Major Time Series Analysis Caveats
SABS 2103 to 2015
Data on purchases for 2012 onwards has been produced using a new Purchases Apportionment Methodology (PAM) for the apportionment from the reporting unit purchases to the local unit level. This new method was introduced to preserve the additivity of purchases components values to the total purchases. This has introduced a discontinuity in the purchases and GVA data series with 2008 to 2011 still calculated using the original method.
The change in the calculation of purchases (and hence GVA) has minimal impact for the large majority of sectors. However, the industry affected to the greatest extent is 'Gambling and Betting (SIC 92)' where there is volatility for a small number of multi-site businesses. The sampled businesses within this Division provide details of the “amounts payable to winning customers” as part of their purchases. The new apportionment method has led to volatile local estimates of payments to winning customers which then also affects other associated variables, like total purchases and GVA. Additional work and analysis is required by ONS to identify further improvements to the methodology, and estimates, to reduce this impact while maintaining the additivity. Until this is completed (which at earliest would impact on publications in 2018), regional Purchases and GVA figures for Division 92 (Gambling and betting activities) should be viewed with caution.
More detail of PAM change can be found in this Information Paper. A detailed 2012 region by industry impact assessment has also been published by ONS.
In line with Office for National Statistics (ONS) ABS releases, data for the Insurance and Reinsurance sector (SIC 65.1 and 65.2) has been removed from the SABS results for all years from 2008 to 2012. The previously published ABS data for the Insurance and Reinsurance sector has been excluded due to on-going volatility while ONS carry out a more detailed quality assessment. This does not affect any other industry sectors and, as the Insurance and Reinsurance sector data has been removed for all years in the series, this does not affect time series analysis.
SABS 2011 (impact on 2008-09 data)
Please note that validation of data returns for 2011 highlighted that a small number of businesses within the Insurance and Reinsurance sector (SIC 65.1 and 65.2) were returning global figures rather than UK only. A revised 2010 estimate (together with new 2011 data) are now available in this publication. However, any global figures reported will remain in 2008 and 2009 published data in line with current Annual Business Survey (ABS) revisions policy. This can have a significant impact on Insurance, reinsurance, professional, scientific and technical activities (Section KM) time series data, as well as overall Totals, in particular local authority areas. Results for this sector are experimental and any comparisons between years should be undertaken with caution.
Further Education (FE) Colleges were reclassified as Central Government entities on 13 October 2010, transferring FE Colleges from the private/business sector to the public sector. This change brings FE Colleges out with scope of ABS survey from 2010 onwards, which represents a step change in the 2010 results for SIC 85 compared to data for earlier years.
The data, from 2008 onwards, are published on a Standard Industrial Classification (SIC) 2007 basis – data pre-2008 were published on a SIC 2003 basis. Updates to the SIC are required to enable it to more accurately reflect the structure of the modern economy - SIC 2003 was replaced by SIC 2007 in January 2008. The revision was motivated by the need to adapt the classification to changes in the world economy. The revised classification reflects the growing importance of service activities in economies over the last fifteen years, mainly due to the developments in information and communications technologies.
More information on SIC (2007) is available at: http://www.gov.scot/Topics/Statistics/Browse/Business/Corporate/sic2007note#top and https://www.ons.gov.uk/methodology/classificationsandstandards/ukstandardindustrialclassificationofeconomicactivities/uksic2007
The employee/employment data published in SABS, from 2008 onwards, are based on Business Register and Employment Survey (BRES) – previous years SABS employee data were based on the employment part of the Annual Business Inquiry (ABI/1). The BRES replaced and integrated the Business Register Survey (BRS) and the ABI/1 in 2009 (note that there was a pilot of BRES in 2008 which allowed 2008 BRES estimates to be produced). For businesses that are surveyed by BRES, the regional and local estimates of employment are more accurate (than those from the ABI/1) because, rather than producing data by a headquarter-based apportionment model, they are obtained directly from data reported at site level (shop, office, factory, etc). More information on BRES is available on the ONS website at: http://www.ons.gov.uk/employmentandlabourmarket/peopleinwork/employmentandemployeetypes/bulletins/businessregisterandemploymentsurveybresprovisionalresults/previousReleases
The 2008 and 2009 ABS followed a similar sample selection criteria (see Sample Design section below) and were both sampled on a SIC 2007 basis. It is important to note that 2009 marked the full introduction of the BRES, this and the change to SIC 2007 may lead to some discontinuity with businesses moving from one SIC to another between years. It is not unusual for some businesses to change SIC code over time, but there are some particular movements between 2008 and 2009 that users should be aware of e.g. a large number of landscape business sites have moved from SIC 01 (Crop and animal production, hunting and related service activities) to SIC 81 (Services to Building and Landscape Activities) between 2008 and 2009 – this is a result of the change from SIC03 (where landscape businesses were included with the agricultural sectors) to SIC07 (where landscape businesses fall under the services sector).
Insurance and Reinsurance (SIC 65.1 and 65.2) data is published for 2008 onwards – this is the first time that these data have been included in SABS. Please note that SIC 65.3 (Pension funding) is still omitted from the SABS results.
ONS improved the sample design for the 2008 survey which means that the results for 2008 onwards cannot be compared with historical trends. Although this temporary loss of comparability over time is a downside - it means that the results for 2008 onwards are more accurate, and comparisons can still be made between sectors across the UK. More details are provided below on the change in sample design due to SIC 2007 and the complete re-allocation of the sample.
Notwithstanding, the change in sampling, the high energy prices in 2008 have had a significant impact on the "Remaining Sectors" results for Scotland in 2008.