Update and Revision History
Indexes for international agricultural total factor production were first published on the ERS website in November 2013, and covered the period from 1961 to 2010. In October 2014, an updated version was posted covering 1961-2011. Each update made in October since has extended the series with one year of data. The latest update (October 2017) covered 1961-2014. With each new posting, historical estimates are revised to reflect newly available data and/or modifications to the estimation procedures. This section briefly describes the updates to the international agricultural TFP indexes.
The data for agricultural outputs and inputs come primarily from the Food and Agriculture Organization (FAO). Other datasets accessed for each update include EUROSTAT, the National Statistical Bureau of China, Asian Development Bank, LABORSTA of the International Labor Organization, and IFADATA from the International Fertilizer Association. The updated TFP estimates use the latest available data from these sources. In addition to updating data for more recent years, these sources sometimes revise data from previous years to reflect more complete information on these series. The updated ERS International Agricultural Productivity database include these revisions to previous years’ data from these sources.
In addition to updating data from other sources, various changes may be introduced in the way data are organized and combined into the TFP series. The following details these changes.
1. The October 2017 data release no longer uses the Hodrick-Prescott filter to smooth the output series. Instead, the agricultural output index is now identical to the FAO index of gross agricultural output. We have further introduced new input cost shares from Rada, Liefert, and Liefert (2017) and applied them to all former east European FSU states and Russia in the post-1991 periods.
2. On January 5, 2017, the file "AgTFPcountrygeographicgroups.xlsx" was replaced to correct an error in the amount of agricultural land for SE Asia.
3. The October 2016 update uses a revised method for estimating agricultural labor for Nigeria. Nigeria has the largest economy and is the largest agricultural producer in Sub-Saharan Africa, although socio-economic data for this country, particular of its population and labor force, remain subject to considerable uncertainty. In previous versions, Nigeria agricultural labor force estimates are from Fuglie and Rada (2013), who used FAO data (2006 version) for 1961-1966 and then extrapolated them to the present assuming a 2 percent annual growth rate (roughly the rate for the rest of Sub-Saharan Africa). This method implied that the share of Nigeria's total adult labor force primarily employed in agriculture fell from 63 percent in 1966 to 55 percent in 2013, compared with a 2013 agricultural employment share of only 23 percent using FAO data. However, recent national labor force surveys indicate a more accurate count may lie in between these estimates. Sackey et al. (2012) cite surveys from the mid- 2000s that find about 40 percent of Nigeria’s labor were primarily employed in agriculture. To derive our new estimate of agricultural labor in Nigeria, we use the latest United Nations and ILO estimates of total population and labor force, and then use census and survey data to determine the share of the labor force in agriculture over time. Multiplying this share by the total labor force gives the size of the agricultural labor force. See Documentation and Methods for a complete description of sources and methods.
4. The October 2015 international productivity database update corrected an error in FAO data on agricultural land in New Zealand prior to 2002. The agricultural land series for New Zealand for 1961-2002 are derived from Statistics New Zealand (2003). We further introduce input cost shares from Rada (2016) for India in the post-1991 periods that are applied to all South Asian countries.
5. The October 2015 international productivity database update includes an explicit measure of animal feed used as an input into agricultural livestock production. The 2013 and 2014 versions of the database assumed that animal feed grew at the same rate of the size of the global livestock herd (measured in cattle-equivalent units). The 2015 update derives actual quantities of animal feed from FAO Commodity Balance Sheets, which provide annual data from 1961 to 2011 (except for former Soviet Socialist Republics (SSRs), for which data are available from 1992 to 2011). Feed quantity is the sum of all crops, crop processing residues (like oilseed meals, sugar, molasses, and bran from flour and grain milling), animal and fish products (including meat, fish, fish meals and whey) in metric tons of dry matter. The dry matter content of each feed type is from the "United States-Canadian Tables of Feed Composition: Nutritional Data for United States and Canadian Feeds, Third Edition," National Research Council, National Academies Press (1982). Dry matter is calculated on the basis of 89 percent dry matter for grain, which is the standard dry matter content of marketed grain. For SSRs prior to 1992, total USSR feed use is apportioned amongst them according to the proportion of total USSR livestock (in cattle-equivalents) in each SSR. Animal feed inputs are extrapolated to 2012 for each country using the average growing rate in animal feed from the previous three years in that country.
6. As FAO reports farm machinery stocks through no later than 2009 (and for some countries the series ends even earlier), various methods are used to extrapolate farm machinery stocks into the future. The original 2013 version simply assumed the same level of machinery stocks in 2010 as in 2009 or the last year for which estimates were available. The October 2014 update revised method for extrapolating estimates of machinery stocks by assuming they grew at the same rate as the ratio of agricultural land to farm workers. This procedure is designed to reflect machinery-labor substitution taking place in some countries as labor leaves the agricultural sector. For the October 2015 update an effort was made to collect new data from national statistical sources on actual machinery stocks. More recent (than FAO) statistics on farm stocks of tractors and combine-harvesters were collected from agricultural census' for Bangladesh, China, Europe, India, Japan, Russia, and the United States (see Data in Documentation and Methods for a listing of sources). The countries account for more than 70 percent of farm machinery stocks globally. For missing data, farm machinery stocks were extrapolated using the average growth rate from the three most recent years of available data.
The October 2016 update introduced two new approaches for extending estimates of agricultural machinery stocks beyond the latest available census or survey years. The first approach employed annual data of new machinery sales, taking into account obsolescence of older machinery. Data on annual sales of new farm tractors and combine-harvesters during 1991-2013 were collected from farm machinery manufacturers (sources: VDMA—Verband Deutscher Maschinen- und Anlagenbau, or the Mechanical Engineering Industry Association—and John Deere corporate reports) for the United States, Canada, Brazil, Argentina, Mexico, South Africa, and European countries. Assuming an average 15-year useful lifespan for new farm machinery, estimates of farm-held machinery stocks were extended from the latest available census year by adding the number of new machinery sales since the census year and subtracting the number of tractors purchased 15 years earlier.
The second approach (for countries for which we lack information on annual sales of new machinery) uses an econometric model to estimate the annual rate of change in farm-held machinery stocks since the latest available census or survey data on farm machinery stocks. The econometric model is based on the Kislev-Peterson model of farm machinery adoption and farm size. Kislev and Peterson (1982) hypothesized that as non-farm wages rose, farm labor would be induced to migrate to the non-farm sector. This would stimulate farm consolidation and mechanization to replace the labor leaving farms. Thus, farm mechanization would be correlated with non-farm wages and farm size. While the Kislev-Peterson model was developed in specific reference to the United States, in a comparative historical assessment of agricultural mechanization, Binswanger (1986) found a "remarkable similarly in the early mechanization experiences of developed and developing countries." See Documentation and Methods for a complete description of the econometric model.
7. The October 2014 and subsequent updates reassign Pacific island states from the Oceania Region to the Southeast Asia Region (now referred to as "Southeast Asia & Pacific").
See References for a complete list of citations mentioned.