RaboResearch - Economic Research

China’s economic recovery: A tale of two speeds


  • China’s economy is recovering from the shock of Covid-19
  • However, the recovery is quite uneven, in multiple dimensions
  • Services and retail are recovering less than production
  • Online retail has picked up some of the slack, but is only a quarter of total retail sales
  • The production side of the economy is doing better but is not without disparities either
  • The production of ‘stay at home’ items such as TVs and smartphones is doing much better than ‘working from the office’ items such as copying machines
  • Covid-19 has increased income inequalities and regional differences
  • If the rising inequality intensifies that could put pressure on social stability
  • Headline inflation figures hide wide disparities as well, especially between food and fuel prices
  • Lopsided trade growth could lead to trade tensions with countries where China exports more than it imports
  • All these imbalances will likely be around for some time as China recovers
  • China’s situation might give a glimpse into imbalances in the recovery of other countries as well

The economy is recovering

China’s economy is recovering from the blow of Covid-19 (Figure 1). However, the headline growth figure disguises the fact that some parts of the Chinese economy are not recovering nearly as well as others. China’s National Bureau of Statistics (NBS) breaks its economy down into three main sectors: Agriculture, Production and Services. China’s rebound has been driven mostly by production, while the services sector has not done as well (Figure 2).

This matters because services (i) are a better indicator of consumer demand than production, (ii) represent 48% of China’s GDP and (iii) employ 46% of its labor force. Therefore the fact that the services sector is not growing as fast as the production sector, indicates that consumer demand is lagging behind production.

Figure 1: GDP growth in China is coming back…
Figure 1: GDP growth in China is coming back…Source: Macrobond, NBS
Figure 2: …however, it is driven more by production than by services
Figure 2: …however, it is driven more by production than by services Source: Macrobond, NBS

Services are not recovering as fast

When we take a closer look within the services sector, we see a clear difference between parts of the economy that are sensitive to containment measures (such as tourism and recreation) and those related to online services. For example, Hotels and Catering and Leasing and Rental services were still down by 5% y/y and 7% y/y by Q3 respectively, while Information transmission, Software and IT services soared back up by 19% y/y (Figure 3). If we take the average of the sectors that were up by Q3 and those that were down, we see that the ‘down’ sectors saw a sharper decline and less of a bounce back (Figure 4). Since the total services sector grew by 4.3% y/y in Q3, the sectors that grew more than made up for the ones that declined. Nonetheless, the differences between the two are stark, showing that there are two stories to tell: one in which the bounce back is strong and one where it really is not.

Figure 3: Internet services have rebounded sharply
Figure 3: Internet services have rebounded sharplySource: Macrobond, NBS
Figure 4: Hotel and rental services have dropped deeper and rebounded less
Figure 4: Hotel and rental services have dropped deeper and rebounded lessSource: Macrobond, NBS

Much of the retail sector is still below normal

Figure 5: Retail sales of most product categories have not fully recovered yet
Figure 5: Retail sales of most product categories have not fully recovered yetSource: Macrobond, NBS

Although improving, consumer demand in the retail sector, has clearly not yet returned to pre-pandemic levels. Overall retail sales were still down by 7.9% ytd y/y in October[1]. In fact, retail sales of most product categories were still negative on a ytd y/y basis in October (Figure 5). Items related to working from or staying at home (such as office appliances bought by consumers) are up, while most categories, especially clothes, shoes and furniture, are still down by 18%, 19% and 37% respectively.

Isn’t the online channel picking up the slack?

As expected, the shift to buying things online has accelerated in China since Covid-19. It is tempting to think that the online channel has picked up the slack, or at least significantly dampened the blow. However, that is not the case. Online sales of consumer goods did indeed increase by 16% ytd y/y in October (Figure 6). However, online goods sales are only 24% of total retail goods sales (Figure 7), and have been unable to turn the tide for the overall retail sector. Moreover, growth in the online channel has been anything but homogenous. Chinese people have been buying more food online, but they are buying less clothes online, so it matters which sector one is looking at. Moreover, online sales are mostly goods. Services make up only 17% of total online sales. This makes sense since many services cannot be delivered online, for example a haircut. This limits the extent to which the online channel can replace the broad retail sector, especially in services.

Figure 6: The growth of online consumer goods sales is driven by daily necessities
Figure 6: The growth of online consumer goods sales is driven by daily necessitiesSource: Macrobond, NBS
Figure 7: Online consumer goods sales are still only a quarter of the total
Figure 7: Online consumer goods sales are still only a quarter of the totalSource: Macrobond, NBS

Manufacturing is doing well, mining isn’t

Similar to the services sector, on close consideration the industrial production part of the economy also tells two different stories, most notably regarding the mining sector. Mining was down ytd y/y by -0.2% (in October), while manufacturing was up by 2.4% (Figure 8). One could argue that mining is a relatively small sector in China (it made up only 4% of China’s total industrial revenues in 2019, while manufacturing made up 88%). However, there are big differences within manufacturing too. The NBS distinguishes 32 different types of industry categories within manufacturing. A look at the top and bottom 5 of these categories (in terms of the growth in valued added) shows that non-essential items and recreational activities-related products are taking big hits, while high-tech electronics and mobile communication are doing well (Figure 9).

Figure 8: Mining is down year to date
Figure 8: Mining is down year to dateSource: Macrobond, NBS
Figure 9: Big differences between the top and bottom 5 industry categories
	Figure 9: Big differences between the top and bottom 5 industry categoriesSource: Macrobond, NBS

An even closer look at the product level corroborates this. The NBS distinguishes 71 industrial products. For more than half of these products (36), the amount produced until October this year is less than it was for the same period last year.

Figure 10: The production of staying home items is up, while that of going out items is down
Figure 10: The production of staying home items is up, while that of going out items is downSource: Macrobond, NBS

Figure 10 shows the ytd y/y percentage change in output for some of these industrial products. The production of items one would associate with people spending more time at home and online (such as freezers, smart phones and TVs) did indeed increase. The production of items related to eating out and going to the office, such as soft drinks and processed sugar (which could be associated with fast-food restaurants) and copying and printing machines (which could be associated with offices), did indeed decline. The increase in the production of plastics (5% ytd y/y) could reflect the increased production of protective equipment, which China has been exporting to other countries dealing with a second wave of coronavirus cases.

Two additional items are worth highlighting here: cars and semiconductors. Car production was down by 6% ytd y/y, while retail car sales seem to be bouncing back (increasing 8% y/y in October). This indicates that most of car sales are coming from current inventory, which could reflect that car manufacturers are still hesitant to increase production and might not be particularly optimistic about the near term prospects for the automobile sector in China. Moreover, it’s worth noting that the recovery of luxury items, such as car sales, could give a biased impression of China’s overall economic recovery since a sizable part of China’s population cannot afford to buy a car.

The production of semiconductors increased by 15% ytd y/y in September, which might be a prelude to China’s increased focus on self-reliance in the technology sector. As we highlighted here, a major theme in China’s recent Five Year Plan seems to be one of self-reliance, especially in high-tech products such as semiconductors.

Inequality has risen further since Covid-19

Urban vs rural areas

Disposable incomes in China have not rebounded nearly as fast as overall GDP. While GDP growth crept up to 4.9% y/y in Q3, per capita disposable incomes in rural areas only increased by 1.6% y/y in the same quarter, and those of urban households even decreased by 0.3% y/y[2] (Figure 11). Moreover, income inequality between China’s urban and rural population has risen since Covid-19, at least in nominal terms. Namely, the difference in average disposable income between people living in urban areas and those living in rural areas, has increased from CNY 20.3K to CNY 20.5K in 2020[3] (Figure 12). Admittedly, the increase is not large, but that overlooks the fact that urban incomes are 2.7 times as high as rural incomes.

Figure 11: Disposable incomes have not rebounded as quickly as overall GDP
Figure 11: Disposable incomes have not rebounded as quickly as overall GDPSource: Macrobond, NBS
Figure 12: The income gap between people living in urban and rural areas has increased further
Figure 12: The income gap between people living in urban and rural areas has increased furtherSource: Macrobond, NBS

Moreover, the difference could widen further. Almost 22% of the total disposable income of people living in rural areas consists of transfers, which is typically money sent home from relatives living in urban areas. These so called ‘rural migrant workers’ are people who live and work in urban areas but are officially registered in rural areas. In many cases rural migrant workers do not receive unemployment benefits and are not counted in official unemployment figures[4]. China has about 290 million workers according to some estimates, which makes them more than a third of China’s labor force. An important implication here is that China’s actual unemployment could be much higher than official unemployment figures: some estimates go as high as three times as much. On the other hand, if these people do go back to work, official unemployment will not drop either. In any case, it is difficult to gauge the pain in China’s labor market by looking at official unemployment figures alone.

If migrant workers that have lost jobs fail to get them back, they will also send back less money to their rural relatives. That might very well be the case for migrant workers in industries that are still struggling. As an example, according to a recent Bloomberg article, the catering industry in China employed 12 million migrant workers.

Inequality within urban areas

Within urban areas, income inequality has increased as well. The urban per capita disposable income of the regions with high incomes is double that of regions with relatively low incomes (Figure 13). And the difference between the two groups increased in the first half of this year by 3%, from CNY 14.6K to 15.1K (Figure 14). Hence income inequality between China’s urban regions has increased; and such inequalities sit alongside the rising gap between the rural and urban populations. The IMF flagged this issue for Asia in general in its recent Asia Outlook, and Xi Jingping has indicated that he wants to tackle inequality, mentioning in a recent speech that China should achieve “common prosperity” by 2035. However, as in all countries, exactly how this can be achieved is likely to prove a challenge.

Figure 13: There is a big income gap between the top and bottom 20% incomes
Figure 13: There is a big income gap between the top and bottom 20% incomesSource: Macrobond, NBS
Figure 14: The income gap has increased since Covid-19
Figure 14: The income gap has increased since Covid-19Source: Macrobond, NBS

Regional growth differences

Even before the corona crisis, there were big regional differences in GDP growth rates in China. These differences have increased. Provinces such as Tibet and Guizhou recovered by 6% ytd y/y and 3% ytd y/y respectively by September, while bigger regions such as Hubei and Shanghai (which together represent 9% of China’s GDP) were down by 10% ytd y/y and 0.3% ytd y/y respectively (Figure 15). In part these differences represent the different sectors within each region. Hubei, for example, has a large automotive sector and chemical sector[5]. Production in these two sectors was down by 6% and 5% ytd y/y by September this year. More broadly speaking, the eastern regions are doing ok (having grown by 2% ytd y/y), whereas the north eastern and south central regions are struggling (down by 0.5% and 0.6% ytd y/y respectively).

What this picture also shows is that region really matters when commenting on whether the economy is recovering or not. And it matters more now than it did before Covid-19. Another way to show this is by looking at the dispersion of growth rates between the different regions. The standard deviation of the different regional (real) growth rates was 1.2 in 2019, and this has increased to 2.7 in 2020 (Figure 16). So the distribution of regional growth rates has become wider.

Figure 15: The headline GDP growth figure masks big regional differences
Figure 15: The headline GDP growth figure masks big regional differencesSource: Macrobond, NBS
Figure 16: And these regional differences have increased since Covid-19
Figure 16: And these regional differences have increased since Covid-19Source: Macrobond, NBS

Headline inflation figure hides disparities

Inflation in China might be underestimated (and purchasing power overestimated), at least in the short run. The reason is that the consumption baskets of Chinese people have not been in line with official CPI weights since the outbreak of the coronavirus. Clothes, for example, have a 9% weight in China’s CPI basket (Figure 17). However, we as we saw earlier in Figure 5, Chinese people are not buying as many clothes as they used to. In addition, food and healthcare related articles have arguably become a larger part of the consumption basket. Meanwhile, food prices have been higher than headline inflation, while clothes and transportation (fuel) have seen deflation (Figure 18).

Admittedly, food price inflation this year was mostly driven by the African swine fever outbreak in 2019 (which led to lower pork supply). These effects were still being felt in 2020, so these numbers should be interpreted with some caution. Also, part of the shift in consumption pattern will, of course, change back. However, there are also likely to be some permanent shifts, for example related to working from home. Until CPI weights are adjusted to reflect the new Chinese consumption basket, CPI figures will give a biased picture of Chinese purchasing power.

Figure 17: CPI basket weights are not representative of what people are buying
Figure 17: CPI basket weights are not representative of what people are buyingSource: Macrobond, NBS
Figure 18: Inflation has been higher in items that are now a larger part of consumption
Figure 18: Inflation has been higher in items that are now a larger part of consumptionSource: Macrobond, NBS

Going forward, there is a risk that inflation will rise quickly next year as China’s economy continues to recover. The PBoC does not officially target inflation, its policy objective is to “maintain the stability of the value of the currency and thereby promote economic growth”. Monetary policy could still tighten next year, for example if inflation stays high long enough to become entrenched in future inflation expectations. The tightening will be modest though, as the PBoC will also take the stability of CNY into account. Tighter monetary policy would be a positive for CNY strength, which the PBoC might not want since it hurts China’s export competitiveness.

In any case, the PBoC will be stuck between a rock and a hard place next year. On the one hand, tightening monetary policy risks halting China’s recovery, while on the other hand, no tightening risks the loss of purchasing power of Chinese consumers in the future.

A lopsided recovery in trade as well

China’s trade account has also shown a lopsided recovery, with imports growing much less than exports. Over the three months from August to October China’s exports (in USD terms) grew on average by 10.3% y/y, while imports only grew by 5.3% y/y. This implies that China’s overall trade surplus is increasing. Even though that is positive from an economic point of view, it might not be that great from a geopolitical perspective. After all, the large trade deficit with China (a surplus from China’s perspective) was one of the key reasons US president Trump mentioned for implementing import tariffs on Chinese goods. If China’s trade surplus rises further, it could form the basis of similar tensions with other countries and further tensions with the US.

Broader implications

China’s unequal recovery highlights that it is becoming more important for businesses and investors to pay attention to the sectors and regions in which they are active in and which socio economic consumer group they target. Some sectors will not go back to their pre-pandemic levels anytime soon, or at all. Online sales, at least in the retail sector, are not fully compensating the loss of sales in physical stores. Moreover, the coronavirus has increased inequality in China. If this persists, it could pose problems for social stability down the line.

From a geopolitical perspective, if China’s economic recovery continues to be stronger on the export side than on in the import side, that could potentially lead to trade tensions with the countries where China has an increasing trade surplus.

Finally, since China’s battle with the coronavirus started about a quarter before the rest of the world, its situation might be a glimpse into the future recovery of other countries, which might harbor similar imbalances. 


[1] When referring to percentage changes in ytd y/y terms, we mean cumulative level of a flow until Q3 (or October if the data is available) compared to its cumulative level until Q3 (or October) a year ago.

[2] Disposable income refers to income that can actually be used for consumption. The official NBS definition is “total income minus personal income tax, household subsidy and expenditure on household sideline production”. Note that this income includes wages, business income, income from property, and transfers (for example money sent home from migrants). Also note that comparing overall GDP to per capita disposable income is not entirely fair. If a country’s population grows, its per capita GDP growth will be less than its overall GDP growth. However, China’s population growth was only 0.3% in 2019 so the distortion is likely small.

[3] These figures are based on the first three quarters of 2019 and 2020, to make them comparable.

[4] This is because of a registration system in China known as Hukou. One of the consequences of this system is that people that work in an area different to where they were born do not receive social welfare services such as unemployment benefits and are also not counted as unemployed.

[5] According to data from the NBS, 10% of China’s total Fertilizer and Chemical Pesticides and 9% of its total vehicle production is in Hubei.

Raphie Hayat
RaboResearch Global Economics & Markets Rabobank KEO
+31 88 726 7864

naar boven