< Previous | Contents | Next >

Scottish Economic Statistics 2002

A3 Estimating the Economic Impact of Tourism in Scotland

Catriona Hayes, Scottish Executive

Introduction

Tourism is very important to the Scottish economy - in 2000, UK and overseas visitors spent an estimated £4.5 billion in Scotland. It is therefore important to be able to understand its economic importance relative to other sectors and to monitor its overall performance over time, to help inform and evaluate Government policy.

There is a wealth of information collected on tourism, from various surveys overseen by VisitScotland and other bodies. There are details of who visits and what they spend and do while they are here. The headline figures concentrate on numbers of visitors, the number of nights they stay and how much they spend in total. This information is invaluable to those involved in tourism - particularly for marketing and planning. However, it is difficult to find a clear and concise picture of the size and make-up of Scottish tourism and how it contributes to the economy as a whole.

The 'tourism industry' is often referred to as if it could be defined like agriculture or manufacturing. However, the main difficulty in measuring the economic impact of tourism is that it is not an 'industry' in the conventional sense. Traditional industries tend to be defined by supply (i.e. by what they do) whereas tourism is better defined by demand (i.e. it depends on the type of customer purchasing the goods and services produced by all industries).

This paper seeks to present some ideas for improving the methodology for the estimation of the economic impact of tourism in Scotland. It sets out the ways that this is currently done, highlights some well-recognised drawbacks of these, and explores the use of an alternative input-output based analysis, for the following two aspects:

• the effect of tourists' expenditure on overall GDP, and
• the employment supported by tourist expenditure.

It should be noted that figures are given in this article for illustrative purposes only. There a number of issues relating to currently available data, which have necessitated the use of several assumptions in the development of the methodology. Many of the issues are unlikely to be a problem in future years, and the assumptions would be revisited if the described approach were to be adopted. Also note that the methodology presented here is
by no means definitive; rather it is hoped that its publication will stimulate further discussion.

Analysis

Tourism GDP

Tourist expenditure is often taken as a proportion of GDP to give a measure of its size in relation to the economy. However, this approach tends to overestimate the importance of tourist demand for the following reasons:

1. It does not allow for the significant proportion of tourist expenditure that leaks out of the economy through imports.

2. The gross expenditure figure also includes inputs and taxes on products, whereas these are not included in GDP - which is a value-added figure.

In an economy like Scotland, which imports a large proportion of goods, this distinction is important as it significantly affects the impression of the importance of any chosen demand category.

The most comprehensive breakdown of tourist expenditure in Scotland currently available is published in the Scottish input-output tables. These provide a sound basis for estimating the contribution of tourism demand to the overall economy. The tables describe the interactions between industries with the final demand categories (including tourism) in a consistent and balanced way.

The direct output of each industry due to tourism is given in the tourist expenditure column of the industry x industry matrix. The indirect and induced output due to tourist expenditure can be estimated from this by applying the appropriate multipliers, and a total output due to tourism can be calculated for each industry. From this, the proportion (pi) of total output that is generated by tourism in each industry can be estimated by taking the ratio of tourism output to total output.

The quarterly GDP(O) index produced by the Scottish Executive is built up from many different weighted output series at a detailed industrial level. These series were aggregated to I-O group level to produce an output index (Gi) and associated weight (wi) for each input-output (I-O) category¹², for each quarter since 1995.

New tourism-adjusted weights can then be calculated for each I-O group from the product of the above weight wi and the estimated proportion of output of I-O group i that is due to tourist expenditure (pi), calculated as described above.

That is t_i = w_i x p_i

The sum of these tourism-adjusted weights (t_i) over all the I-O categories gives an estimate of the relative importance of tourism to the Scottish economy. This results in a figure of 4.2 percent. If the induced effects were not included, it would be slightly less, at 3.8 percent of GDP. If no multiplier effects were included, direct tourist expenditure would result in an estimate of 2.6 percent. These estimates are effectively based on 1995 data, the year the weights were assigned. The introduction of chain-linking will allow new tourism-adjusted weights to be calculated for each year.

A Potential Tourism Index

For any quarter, the contribution of tourism to total GDP can be calculated from:

where

Gⁱ is the output index for I-O group i
and
t_i is its associated tourism-adjusted weight, as described above.

The GDPT calculated above can then be transformed into an index (1995=100), in order to track the growth of tourism related output over time. This index reflects the change in output of industries supplying tourists, and is based on assuming a constant proportion of output driven by tourist demand in each industry over the time period. There would be potential in future development of the work to incorporate different proportions for each year based on annual input-output tables and tourist expenditure, as well as the annual GDP weights resulting from chain-linking.

Chart A3.1 compares the quarterly tourism index with the published GDP index. This shows that output generated by tourist demand was increasing faster than the economy as a whole up until 1998 but that growth slowed in 1999 before picking up again in 2000.

Chart A3.1: Estimated quarterly tourism index compared with GDP series

Tourism Employment

Tourism employment is currently estimated from a definition of tourism based on supply. All employees in certain tourism related SIC categories are counted, plus an estimate for the number of self-employed. Table A3.1 shows the constituents of the estimate for 1999, the latest to be published by VisitScotland, which is based on Annual Business Inquiry data. The total is over 8 per cent of Scottish employment.

Another characteristic of tourism related employment is that it can often be seasonal. The ABI survey is carried out in December each year, providing a 'snapshot' of employment. Therefore these figures might well underestimate the totals working in tourism throughout the year.

While tourism is most easily understood as defined by demand, the supply side approach shown above is straightforward and allows for comparisons between different areas of the UK. However the above supply definition has the following problems associated with it:

1. Most of the categories included are likely to also include a large amount of local, non-tourist activity (e.g. bars and restaurants, recreation facilities), which could obscure any trends due to tourism.

2. Some other areas which are likely to be affected by tourism (e.g. transport) are not included.

For these reasons, a more robust definition is desirable. One approach, considering employment in isolation, might be to refine the definition by estimating a proportion of each industry's activity that relies on tourism demand and applying this to the employment figures. Estimating such proportions directly would not be easy, and businesses may not be able to provide information as to how much of their custom was from tourists. Even if it was possible for industries that provide goods and services directly to tourists, it would be impossible for industries where tourist expenditure has an indirect impact. However, we can estimate this indirectly by considering tourist expenditure as a proportion of output.

Thus the approach suggested here is to apply the method described above for GDP to employment data. That is, for each industry, to multiply the proportion of output that is tourism-related by the total employment in that industry. This should give an estimate of tourism employment based on direct, indirect and induced effects that is compatible with the GDP estimate, i.e. uses a consistent definition of tourism. For example, if half of the total output of hotels, catering and bars is related to tourism, we assume that half of the jobs are also related to tourism. This means that we are in effect assuming that productivity is consistent within input-output groups, e.g. that average output of workers in catering is the same as that of workers in bars. Although there will be variations in reality, it does not seem too unreasonable to assume this in order to simplify the analysis.

Numbers of employees by input-output categories can be obtained from ABI data. Estimation of numbers of self-employed is more difficult and for the purposes of this preliminary research the estimates were fairly rough. The Labour Force Survey is the main data source, and gives estimates of numbers of self-employed and family workers by broad industrial classification. By making some assumptions totals can be estimated for some input-output categories or groups of categories. Data from other sources such as the agricultural census can also be incorporated. The tourism-related employment figures can then be calculated using these estimates and the proportion of tourism-related output obtained from input-output tables.

The resulting estimates gave a figure that comprised 6.7 per cent of total Scottish employment in 1999 when based on the data as described above - slightly less than the 8 per cent of employment in 'tourism related' industries. Considering FTE employees as described in the box above, this method would estimate just under 6 per cent of Scottish employment being driven by tourist demand.

Conclusion

The preliminary analysis described in this paper indicates that an index of the performance of industries affected by tourist demand in the Scottish economy, derived from input-output tables and the components of Scottish GDP, could be produced on a quarterly basis. This would fill a perceived 'gap' in tourism statistics.

The tourism expenditure estimates in the Scottish input-output tables will be reviewed as part of the ongoing programme of improvements. This will result in updated methods leading to data quality improvements. Such improvements would have a knock-on effect on the quality, and hence the potential value, of the analysis described in this paper. In addition, the introduction of chain-linking of GDP estimates will lead to more accurate and relevant assessments of the 'size' of tourism in the economy, and would greatly improve the usefulness of the proposed tourism indicator.

The input-output groupings are not ideal for the analysis of tourism. For example, the industries included in category 92 (hotels, catering etc.) are obviously not homogeneous with respect to output, employment and proportion driven by tourist expenditure. Some as yet unpublished work in Wales has studied this in more detail.

Further refinement of the approach may depend on the status of other potential developments in the area. In particular, if the Department for Culture, Media and Sport decide to go ahead with the production of UK tourism satellite accounts (TSA's), it may be that the Scottish Executive no longer has to produce its own analyses. However, since TSA's are an extension to the input-output tables involving more detailed information, this work could equally feed into their development. In addition, their development would help to resolve the above issue relating to lack of detail in the input-output categories.

If the DCMS work goes ahead, it is likely to be several years until sub-UK tables are developed and so the method described in this paper could fill the Scottish information gap in the interim. The Executive will therefore continue to consider the range of available data to ensure that analysts and policy makers have the best possible information on which to monitor and develop policy in Scotland.

Comments and suggestions relating to issues raised in this article would be welcomed. Please contact catriona.hayes@scotland.gsi.gov.uk

< Previous | Contents | Next >