Scottish Government

Scottish Economic Statistics 2004

A2 Development of a Tourism Satellite Account for Scotland

Catriona Hayes and Claire Boag, Scottish Executive

Introduction

There is increasing demand for the quantification of the contribution of tourism to the economy, on a standardised basis. Tourism is seen as a key factor in economic development in Scotland, so its progress needs to be benchmarked and tracked to help assess the impact of policy changes. In order for development and marketing agencies to work effectively with industry, the sources of growth must be understood and supported. It is also important to understand the linkages that tourism activity has with the wider economy, and therefore to gain insight into the extent that various Scottish industries are dependent on tourists. This is linked to the need to further analyse tourism employment, and thus the implications of any changes in tourism demand on the labour market across the various industrial sectors in Scotland.

Scottish Economic Statistics 2002 contained an article presenting methods for analysing the size of tourism in the Scottish economy based on the Scottish input-output (IO) tables. The article outlined some of the issues involved in assessing tourism's economic impact, and highlighted the need to improve the tourist expenditure data used in the input-output tables. This article seeks to describe progress since that time, and to present an experimental Tourism Satellite Account (TSA) for Scotland for 2000.

A TSA provides a means of quantifying the economic impact of tourism within a National Accounting framework, and its construction benefits from a set of international guidelines (the Recommended Methodological Framework (RMF), published by the UN in 2001). A satellite account is an extension to the main accounts, and concentrates on an area that is not well defined within the standard framework. The concept is also being applied to other difficult to define areas, such as health and the environment. Background information on TSAs can be found via the following Scottish Executive website: http://www.scotland.gov.uk/about/ASD/OCEA/00018300/TSAlinks.aspx.

The developmental work described here for Scotland has been carried out by the Scottish Executive in partnership with VisitScotland and Scottish Enterprise. Similar projects are being undertaken by the University of Wales, Nottingham University, Strathclyde University and the University of Limerick to develop TSAs for the UK, English Regions and for Ireland. Collaboration between project teams allowed for the sharing of ideas and information and consistency of approach. The initial report for the UK was published recently and can be accessed on the Department for Culture, Media and Sport (DCMS) website ( http://www.culture.gov.uk/global/research/statistics-outputs/uk-tsa-fsp.htm).

The main stages in the development of the preliminary Scottish TSA were as follows:

1. Initial scoping study carried out by Professor Iain McNicoll in 2003. The report is available via http://www.scotland.gov.uk/about/ASD/OCEA/00018300/TSAproject.aspx.

2. Detailed analysis of available data on tourist expenditure.

3. Creation of a new input-output tourist expenditure column based on available information for Scotland.

4. Rebalancing of the 2000 input-output tables to incorporate the improved tourism information.

5. Interpretation and development of the TSA RMF applied to Scotland.

6. Analysis of the input-output tables to produce Scottish TSA tables.

Readers should note that the data presented here are based on an experimental analysis, and so estimates should not be taken as definitive. The authors would welcome feedback on the methods used, and the validity of the assumptions made (email addresses are provided at the end of this article).

Development of the input-output tourism column

The existing tourist expenditure column in the Scottish input-output tables, while fit for the purpose of general economic analysis for Scotland, was not suitable for specific, detailed analysis of tourism. This was mainly because it relied heavily on the use of UK data for calculating Scottish estimates. In order to improve the tourism information contained within the tables, detailed analysis of the available data was carried out, ensuring that actual Scottish data were used as the starting point where possible. The sources used were:

• UK Tourism Survey (UKTS) 2000-2002
• International Passenger Survey (IPS) 2001, 1997 expenditure trailer
• Expenditure and Food Survey (EFS) 2001/02
• UK Day Visits Survey (DVS) 2002

Although these are UK wide surveys, they all collect separate data on spending in Scotland. There are recognised problems with the accuracy of the tourism surveys, particularly for sub-UK level. These issues were highlighted in the DCMS review of tourism statistics, published at http://www.culture.gov.uk/global/publications/archive-2004/Review-Tourism-Statistics.htm in June 2004. Nonetheless these are the only available sources and are used currently to monitor tourism activity. The quality of the TSA will benefit from any improvements in the data sources that result from the implementation of the recommendations of the review.

Expenditure data by broad category from the UKTS and IPS was combined with EFS spending patterns and UKTS information on means of transport, type of accommodation used etc. to produce a detailed expenditure breakdown for different types of tourists, as summarised in the table below.

Ta ble A2.1: Summary of estimated expenditure by type of tourist, 2001

Some of the expenditure data that could be derived from the surveys, for example on accommodation, contained more detail than required for the input-output tables, but most required further disaggregation to make them fit into IO Categories (IOC). This necessarily required some assumptions, upon which feedback from interested readers would be welcomed. It is beyond the scope of this article to provide a complete list of assumptions made, but the table below summarises some of the main assumptions used in splitting expenditure by IOC. For example, it was estimated that holiday tourists from the rest of the UK spent a total of 3.8 million in 2001 on "glassware, tableware and household utensils". This category spans a number of IOCs and was split as shown in table A2.2, based on information from the UK input-output tables along with consideration of likely purchases made by tourists in Scotland.

T able A2.2: Examples of assumptions used to split expenditure categories to fit input-output categories

In TSA construction, travel agent turnover is treated net of constituent package costs, which are reallocated from the agent to the relevant industry/commodity and reference economy. Information was available on total spending on packages, but not specifically on margins paid to travel agents. Based on discussions with the Scottish Tourism industry, it was estimated that 10% of the spending by inbound and domestic tourists on a package holiday was margins. For domestic tourists, 100% of this was assumed to be incurred in Scotland, for rest UK tourists, 40% was assumed to be either incurred in Scotland, or paid to a Scottish company, and for overseas tourists, only 20% was assumed to be incurred in Scotland, or paid to a Scottish company. Additionally, small portions of spend on "entertainment services" and "other services and professional fees" were allocated to this category to account for the activities of tour guides. Further methodological details will be made available on the Scottish Executive website at http://www.scotland.gov.uk/about/ASD/OCEA/00018300/TSAintro.aspx

A detailed breakdown of expenditure by type of tourist was produced in order to carry out the analysis to produce a TSA at the most detailed level possible. This will also allow impact analysis to be carried out separately for the different types of tourists shown in table A2.1. Additionally, future updates of the tables can be produced by changing the constraining totals each year for each type of tourist, and the estimates will account for any changes in the types of tourism occurring. It will be assumed that the expenditure patterns for each type of tourist stay similar for the next few years - at least until better information becomes available. Note that while these estimates are the best available currently, they are dependent on the quality of the survey data and are subject to error due to the problems outlined above.

Creating the input-output column for 2000

The expenditure breakdown used to create the IO column for 2001 was applied to the published 2000 tourism expenditure totals to create a tourist expenditure breakdown for 2000. The tourism final demand column in the IO tables comprises all tourist expenditure by rest of UK and overseas tourists on domestically produced products. Domestic tourists' and day visitors' expenditure are subsumed within the household final demand in the input-output tables, and the estimates are extracted from there in order to create a TSA. Incorporating this new tourism data into the household final demand column resulted in a slightly different 2000 column from that published in the original version of the 2000 tables.

Tourist import propensities are based mainly on household import propensities, as it was considered that for most purchases where items are not 'typically' Scottish, tourists would consume based on availability of goods and would therefore purchase similar levels of imports to households. However, adjustments were made where it was considered that tourists would be more likely to buy Scottish products, for example - spirits & wines, knitted goods, glassware & ceramics, and certain types of foods. These adjustments were discussed with, and supported by, VisitScotland and Scottish Enterprise.

Once the new household and tourism columns were created, the 2000 IO tables were rebalanced, so that demand could be reconciled with supply. Assuming that any change to the household column is due to the change in the derivation of the domestic tourist and day visitor components, the revised tables showed higher overall tourist spend on fuel, knitwear, hotels, bars and restaurants, alcohol, transport, communications and recreational services than previous estimates. Slightly reduced figures for plastic goods, financial services, health services, real estate and renting of machinery also resulted.

Tourism Satellite Account

The input-output tables, including the tourist column described above, along with compatible expenditure breakdowns for Scottish domestic tourists provide a basis for development of a full TSA. (For background information on input-output please go to www.scotland.gov.uk/input-output.) Tables A2.3 - A2.8 in the annex to this article represent a first attempt at a Scottish TSA. The tables are also numbered TSA1 - TSA7 in line with the tables in the international TSA Recommended Methodological Framework (TSARMF). Table 5 is available for Scotland but is not reproduced here because all the information it contains is also in table 6. Tables 8 to 10 of the TSARMF, covering gross fixed capital formation, collective consumption by functions of government, and non-monetary indicators, have not yet been created. In fact, no country has yet produced a full set of all ten tables to the standard suggested in the framework, and some of the methods have not yet been finalised and agreed at the international level. Additionally, the detailed methodology is often not fully specified in the RMF so, while taking into consideration the progress made in other countries, we have had to develop our own approach to some extent, while keeping within the international guidelines. The methodology used to produce the Scottish TSA is outlined as follows:

1. The make matrix at basic prices shows the output of products by each industry, and provides most of the information needed for table 6 of the TSARMF, the central table of the TSA. The matrix is not published in its entirety due to the disclosive nature of the data, but has been used for this analysis. Within the make matrix, it was necessary to split the tourism related industries and commodities into more detailed categories. The industry splits of hotels and catering, travel agents from transport services and visitor attractions from recreation were created using total turnover on the Scottish Inter-departmental Business Register (IDBR) for the relevant SICs. In terms of the various commodities, more assumptions had to be made. In the absence of available data, proportions of each product produced by each of the separated industries had to be derived. These proportions are detailed below for the Hotel & Catering industries, and comments on their validity would be appreciated.

Assumed breakdown of hotel & catering commodities produced by hotel & catering industries

2. Starting with the make matrix described in (1), a make matrix for tourism only was derived. This required some assumptions about production of products for tourists by industries in order to split up the IOCs that are important to tourism. Again, the industry totals were mostly split in terms of IDBR turnover, but in this case the commodities were split in terms of tourist expenditure on products. The tourist expenditure columns from the symmetric matrices provided constraining totals for the rows (commodities) and columns (industries) of the matrix. The matrix was then balanced iteratively to reconcile industry output for tourists with tourist expenditure on commodities.

3. The tourism outputs in (2) were applied to the total outputs in (1) to give tourism shares for each commodity within each industry. The industries and commodities were then grouped into 'tourism characteristic' (agreed standard industries or products as used in the UK TSA, plus any with tourism shares of >10%), tourism related (those with some links to tourism - tourism shares of 5-10%) and non-specific (everything else).

4. Intermediate consumption was obtained from the Use matrix, with the relevant sectors' purchases being split up using size of the industry output adjusted to account for different purchasing e.g. of food, and also using Annual Business Inquiry (ABI) purchase information where available. The intermediate consumption to produce tourism products was assumed to be in proportion to the amount of tourism products produced by an industry (i.e. it was estimated using the industry tourism output ratio).

5. Tourism Gross Value Added (GVA) for each industry was then calculated by subtracting intermediate consumption to produce tourism products from tourism output.

Provisional Results

Tables A2.3 to A2.8 (TSA1 - TSA7) in the Annex show the provisional results obtained for Scotland using the above methodology. Note that the account does not include all consumption as defined in the TSARMF, in line with the approach taken for the UK TSA project. Currently excluded from total consumption estimates are the following:

• Tourism Social Transfers in Kind (Government expenditure on tourism policy, promotion and facilities. This is difficult to define and so it is currently recommended that it is excluded from international comparisons.)
• Visitor final consumption in kind. (Except for imputed rent on second homes which has been included. This would also include expenditure of residents on behalf of visiting relatives for example.)
• Business tourism expenses (Intermediate consumption where companies, Government etc. have paid for transport or accommodation (but not for meals or other types of travel expenses) for employees and guests on business trips. Most such expenditure should be covered in the final consumption data however, as the surveys ask for expenditure on behalf of respondents as well as for their own outlays on a trip.)
• Purchases of consumer durables for use in tourism and pre- and post-trip expenditure on items for the trip, such as clothes, immunisations, insurance, sunscreen, film and film processing etc.

Table TSA1 (A2.3) shows estimates of inbound tourism consumption that have been reconciled with supply of products using the input-output balancing process. Because they are balanced, and also due to some definitional differences, the totals differ slightly from expenditure estimates directly obtained from surveys. Inbound tourism consumption consists of expenditure of visitors from the rest of the UK and overseas within Scotland and totalled 2.8 billion in 2000. Of this, 60 per cent was on tourism specific or related products.

Table TSA2 (A2.4) shows the equivalent estimates for Scottish tourists, along with the other components of domestic tourism. It should be noted that the TSA definition of domestic tourism is wider than the one which is often used, as it includes the expenditure within Scotland of outbound tourists to both the rest of the UK and overseas (column 2.6). For example, this covers purchases of package tours, expenditure on getting to the border or the airport and any meals or hotels taken on the way. In theory it could also include any purchases made specifically for tourism trips, but this has not been estimated, so the majority of the outbound expenditure is on tourism products. Note that expenditure by domestic day visitors has also been included. This incorporates leisure trips as defined in the UK DVS, but not business day trips as there are no data on them. Total domestic tourism consumption has been estimated as 4.6 billion.

In table TSA3 (A2.5), estimates of outbound expenditure by Scottish tourists are provided. These are expenditure in other countries by residents on tourism trips, which are included in imports in the national accounts. Any items paid for in advance of the trip are not included here, as they are part of domestic tourism expenditure and are shown in table 2. These estimates are subject to review as the methodology is still in development.

Table TSA4 (A2.6) brings together the information from tables 1 and 2, showing total internal tourism, that is all inbound plus domestic tourism consumption. This totals 7.4 billion, although a large part (over 3 billion) of this is on imported goods and services.

Tourism GVA was estimated as 2,053 million, as can be seen in table TSA6 (A2.7). This was 3 per cent of total Scottish GVA in 2000. This compares, for example with agriculture, forestry & fishing which made up 2 per cent of GVA, and manufacture of electronics, which contributed 4.5 per cent. The industries that accounted for the largest shares of Tourism GVA were hotels (18% of total), restaurants (17%) and recreational services (12%). TSA6 also shows the shares of each product produced by each industry that have been estimated to be produced for tourism consumption, for example it is estimated that 47 per cent of the output of restaurant & catering services provided by visitor attractions was consumed by tourists. The tourism ratio on supply (final column) shows the percentage of total supply of products (including imports) within Scotland that is accounted for by tourism consumption, and has been estimated as 4 per cent for the whole economy.

It can also be seen in table TSA6 that although the total Value Added of tourism industries was 7.7 billion, only 20% of this was due to tourism consumption as currently defined. The rest could be due to consumption by residents or non-residents for non-tourism purposes, or by business tourism that has not been fully captured.

Table TSA7 (A2.8) shows all Scottish employment in the tourism and non-tourism industries. Data on employees was taken from the Annual Business Inquiry, and combined with estimates of the self-employed from the Labour Force Survey. The table also shows estimates of 'tourism employment'. This has been calculated by applying the tourism output ratios from each industry to total employment in that industry. This differs from the usual estimate of tourism related employment that counts all employment in particular industry categories (equivalent to the tourism characteristic industries excluding transport and parts of recreational services). In 2000, the estimate of total tourism related employment was 214,000 or almost 9 per cent of jobs, based on the same data sources. However these include jobs in businesses that also depend significantly on local custom, for example bars, cafes and sports centres. The table shows that an estimated 130,000 jobs or 5 per cent of employment in Scotland, were directly supported by tourism. Note that this refers to all employment and not full time equivalents. It can be seen that while in the whole economy 30 per cent of jobs are part time, an estimated 47 per cent of tourism dependent jobs are part time, therefore the percentage of FTE employment would be lower. It is also notable that tourism supports a higher proportion of females (58%) than are employed overall (48%).

The tourism employment figure provides an estimate of the number of jobs that directly depend on tourist expenditure. It does not, therefore, currently include employment providing non-market services to tourists or public sector jobs developing or promoting tourism.

Future developments

After feedback from readers of this article and others with an interest in the area, the methodology and assumptions used in producing the expenditure breakdowns and in deriving the TSA tables will be reviewed. These will then feed into the development of future updates of the TSA. Feedback on demand and priorities for developing estimates of the currently excluded items listed above will also be welcomed. These inputs along with consideration of developments at the UK level will determine future research in the area.

Later this year, the third in a series of workshops with TSA experts will be held, at which these Scottish TSA results will be presented, along with those for the UK, Wales, Ireland and the English Regions.

In line with a commitment in the 2002 Spending Review, the Scottish Executive will produce a tourism output index by the end of 2004. The tourism index will effectively be a re-weighting of the Scottish GDP index based on the Gross Value Added attributable to tourist expenditure. This will track the growth in tourism on a quarterly basis, summarising trends in the combination of industry groups that tourists spend money in. The initial index will use 2001 as a base year, with industry tourism ratios from the TSA applied to the detailed industry GVA weights. The output from this TSA work will also be used in the development of VisitScotland's "Forecasting, Trends and Scenario Planning Model" (Moffat Model) and Scottish Enterprise's approach to tourism cluster mapping.

If you have any comments or queries on this work, please contact either of the authors: Claire.Boag@scotland.gsi.gov.uk or Catriona.Hayes@scotland.gsi.gov.uk.

Table A2.3: TSA1 - Inbound Tourism consumption by products and categories of visitors, 2000