Measuring European Construction Output: Problems and possible solutions

Transcription

1 Measuring European Construction Output: Problems and possible solutions JAMES L MEIKLE AND MAURIZIO T GRILLI Davis Langdon Consultancy United Kingdom Abstract Construction industries are often characterised as large or small by reference to the proportion which gross construction output represents of the national economy. While this may be adequate in developed western economies, it is not so helpful in developing economies, or economies in transition, where data on both construction output and the national economy are particularly uncertain. This paper presents available data, discusses evidence of inaccuracy and examines a number of methods for improving the quality and reliability of that data. The paper concludes that this is an important topic and that there are methods of addressing it, including establishing internationally agreed definitions for comparable international values, specifically for construction. Keywords: Construction, output, measurement, undercounting, international. INTRODUCTION Construction output is difficult to quantify with any accuracy. This is partly because of difficulties in definition: what should be included, what should be excluded? But it is also because it is intrinsically difficult to measure. Construction output in this paper refers to gross output: the total value of all inputs to construction work and the value added by the construction industry itself. In simple terms this is the price paid for construction work by the industry s ultimate customers. Typically, in national accounts, construction output is net output, the value added by the construction industry. Inputs - mainly building products and materials - will generally appear as outputs from manufacturing industry or mining. Gross output is usually measured; value-added is always an estimate. If construction is measured gross, in most countries it would be the single most important economic sector. The usual bases of measurement for gross construction output include surveys of contractors, estimates of work approved or completed, and financial returns by major clients of the industry. All of these, however, have their shortcomings. Construction work is geographically dispersed and takes time to complete; much of it is small scale and undertaken by small firms or individual tradesmen (who may be missed by surveys); and, when it is large scale, much of the work is often subcontracted (and then there is danger of double counting). Often physical measures such as m² built area are adjusted to value measures using dubious conversion factors. In addition, in many countries a significant proportion of construction work is informal, if not illegal, and, more often than not, is unrecorded or under-recorded. It can also be difficult to estimate the importance of do-it-yourself (DIY) and self-build activity. The usual units of measurement are, of course, national financial units. This is because these are the only common units we have to add together houses, roads, dams, schools, et cetera. But construction value is not necessarily a good indicator of construction volume particularly for international comparisons. For example, different (space and quality) standards and different conditions prevail in different countries and

2 price levels in some countries can be significantly higher or lower than those in others. Moreover, methods of measuring construction output are different across countries and no standard international procedures have been agreed (Eurostat, 1995 and United Nations, 1997). Even when physical measures of output (e.g., cement) are used, the differences in the use of the item measured are only partially related to the actual volume of construction. Construction output is quite heterogeneous in nature and country-specific construction processes can be important in relating resource inputs to construction output. The purpose of this paper is to review the problems in obtaining data reflecting real national construction output in selected countries and which is comparable between countries. The paper is in six parts including this introduction. Part 2 reviews published data on construction output for selected countries in Western and Eastern Europe. Part 3 discusses the evidence for inaccuracy. Part 4 looks at volume data as an indicator of construction output. Part 5 attempts to reconcile different data sets and Part 6 presents some conclusions and recommendations. CONSTRUCTION OUTPUT AND ITS SIGNIFICANCE Table 1 summarises key data for eighteen countries in Western and Eastern Europe. The first three data columns provide contextual information on the relative sizes of the countries population, economy and construction industry. The last two columns provide normalised comparable data. Output per capita is relatively high in Switzerland and Germany and relatively low in Spain and the UK in Western Europe, while it is relatively high in Slovenia and relatively low in Romania in Eastern Europe. The last column indicates the importance of gross construction output in the economy. Country Table 1: Indicators of construction output in selected countries, 1997 Population (millions) GDP ($US bn) Gross construction output ($US bn) Gross construction output per capita ($US) Gross construction output as a proportion of GDP (%) Finland , France , , Germany , , Italy , , Netherlands , Spain , Sweden , Switzerland , UK , , Weighted average , Czech Republic Estonia Hungary Latvia Poland Romania Russia Slovak Republic Slovenia , Weighted average Sources: based on IMF International Financial Statistics and Euroconstruct conference paper, Berlin, December 1998.

3 Table 1 indicates wide variations even in the last two columns. Relative to the Eastern European group of countries, the Western European countries tend to have high construction output per capita but low construction output as a proportion of GDP. In general terms the more mature economies have relatively high output per capita, in line with relatively high GDP per capita, but relatively low output as a proportion of GDP, largely because they are well provided with built infrastructure. Italy has close to archetypal Western European output characteristics; the Slovak Republic is a close to archetypal Eastern European economy. Slovenia is the Eastern European country most similar to the Western European ones. It is important to note, however, that the data in Table 1 are not necessary very reliable. The next three sections of this paper address and comment on this issue. EVIDENCE OF UNDERCOUNTING Construction output data are not consistent in content. There is no generally accepted standard international definition and what is included and excluded varies from country to country. The Statistical Office of the European Communities (Eurostat) has, however, suggested such a definition (Eurostat, 1995). According to Eurostat, construction is defined as an industry, not as an activity including households. It follows, therefore, that it is not possible or desirable to attempt to measure the construction activity of households. Another Eurostat conclusion is that industry output should be measured through sample surveys of construction firms including small firms and self-employed tradesmen. The authors have reviewed construction output data for two of the countries included in table 1 Finland and the UK. These two countries were selected because the relevant information was available. Table 1 indicates that Finland s construction output per capita and its construction output as a proportion of GDP are 50% and 42% respectively higher than UK s. Finland s figures are higher than the Western European average; UK s are significantly lower. There is evidence, however, that Finnish output tends to be overstated and UK output understated. Table 2 summarises what is known to be included and excluded in the two countries official construction output figures. Table 2: Construction output in Finland and UK Components of output Finland UK Contractors output (including sub and specialist contractors) included included Direct works organisations output * included included Unrecorded output not known not known Construction services included excluded Black economy in construction included largely excluded DIY materials included excluded DIY labour included excluded Taxes (including VAT) probably excluded VAT excluded * UK output explicitly includes public direct works output and explicitly excludes private direct works output. Finland includes public output; it is not clear about private output. Source: Euroconstruct, It is commonly agreed that construction services (design, management and other consultancy work) should be included in construction output. There is rather less agreement about the DIY sector: probably DIY labour should be excluded but possibly DIY material should be included. Similarly, treatment of the black economy varies from country to country. Generally, taxation should be included if it is paid by the industry s customers.

4 A recent report for the EU (Deloitte and Touche, 1997) suggests that the informal economy accounts for 12% of GDP in the UK. The percentage is considered to be around 35% in Greece, 25% in Spain and Italy and around 13% in France and German. The percentage of informal output in the construction sector is likely to be greater than that in the economy generally. Research from the University of Michigan Business School (1997) estimates the contribution of the illegal economy to GDP in Eastern Europe from 1989 to A comparison of their estimation for 1995 with Euroconstruct estimates of illegal work in construction in 1996 gave the following results: in the Czech Republic it was estimated to be around 11% in the total economy against 13% in construction; the figures for Hungary were 29% and 28%; the figures for Poland were 13% and 20%; and the figures for the Slovak Republic were 6% and 9%. In the former Soviet Union states the share of the black economy in GDP is reckoned to be even higher, on average 40% of GDP in 1995, from a low of 11% in Estonia to a high of 63% in Georgia. Recent work in the UK (Davis Langdon Consultancy, Construction Forecasting and Research, 1997) suggests that construction services, DIY work and the black economy in the construction industry account for 21% of total output. Hence the share is equivalent to more than 1.5% of GDP (21% of 7.4%) or around US$20 billion. Allowing for construction services and the black economy, but ignoring DIY activity, UK construction output might be as much as, say, 8.9% of GDP or US$1,966 per capita. Reducing Finnish construction output by, say, 2% to allow for DIY materials and labour would make it 8.5% or US$1,984 per capita. Hence, both output per capita and output as a proportion of GDP in Finland become lower than the UK. This example is illustrative only but it serves to support the argument that official output data even when it is considered reliable, may not be comparable. A review of definitions of construction output in European countries (Euroconstruct, 1997) suggests that new construction work by contractors is relatively well recorded but that data on work to existing buildings are rather less reliable. Construction services, DIY activity and the black economy may be included to at least some extent but are often based on estimates, rather than formal measurement; VAT is normally excluded. National construction output data are known to have been revised due to undercounting. In the UK, for example, there have been at least two occasions when output data were revised upwards - to take account of activity by firms missed by surveys (Hillebrandt, 1984). And in Egypt in the early 1980 s the authors know that national construction output data were adjusted to take account of unrecorded activity (World Bank, 1981). As suggested above, the purchasing power of different currencies can vary. Purchasing power parity is an alternative approach to the determination of exchange rates. The underlying economic concept is the law of one price, which says that a good should have only one price in a competitive environment. Purchasing power parities generalise the law of one price to the general price level of a country. According to this, the exchange rate between, say, British pounds and US dollars should equate the general British price level with the general US level. Table 3 lists GDP per capita in nominal exchange rate terms and in PPP (Purchasing Power Parity) terms for the same countries as Table 1 as calculated by the World Bank.

5 Table 3: GDP in nominal and PPP terms for selected countries, 1997 Country Per capita nominal GDP ($US th) Per capita GDP (PPP) ($US th) Ratio nominal to PPP Gross construction output per capita in PPP terms ($US) Finland ,078.6 France ,770.7 Germany ,245.6 Italy ,102.1 Netherlands ,389.6 Spain ,871.2 Sweden ,968.9 Switzerland ,424.5 UK ,567.7 Weighted average ,973.9 Czech Republic ,418.2 Estonia ,425.0 Hungary Latvia Poland Romania Russia Slovak Republic Slovenia ,273.9 Weighted average Sources: based on IMF International Financial Statistics, CIA World Factbook 1997 and Euroconstruct conference paper, Berlin, December GDP per capita in nominal exchange rates and in PPP terms tend to be similar in developed countries - they are identical in USA, because it is the base country. They diverge significantly in less developed and transition economies: the Czech Republic, Estonia and Poland have purchasing powers for their national currency more than double that obtained by using commercial exchange rates. The estimation of PPPs for the construction industry is difficult because output can be very different in content across countries. Variations in design, materials and standards can make comparisons problematic. Moreover, on average, construction generally relies on national resources and is non-traded. Vermande and van Mulligen (1999) show that the PPP methodology (applied by OECD/Eurostat) for the construction sector is not foolproof. If standard, identical buildings are compared there is a loss of representivity. On the other hand, if functionally similar buildings are used, the gain in representivity is offset by more heterogeneity and comparability will suffer. In this paper we approximate PPPs in construction with PPPs in the general economy. The implication is that the same amount of US$ buys about three times as much construction in Poland as it does in Switzerland. The evidence is that this is probably true for small-scale domestic construction but rather less true for major sophisticated foreign investment projects (where much of the material and some labour are likely to be imported). This section demonstrates that not only can definitions of construction output distort international comparisons but so also can international exchange rates. Table 3 suggests that general price levels in Finland are some 13% higher than those in the UK, hence the relative difference between construction volume in the UK and Finland should be adjusted again.

6 CEMENT CONSUMPTION AND DWELLING COMPLETIONS AS INDICATORS OF OUTPUT Other research (Meikle and Hillebrandt, 1988) suggests that some construction resources can be used as indicators of construction output. Cement is an example. It is usually manufactured in large-scale plants and reasonably reliable production data are maintained. Data are usually also available on exports and imports and, since cement does not have an extended shelf life, annual consumption figures can be calculated with reasonable confidence. Table 4 is compiled from data obtained from the international trade association. It presents data for fourteen countries. Excluding the highest and lowest figures in each group (Spain and Sweden for Western Europe and Hungary and Estonia for Eastern Europe), the range of national figures for cement consumption per capita is relatively narrow. It is narrowed even further when PPP s are used. Country Table 4: Cement consumption data for selected countries, 1997 Cement Cement consumption consumption per 000$ per capita (kg) construction output (kg) Cement consumption (000 tonnes) Cement consumption per 000 $ construction output in PPP terms (kg) Finland 1, France 18, Germany 34, Italy 33, Netherlands 5, Spain 26, Sweden 1, Switzerland 3, UK 12, Czech Republic 4, Estonia Hungary 2, , Poland 12, Slovak Republic 1, Sources: based on IMF International Financial Statistics, CIA World Factbook 1997 and DLC estimates based on Cembureau, Brussels. Cement consumption per unit of construction output in Eastern and Central Europe was formerly probably at least double that in Western Europe. This was because of the highly concrete-intensive technology used, for example, for high-rise multi-storey flats and factory buildings. It may also have been due to the low price put on construction output which would make cement, expressed in real terms, high in relation to undervalued output. The countries which have high apparent consumption of cement are often those where the GDP in PPP terms is higher than the money GDP and those with low consumption of cement are those where GDP in PPP terms is lower than the money GDP. Since the collapse of the command economy in Eastern and Central Europe, less cement-intensive methods of construction are being used so that some of the differences in apparent cement intensity may disappear in the future. Cement may be a useful indicator of the real volume of construction output. Table 4, for example, tends to support the earlier conclusions that, despite the official data, construction output in the UK may be higher, rather than lower, than construction output in Finland. The shortcomings, however, of cement as an indicator of construction volume include taking account of the types of technologies used: some countries use or used highly cement intensive construction technologies (World Bank, 1992). There can also be problems with data: while production figures are usually reliable, figures for exports and imports are often

7 less so. Cement data can, therefore, be a useful guide but cannot be a substitute for direct measures of construction output. Data on housing completions give some indication of one albeit a significant sector of construction output. However, they cannot be considered as a reliable indicator of the total amount of construction work. In the past the proportion of housing in total work was fairly stable in Eastern and Central Europe; now, however, disruption in the economies means that such relationships are much less reliable. Table 5 presents data on dwelling completions for thirteen countries. The data in the second column confirm that, generally, the construction industry builds more dwellings per capita in Western Europe. In the Western European countries only Sweden and, to a certain extent, the UK show low output. Conversely, the data in the third column, and to a lower extent in the fourth column, show a different picture. When normalised by output, housing completions are relatively higher in Eastern countries than Western countries. Table 5: Dwelling completions data for selected countries, 1997 Country Dwelling completions (000 s) Dwelling completions per 000 population Dwelling completions per m $ housing output Dwelling completions per m $ housing output in PPP terms Finland France Germany Italy Netherlands Spain Sweden Switzerland UK Czech Republic Hungary Poland Slovak Republic Sources: DLC estimates based on Euroconstruct conference paper, Berlin, December starts Generally, in housing statistics, a building is classified as residential when more than half its gross floor area is used for dwelling purposes. It is unclear whether this distinction for number of dwellings also applies to value of output. Some countries collect data on number of dwellings only in these residential buildings and there may or may not be a substantial number of dwellings in other buildings. There are also problems the other way round. It is common in some countries for commercial and professional offices to be located in primarily residential buildings so that the statistics of non-residential building may not reflect, and therefore understate, the actual new occupation of office space. Labour is an indicator which, in theory, should provide a reasonable estimate of construction output. Unfortunately, labour figures are often not reliable and can distort both inter-industry and international comparisons. One problem is the extensive use of part-time and self-employed workers. But the main problem is that a substantial number of construction workers undertake work on a piece-work basis and, therefore, their employment is not recorded. Brick consumption (or production) is another possible indicator. Again, substantial differences in building technology across countries and the ability to stockpile bricks indefinitely reduce the reliability of this indicator.

8 The next section brings together and compares some of the different approaches to assessment of construction output. RECONCILING ESTIMATES OF CONSTRUCTION OUTPUT AND OTHER INDICATORS This paper reviewed definitions of construction output and the differences these can make to the available data; it has also discussed how purchasing power parities can help to adjust nominal exchange rates value measures of output to a more comparable basis; and it has considered how physical resource units such as cement consumption can help to indicate output. Table 6 presents output data in nominal and PPP terms and cement consumption and dwelling data for fifteen countries. All output are expressed in terms of indices with the All Europe Average being equivalent to 100. Country Table 6: Comparisons of measures of construction output Gross construction Gross construction Cement consumption output per capita output per capita per capita in nominal terms in PPP terms index* index* index* Dwelling completions per capita index* Finland France Germany Italy Netherlands Spain Sweden Switzerland UK Average Czech Republic Estonia NA Hungary Latvia NA NA Poland Romania NA NA Russia Slovak Republic Slovenia NA NA Average Average - All Europe Index based on each column s average value for all countries listed NA = not available All of the data in Table 6 are presented in per capita terms but the output columns take no account of differences in definition. It is interesting to see that some columns indicate that some countries are very similar while other columns show very different figures. For example, Germany and Switzerland score high on every single indicator whereas Finland and Estonia score low only on cement consumption but this may

9 well be due to the particular construction techniques adopted in those countries. Gross construction output per capita in nominal terms is generally higher in Western Europe and always lower in Eastern Europe than the other measures. Generally, the physical measures (cement and dwellings) are as good if not better indicators of output per capita in PPP terms than output per capita in nominal terms. The table does not prove that cement or dwelling completions are consistently reliable indicators but suggests they may be used along with other measures. Presentation of the data in this way raises more questions than it answers. CONCLUSIONS AND RECOMMENDATIONS In order to demonstrate clearly the relative importance of construction output within and between countries it is essential that a standard international definition is established. This should be in current financial terms in national currency and should probably include: all construction work undertaken by contractors, tradesmen or direct works organisations at the value charged to the customer including tax, if appropriate. all technical professional services related to construction work (whether or not the work is undertaken) at the value charged to the customer including tax, if appropriate. This will include design and other technical services and project management services but will generally exclude financial and legal services unless directly associated with the construction work. Professional services should be included because increasingly the boundaries between work done by contracting and consulting companies are becoming blurred. In package deal and design and build contracts, for example, design and management charges are usually part of the total contract value and hence total construction output. Care should be taken to allow for all work whether formal or informal. If, however, construction output is not included in recorded economic activity so that both construction and total national output are underrecorded, then this should be taken into account when assessing the contribution construction makes to the economy. It is likely that there is rather more informal activity in construction than there is informal activity in the economy generally. A recent proposal from the European Commission suggested that lowering value added tax (VAT) on construction could help to increase revenue. It could also, incidentally, by reducing the volume of informal construction, improve construction statistics. Improved data on construction output could influence allocations of funding within the European Union and between the Union and other countries. Recognising that value measures of construction output are not necessarily good indicators of construction volume, more work needs to be done on international construction comparisons. PPPs need to be developed specifically for construction work. Initial work on construction cost comparison methodologies in European countries has been undertaken (Eurostat 1995; United Nations, 1997); this will need to be extended to less developed and transition economies. Having established that construction output is generally undercounted, does it matter? The authors of this paper believe that it does. It matters because it understates and, by implication, underepresents the significance of construction in national economies and in the world economy. It also matters because it distorts comparisons among countries. But, perhaps most of all, it matters because what is undercounted tends to be under-valued in national policies. The authors would welcome comments on the issues raised in this paper.

10 REFERENCES 1. Davis Langdon Consultancy and Construction Forecasting and Research (1997) Survey of UK Construction Professional Services. Construction Industry Council and Department of the Environment, London. 2. Davis Langdon Consultancy et al (1999), Framework for International Construction Cost Comparisons. 3. Deloitte & Touche (1997) The Black Economy in Europe (unpublished report for the European Union). 4. Euroconstruct (1997) European Construction Markets: New directions in European regional policies. Barcelona. 5. Eurostat (1995) Methodological Handbook for the Construction Sector Version 4, ISTI Short Term Indicators. Eurostat. 6. Hillebrandt, PM (1984) Analysis of the British Construction Industry, Macmillan. 7. Meikle, JL and Hillebrandt, PM (1988) The potential for construction resource planning, Habitat International. 8. United Nations (1997) International Recommendations for Construction Statistics. Department of Economics and Social Information Statistics Division, Statistical Papers, Series M, 47. New York. 9. The William Davidson Institute, University of Michigan (1997) Politics and Entrepreneurship in the Transition Economies, Working Paper no Vermande, HM and Van Mulligen, PH (1999) Construction Costs in the Netherlands in an International Context. Construction Management and Economics, 17, World Bank (1981) Arab Republic of Egypt, Construction/Contracting Industry Study (unpublished report). 12. World Bank (1992) Housing and the Construction Industry in the Russian Federation and Moscow, (unpublished report).