“domestic demand"the word "has once again run through the main lines of various policies and the headlines of major media. from the china national knowledge infrastructure, statistics of ""domestic demandthe proportion of times the words "external demand" and "external demand" appear in academics and newspapers can be drawn as follows:

as you can see, last timelots of references to "domestic demand", from 2008 to 2010. at that timeseveral years have passed since china joined the wto. the global financial crisis has caused demand to shrink around the world. scholars have therefore turned their attention to domestic demand, hoping to reconstruct the driving force of economic growth.

with the rapid advancement of large-scale infrastructure construction and urbanization, first- and second-tieroutsidefrom then on, the city began to embark on the fast track of development. you know, before 2010, most third-tier cities and below did not have hypermarkets, movie theaters, high-end office buildings, or five-star hotels.. at the same time, the boom in e-commerce brought about by falling logistics costs has made it easier for various commodities to reach thousands of households, releasing the spending power of hundreds of millions of people.

it can also be seen from the figure above that from 2011 to 2018, the frequency of "domestic demand" being mentioned has significantly reduced - domestic demand is being rapidly released in reality, effectively stimulating economic growth, and naturally there is no need to mention it more.

but starting from 2018, the word frequency of “domestic demand” has increased year by year again. in 2023, the ratio of the frequency with which "domestic demand" is mentioned to "external demand" has exceeded that in 2009, and is second only to 2008, the peak of the crisis.

in 2024, boosting domestic demand is an urgent task, and the concentrated release of policies before the national day echoes such needs. however, market fluctuations are only short-term, and we still need to think about - how will china's domestic demand grow in the medium and long term? where are the possible drivers?

▍measuring the consumption power of different groups: are men inferior to dogs?

you must have seen this inequality in many articles - men's consumption power is not as good as that of dogs.

many readers may have seen the subsequent development of this inequality - men's consumption power finally surpassed that of dogs.

it is not difficult to hype this kind of gimmick, but it is really difficult to calculate the consumption power of a specific gender and age group. because consumption cannot be attributed to individuals, whether it is defined by who pays for it or who ultimately uses it, it is not particularly accurate. therefore, in all micro-survey data, consumption can often onlyasked at the household level -how much money did your family spend last month?

therefore, we must break down household consumption to each individual based on a series of assumptions.

let’s take the simplest example, with data from three families:

family a, a 30-year-old man living alone, consumes 5,000 yuan per month.

family b, a 30-year-old man and a 30-year-old woman live together, and their monthly consumption is 11,000 yuan.

‍family c, a 30-year-old man, a 30-year-old woman, and a 3-year-old child live together, and their monthly consumption is 15,000 yuan.

hypothesisthe 30-year-old men in families a, b, and c have the same consumption, and the 30-year-old women in families b and c have the same consumption. we can naturally draw the conclusion:

a 30-year-old man consumes 5,000 yuan a month, a 30-year-old woman consumes 6,000 yuan a month, and a 3-year-old child consumes 4,000 yuan a month.

using a similar method, the population of tens of thousands of surveyed households by gender and age are used as independent variables, and the least squares method is used to calculate the consumption of these households to calculate the average consumption of the population of each age group and gender. take the results of cfps data calculation in 2020 as an example, see the figure below:

the two curves in the above figure represent the average consumption of men and women in each age group in 2020. for example, in the picture above, the corresponding value for women aged 35-39 is 50,000 yuan, which means:every time a 35-39-year-old woman is added to a family, the family's total consumption will increase by 50,000 yuan.

it can be seen that in most age groups, the consumption power of women is significantly higher than that of men, and this is true from 25 to 60 years old. however, women's consumption has a significant inverted u shape. in 2020, the per capita consumption of women aged 35-39 is 50,000 yuan, while that of men of the same age is only 30,000 yuan; while the per capita consumption of women over 75 is only 20,000 yuan. ten thousand yuan, but for men it is still close to 30,000 yuan.

therefore, the statement that women have higher spending power than men,there doesn't seem to be any problem. only after the age of 70 can men overtake women with their ever-increasing medical and health consumption.

however, it does not seem correct that men consume less than children - from the figure, men consume much more than children. children do not cost as much as we think. for the vast majority of rural families covered by cfps data, it is still a situation of "adding more children means more food".

we calculated using the same method using data from many other countries, including:

u.s. consumer expenditure survey (Consumer Expenditure Surveys）, including four quarterly surveys each year, with approximately 10,000 households surveyed each time, totaling 40,000 households.

uk cost of living and food survey (Living Costs and Food Survey）, surveying 5,000-7,000 households every year.

japanese household panel survey (jhps and khps), surveying 4,000-6,000 households every year.

korea labor and income panel survey (Korean Labor & Income Panel Study）, surveying 15,000 households every year.

the 2020 consumption curves by age and gender in the united states, the united kingdom, japan, and south korea are as follows:

compared with china, the consumption curves of various genders and age groups in britain, the united states, japan and south korea have one thing in common - they are basically inverted v-shaped or inverted u-shaped, with middle-aged people being taller and the elderly and children being lower. it seems that the spending power of middle-aged men is significantly higher than that of children, and this is the same all over the world. 

there are two differences.

1，chinese women’s consumption far exceeds that of men in most age groups(the highest amplitude is about 60%), whilethe difference in consumption between men and women of all ages in the uk, the us, japan and south korea is not obvious, the only difference appears among the middle-aged and older populations in japan and the united states, and in this age group, toomen consume more than women。

2，peak consumption occurs at different times：

it can be seen that the age corresponding to the peak consumption in china is younger than that in other developed countries. this gap is precisely about "silver economy"theoretical sources of various expectations——

ifthe age group with the highest consumption in china is not around 40 years old, but over 50 years old. even like japan, male consumption remains at a high point from 50 to 74 years old. so, wouldn’t this wave of aging in china bring about new trends? a wave of consumption growth?

▍“silver economy”: what’s the future?

before discussing whether aging can bring about a silver-haired economy, we must clearly realize that the curves in the above figures show not only age effects, but also intergenerational effects.

age effect, which can be understood as whether you arepost-80s, post-90s or post-00s,do what you should do at that age.

intergenerational effect, that is, you have the characteristics of that generation, no matter how old you are.

in the picture abovetaking the average consumption of women in various age groups in 2020 as an example, it actually mixes age effects and intergenerational effects. the consumption of people in their 40s is higher than that of people in their 60s. on the one hand, this is because 40-year-olds are in their prime and have stronger spending power than 60-year-olds; on the other hand, because 40-year-olds were born in 1980 years, while 60-year-olds were born in 1960. the average years of education of the former is much higher than that of the latter. they also enjoyed longer economic growth dividends during the human capital accumulation stage, so their incomes are also higher.

age effects and generational effects are mixed together to get each of the curves we saw above for every country. the higher consumption of japan’s 60-year-old population is the result of the intergenerational effect of the japanese population born in 1960 and the age effect of the 60-year-old population. this conclusion cannot be directly applied to chinese data.

to calculate how much consumption china's 60-year-old population will have after they reach the age of 60, we must separate the age effect from the intergenerational effect.

in the above, we drew the consumption curves of each gender and age group in 2020. after doing the same calculation using the 2015 survey team data, we put the two female consumption curves together to get the following figure -

it can be seen that from 2015 to 2020, the highest point of women's age-specific consumption curve moved from 35-39 years old to 40-44 years old, a full 5-year shift to the right. this is easy to understand - the n-year-old population in 2015 and the n+5-year-old population in 2020 are exactly the same group of people. draw a small arrow between the adjacent ages of adjacent curves - for example, 20 years old, 35 years old, and 50 years old in 2015, connected to 25 years old, 40 years old, and 55 years old in 2020 respectively. the length of each small arrow , represents the consumption growth rate of the same group of people after they grow up to 5 years old. since both ends of each arrow point to the same generation, this method canremove the intergenerational effect and only retain the age effect of consumption growth。

we use micro-data from the survey corps from 2001 to 2015, and cfps data from 2016, 2018, and 2020(peking university family tracking survey)and chfs data for 2017, 2019, and 2021(southwestern university of finance and economics household finance survey), a summary of up to21 yearshousehold consumptionqueuedata, using this data, we can do a calculation every 5 years.

for example, if the population aged 25-29 years old grows for five years, how much will their consumption increase when they are 30-34 years old? using the population aged 25-29 in 2001 and the population aged 30-34 in 2006, it can be calculated once; the data in 2002 and 2007; the data in 2003 and 2008... until the data in 2016 and 2021, all can be calculated once. at this time, for each gender and age group of the population, we have 16 sets of consumption growth rate values. putting together the 16 sets of values ​​for men and women of each age group, there are 32 growth rate values ​​for one age. the 32 growth rates for each age group are as shown in the figure below:

it can be seen that the points in the figure above form an obvious growth curve by age - from 20-24 years old to 25-29 years old, in most years, it is the age group with the fastest growth in consumption. after the age of 30, the consumption growth rate declines until around the age of 50-60, reaching the lowest point of consumption growth. after the age of 65, it picks up slightly again - but still far below the growth seen between the ages of 25 and 29.

another piece of information that can be seen from the picture above is that people born in different years are in the same age group.consumption growth modelit's very close.

this pattern has little to do with people's birth years. people often use different years of birth to classify people into different generations, and give them all kinds of cool names.but 20 years of consumption panel data tells us thatmaybebecausetechnologydevelopleveldifferent,differentgenerationsconsumptionofcommoditythere is a difference，butfromconsumptionincreasesuperiorcome and see，doing something when you reach a certain age is the mainstream that humans have repeated for hundreds of generations.。

similar phenomena can also be seen in the data of the united states, the united kingdom, japan, and south korea. the u.s. data has the longest time series, with micro-public data provided since 1980, and there are already 44 samples. we use the earliest sample (1985) and the latest sample (2022) that can be calculated for comparison. the results are as shown in the figure below:

it can be seen that the 20-year-old population in 1985 and the 20-year-old population in 2022, the former was born in 1965, and the latter was born 37 years later in 2002, but their consumption growth rate when they were 20 years old was exactly the same. the shapes of the two curves are almost exactly the same - not only are they consistent in time, but they are also almost consistent with china's curve. they both reach the lowest point of consumption growth in life around the age of 50-60, but the highest consumption growth rate is in the united states. it appears at the age of 20, and china appears at the age of 25. there is a slight difference between the two.

the five-year consumption growth rates of various age groups in the united kingdom, south korea, and japan are shown in the following three figures:

it can be seen that the five-year consumption growth rates by age in these countries are also similar.

from the above data spread across time and space, the following inferences can be drawn:

1，around the age of 20-30, people always reach the peak of consumption growth in their lives.

2，around the age of 50-60, people always reach the lowest point in consumption growth.

3，the above conclusion is equally valid whether in china or the united states, whether born in the 1960s or the 00s.

this istwo windows for consumption.20-30 years old, is a high window for consumption;50-60 years old, is the low window for consumption.

existhigh window for consumption, young people walk out of college and enter the society. they begin to have their first income, have their own life plans, and want to live a better life.populations born in different countries and at different times have all experienced accelerated consumption growth in this range.

existlow window for consumption, people's physical strength, intelligence, and professional competitiveness have begun to decrease, income growth has slowed down, and retirement is approaching, and children are generally able to stand on their own feet, so they can breathe a sigh of relief.populations born in different countries and at different times have all experienced stagnation in consumption growth during this period.

between the two windows, from 30 to 50 years old, there is a period of decline in consumption growth. every time one age increases, the absolute level of consumption may increase, but the growth rate declines year by year.

therefore,the post-60s generation, whose current consumption is not high, is unlikely to bring consumption growth in the next few years because they are in the age group with the lowest consumption growth rate in their lives.

some readers may ask, since the consumption growth rate of 50-60 year olds is so low, why is the consumption of japanese people aged 60-70 still so high?

the reason is very simple. people in japan in their 60s and 70s were born during japan’s economic take-off after world war ii. they are the generation with the fastest growing educational level in japan. when this generation enters the labor market, they have gained more than theirs. higher wages and consumption levels of predecessors - consumptionintergenerational effectmuch higher than its predecessors, so at the sameage effectunder the influence, consumption levels can remain at the highest value for longer.

there may be absolute wage differences between one generation (intergenerational effect), but their life paths have not changed much (age effect). therefore, we naturally cannot expect that the post-60s generation, whose intergenerational consumption level is relatively low, will suddenly experience consumption growth that exceeds the age effect.the prospects for the silver economy are not ideal。

▍looking at domestic demand from the perspective of population: an unoptimistic future

under the stable age effect, we use the population of different age groups, multiply them by the total consumption of this age group, and then multiply them by the consumption growth rate of this age group to calculate the estimated average consumption growth rate of the entire society.

can the consumption growth rate calculated in this way match the consumption growth rate in the real world?

we use the population of each country by gender and age group from 2015 to 2019 given by the unpopulation project, multiply it by the consumption growth rate, calculate the average, and compare it with the consumption statistics of various countries in the penn world table from 2015 to 2019. , get the following picture——

the horizontal axis of the above figure is the real growth rate, and the vertical axis is inferred based only on population characteristics.growth rate. it can be seen that there is a high correlation between the real growth rate and the inferred growth rate calculated using the consumption growth rate of each age group. using only the single variable of population, it has reached a statistically significant level of 0.1%. the overall goodness of fit also reached 0.4.using a fixed consumption growth rate of the population of each age and multiplying it by the age structure of each country, we can well explain the current differences in consumption growth rates between countries!

then, if we continue to use this stable consumption growth rate of each age and multiply it by the predicted population of each age group and gender, we can backward predict the future consumption growth rate of each country and region.

median forecast version in 2024 version using unpopulation, you can draw the following figure, which lists the consumption growth curves of china, the united states, britain and japan from 2000 to 2100:

it can be seen that china’s estimated consumption growth rate before 2020 is the highest among these four countries, but theextrapolatethe growth rate will decline rapidly from around 2020, and will be lower than that of the united states and the united kingdom. there will be a wave of rebound for more than ten years starting from 2032, and the second wave of rapid decline will begin around 2043, and will be lower than japan around 2050.

the reason for the decline in china's projected growth rate can be seen in this chart -

as time goes by, during the high consumption window, the young population will continue to decrease and stay at the trough for a few years.

during the low window of consumption, the baby boomers who rushed in before delaying retirement gradually reached retirement age.

on the slippery slope of consumption, the post-80s generation, which has the largest population, has all left the high consumption level.the growth window is gradually moving from youth to middle age, and is slipping towards the low window of consumption.

in the final analysis, if the number of people in a region with high consumption windows is much greater than that with low consumption windows, then domestic demand in this region will most likely grow rapidly; conversely, domestic demand growth will be difficult to avoid entering a sluggish state.

all data currently point to the same result: from a demographic perspective, china's domestic demand will be extremely difficult to pick up before 2032.

▍when three mountains pass through two windows

it is not difficult to find that the estimated growth rate of domestic demand in china is not consistent with the real growth rate of domestic demand. if the age pattern of population consumption were completely followed, china's growth should have started to decline since the beginning of this century, but the fact is exactly the opposite.

let the real world growth be higher than the growth using population projections. what did we do right in the past?

let's go back to the picture above. there are three "high mountains" in china's population. those born around 1963-69 are on the first mountain; those born around 85-90 are on the second mountain; those born between 2012 and 2017 are on the third mountain (although the height is much lower than the first two) .

every wave of challenges in china's economy occurs exactly when every mountain leaves the window of high consumption.。

when the first high mountain born in 1963-1969 left the high window, it was in the mid to late 1990s. the decline in demand is intertwined with the transformation of the domestic economic structure. the first wave of real estate boom has waned, and state-owned enterprises have a serious backlog of products. at this time, china used large-scale layoffs, reform of state-owned enterprises, and entry into the wto to introduce external demand to overcome this difficult period.

when the second high mountain born in 1985-1990 leaves the high window, it is the period before 2015 to 2020. at this time, china launched a vigorous downward expansion of consumption, from real estate shantytown renovation, to the decline of home appliances, to various urban consumption scenarios settling in counties and villages. the growth during this period is defined as "consumption sinking", which has expanded china's consumption plate.

in the past, when the demographic structure brought challenges, we responded by expanding external demand and expanding the scope of domestic demand, which produced a strong "intergenerational effect" and brought about actual domestic demand growth that was higher than the projected growth rate.

but now, it seems difficult to go further in either direction.

in the complex international political environment, the growth of external demand is no longer what it was 20 years ago. as long as it is cheap and easy to use, you can occupy the international market.

the space for consumption sinking within china has basically been exhausted. as we said in "mentioned in the article,in slightly more developed areas, box office growth has long since stagnated.now only the areas with the lowest population density and the most significant population outflow are still experiencing growth.

the second population peak, those born between 1985 and 1990, has completely left the window of high consumption and is entering the track of declining growth. however, the third population peak - the population born between 2012 and 2017 - still has many years to go before it reaches the peak. consumption window.

according to the calculation above, until2032, the impact of population growth on domestic demand growth will turn from negative to positive. before that, the growth rate of domestic demand will only continue to decline under the influence of demographic structure.

how to spend the next 8 years? how to find new driving forces?

in the short term, this driverit may be that the trade-in policy subsidy accelerates households' replacement of durable goods; it stops household deleveraging, increases the wealth effect of the stock market and real estate market, and stimulates consumption.

but in the longer term, we still need to rely on more in-depth reforms. internally, we will adjust household registration, employment security and other systems that cannot adapt to the current situation and promote the construction of a unified market. externally, look for new sources of external demand outside the united states and europe. only by taking a multi-pronged approach and finding new growth points can we fundamentally ensure the stability of people's employment and the expected recovery of income growth.

in fact, consumers’ wishes are very simple:

hope tomorrow is better than today.