people believe China's consumption rate is too low. Looking at the data, taking 2022 as an example, China's final consumption was 43% of the US (exchange rate 1:7). In 2022, China's final consumption (63.8 trillion yuan) accounted for 53.2% of GDP (=63.8/120); the same year, US final consumption ($21.08 trillion) accounted for 82.9% of GDP (=21.08/25.44). In 2022, China's retail goods expenditure was 44 trillion yuan, or $6.29 trillion at a 1:7 exchange rate. The same year, US goods consumption and food services & accommodations were $5.94 trillion and $1.25 trillion respectively, totaling $7.2 trillion. China's retail goods expenditure was 87.4% of the US (=6.29/7.2) (not entirely directly comparable).
The main reason for the huge gap between China's final consumption as a percentage of US final consumption versus China's retail goods expenditure as a percentage of US goods + dining expenditure (43% vs 87.4%) is that service consumption accounts for a much higher proportion of final consumption in the US than in China, and US service prices are much higher than China's. In the US, service consumption typically accounts for 60-70% of final consumption. Before COVID, US service consumption accounted for 69%, significantly higher than goods consumption. [Consumption rate = (goods consumption + service consumption)/GDP]
China's consumption structure is completely different. In 2022, national per capita consumption expenditure was 24,538 yuan, including per capita service consumption expenditure of 10,590 yuan, accounting for 43.2% of per capita consumption expenditure. China's service consumption expenditure as a share of GDP is significantly smaller than the US ratio. While further analysis is needed to determine how much is due to quantity versus price, it's well known that US service consumption prices are much higher than China's. Our own experience also shows that services in the US and Western countries (such as domestic help, home repairs, delivery, theaters, lawyer fees, etc.) are significantly more expensive than in China. For example, ambulance service fees in Los Angeles average about $1,200-$1,300 per trip. In Guangzhou, a single ambulance trip (including about 5 km transport, one doctor and one nurse, two stretcher bearers, and pre-hospital emergency care) costs 148 yuan.
From the perspective of consumer goods expenditure as a share of GDP, China significantly exceeds the US. In 2022, China's GDP was about 70% of US GDP; retail sales were 87.4% of US goods consumption (including dining). In other words, from the perspective of owning consumer goods (cars, TVs, air conditioners, etc.), China's consumption expenditure as a share of GDP significantly exceeds the US. Additionally, there are many statistical scope and methodological issues that tend to underestimate China's final consumption as a share of GDP.
Therefore, after considering consumption structure and price differences, the gap in consumption expenditure as a share of GDP between China and the US shown statistically is actually far less significant than the data suggests. Rough estimates show that measured by goods consumption as a share of GDP, China is about 1.25 times the US.
In short, Chinese residents are more inclined to purchase physical consumer goods like cars and TVs, while American residents are more inclined toward service consumption like legal services. Therefore, from the perspective of consumer goods ownership, China's consumption as a share of GDP may actually exceed the US.
Finally, a fact that shouldn't be overlooked is that China's real estate investment as a share of GDP has long far exceeded that of most countries, but a large portion of real estate investment should be considered as meeting homebuyers' consumption needs rather than investment needs. If this type of real estate consumption is counted as household consumption, China's adjusted consumption rate would be much higher than the figures currently published by the statistics bureau.
Different ethnicities, cultures, institutions, and generations have considerably different time preferences (or consumption discount rates), resulting in different savings rates. For example, East Asian countries generally have high savings rates, which is one of the important reasons for the East Asian miracle. A high savings rate (investment rate) and low consumption rate are not necessarily problems in themselves. On the contrary, low consumption rates and high savings rates may be a huge advantage for developing countries to catch up with developed countries. Everything requires specific analysis rather than generalizations.
First, savings should not be forced savings by the government but autonomous decisions by residents. If the public believes China's current savings rate is too high (consumption rate too low) and wishes to increase the consumption rate, there's no problem. Conversely, if the public wishes to maintain China's current savings rate, there should also be no problem.
Second, the ratio of household disposable income to GDP should be appropriate—neither too high nor too low. But what is appropriate is not only an economic issue but also a geopolitical one.
Third, the savings rate can be decomposed into a basic savings rate determined by culture, society, and economic growth trends, and savings rate fluctuations caused by economic cycles. If the savings rate is high because the basic savings rate is higher than in other countries, the government need not deliberately lower it. If the increase in savings rate is due to "cyclical" reasons (such as economic shocks, declining growth, deteriorating income expectations, reduced consumption—Keynes's Paradox of Thrift), the government needs a counter-cyclical adjustment to increase the consumption rate and reduce the savings rate.
4. Is China's Household Disposable Income as a Share of GDP Too Low? It's Exaggerated
"China's household disposable income is too low" seems to be another consensus in academic circles. The commonly cited data is that China's household disposable income accounted for 43% of GDP in 2022, while this ratio is generally above 60% in other countries. In fact, the National Bureau of Statistics publishes two sets of household disposable income data: one based on household surveys, another based on flow of funds accounts. In 2022, based on flow of funds calculations, household sector disposable income accounted for 59.3% of GDP—a significant difference. Household survey data has systematic bias because surveys require questionnaires with limited subsidies, making low-income groups more willing to participate, resulting in more low-income respondents. Additionally, some residents tend to underreport when filling out forms.
Compared with other countries, Japan's household disposable income accounted for 56.22% of GDP in 2022, Denmark 46.1%, both lower than China's 59.3% calculated by flow of funds. The UK was 61.47%, only slightly higher than China. Different countries have different conditions, and simple comparisons don't tell the whole story. While China's household disposable income level is indeed somewhat low, it's not as severely low as some claim. Given the inherent inaccuracies in statistical data, we shouldn't over-interpret some statistical results.
Compared with other countries, China's household disposable income as a share of GDP is low, but government revenue as a share of GDP is also low. The most prominent feature of China's national income distribution should be that corporate undistributed profits as a share of GDP are significantly higher than other countries. If we want to increase household disposable income as a share of GDP, obviously we cannot rely on reducing government revenue as a share of GDP. In this case, the only possible option is to reduce corporate undistributed profits as a share of GDP.
Corporate disposable income = after-tax profits + depreciation - dividends. China's high corporate undistributed profits as a share of GDP is determined by multiple factors including China's economic system, tax system, and counter-cyclical macroeconomic policies. Chinese companies' high retained earnings may also mean companies need sufficient funds for technological innovation (though there's also the issue of listed companies not paying dividends). China's recent major technological breakthroughs by numerous companies are inseparable from their substantial investments.
At the same time, Chinese corporate debt ratios are significantly higher than in most countries. In this situation, whether and how China's corporate undistributed profits as a share of GDP should be adjusted is a complex issue requiring separate analysis. Claims that China's household disposable income as a share of GDP is too low, implying government revenue as a share of GDP is too high, lack sufficient evidence. In current circumstances, adjusting household disposable income as a share of GDP is not our most urgent priority.
Comparing physical consumption, China's life expectancy exceeds the US (78.2 years vs 76.1 years in 2021); in physical indicators, Chinese household consumption is not lower than developed countries and even leads the US in some areas. For example: calorie intake, protein intake, children's height, urban living space, homeownership rate, years of education, luxury goods sales, per capita meat consumption, etc. When considering social transfers in kind (STIK), World Bank 2021 ICP prices show China's consumption in housing, education, leisure, and healthcare is more than double that measured by market exchange rates. Considering China's quasi-deflation over the past three years, China's consumption in these areas is even higher.
China's biggest current problem is excessive income distribution gaps. Although the Gini coefficient has declined somewhat, it remains at a high level. China Merchants Bank data shows 2.4% of Sunflower VIP clients （Individuals with monthly average total assets exceeding 500,000 RMB） own 81.8% of savings assets. According to Professor Li Shi's research, over the past 20 years, China's income gap first rose, then fell, and now remains at a relatively stable high level. The Gini coefficient in 2008 was 0.491, approaching 0.5. If a country's income gap Gini coefficient exceeds 0.5, it's considered by the international community as extremely unequal in income distribution. China's Gini coefficient turned before reaching 0.5, entering a period of slow decline, but the decline was modest—less than 3 percentage points over 7 years. Since 2016, China's Gini coefficient has fluctuated between 0.46-0.47, meaning income gaps haven't narrowed further since 2016 and remain at a high level. Narrowing income gaps as much as possible is not only about social fairness and justice but also helps increase society's overall marginal propensity to consume.
In summary, from the perspective of consumption as a share of GDP, China's gap with most countries isn't as significant as some self-media claim, but we indeed have much work to do in pursuing common prosperity.
5. How Effective Are Current Consumption Stimulus Measures? More Research Needed
China's final consumption (including household and government consumption) as a share of GDP averaged about 54% over the past decade (2014-2023). In 2010, China's final consumption rate (final consumption as a share of GDP) fell to a post-reform low of 49.1%. It recovered to 53.7% in 2023. In contrast, the global average final consumption rate is about 56.5%, with high-income countries generally above 70%. Simple international comparisons can only serve as references and don't explain much. For example, China's final consumption rate of only 49.1% in 2010 was largely because China was implementing an expansionary fiscal stimulus plan—GDP growth was 10.6% in 2010, while fixed asset investment growth was as high as 23.8%. China's consumption rate falling to 49.1% in 2010 reflected the consequences of expansionary fiscal policy, not any change in Chinese residents' consumption propensity.
After the global financial crisis, China's consumption as a share of GDP has trended upward; investment and net exports as shares of GDP have trended downward. Under conditions of insufficient effective demand, the shifting shares of consumption, investment, and net exports in GDP reflect endogenous changes in macroeconomic variables and the influence of government macroeconomic policies.
According to my own estimates, in 2024, final consumption, capital formation, and net exports accounted for 56.2%, 40.4%, and 3.4% of GDP respectively. Compared with 2023, the main changes in demand structure in 2024 were that net export growth greatly exceeded GDP growth, thereby increasing net exports as a share of GDP; real estate investment growth plummeted sharply, and increased infrastructure investment growth was insufficient to offset the drag on investment growth from declining real estate investment, so investment growth was lower than GDP growth, with its share of GDP declining somewhat. When formulating macroeconomic policies, there's no need to focus excessively on changes in these shares. What's important is identifying demand variables that are more controllable and most helpful for achieving economic growth targets.
Recently, the biggest bright spot in China's economy has been Chinese manufacturing continuously breaking through US technology blockades, making China's industrial chain increasingly autonomous and secure. Chinese manufacturing companies' maintenance of 7-8% average annual investment growth over the past decade has been crucial. The government should minimize intervention in corporate investment decisions and trust that companies best understand their industry and sector development prospects and are most capable of making correct investment decisions.
China has set a 5% economic growth target for 2025, but net exports' contribution to GDP growth may fall from 1.5 percentage points in 2024 to 0 or even lower. Consumption accounts for nearly 60% of GDP. If consumption growth is below 5%, achieving the annual 5% economic growth target will face significant difficulties, substantially increasing pressure to raise investment growth. Therefore, as an important part of short-term macroeconomic policy, taking various measures to vigorously promote consumption growth is entirely necessary.
The problem is that consumption is a function of income, income expectations, and wealth or permanent income. From individual households' perspective, to increase consumption, income must first increase. From the perspective of the household sector as a whole, increasing consumption requires equalizing income distribution—further achieving common prosperity. The more equal an economy's income distribution, the higher the economy's overall marginal propensity to consume.
Regarding specific consumption promotion measures, scholars' suggestions mainly include issuing consumption vouchers, reducing personal income tax, and reforming the social security system. Proposals to stimulate economic growth through consumption stimulus often fall into circular reasoning: to increase economic growth, we need to expand consumption; to expand consumption, we need to increase economic growth. So where should we start?
The proposal to issue consumption vouchers is logically coherent. Methods like giving money or consumption vouchers can temporarily increase income, but the actual effect and sustainability of this approach remain questionable. From January-November 2024, consumer goods categories with growth below 3.5%, listed from lowest to highest, were: gold, silver & jewelry (-3.3%), building & decoration materials (-2.3%), cosmetics (-1.3%), cultural & office supplies (-1.3%), automobiles (-0.7%), clothing, shoes, hats & textiles (0.4%), petroleum & products (0.6%), daily necessities (2.7%), furniture (2.9%). Categories with growth above 3.5% were: pharmaceuticals, food, beverages & tobacco, communication equipment, household appliances & AV equipment, and sports & entertainment goods. It's clear that luxury goods saw the biggest decline in growth, followed by non-necessities, while necessities grew basically in line with or above GDP growth. In this context, whether consumption vouchers can significantly boost consumption growth is uncertain.
First, reduced consumption demand may mainly result from decreased demand for high-end consumer goods among middle and high-income groups, which may be because their high-end consumption demand is more affected by asset price changes (like falling stock and real estate prices). Second, consumption demand growth among low and middle-income groups is relatively stable. Their consumption demand is less affected by income changes—even if income falls, they must maintain necessary consumption levels. In summary, consumption vouchers are unlikely to significantly affect spending behavior among high and middle-to-high income groups. For low and middle-income groups, concerns about future uncertainty make them more inclined to save rather than consume additional funds received. Therefore, consumption voucher distribution may not significantly impact overall consumption levels and growth.
Regarding tax cuts, the main problem with reducing personal income tax is that China's tax system is primarily based on value-added tax, with limited total personal income tax. In 2023, China's total personal income tax was 1.4 trillion yuan. Lowering tax rates and raising tax thresholds would have limited effects on consumption growth. However, tax cuts should be considered for enterprises, especially small and medium-sized enterprises.
Regarding the social security system, China's five social insurances (pension, medical, maternity, unemployment, work injury) are basically actuarially based. In principle, we shouldn't change various insurance contribution and payment rules just because we need expansionary fiscal policy. However, some components are closely related to fiscal expenditure. For example, 85% of funding for urban and rural resident pension insurance (mainly for farmers) in China's pension system comes from fiscal transfers. We should increase urban and rural resident pension insurance payment standards and gradually reduce the urban-rural pension insurance gap.
Additionally, the minimum living security coverage and standards in the social security system can be further improved. By the end of 2023, minimum living security benefited 6.636 million urban residents and 33.997 million rural residents. The national urban minimum living security average standard was 785.9 yuan per person per month; the national rural minimum living security average standard was 621.3 yuan per person per month. While measures to improve the social security system are mainly aimed at reducing wealth gaps and maintaining social fairness and justice, with no direct relationship to macroeconomic regulation, these measures can play a role in increasing household consumption due to low-income groups' high marginal propensity to consume. The government should also consider the possibility of providing childbirth subsidies.
"Trade-in" policies are similar in nature to issuing consumption vouchers and will certainly play a positive role in stimulating consumption. But how effective these policies are deserves further study. Moreover, "trade-in" is also an industrial policy, and the potential negative impacts of industrial policies must be noted.
6. Vigorously Leverage Infrastructure Investment to Stabilize the Economy and Promote Growth
To avoid the circular reasoning of "to increase income we need to increase consumption, to increase consumption we need to increase income," we must find a "prime mover." This prime mover can only be government-financed infrastructure investment. Infrastructure investment immediately generates equivalent income, which then generates derivative income through new investment and consumption. For consumption, this creates a virtuous cycle of "infrastructure investment leads to increased income→increased income leads to increased consumption→increased consumption leads to increased income." The initial infrastructure investment will ultimately generate income multiple times the initial investment. Increasing infrastructure investment to stimulate consumption may be more effective than methods like issuing consumption vouchers, and will also increase potential economic growth. Besides infrastructure investment, the government can also support corporate investment, especially by innovative companies, through industrial policies.
From a "consumption promotion" perspective, infrastructure investment areas to consider include: First, infrastructure investment not directly related to consumption. While not directly related to consumption, such investment can increase residents' permanent income, thereby increasing consumption. Second, infrastructure investment to meet future consumption demands, such as people-centered new urbanization construction, various types of elderly care facility construction, professional caregiver training and wage subsidies, hospital construction, and nurseries. This forward-looking investment for future consumption not only helps compensate for current insufficient aggregate demand but can also increase the share of service consumption (no elderly care facilities means no elderly care expenditure) in consumption; it can increase not only current employment but also future employment (increasing the proportion of labor-intensive industries). Third, improving the supply of consumer goods and services (like dining, tourism, etc.) from the supply side—such supply can create demand.
In China's specific institutional environment, infrastructure investment is a policy variable that macroeconomic authorities can directly control. Besides compensating for insufficient aggregate demand currently, it can also increase China's potential economic growth. Using infrastructure investment rather than specific manufacturing sector investment as a policy variable is determined by infrastructure's "fundamental, public welfare, and long-term nature."
Long-term economic growth is inseparable from investment. While policymakers can make general judgments about future industrial development trends based on available information, except in individual cases, it's difficult to specifically decide how to allocate investment funds for the development of "high-tech industries" (such as semiconductors & chips, high-end machine tools, large aircraft, big data, artificial intelligence & robotics, new energy, big health & biotechnology, quantum computing & space economy, cross-border e-commerce & globalized services) and "traditional industries" that account for 80% of manufacturing output (such as machinery, electronics, chemicals, metallurgy, building materials, light industry, textiles) and corresponding specific product production.
Government-led infrastructure investment can not only provide important support for these industries' development but also avoid "overcapacity" in industries and products due to wrong decisions. For example, the government might not have known in advance that artificial intelligence would become the most important emerging industry, nor that electric vehicles would replace fuel vehicles as the mainstream in the automotive industry, but the government knows the importance of electricity supply for economic development. So it has long insisted on investing in new and old infrastructure like thermal power, hydropower, nuclear power, wind and solar, ultra-high voltage, charging stations, and comprehensive energy. This massive electricity infrastructure investment has laid a solid foundation for Chinese manufacturing development—regardless of what form that development takes—that other countries find hard to match.
An important "reason" for opposing government promotion of economic growth through increasing infrastructure investment is that China's infrastructure is "nearly saturated." This view is debatable. The CF40 research team estimates that China has at least 31 trillion yuan of incremental public investment space over the next 5 years. A report from an authoritative research institution indicates that infrastructure investment needed just for urban underground drainage systems amounts to 4.5 trillion yuan.
The December 12, 2024 Central Economic Work Conference proposed to "improve investment efficiency." Investment efficiency includes both commercial returns and economic and social benefits. Due to infrastructure's "public welfare, long-term, and fundamental nature," whether infrastructure investment projects succeed depends on their spillover effects.
Taking 5G as an example, China's 5G construction has accumulated investment exceeding 730 billion yuan. By the end of July 2024, there were 4.042 million 5G base stations and 966 million 5G mobile phone users. Investment was mainly borne by China Mobile, China Unicom, and China Telecom (some internet companies and equipment manufacturers also participated in 5G base station investment), with the government also providing some support. Without base station construction, the development of Alibaba, Tencent, and Huawei would have been impossible; without massive power generation capacity, where would big data computing power come from? Where would related startups develop?
We cannot require public investment projects to achieve commercial returns in the short term. The three major telecom operators engaged in base station construction are profitable. Some other projects, like high-speed rail, showed serious losses before 2023 according to China Railway financial reports, but the high-speed rail network has transformed China's entire economic and social life, with social benefits that are difficult to estimate.
China's infrastructure investment and construction capability is China's institutional advantage. Periods of insufficient aggregate demand are precisely when this advantage should be maximized. After the economy fully recovers autonomous growth, government-led infrastructure investment can gradually withdraw along with the exit of macroeconomic stimulus policies.
Undeniably, past infrastructure investment has indeed had serious problems with waste and redundant construction. In December 2024, the State Council office released a list of prohibited local government special bond projects. Such lists are very necessary and should be used as important criteria for infrastructure project approval. Large-scale infrastructure investment projects may even breed corruption. But such problems can be solved through legal and political means; we shouldn't throw out the baby with the bathwater.
7. The 15th Five-Year Plan Should Arrange Major Projects Similar to the "4 Trillion" Stimulus
In a 2016 article, former US President's Council of Economic Advisers Chairman Furman pointed out that compared with 1994-2004, almost all G7 countries saw significant declines in capital deepening rates and total factor productivity growth from 2004-2014. According to Solow's growth theory analytical framework, labor productivity is determined by capital intensity (or per capita capital equipment) and total factor productivity. The simultaneous decline in aggregate demand and labor productivity growth in G7 countries was not accidental. Insufficient demand led to serious underinvestment. The decline in capital intensity growth was the most important reason for declining labor productivity growth in G7 countries. Japan saw the largest decline in capital intensity growth at 3%.
Furman's research shows that increasing aggregate demand rather than reducing investment growth is a necessary condition for improving labor productivity (i.e., per capita GDP).
The government has already formulated fiscal policy for 2025 with expansion far greater than before, with the general public budget deficit ratio set at 4% for the first time. But given the severe external challenges China faces in 2025, whether China's fiscal policy expansion is sufficiently large seems open to discussion. During the 15th Five-Year Plan period, could the government arrange some major projects similar to the 4 trillion stimulus plan period?
For example, further advancing the "Western Development Strategy," conducting large-scale infrastructure investment along the Hexi Corridor toward Central Asia, and establishing corresponding economic corridors. Central Asia is located in the center of the Eurasian continent, a key region connecting the heartland and periphery of Eurasia. Linking Western development with building Central Asian economic corridors has important significance for stimulating domestic demand and strengthening China's domestic security and geopolitical position.
There are numerous investment opportunities in the West. Taking Xinjiang as an example, Xinjiang's railway operating mileage reached 9,252 kilometers (end of 2023), but network density is still only 45% of the national average. The China-Kyrgyzstan-Uzbekistan Railway (Xinjiang section) also has huge investment gaps; in 2024, Xinjiang's wind and solar installations exceeded 60GW (18% of national total), but supporting energy storage was only 2.4GW/4.8GWh; for communications, Xinjiang's administrative village 5G coverage rate is 76% (national average 92%), with the remaining 24% concentrated in border counties; Xinjiang's compulsory education school standardization rate is 85% (national average 95%), with huge human and financial gaps in southern Xinjiang for teachers, school building renovation, and grassroots practicing physicians and medical equipment updates, etc.
If not now, when should we undertake such multi-benefit infrastructure investments? We look forward to the NDRC quickly releasing a project list for the 15th Five-Year Plan period.