Scottish agricultural survey: December 2017

Annual agricultural survey based on data from larger agricultural holdings together with estimates for smaller farms.

11. Notes

11.1 Background

This publication contains results for the December Agricultural Survey for 2017 and includes trends for the last ten years where available. Where appropriate, comparisons have been made between results of the December Survey and the June Agricultural Census.

11.2 Uses of the information

The December survey is conducted for a range of purposes. The statistics help the government to form, monitor and evaluate policy, and to assess the economic well-being of the agricultural sector.

Most of the data collected is required by the Statistical Office of the European Communities, specifically Council Regulation No 1165/2008 which sets out requirements for provision of cattle, pig, sheep and goat statistics in both May/June and November/December. It defines the category, age or weight of livestock for which statistics are to be provided and specifies the provision of quarter-year or half-year production forecasts. There is also a separate EC Regulation covering the provision of winter crops. This information is collated by Defra (Department for Environment, Food and Rural Affairs) for submission at member state ( UK) level.

December Survey results are not as widely used as results from the June Census as the survey only covers larger holdings, generally of at least one hectare, whereas the June Census is representative of all agricultural holdings in Scotland. However, December results supply supplementary information not available through the June census on rent values, machinery, winter livestock levels and grass sown as well as detail on hay and silage production.

Some examples detailing how the December Survey data are or have been used are:

  • Estimates of Total Income From Farming ( TIFF), which are used to measure the value of agricultural productivity in Scotland. The December Survey, which gives approximately end-year livestock numbers, are more useful for the calculation of calendar-year production. For example, although the June Census records the number of lambs present in summer each year, it does not (on its own) give an indication of the volumes of finished sheep and lambs that are being processed within the calendar year. The rents reported in the December Survey for each category are grossed up to the total areas reported in the census, to calculate the total amount of rent paid.
  • It is also useful to monitor livestock maintained for the next breeding season and winter crops in December so that the farming industry can better understand what to plan for in the coming year.
  • The data on machinery that is collected on the December Survey is also used to help estimate some of the input costs incurred within Scottish agriculture (for example, machinery repairs, depreciation, fuel and asset worth).
  • The information on rents is used to monitor the cost of land rental in different categories of land.

Results from the December survey are available to the public as follows:

This statistical publication is available for download from the Scottish Government website along with previous releases of December Survey results:

Headline results for TIFF (mentioned above) are published each January. These can be accessed as follows:

Economic Report on Scottish Agriculture ( ERSA) is a compendium publication containing detailed statistics on Scottish agriculture, combining further information from Total Income From Farming ( TIFF - see above for more details), Farm Accounts analysis (income and expenditure statistics by different farm types) and additional statistics/analysis from the June census.

11.3 Methodology - Data collection

Since 2015, the December Survey has incorporated collection of data for the Sheep and Goat Annual Inventory ( SGAI). This involves 14,700 December Survey forms requesting data on land, livestock and machinery data (including questions on sheep and goats) in addition to 10,800 holdings requesting only sheep and goat information. Holdings completing December Survey forms were selected using stratified random sampling where the sampling frame comprised of a list of all the larger (generally over one hectare) holdings in Scotland stratified by farm size and region as measured through the 2017 June Census. This spread is intended to ensure a good representation across the country and by farm size. Optimal allocation was used to calculate the sample size required in each strata in order to maximise precision of results. Following this, a random sample is selected from each strata. The sample was topped up with holdings which were included in the previous December survey (2016) but didn't respond.

The results are based on information returned from approximately 8,900 holdings, providing a response rate of 60 per cent for the December Survey. There were also 8,600 returns from holdings receiving the SGAI form only, providing a response rate of 79 per cent.

Since 2015, respondents have been able to complete their December Survey (and SGAI) online. In 2015 there were 2,200 online responses, which increased in 2016 to 6,100, and to 7,600 in 2017, accounting for 44 per cent of all survey returns.

Among the respondents, 2,910 holdings reported renting-in land in 4,540 leases (including seasonal lets). About 400 tenancies paid entirely or partly 'in kind' were reported, which have been excluded from the analysis. The following shows the approximate number of returns with non-seasonal rented land, by stratum, received in the 2017 survey. The random nature of the sampling within strata resulted in between 40 (Orkney) and 350 (North East Scotland) tenancies arrangements per sub-region.

0-<50ha 50-<200ha 200-<500ha 500-<1000ha 1000ha+
North East 119 172 45 5 8
North West 479 109 38 15 20
South East 142 183 90 34 18
South West 179 220 80 54 43

11.4 Methodology - Non-response

In Scotland there are around 51,100 agricultural holdings registered with the Scottish Government. We use these register details to maintain a full holding-level data set of Scottish agriculture for statistical purposes. This provides us with virtually complete coverage of agricultural activity in Scotland. However, please note that:

  • we very rarely conduct a full census of holdings as this would place an unnecessary burden on farmers;
  • for the selected holdings that are surveyed, not all farmers return data to us;
  • where we have gaps in our holding level data set, we maintain records by producing estimates.

The December Survey is representative of larger holdings (generally over one hectare in size around), of which there were 26,300 at June 2017. Estimates are produced for those holdings which were (a) large enough but not sampled, (b) surveyed but did not provide a response, and, for some variables, (c) smaller holdings.

Two stages of estimation are undertaken to calculate the December results where holdings are not included in the sample, or do not return data:

(i) For items collected both in the June Census and December Survey (livestock items and winter crops), a trending technique is applied to estimate the current year December values. The holdings are divided into strata using farm size and region. Where holdings have reported for both surveys, the total change between June and December for holdings within individual stratum are calculated. These rates of change are then applied to June Census results.

From 2013, the trending methodology was refined to provide improved estimates to account for the fact that holdings often report farming a particular crop or livestock in either the December Survey or June Census only. The previous method will have partially suppressed these trends. A time series for all items has been provided back to 2008, calculated using the new methodology. For years prior to 2008, the previous estimates have been adjusted by the percentage difference between the old and new methodologies.

(ii) For items only collected in December, such as machinery, arable silage production and grass sown, data in each strata are simply scaled up proportionally to account for non-response/inclusion in order to calculate estimates for all of those larger holdings within the scope of the survey. Note that the number of holdings classified as larger holdings will change from year to year, which will affect the scaled up figure. We are unable to scale figures up for smaller holdings for these items as we do not have a proxy measure to use from the June Census .However, for hay and grass silage/haylage this is possible, based on proportions of grass grown recorded in the June Census.

11.5 Methodology - Rents

Average rental value is calculated by weighting the survey data using farm-type and size, and land-type. The sample is not originally stratified by farm-type, however the following table shows the breakdown of land-area used in calculations. It is clear that the overall average rate is heavily dependent on the rental value of LFA cattle & sheep farms.

Prior to 2013, calculations were based on the much smaller Tenanted Land Survey (see the previous versions of this publication [8] ), with values often built up from five-year averages or best estimates for those farm-types with only small representation within the sample.

Area of rented land in 1,000 hectares

LFA non- LFA
0-<50ha 50-<200ha 200-<500ha 500-<1000ha 1000ha+ 0-<50ha 50-<200ha 200-<500ha 500-<1000ha 1000ha+
Cereals 1.0 3.2 1.6 1.4 - 3.8 21.2 10.0 2.1 2.9
Gen Cropping 0.7 2.0 1.6 - - 2.3 21.1 19.1 1.2 4.3
Horticulture 0.6 0.2 - - - 0.4 1.5 1.0 - -
Pigs and poultry 0.6 0.3 - - - 0.2 0.4 0.5 - -
Dairy 0.8 9.2 2.1 - 1.4 0.4 3.1 3.1 - -
Cattle & Sheep 54.7 106.3 142.7 170.4 528.0 3.9 10.9 3.2 1.8 6.7
Mixed 3.3 9.4 10.2 5.8 3.2 2.1 15.9 13.8 3.4 1.2
Forage/other 31.7 17.1 13.4 10.0 34.2 2.7 3.0 0.5 - -

For 2013 onwards, farm-type, including the LFA/non- LFA split, and size band were again used to stratify the data. The larger dataset means that average rents based on a single year can be used for a greater number of cells within the stratification, with a best estimate provided for others. Results for previous years are then often revised using rolling averages. However, all of the strata covering the largest amounts of rent have useable one-year data.

11.6 Collection of Cattle Data through the Cattle Tracing System

Statistical data on cattle populations have historically been collected through the June Census and December Survey in Scotland. In order to reduce the burden on survey respondents' data for the December Survey has been obtained through the Cattle Tracing System ( CTS), an administrative data source held by the British Cattle Movement Service ( BCMS) which holds records of cattle numbers and movements across Great Britain. These were used for the first time in Scotland in the publication of results from the 2013 June Agricultural Census.

11.7 Collection of Sheep and Goat Data through the Annual Sheep and Goat Inventory

In order to reduce the burden on survey respondents, data collection for the December Survey and the Sheep and Goat Annual Inventory ( SGAI) were merged for the first time in 2015. A section requesting sheep and goat data was incorporated into the December Survey form, while shorter forms asking just about sheep and goats were used for remaining businesses understood to keep sheep. Use of SGAI data allows for a more complete data collection and eliminates the need for separate data collections.

11.8 Data Quality

Data undergo several validation processes as follows; (i) checking for any obvious errors on the paper census forms upon receipt, (ii) auto-checking and identifying any internal inconsistencies once loaded onto the initial database, (iii) auto-checking for any sudden changes in comparison with previous annual returns and other holdings, (iv) assessing any trends or switches in item areas or quantities that look unreasonable.

If necessary, farmers are contacted to ensure data are correct. Additional quality assurance is provided at the later stages by utilising expert knowledge within the Scottish Government and the agriculture industry. The opportunity to complete the December Survey and SGAI form online was made available for the first time in 2015, incoroporating in-form validation in order to minimise errors in completion.

11.9 Main sources of bias and other error

The December Survey will be subject to measurement bias since we are reliant on farmers completing the form accurately. Ideally livestock counts should be undertaken to ascertain precise numbers of animals but, given time constraints, exact numbers of livestock are likely to be estimated. This bias will impact particularly on sub categories of livestock ( e.g. weight categories for pigs or ages of cattle) rather than the total population for a livestock type. Other categories likely to be estimated by farmers include the tonnage of hay and silage produced in the year.

Guidance notes detailing what to include on the form are supplied to avoid farmers misreporting information. With regards to livestock, we require farmers to report those animals located on the holding that are either owned by the farmer or animals that are owned by someone else but are held under formal contract. It has been noted that animals are sometimes double counted in situations where animals are held under contract with both the owner of the livestock and the farmer looking after the livestock reporting the animals. To avoid this double counting we have added specific guidance on the form itself in attempt to avoid this reporting bias.

The survey may also be subject to an element of non-response bias with farmers on certain farm types being more likely to respond to the survey than others. This means that we need use older information to estimate values for farm types less likely to supply us with current information.

A stratified random sample, grouped by farm size and region, is used to select holdings for the December survey. Individual strata are sampled to different extents. However, in estimating the results we weight by strata in order to produce a full dataset and to counteract the effects of some strata being sampled to a greater degree than others. This helps to address any sampling bias that is inherent in the sample design.

11.10 Survey burden

In December 2011, a representative sample of around 110 farmers participated in a telephone survey in order to calculate the burden of participating in the December survey. It was not considered beneficial to repeat this survey each year, however we do have updated figures for hourly rates [9] which we can apply to the time data from the 2011 survey. These give a total compliance cost for the December Survey of £60,300. It should however be noted that since the 2011 survey there have been several changes, namely the removal of the requirement to report cattle data on the form, reducing the burden for approximately 5,700 holdings, but added information on tenancy for approximately 4,300 holdings [10] .

Please refer to the December 2011 publication [11] for details on how this figure was calculated and the range of times reported.

11.11 Other publications

The next large agricultural survey is the June Census of agricultural holdings. This is a larger exercise which surveys around 33,000 holdings with results scheduled for publication in October 2018. Results for the 2018 December survey will be released in Spring 2019.

Results from all Scottish Government agricultural surveys can be accessed here:

Results from previous June Censuses can be accessed here:

Publications relating to cereal and oilseed rape production can be accessed here:

Agricultural Facts and Figures pocketbook. This provides a useful summary of the key statistics in the Scottish agriculture and food sector in a convenient pocketbook format.



Back to top