SCOTTISH EXECUTIVE

The official series on Scottish Manufactured Exports was first published in May 1999. This quarterly series, which is broken down by ten Standard Industrial Classification (SIC) based industry groups, gives total export turnover of goods manufactured in Scotland at current and constant 1995 prices. The series only measures the value of sales of finished goods, with no adjustments made for the value of components used to produce those goods. Such adjustments can be made using the Input Output Tables for Scotland, which are produced annually by the Scottish Executive. This note gives details of the data sources and methodology used to produce the series.

Data Sources

The source of exports data used to produce the manufactured exports series is the Scottish extract of the Monthly Production Inquiry (MPI), a survey which is conducted by the Office for National Statistics (ONS). The Scottish MPI extract contains total export turnover for all ONS reporting units sampled within the MPI which include some Scottish activity. ONS reporting units are administrative groupings of one or more local units (i.e. sites where activity occurs). ONS surveys are generally carried out at the reporting unit level in order to improve accuracy and efficiency.

There are cases where a reporting unit on the Scottish MPI extract contains non-Scottish local units. In such an example, the share of total reporting unit export turnover allocable to Scotland is assumed to be the same as the Scottish share of employment within the reporting unit. Although this will not always be the case, it is an assumption made in the face of the lack of MPI export data at the local unit level. Any negative impact of this assumption is softened by the knowledge that in most cases reporting units are made up of Scottish local units only. This is particularly true of larger enterprises, which dominate Scottish manufacturing export activity.

UK Export Producer Price Indices (EPPIs) are used to deflate Scottish Manufactured Exports to 1995 prices. It is assumed that export prices are generally the same for Scottish manufactured goods as those for goods manufactured in the UK as a whole. The EPPIs used in the production of Scottish Manufactured Exports are supplied by the ONS.

Two sources of data are used to gross up the export turnover figures collected from the sample of manufacturing units covered by the MPI to totals for the overall population of Scottish manufacturing units. These are the Annual Census of Production (ACOP) (this has now been superseded by the Annual Business Inquiry (ABI), conducted by ONS, which covers output and employment in all sectors of the economy) and the Inter Departmental Business Register (IDBR). Both of these sources are supplied for Scotland by the ONS. Two sources are used because there is no single source of information on the population of Scottish manufacturing units throughout the period covered by the series (1995-1999).

Methodology

The Scottish percentage of each reporting unit's employment is applied to both export turnover and overall turnover. For reporting units which do not report for all three months in a quarter, estimates are made pro rata for missing months using the months for which data have been returned. All non-manufacturing reporting units covered by the MPI are excluded.

MPI export turnover totals are deflated at the 5 digit SIC level. Deflation takes place at this level to cover price change differences within published industry groups. ONS provide EPPIs for most of the 5 digit SIC levels covered by the MPI. Estimates are made for those 5 digit SIC levels for which no EPPI is available. These estimates are based on those EPPIs relating to a range of similar industrial sectors. When an EPPI is in place for all 5 digit SIC levels for all quarters, deflation of the export totals to 1995 prices takes place.

MPI export turnover is grossed up to a population total using the ratio of overall population turnover to overall MPI turnover. (Overall turnover is defined as the sum of export and domestic turnover). As mentioned above, overall population turnover is based on ACOP and IDBR totals for different periods.

For 1995 and 1996, the ACOP-based overall turnover is calculated on an annual basis for each of the ten industry groups published. Since ACOP only gives annual totals, these are converted quarterly estimates. No adjustments are made to build seasonal fluctuations into these quarterly estimates.

From 1997 onwards, quarterly population estimates of overall turnover are based on the quarterly IDBR turnover total for each published industry group. This stage in the process is fairly straightforward with the exception that reliable IDBR data for Scotland are only available from the second quarter of 1997 onwards. Estimates are made for the first quarter of 1997 based on the other 3 quarters of 1997.

Both deflated and undeflated MPI-based export totals are summarised from the 5 digit SIC level to the published level. There are examples where the total MPI-based turnover for an industry group exceeds the total ACOP/IDBR-based population turnover for the industry group. In such an example, the grossing factor (population turnover divided by sample turnover) would be less than 1, its theoretical minimum occurring only when the MPI includes all units in the population for that industry group.

An MPI-based turnover figure can exceed the equivalent ACOP/IDBR-based turnover figure in a single quarter due to the way that population turnover is calculated, particularly for ACOP-based population estimate. As mentioned above, annual ACOP data are converted to quarterly estimates with no adjustment being made for seasonal fluctuation. Over the course of a year, ACOP/IDBR totals should be greater than or equal to the MPI totals. However, in a particular quarter, the MPI total can be larger. In such examples, scaling factors are reset to 1.

The scaling factors calculated for the quarters currently available are low and generally stable. This suggests that, at the industry group level at which Scottish Manufactured Exports are published, the samples given by the MPI consistently represent a significant proportion of the overall population in terms of total turnover.

Each industry group has its own scaling factor for each quarter. These are multiplied by the MPI-based export turnover at both current and constant prices for that industry group in that quarter to give an estimate of the export turnover in the population, also at current and constant prices.

Further development

Further developments are planned for the series over the next year. Most significantly Scottish Manufactured Exports will be seasonally adjusted from the first quarter of 2000. A minimum of 20 quarters of data is required before the Scottish Executive's quarterly economic series are considered for seasonal adjustment. The series for each industrial group will be tested for seasonality and, if significant, will be seasonally adjusted, probably using the standard X11 process used in the seasonal adjustment of other economic series. Other smoothing techniques will also be considered for erratic series of a non-seasonal nature.

At present, ACOP population totals are used to gross up 1995 and 1996 sample data while IDBR population totals are used to gross up subsequent quarters. There is a clear potential for discontinuity using this approach. Any discontinuities will be examined and rectified when 1997 data from the new ABI becomes available at the end of 1999.

For further information contact Hugh McAloon (hugh.mcaloon@scotland.gov.uk)