Monday, January 2, 2017

China Sow Inventory Statistics Explained

An official explained how the Ministry of Agriculture generates its data series on monthly swine inventories--the first explanation of the widely-disseminated series since the data was first released seven years ago.

The explanation was given to a reporter from Southern Rural News who explained that the accuracy of the data is scrutinized by people in the industry who watch the monthly releases for clues about the future course of hog prices. There are differing interpretations that stem from differing understandings of how the data series is constructed. Analysts and industry participants would like to be able to judge the accuracy of the statistics, he says. The dimsums blog described problems with the Ministry of Agriculture's livestock statistical system in a July 2017 post.

The deputy director of the Ministry's Livestock Industry Office explains that the monthly monitoring of changes in the number of hogs and the number of sows was launched after a pork price spike got the attention of government leaders during 2007. The annual estimates of swine inventories, slaughter and pork production were not timely enough to anticipate shortages and surpluses, so authorities commissioned several government departments to supply regular data on the hog and pork markets. A reporting system was set up and began issuing monthly data in January 2009. Numbers are posted on an official Chinese government site.

The swine inventory numbers are derived from reports on hog and swine numbers filed monthly by 4000 villages around the country. (The official's explanation implies that the villages in the survey are rotated from time to time.) Veterinary personnel or village accountants in each village count all swine within the village's boundaries, whether held by rural households, privately owned farms, or companies. Reports are sent directly to Beijing, bypassing local officials who might be tempted to inflate or otherwise distort the data. With direct reporting, the office in Beijing can check with the village clerks to verify information. After compiling the reports each month, the Ministry publishes the percentage change in sow and hog numbers from the previous month and from the same month a year earlier.

The Ministry only reports percentage changes. The official explained that the Statistics Law designates the National Bureau of Statistics as the official source of all statistics. No organization can issue statistics that duplicate or conflict with the Bureau's numbers. The official says the National Bureau of Statistics only reports hog inventories at the end of the year while the Ministry reports monthly changes in the herd. (In fact, some of the Bureau's quarterly reports include hog numbers that are inconsistent with the Ministry's numbers.)

The agriculture official says the Ministry's swine statistics may not be entirely consistent with the National Bureau of Statistics numbers since they use different methods. The Ministry has not conducted a census, so it cannot accurately estimate the national number, the official said (In fact, the Ministry reported monthly hog and sow numbers from January 2009 until March 2012. The National Bureau of Statistics has not conducted a census either since ten years ago--a new one started this month.)


The village-based monitoring may be outdated now that swine production is shifting away from small-scale household-based production to large-scale farms operated by companies. In theory, the village reporting system captures all types of farms within its boundaries, but it covers only 4000 out of more than 700,000 villages in the country. Hogs--and especially sows--are becoming concentrated in large farms that are less likely to be under the purview of a village accountant.

The Ministry is adapting its monitoring system to the new farm structure. The official acknowledges that large numbers of village-based household farms have quit the business, and explains that the Ministry is now preparing a "scale livestock and poultry farm cloud reporting system" to supplement the village reporting system. The new system has been bid out to a company and is still in the trial stage. If the share of livestock on "scale" farms reaches 80%, they will give more weight to the new farm reporting system.

Private sector analysts consulted by the Southern Rural News reporter have differing opinions on the usefulness of the Ministry of Agriculture data. The general conclusion is that the sow inventory change reflects the general trend in the industry. Sows at reproductive age represent the production capacity of the sector: they produce 8-12 piglets that will be raised to market weight (or held back to add to the sow herd). Thus, a shrinkage in the number of sows can portend tight supplies of hogs 6-12 months in the future, and an expansion could mean abundant supplies on the horizon.

The chart below shows the correlation between the monthly change in sow inventories and the ratio of hog to corn prices--a measure of profitability. Some of the monthly changes of 2-3% in a single month seem improbable, and the magnitude of the decline during 2014-15 also seems hard to believe. Overlooking statistical blips, the changes seem to correlate to changes in profitability. Low hog-corn price ratios below 6:1 tend to coincide with shrinking sow numbers while a high hog-corn ratio during 2011 coincided with an expansion of sow numbers. The Ministry's numbers show a massive shrinkage of sow numbers that continued from 2014 through 2016. The number of sows has failed to increase during 2016, despite a record-high hog-corn price ratio as high as 10:1. Is the Ministry's village monitoring missing an expansion by large farms that are outside its villages? Or is expansion of China's swine industry constrained by environmental regulations, lack of managerial talent, or high costs?

Does the expansion and contraction of sow numbers predict the future course of hog prices in China? The relationship is not very precise, but increases in price do appear to follow contractions in sow numbers. The large contraction in sow numbers during 2014-15 was followed by increases in hog prices that reached a record level in mid-2016.
The Ministry's numbers on the hog inventory raise serious questions, however. The number of hogs has been almost perfectly flat every month since January 2015 (did someone fall asleep at the switch?). The chart below displays estimates of the Ministry's hog inventory calculated based on monthly percentage changes. The National Bureau of Statistics has reported quarterly inventory numbers intermittently since then which show fluctuations in hog inventories, including an apparent recovery during the third quarter of 2016. Later this month, the Bureau will likely report its December 2016 hog inventory statistic.

No comments: