Grossing of Global Connections Survey data for 2004 provisional and 2002 - 2003 revised Estimates
1. Sample data
Data from the surveys were supplemented by information from the Office for National Statistics Monthly Production Inquiry (MPI) and International Trade In Services inquiry (ITIS), as well as a small amount of information from trade flow surveys carried out by Scottish Executive to inform the input-output tables. Partial information was also obtained directly from some major companies that did not return forms.
2. Population
The population data set is based on the Inter Departmental Business Register. The data from the survey have been grossed up to the IDBR data based on company turnover. This has been done at reporting unit level, i.e. based on the sum of activity of all Scottish local units. Companies have been classified to an industry based on the SIC of the local unit with the most employees. With the exception of sea-fishing [SIC 0501] (for which an alternative data source was used - see below) export estimates are constrained to IDBR turnover.
3. Grossing procedures
Firstly, total exports for each company were estimated using a process of iterative stratification based on combinations of the following factors:
- Employment sizeband
- Standard Industrial Classification
- Prior knowledge that the company exported (as used in sampling frame)
- Local Enterprise Company Area
The initial step was to classify companies at the most detailed level of each of these factors. For any group containing 10 or more companies in the sample, the ratio of exports to turnover was calculated and applied to companies in the population with the same properties.
The process then involved collapsing categories (firstly employment was collapsed from 6 bands into 5, then at the next iteration SIC from 5-digit to 4-digit, then LEC to HIE/Scottish Enterprise area etc.); and at each stage estimating the proportions of goods and services exported for each company in the population based on the sample data for companies in that particular group. At the final stage, all remaining companies were classified by 3-digit SIC.
A similar process was followed to split total exports into EU/ non-EU. Proportion of total exports to each area for each company was estimated and applied to total exports. This procedure was based only on employment size-band and SIC for exporting companies and took 9 iterations. Sample proportions were then applied to the estimated EU total to estimate amounts for individual countries. This analysis was stratified by industry and employment - firstly 5-digit SIC and small/ large companies, going down to 2-digit SIC only. The non-EU totals were firstly split into regional totals using a similar methodology, before these were apportioned to country of destination based on the sample data.
4. Adjustments
Tourism
Expenditure by tourists from outwith the UK counts as an export of the good or service purchased. Data from the sample survey on hotels and restaurants and retail have been supplemented with information from the International Passenger Survey on tourist expenditure. This was done because these businesses often cannot identify the origin of their customers. Estimated tourist expenditure on 'accommodation' and 'food and drink' was compared with total turnover of SIC 55 on the IDBR by LEC area in order to estimate export proportions for companies in this sector. For 'shopping' and 'buying clothes', adjustments were made at the Scotland level. Additional adjustments were made for vehicle rental companies. Country breakdowns of total expenditure from the IPS were used to estimate destination totals.
Banking
Information on total exports from the four Scottish clearing banks was received from the Committee of Scottish Clearing Bankers, and information was also received from other banks. It was apparent that IDBR turnover in this sector was much higher than was being reported by the banks (there are known problems with financial sectors, and the Scottish Corporate Sector Analyses do not produce turnover for this sector). The average ratio of reported turnover to IDBR turnover was used to adjust the IDBR totals for all banks. No information was available on destinations of exports for this sector so they could not be included in the country analyses.
Fishing
Information on exports of fishing (direct lands of fish abroad) was provided by the Scottish Executive Rural Affairs Department (SERAD).
5. Revisions
The revised estimates for 2002 and 2003 reflect a number of improvements to the analysis since that time. Where new information was obtained for companies responding for the first time, this was fed back to make better estimates for previous years. Other improvements are outlined below;
Alignment with quarterly Index of Manufactured Exports (IME)
Adjustments have been made to the industry classifications and exports values of some large companies in order that the results of this survey correspond with trends in the IME series. Where, for example, companies are moved from a service to a manufacturing SIC this affects the grossing and results in both categories.
Addition of ITIS data
During 2004, data from the International Trade In Services (ITIS) inquiry was obtained from ONS for 2002 and 2003. This was used to supplement the GCS where no data were received from companies resulting in improved service sector estimation.
Further information
For more background on the issues involved in estimating exports figures for Scotland, please see the article in this years' Scottish Economic Statistics at www.scotland.gov.uk/publications/2005/11/2485808/58104.