Wednesday, December 20, 2017

China's New Ag Census: Statistical Fog Remains

Initial results of China's 2016 agricultural census confirm that the country's farming sector remains shrouded in a statistical fog where numbers reveal only gray indistinct shapes whose details cannot be discerned with any precision. The skimpy initial release of statistics reveals nothing about what is produced, but they suggest China still has lots of people in the countryside squatting on plots of land they are not allowed to sell. Livestock and aquaculture farming--where land is less of a constraint--have proceeded much faster in commercialization and specialization than crop production. Scaled-up farms are increasing in number, but they have a long way to go to transform farming in China.

Last week, the National Bureau of Statistics released a series of five communiques, a graphical summary of changes over the last 10 years and two essays by communist party go-to academics which amount to cheerleading about how much life in the countryside has improved because of the wise policies of socialism with Chinese characteristics under Xi Jinping's leadership since the 18th communist party congress. The monumental effort reportedly engaged 4 million enumerators, drones, and satellite imagery to canvas the countryside's 230 million households, 600,000 villages, 40,000 towns, and 2 million agricultural business units to produce an inventory of rural China as of December 31, 2016. Statisticians claim to have covered all but 0.19 percent of China's farms.

While the census is hailed as "scientific," the initial reporting of results is highly political in nature. A theme of the reports is, "Look how much things have changed in the countryside," by emphasizing statistics on rural infrastructure, farm equipment, greenhouses, and scaled-up farms. There is nothing in these initial reports about what is produced, planted or sold--although the census did collect these numbers--and there is no effort shown to determine whether statistics have deviated from reality with so much change occurring in China's countryside. Detailed statistical volumes will be released later, which hopefully will reveal more about land use, sales value, and animal numbers.

When compared with the previous agricultural censuses conducted in 1996 and 2006, the 2016 census reveals a major outflow of people and consolidation of villages and townships. There has been massive migration to cities--agricultural employment reported by the censuses fell from 433.5 million to 314.2 million during 1996-2016. The number of administrative villages fell from 748,320 to 596,450, and the number of townships declined from 43,112 to 31,925 during 1996-2016.

Basic agricultural statistics in China's three censuses
Farming households
Other farming businesses
Cultivated land
Agricultural employment
Million hectares

Yet a comparison across censuses reveals that the number of agricultural households has actually increased gradually from 193 million in 1996 to 207.4 million in 2016. This reflects the continued formation of new households through population growth. As some family members move to cities, most families hold on to their allocation of village land since they can't sell it under China's collective land ownership system. Elderly parents are often left behind to tend the plots, watch grandchildren, and play mah jong. While many young people have no interest in farming, enough have stayed behind to take over land-holdings as elderly patriarchs pass on in order to keep the number of farming families steady or increasing slightly. The amount of land available is static or shrinking, so the allocation of land per household is also small and getting smaller.

Not all farms are operated by rural families, and the new census shows a big increase in the number of non-family farming units. In the 1996 and 2006 censuses, these other farming businesses included mainly collectives, state farms, government research farms, and probably prison camps and military farms. The increase in nonfamily farming units from 395,000 to over 2 million between 2006 and 2016 reflects a big jump in formation of farmer cooperatives after the 2007 cooperative law and formation of private agricultural enterprises (most of the cooperatives are aggregations of farmers who cultivate their own land). In fact, one of the few interesting results of the census is its tacit admission that only half of the 1.8 million farmers cooperatives officials brag about actually exist. The census notes that 1.78 million farmer cooperatives are registered with commercial bureaus, but census enumerators counted only 910,000 cooperatives. This confirms reports from the field that many cooperatives were set up as bogus shells to collect subsidies or to pad statistics, and many others just failed.

The most disappointing aspect of the census is its failure to update China's estimate of cropland area. The new census endorses a Ministry of Land Resources survey from 2008 that reported 135 million hectares of cultivated land, nine years ago. Uncertainty about the area of land under cultivation dates back centuries--landlords have always under-reported land holdings to evade taxes, a practice that continued under communist authorities until land-based taxes were eliminated 10 years ago.

Gyrations in estimates of China's farmland since the 1990s are an indicator that no one really knows exactly how much land China cultivates. Statistics reported a steadily declining stock of cultivated land starting in the early 1960s that reached 95 million hectares in 1995. At the time, everyone knew this was an understatement, which was confirmed by the 130 million hectares reported by the first agricultural census for 1996 (it took several years to report this number, probably because of bickering behind the scenes). The second agricultural census found a seemingly plausible shrinkage in farmland to 122 million hectares in 2006. However, the Ministry of Land Resources survey in 2008 reported an even larger 135 million-hectare estimate of farmland which has been reported as the cultivated land area in each statistical yearbook since then.

Uncertainty about the size of China's cropland base means that no one really knows how much is produced in the country. Crop output is based on the annual estimate of area planted in crops which exceeds the area of cultivated land, due to double-cropping, and perhaps planting of crops on grassland, hillsides, riverbeds, etc. that are not classified as "cultivated land." Statistics on the area planted in crops have been unchanged during the wild gyrations in cultivated land statistics. Between 1995 and 2016, crop area planted increased from 150 million hectares to 165.65 million hectares. The area planted in crops has increased 9.5 percent over the most recent 10 years--an implausible trend in view of the amount of land gobbled up by real estate, roads, and ecological projects. No estimate of area planted in crops was reported in this initial set of communiques on the 2016 census.

(In comparison, the area of cropland reported by U.S. agricultural censuses declined from 180 million hectares in 1997 to 158 million in 2012--the most recent U.S. census).

China's census suggests that there are more people engaged in farming than other statistics report. The 2016 census reported that 314 million people were employed in agriculture, which equals 40 percent of total national employment (776 million) reported for that year in other publications. The estimate of employment in the "primary" sector published in Statistical Yearbooks is much lower, at 215 million (28 percent of all employment). Agriculture's share of GDP is much smaller than its share of employment--less than 9 percent of GDP in 2016, according to macroeconomic statistics.

As noted above, the census agricultural employment totals have fallen over time, but the rate of decline differs from that shown by annual employment statistics. The censuses report a decline in agricultural employment of over 100 million between 1996 and 2006, but only a 28-million decline between 2006 and 2016. Annual estimates of employment in "primary industry" reported by the statistics bureau show a much steeper decline in ag employment of 104 million during 2006-16--four times faster than the decline reported by the censuses.

China's agricultural labor force is aging. The 2016 census reported that 33.6 percent of agricultural workers were age 55 or older and 19.2 percent were under age 35. The 1996 census reported a much younger rural labor force with only 11.8 percent of rural workers 55 and older and 53.5 percent under age 35 (the 1996 census did not tabulate agricultural employment separately, but most rural people were employed in farming at that time). The youthful labor force in the 1990s reflected a bulge of young people born in the early 1970s--a cohort that fueled China's labor-intensive manufacturing boom during the 2000s as they migrated to factories. (The 2006 census used a different set of age-groupings, reporting that 30 percent of ag workers were over age 50.) The percentage of farmers 55-and-older translates to over 100 million people who will probably never switch to another profession or move to a city.

(An aging farmer population is not unique to China. The average age of farmers reported by U.S. ag censuses has been over 50 since the 1970s. The average age of U.S. farmers rose from 54 in 1997 to 58.3 years old in 2012.)

Education-level tabulations indicate that China's agriculture sector is not attracting the best and the brightest. Education levels have gradually improved. The share of unschooled farmers has fallen from 14 percent in 1996 to just 6.4 percent in 2016, and farmers with just a primary school education have fallen from 42 to 37 percent of all farmers. Lower middle school (9 years of schooling) is now the predominant education level of farmers, comprising 48.4 percent of farmers. The share of farmers with a secondary school or higher education level is increasing, but highly educated farmers are still rare in China. In cities, secondary school is the most common education level, and a much higher proportion have higher education.

China's agricultural laborers by education level (percent)
Primary school
Lower middle
Secondary, vocational
Technical school or higher
*1996 data are for all rural workers.

The 2016 census is the first to profile scaled-up farms. The census found 3.98 million scaled-up farms, just under 2 percent of all farms. The census found that 28.6 percent of cultivated land was in scaled-up operations (the definition varied: 6.7 ha or more of land in areas where a single crop is grown; 3.35 ha in areas where two crops are grown; 1.67 ha for farms where crops are grown in plastic-covered sheds or greenhouses). There were 12.9 million people working on scaled-up farms, but this was just 4.1 percent of all agricultural employment.

The scaled-up farms had a smaller proportion of older workers, but only a slightly higher proportion of young workers under age 35, compared with the overall agricultural labor force. Thus, scaled-up farms had a higher proportion of workers at peak working ages of 35-54 than the general farm labor force. The scaled-up farmers were more likely to have a junior-middle school education and less likely to be unschooled or have a primary school education compared with the overall agricultural labor force.

The most striking feature of scaled-up farms is that they are much more likely to be engaged in livestock and aquaculture production than other farmers. Over 28 percent of scaled-up farmers are engaged in either livestock or fishing, compared with 4.3 percent of the overall agricultural labor force. The census also reported that 62 percent of the swine inventory was on scaled-up farms (slaughter 200 head or more--smaller than the 500-head definition used by the Ministry of Ag) and 73 percent of chickens (slaughter 10,000 birds or more) were on scaled-up farms. Again, the preponderance of tiny crop farms reflects the incentives and barriers imposed by the land tenure system. Land is collectively owned, but animals and fish are private property.

Characteristics of scaled up agricultural businesses
Scale farms All farms
People engaged (million) 12.9 314.2
Male 52.8 52.5
Age 35-less 21.1 19.2
Age 55-older 20.7 33.6
None 3.6 6.4
Primary school 30.6 37.0
Lower middle 55.5 48.4
Secondary, vocational 8.9 7.1
Technical or higher 1.5 1.2
Employment by sector:
Crops 67.7 92.9
Forestry 2.7 2.2
Livestock 21.3 3.5
Fishing 6.4 0.8
Services 1.9 0.6

Irrigation is used on 61.89 million hectares of China's farmland, 46 percent of the total. 30.5 percent of farms pump water from underground for irrigation and 69.5 percent use surface water from reservoirs, rivers, and canals.

Permanent greenhouses constructed of steel and glass cover 334,000 ha (up from 81,000 ha in 2006), and structures covered with plastic sheeting occupy 981,000 ha (up from 465,000 ha in 2006).

The census also conducted a more extensive inventory of farm equipment than in previous censuses. The number of tractors increased by 1.4 million between 2006 and 2016, but the composition of tractors was not reported--it likely shifted from small to bigger machines. The total of 26,900 tractors is one for every 7.7 farms or one tractor for every 5 hectares of land. The number of combine harvesters doubled between 2006 and 2016, to reach 1.14 million.

China's inventory of farm equipment
Type of equipment 2016 2006 1996
Tractors 26,900 25,500 11,790
Tillage equipment 5,130
Rotary cultivators 8,250
Planters 6,520
Rice transplanters 680
Combine harvesters 1,140 550 113
Mechanical threshers 10,310 7,519
Automated irrigation machinery 14,310
Hay processing machinery 4,090
Milking machines 100
Wool shearing machines 50
Aerators 1,940
Fruit tree pruners 490
Inland motorized fishing boats 280 467
Sea-going fishing boats 250


Shan Guo said...

MOA has been conducting household land registration since 2013, and their latest number shows cultivated land increased from 135 million to 152 million. See their press conference 11/19 below. The statistics on rural property land is also interesting

dimsums said...

Thanks for pointing out this press conference on the land registration and certification program. MOA has verified 1.52 BILLION MU of land, which equals 100 million hectares. Note: this is a little more than the 95 million ha number reported as cultivated land until 1995, which was derived from administrative records of the amount of collective land contracted to rural households.