growth and productivity in the finnish trade industry ...euklems.net/pub/no2.pdf · growth and...

Growth and Productivity in the Finnish Trade Industry, 1975-2003: A National Comparative Perspective*

Jukka Jalava**

Abstract: We analyse the growth and productivity in trade in comparison with the rest of the Finnish economy. Economic growth shifted into a faster gear in the post-1995 era compared to earlier periods. Aggre-gate labour productivity change slowed down due to lacking contributions from capital deepening but aggregate multi-factor productivity growth stayed level. For the trade industries the post-recession period looked good. Trade joined the growth clubs in value added, labour productivity and multi-factor productivity. Unit labour cost growth was moderate and profits were high. However, vis-à-vis the level of labour productivity only wholesale trade was above the national average. Overall, the post-1995 productivity change was more concentrated than before as fewer industries than previously contributed to aggregate productivity growth.

Key words: growth, productivity, trade

* Financial support from Kesko is gratefully acknowledged. The opinions expressed are those of the author and

do not necessarily represent the views of Statistics Finland. I thank Matti Pohjola and seminar participants in Jyväskylä for helpful comments without implicating them for any remaining errors.

**Economic Statistics, Statistics Finland, BOX 6C, FIN-00022 Statistics Finland, Finland. E-mail: [email protected]

1. Introduction

What is a service? The casual answer to this question is that a service is something that is not a good (which is perceived as tangible), i.e. a service is something intangible. This is a misconception. In fact goods can be both tangible and intangible. Intangible goods are the results of creative and innovative endeavors, such as literary, scientific, and educational writings, musical compositions, computer software etc. (Hill, 1999). We find that Hill (1977) comprehensively defines what a service is:

‘A service may be defined as a change in the condition of a person, or of a good belong-ing to some economic unit, which is brought about as the result of the activity of some other economic unit, with the prior agreement of the former person or economic unit’.

Finland trails leading economies in the productivity of services. The labour productivity (LP) in wholesale and retail trade was according to some calculations (van Ark and Timmer, 2001) only 56 per cent and in transport and communication 86 per cent of the US level in 1990. It is, however, un-clear how much (and for what reasons) we presently are behind leading countries in services produc-tivity. Are there barriers to adopting best practices due to a lack of competition or regulations? The objective of this paper is to compare the productivity development in trade with that of the rest of the economy. This topic is important since the share of services in production is ever increasing and the so called Baumol’s disease predicts that overall productivity growth will slow down as resources shift to service industries whose productivity growth is slower than that of e.g. manufacturing. In the U.S. close to three quarters of GDP stems from total services, whereas the share of services in Finnish GDP was almost two thirds in 2001 (OECD, 2003). The good news is that Mankinen, Rouvinen and Ylä-Anttila (2002) found in an international comparison that - unlike in manufacturing – there is not necessarily a negative correlation between the share of employment and productivity in services. The detailed quantitative account of productivity in the Finnish trade sector is, however, still missing. A gap that this paper tries to fill. We start by taking a broad look at the growth in trade and the rest of the economy within the framework of national accounts. We also calculate the changes in and levels of labour productivity. Then the impact of structural change on LP is determined and finally the growth of multi-factor productivity (MFP) is estimated. With structural change is in this context meant the shift of labour into industries with either a higher level of or higher growth rate of LP. In manufacturing the impact of this creative destruction has been significant in Finland since the mid-1980s (Maliranta, 2003). In the case of services it is unclear what the impact of creative destruction has been. Perhaps it is still to happen? What are the reasons for the low productivity in Finnish ser-vices? The reasons cannot be connected to education or infrastructure, since these factors also affect secondary production. Therefore the reason – or culprit – must be the level of efficiency and MFP. It would seem that, for some reason or other Finnish service companies do not use the best possible technology. Since this technology – at least in principle – is readily available, there must be some practical obstacles preventing its adoption. It might simply be the case that companies do not have incentives or compulsion to adopt best technology. Lack of competition might also be an explanation, which could be due to the smallness of the market. Lately competition has become more intense as Kappahl (in 1990), IKEA (in 1996), H&M (in 1997), Bauhaus (in 2000), Clas Ohlson (in 2002) and Lidl (in 2002) among others have entered the Finnish market.

Productivity measurement requires the decomposition of production into value, quantity and price.

2

This is particularly difficult when studying productivity in services, since for instance average earn-ings indexes are being used in some industries as proxies for price changes of service output. The development work on service statistics is ongoing both in international fora and in Statistics Finland. The focus of the development work is on the methods used compiling producer price indexes for the service sector1, on the classification of service products and on statistics on service production by products. In this paper the main focus is in a national comparative perspective on the trade industry, which is an interesting industry that is dependent on the production and import of tradeable goods on one hand and the demand for these goods on the other hand. As wholesalers and retailers are service providers the gross output of trade industries is the trade margin which is the difference between the price paid and the price at which a good or service is sold. And not as in the usual case where the gross output of an industry consists of the goods or services produced by its establishments and the prices paid for goods or services are separately recorded as intermediate consumption. In this paper the trade sector is divided into three sub-industries: industry ISIC 50 Sale, repair and maintenance of motor vehicles; service stations, industry ISIC 51 Wholesale trade and commission trade and industry ISIC 52 Retail trade; repair of household goods.

The outline of the paper is the following. In section 2 we briefly survey previous research. Sections 3 and 4 look at the growth and productivity in trade and the rest of the economy. Section 5 concludes.

1 In October 2004 Statistics Finland will start publishing producer price indexes for: hotelling services, legal activities, accounting, bookkeeping and auditing activities, letting of business premises, renting and mainte-nance of textiles and technical testing and analysis.

3

2. Previous research

Baumol’s (1967) hypothesis of unbalanced growth is central in any discussion of productivity in ser-vices. According to Baumol productivity growth in the whole economy will slow down as resources shift to service industries whose productivity growth is slower than that of e.g. manufacturing. Oulton (2001), showed that this is only the case for those service industries that produce final goods, and not for the industries producing intermediate goods. Another important hypothesis was Griliches’ (1992) concern that when the difficult to measure industries share of the economy grows, it will lead to a slowdown in aggregate productivity. In a recent paper Gordon (2002) states that the post-1972 slow-down in US productivity is not due to the increasing share of difficult to measure industries in the economy, in addition Gordon is of the opinion that the step-up in US productivity in the late 1990s is not connected to output measurement problems. The large share of services in US output is not neces-sarily a drag on aggregate growth as Baily and Lawrence (2001) show that industries using informa-tion and communication technology (ICT) have performed better than the private sector on average in the US in 1995-99. In a similar vein van Ark (2001) noted in an analysis of ten industrialized coun-tries, that the contribution to aggregate LP of service industries using ICT increased in all countries (except Japan) from the early 1990s to 1995-99. Jalava and Pohjola (2002) observed that ICT-producing industries contributed one third of Finnish market sector output growth in 1995-99 and Jalava (2003) pointed out that the level of labour productivity declined from 1975 to 2001 in those Finnish industries that neither produced nor used ICT. What is new in international productivity com-parisons of service branches is represented by Baily and Zitzewitz (2001), who report of the McKinsey Global Institutes projects (e.g. McKinsey, 1992) which quite innovatively strive to deter-mine the international levels of productivity in service industries. For instance in retailing Baily and Zitzewitz calculated productivity by weighing the absolute productivity of different retailing formats (mass merchandising, out-of-town specialized chains, in-town specialized chains, department stores, mail order and traditional stores) with their share of employment. This takes into account the fact that the trend in retailing is that specialized chains come up with innovations that discounters copy and offer at lower prices to consumers.

The classic account of growth and productivity in the Finnish trade industry is Forssell (1979). Forssell’s work was part of the historical national accounting effort that was carried out under the auspices of the Bank of Finland’s Growth Studies Committee. Forssell (1979) was the tenth mono-graph in a series that started in 1966 with a study on Finnish agricultural production and ended in 1988 with monograph number thirteen that presented a consistent historical national accounting view on Finnish economic growth and structural change from 1860 onwards (Hjerppe, 1988). Forssell es-timated the volume of value added in 1860-1900 based on the number of persons engaged in trade. The series from 1900-1960 are more complete and contain data on employment, average wages and salaries, volume of sales, price of sales and labour productivity in trade. Forssell was not content with simply presenting the results of the arduous compilation of the historical series. He also used regres-sion analysis to model developments in trade sector value added. Explanatory variables were average population, share of urban population in total population and real GDP per capita.

A more recent look at Finnish service sector productivity is Mankinen, Rouvinen and Ylä-Anttila (2002). They used growth accounting tools to analyze how services have performed in a national per-spective. The good news is that they found a positive correlation between the level of productivity and

4

the employment share in services, quite contrary to manufacturing. Thus a post-industrial country need not necessarily face lower productivity growth. For both wholesale trade and retail trade Manki-nen, Rouvinen and Ylä-Anttila (2002) reported a step-up in both volume of value added and multi-factor productivity in 1995-2000 compared to the 1975-95 period.

5

3. Growth…

The importance of services to Finnish GDP increased rapidly in the latter part of the 19th century as trade was liberalized, railways were built and Finland was electrified. For commerce the important milestones were the permission to open up stores in the countryside in 1859, the reduced regulation of trading licenses in 1868 and the full freedom of trade in 1879 (Forssell, 1979). Hence the share of trade in GDP increased from a meager 3 per cent in 1860 to almost 7 per cent in 1900 and 10 per cent in 1950, a level which has remained constant for half a century. Characteristic to the Finnish long run economic transformation was that industrialization started late and that services increased directly at the expense of primary production, as the share of secondary production in GDP did not decrease until the 1970s (figure 1). This is in contrast with the classical view of historical development in many developed countries where the main contributor to economic growth first shifts from primary produc-tion to secondary production during the process of industrialization, and subsequently from secondary production to tertiary production as the post-industrial stage is entered.2

Figure 1 Value added as a share of GDP, 1860-2003*

0 %

20 %

40 %

60 %

80 %

100 %

1860

1872

1884

1896

1908

1920

1932

1944

1956

1968

1980

1992

TERTIARYPRODUCTIONSECONDARYPRODUCTIONPRIMARYPRODUCTION

*=Preliminary estimate. Sources: Hjerppe (1988); Statistics Finland.

3.1 From gross output to value added In this paper the trade industry is broken down into three sub-industries. Industry ISIC 50 Sale, repair and maintenance of motor vehicles; service stations contains all wholesale, brokerage and retail sale of new and used motor vehicles and retail sales of fuel and lubricants as well as the maintenance and repair of motor vehicles and the installation, repair and studding of tires other than during manufac-ture. Industry ISIC 51 Wholesale trade and commission trade consists of the resale of new and used goods to retailers and other wholesalers and allows the industry both to acquire ownership of the goods, which is wholesaling, or not, which is commission trade. Industry ISIC 52 Retail trade; repair of household goods consists of the sale to consumers for personal and household use of new and sec-ond-hand goods as well as the repair as a principal activity of footwear, household electrical appli-

2 This view has been challenged by Broadberry (1998) who argues that Germany and the United States sur-

passed Britain’s level of aggregate labour productivity by shifting resources out of agriculture and improving the productivity of services rather than manufacturing.

6

ances, watches, clocks, jewelry, etc. The gross output of trade industries is the trade margin which is the difference between the price paid and the price at which a good or service is sold. The ideology is that wholesalers and retailers are thought of as service providers (SNA93, para. 6.110). The goods purchased for resale are not treated as intermediate consumption but are directly deducted from the gross output. Included in intermediate consumption is external services, leasing and other rents as well as other fixed or variable expenses. The output and intermediate consumption of trade in Finnish national accounts is calculated by combining data from the business register3 and the structural busi-ness statistics4. The structure of estimates is obtained from structural business statistics (e.g. the per-centage of output in turnover) and the levels are obtained from the business register (turnover and wage and salary levels) with the establishment as the unit.

The statistical unit in structural business statistics is the enterprise as the statistical unit used by na-tional accounts is the establishment. An establishment is a production unit belonging to an individual non-financial corporation or quasi-corporation, situated in a single place and mainly producing one type of good or service. Thus an enterprise with many establishments can have its economic activity recorded in various industries depending on the main activity of each establishment. Establishments are sometimes called local kind of activity units.

In order to ascertain changes in the structure of gross output in trade we look at the input coefficients from the most recent input-output studies.5 The input coefficients are from the part of the symmetric industry by industry input-output tables that record the direct inputs from other industries required to produce one unit of gross output in each industry. To save space and to clarify the analysis we only show the columns for trade in table 1. Thus e.g. for the gross output of 20,778 million Euros in trading 142 million Euros worth of outputs from hotels & restaurants were needed as inputs in 2000. This means that the input coefficient is 0.7 per cent (=142/20,778). From 1980 to 1989 the share of inter-mediate consumption in the trade sector gross output decreased from 40 to 33 per cent. The decline stemmed from the use of domestic products as all domestic sectors’, except the financial sector’s and real estate & business services’, relative shares decreased. The step-up in the share of value added went to the compensation of employees (from 45 to 50 per cent) and consumption of fixed capital (from 5 to 8 per cent) as the net operating surplus actually shrank to 8 per cent (having been 11 per cent). In 1989 the use of domestic products constituted 30 per cent of gross output. Real estate & business services supplied 11 percentage points of the domestic inputs, transport & communications 10 percentage points, total manufacturing 3 percentage points and the financial sector and trade itself 2 percentage points, respectively. The share of imports was only 3 per cent. In 2000 the share of the use of domestic products in the trade sector’s gross output was 36 per cent. Total manufacturing con-tributed 10 percentage points of it as did transport & communication. Trade and real estate & business

3 The business register covers all enterprises, self-employed persons and non-profit corporations in the capacity of employers, recorded in the Value Added Tax Payment Register or the PAYE Register (pay-as-you-earn/Employee’s Advance Tax Declaration Register). 4 The structural business statistics contains combined enterprise data from the business register, the business tax register and direct survey data. 5 They are for the statistical years 1980, 1982, 1985, 1989, 1992, 1993, 1995, and 2000. As these studies were made 1 to 21 years ago the earlier ones do not exactly match the current national accounting data due to changes in the international national accounting standard. E.g. the earlier years’ shares of value added in gross output shown in table 3 are not exactly the same as those calculated using the most recent data in table 2.

7

services supplied inputs worth 6 percent of gross output, respectively. As also imports increased to 8 per cent the share of intermediate consumption was 45 per cent which is the highest observation throughout our observation period. The high share of intermediate consumption naturally implies a low share of value added. However, as especially the share of compensation of employees but also consumption of fixed capital actually declined (to 33 per cent and 6 per cent) the net operating surplus was a record 15 per cent of output.

In addition to the direct input requirements as portrayed by the input coefficients there are also indi-rect impacts. The indirect impacts stem from the fact that in addition to the primary inputs required to produce the gross output secondary inputs are required to produce the primary inputs, tertiary inputs to produce the secondary ones and so on ad infinitum. The Leontief inverse matrix L takes into ac-count all these direct and indirect effects. It is calculated by subtracting from the identity matrix I the input coefficient matrix IC and taking the inverse of the thus calculated matrix (see the Appendix for a clarifying example and Statistics Finland, 2003, for the exact definition):

(1) L = (I-IC)-1.

Table 1 also contains the column for trade from the Leontief inverse matrix. The column total shows the total domestic inputs needed to produce one unit of output. As the columns in the inverse matrix show the multiplier effects backwards they are often used for analytical purposes. The rows would have shown also the downstream impacts but rows have been omitted for the economy of exposition. In 1980 were 1.562 units of inputs needed to produce one unit of trade output. In 1989 only 1.447 inputs were needed per trading output. Now it only took 0.1 units of manufacturing goods to produce one unit of trade goods compared to 0.2 in 1980. In 2000 more inputs (1.571) were needed per output. It took 0.01 unit of primary production, 0.183 units of total manufacturing, 0.011 units of construc-tion, 1.086 units of trade, 0.011 units of hotels & restaurants, 0.131 units of transport & communica-tions, 0.015 units of financial sector inputs, 0.088 real estate & business services inputs and 0.036 other services inputs to produce one unit of trade output. Alternatively we could think that if trade output was to increase by one unit these would be the impacts on the industries supplying inputs for trade.

8

Table 1 Input coefficients and inverse matrixes for trade in selected years, %

Source: Own calculations, data from Statistics Finland (1983, 1985, 1988, 1992, 1996, and 2003).

Table 2 Value added by industry as a share of gross output, %-share

*=Preliminary estimate. Source: Own calculations, data from Statistics Finland.

9

The aggregate share of value added in gross output increased from 1975 to 1990, as agriculture, total manufacturing, trade, and transport & communication managed to extract more value added per unit of gross output (table 2). During the 1990s an increasing share of the economic activity was out-sourced either to other domestic producers or substituted with imports. This was the case also in the trade industries where the share of intermediate consumption in gross output increased by 3 to 6 per cent when comparing 2003 with 1975.

In 1975 Finland’s gross domestic product (GDP) at basic prices, or the sum of all industries’ value added, was 16.1 billion Euros (Table 3). Half of the value of GDP was still generated in primary and secondary production. The value of the goods and services generated in the whole economy increased by a factor of 7.7 from 1975 as GDP was 123.3 billion Euro in 2003. By 2003 the combined share of primary and secondary production dwindled to one third with especially agriculture and construction but also total manufacturing losing ground. Among the service industries real estate & business activi-ties significantly increased its share from more than 10 per cent in 1975 to 19 per cent in 2003. Other big gainers were the transport & communication industry and health & social work. All service indus-tries increased their share of the economy with the exception of the trade industry and hotels & restau-rants. Within trade wholesale trade and the sale of motor vehicles (including service stations) held their ground but retail trade decreased by more than a percentage point. FISIM, or financial interme-diation services indirectly measured, is shown separately as it has not been allocated by industry as yet.6 Nominal value added in service industries grew the fastest. That is, in all service industries ex-cept the sale of motor vehicles, retail trade and hotels and restaurants. Trade matched the national growth of current price value added until 1989-90. The recession was especially severe in the sale of motor vehicles and wholesale trade. Wholesale and retail trade regained their 1990 levels in 1996 as did the sale of motor vehicles the following year. However, of the trade industries only wholesale trade could reclaim its position relative to the national average. That happened as late as 2002.

The growth in nominal gross value added can be decomposed into price7 and volume components (table 4). Thus, e.g. the growth of 5.4 p.p.a. in nominal value added in the trade industry in 1995-2003 can be decomposed into a price component of 1.0 p.p.a. and a volume component of 4.4 p.p.a. as shown in table 6 (due to rounding and averages the components do not always exactly sum to the totals). The implication of a slower price growth is that given a nominal growth rate the slower the price growth is the higher the growth of real value added will be. In 2003 the quantity of GDP is twice the size it was in 1975. In the 1975-1990 period seven industries managed to match or surpass the average growth rate, wholesale trade was one of them, but in 1995-2003 only six as health & social and other services work fell behind. Retail trade managed to join the rapidly growing industries.

6 FISIM is services provided by financial intermediaries to their customers that they only indirectly are charged for by receiving a lower interest on deposits than the interest rate charged for loans. The Commission (Regula-tion no. 1889/2002) stipulated that FISIM shall be allocated by industry starting from the statistical year 1995. When this allocation is performed the value added of the industries will change depending on which institutional sector is the user of the FISIM. For companies FISIM is intermediate consumption, for households and govern-ment FISIM is final consumption and for non-residents it is exports. 7 As the prices used in national accounts ideally are indicators of pure price changes the quality improvements are reflected in the volume growth.

10

The price index of GDP grew during the observation period at an annual average rate of 4.7 per cent8 which means that prices were 3.7 times higher in 2003 than they were in 1975. Of the service indus-tries only the prices in sale of motor vehicles, retail trade and the financial sector grew slower than the national mean. The GDP inflation rate slowed down in the 1990s as the average growth of the price of GDP slowed down from 7.2 per cent per annum (p.p.a.) in 1975-90 to 2.4 p.p.a. in 1990-95 and to 1.4 p.p.a. in 1995-2003. The prices in trade grew at an average rate in 1975-90, above average in 1990-95 (3.8 p.p.a. due to the 5.7 p.p.a. growth in wholesale trade) and below average in 1995-2003 (1.0 p.p.a., with retail trade showing price decreases of 0.1 p.p.a.). Price growth was only in primary and secon-dary production and the financial sector slower post-1995 than in retail trade.

8 In this paper the growth rates are expressed logarithmically. Thus growth is defined as: 100*[ln(xt )/ln(xt-1 )].

11

Table 3 Value added by industry as a share of GDP, %-share and value in million Euro


12

Table 4 Average growth of value, quantity and price of value added by industry, 2000=100, ln-% Value Quantity Price 1975-1990 1990-1995 1995-2003* 1975-1990 1990-1995 1995-2003* 1975-

1990 1990-1995 1995-2003*

A Agriculture, forestry & hunt. 7.4 -5.2 1.3 0.4 -0.6 2.0 7.0 -4.6 -0.7B Fishing 6.7 2.5 -3.3 1.8 6.5 -1.4 4.9 -3.9 -1.9CDE Total manufacturing 9.4 4.1 3.4 3.7 2.2 4.6 5.7 1.9 -1.2F Construction 9.2 -10.6 6.9 1.5 -9.9 2.9 7.7 -0.7 4.0G Trade 10.1 -0.6 5.4 3.0 -4.4 4.4 7.2 3.8 1.0 Sale of motor veh.; serv. stat. 10.7 -3.0 6.0 2.6 -2.9 5.1 8.1 0.0 0.9 Wholesale trade 10.4 -0.2 6.3 3.5 -5.9 4.5 7.0 5.7 1.8 Retail trade 9.5 -0.1 3.8 2.3 -2.2 3.8 7.3 2.1 -0.1 H Hotels & restaurants 11.4 -1.7 4.7 2.7 -2.1 2.8 8.8 0.3 1.9I Transport, storage & comm. 11.3 3.0 6.3 3.4 1.6 4.9 7.9 1.3 1.5J Financial interm. & insurance 13.6 -1.7 2.0 5.7 -7.0 4.2 7.9 5.3 -2.2K Real estate & business act. 11.6 5.9 7.2 4.2 2.1 4.0 7.5 3.9 3.2L Admin., comp. soc. serv. 11.5 1.9 4.1 2.4 -0.8 1.1 9.0 2.7 3.0M Education 11.1 2.7 4.4 2.1 0.2 1.8 9.0 2.6 2.5N Health & social work 13.3 2.4 4.8 4.0 -1.6 2.0 9.3 4.0 2.8O Other comm., soc. & pers. s. 13.1 1.7 4.6 4.0 -1.0 2.5 9.1 2.7 2.0P Household service activities 7.9 8.9 11.4 -5.3 9.4 10.1 13.2 -0.5 1.4FISIM 13.2 -2.9 1.5 5.7 -3.8 6.4 7.5 0.9 -4.9GDP at bp 10.4 1.7 4.9 3.1 -0.7 3.6 7.2 2.4 1.4*=Preliminary estimate. Source: Own calculations, data from Statistics Finland.

13

3.2 Hours worked and employment The hours worked in the Finnish economy have declined from 4.4 billion hours in 1975 to 4.1 billion in 2003 (table 5). During the early 1990s economic recession the labour input went as low as 3.6 bil-lion hours in 1993-4. As was the case with value added also the labour input of primary and secondary production declined the most while nearly all service industries increased their hours worked. In 1975 more than half of the hours worked of the whole economy were in primary and secondary production but in 2003 only one third. Agriculture suffered the biggest loss with the share of hours worked de-clining by 12 percentage points to 7.0 per cent. Total manufacturing lost 6 percentage points ending at 18.4 per cent. Other industries whose shares declined were construction (from 10.3 to 8.1 per cent), trade (from 13.7 to 12.8 per cent, due to decreases in retail trade), financial intermediation & insur-ance (from 1.9 per cent to 1.5 per cent) and fishing (from 0.7 to 0.4 per cent). The largest gainers were health & social work and real estate & business activities that ended up at 12.9 and 9.9 per cent of the hours worked in 2003, respectively.

A good measure of the effectiveness of labour input in a time-series perspective is how many worked hours are required per one million Euro at constant year 2000 prices. As expected the labour input required in primary production to generate one million Euro of value added was markedly higher than in either secondary or tertiary production (table 6). The most effective industry in 1975 was the real estate & business services industry that needed only 19,500 hours per unit of value added. The second and third most effective industries were the education and the financial intermediation & insurance industries that needed 39,000 and 43,300 hours to achieve the same feat. In 1975 the trade industry was more effective with 84,800 hours worked per unit of value added than the 89,400 hours required in total manufacturing but less effective than construction (67,100 hours). Wholesale trade was the most effective trade industry (50,900 hours) followed by the sale of motor vehicles (84,600 hours) and retail trade (143,000 hours). In 2003 the hours worked required for one million of value added at the level of the whole economy were 34,200. Primary production, hotels & restaurants and household service activities still failed to reach the national average. A new feature was that administration and personal services (industries L through O) and construction and trade failed to do so too. The trade industry was not far off the mark (41,000 hours) with wholesale trade reaching 26,400 hours. The most effective industries were financial intermediation & insurance (14,000 hours), real estate & business activities (19,200 hours), total manufacturing (22,200 hours) and transport & communication (24,400 hours).

14

Table 5 Hours worked by industry as a share of total, %-share and quantity in 100,000h


Table 6 Hours worked per one million Euro of value added at 2000 prices


15

Total employment, the aggregate number of employees and self-employed, has fluctuated during our observation period between 2 and 2.5 million persons (table 7), when the population of Finland has increased from 4.7 million persons to 5.2 million persons. In 1975 employment was 2.3 million and the unemployment rate was at a record low level: 2.6 per cent. Employment increased steadily – with the exception of a temporary hiccup at the end of the 1970s – reaching the 2.5 million mark at the close of the 1980s, when unemployment went close to 3 per cent. As a result of the early 1990s reces-sion the amount of employees and self-employed dwindled to 2.0 million by 1994 and unemployment soared to 16.6 per cent. The post-1995 economic boom increased employment which rebounded to 2.4 million in 2003, but the unemployment rate only decreased to 9.0 per cent. Primary and secondary productions were big losers in employment. Agriculture’s share declined from 14.9 per cent in 1975 to 5.1 per cent in 2003, total manufacturing’s share from 26.2 to 19.6 per cent and construction’s share from 9.3 to 6.5 per cent. Of the service industries trade lost almost one percentage point (due to a reduction in retail trade) ending up at 12.8 per cent and financial intermediation & insurance shed half a percentage point. All other service industries increased their relative shares. More than 6 per-centage points gains were experienced in real estate & business activities and health & social work.

When looking at employees and self-employed separately the main difference with the development in total employment is that the self-employed failed to regain their pre-recession level in absolute terms as there were 430,000 self-employed in 1975 and 280,000 in 2003 (tables 8 and 9). Another difference in self-employment, when contrasting with total employment and the case of the employ-ees, is that it did not peak prior to the 1990s recession. In self-employment the main story is one of a massive decline in agricultural self-employment while all other industries except fishing and the three that do not have self-employment (financial intermediation & insurance, administration and house-hold service activities) have increased their relative shares. The trade industry increased its share by three percentage points to 10.2 per cent of all self-employed in 2003. The share of services in self-employment is likely to keep increasing as Rouvinen and Ylä-Anttila (2004) found that in 2003 three quarters of starting entrepreneurs were in services.

In 1975 there were 1.9 million employees. The number of employees peaked in 1989-90 at 2.1-2.2 million and was at its lowest in 1994 at 1.7 million. By 2003 the amount of employees rose back to 2.1 million persons. Total manufacturing’s share of the number of employees declined during our observation period by 10 percentage points to 21.2 per cent in 2003. Other losers were construction, trade (that ended up at 13.1 per cent with a decline of nearly two percentage points mostly due to re-tail trade but also wholesale trade), agriculture and the financial sector. The share of employees stayed close to level in hotels & restaurants, transport & communication and household service activities. The largest gainers were health & social work, real estate & business activities and education.

16

Table 7 Employees and self-employed by industry, %-share and amount in 100 persons


Table 8 Self-employed by industry, %-share and amount in 100 persons


17

Table 9 Employees by industry, %-share and amount in 100 persons


Table 10a Hours worked per employee and self-employed


18

Table 10b Hours worked per self-employed


Table 10c Hours worked per employee


19

The average amount of work done by employees and self-employed decreased from 1,900 to 1,700 hours per year during our observation period (table 10a). The longest annual hours were performed in primary production and the shortest working time was in education: 1,400 hours per year in 2003. Below average hours were worked also in health & social work, other community, social & personal services and household services with 1,500 hours per year, respectively, and financial intermediation & insurance and total manufacturing with 1,600 hours in 2003. The working time was average in trade and real estate & business activities, administration, and more than average in construction, ho-tels & restaurants, transport & communications, and primary production. Self-employed persons worked more than employees: 2,400 hours per year (table 10b). Entrepreneurs in trade managed to cut their hours from 2,600 per year in 1975 to 2,300 per year in 2003. The amount of work performed by employees annually declined from 1,800 to 1,600 hours during our observation period (table 10c). Longer than average hours were worked in primary production, construction, sale of motor vehicles, wholesale trade, transport & communication and administration.

3.3 Wages, average earnings and consumption Employees’ nominal average earnings from wages and salaries grew from 4,400 Euro per year to 27,100 Euro per year, or 6.2 times, from 1975 to 2003 (table 11). During the same time the cost of living index (1951=100) grew from 418 points in December 1975 to 1577 points in December 2003 or by a factor of 3.8 meaning that real wages were 2.4 times higher at the end of our observation period than at the beginning. The increased purchasing power of households resulted in a 7.7 fold increase in the consumption expenditure of households in Finland from 1975 to 2003 (table 12). Households’ propensity to save declined as the savings ratio (savings per disposable income) changed from a pre-recession arithmetic average of 3.2 per cent to a post-recession average of 0.8 per cent. The debt ratio (disposable income per stock of credit) of households was 37.9 per cent in 1975 but as much as 74.6 per cent in 2003. The structure of consumption also changed as durable goods, semi-durable goods and especially non-durable goods lost ground to services. The shares of more traditional consumption groups such as food and beverages, alcohol and tobacco, clothing, furnishings and transport declined, whereas housing, health, communications, recreation & culture, education and miscellaneous goods and services increased their shares. In the 1990s housing expenditures increased to a full quarter of the total.

In 1975 the average wage in the trade industry was 4,000 Euro per year (2,700 in retail trade, 3,700 in sale of motor vehicles and 6,000 in wholesale trade) which was 90.9 (=4,000/4,400) per cent of the national average. In 2003 the average wage in the trade industry was 24,000 Euro which was 88.9 per cent of the national average. The highest annual wages and salaries were in financial intermediation & insurance (36,500 Euro), wholesale trade (33,500 Euro), real estate & business services (31,300 Euro) and total manufacturing (31,200 Euro). Eye-ball econometrics tells us that in 1975 the average earn-ings were average or more than average in eight of the shown industries, but in seven industries in 1990 and in 2003. To ascertain whether the relative variation of wages around the mean has increased or decreased we calculated the coefficient of variation (CV) in these years. The coefficient of variation was obtained by dividing the standard deviation s (which is equal to the square root of the arithmetic

mean of the squared deviations from the mean) with the mean X (see Feinstein and Thomas, 2002).

(2) XsCV /= , where

20

(3) ( )

nXX

s i∑ −=

2

.

Using the data in table 13 (excluding sale of motor vehicles, wholesale trade and retail trade as they are included in the aggregate trade industry) we found that the standard deviations in 1975, 1990 and 2003 were 1.0, 3.9 and 6.1 respectively. When dividing these standard deviations with the years’ means 4.4, 18.2 and 26.1 the coefficients of variation were 0.227, 0.214 and 0.223 which show a de-crease in the relative variation of wages in 1990 compared with 1975 and an increase from 1990 to 2003.

Table 4 showed the interdependence between the value, price and quantity of value added. A similar relationship exists between the average growth of wages and salaries, employees and average earnings by industry as can be seen in table 13. For instance, the growth in the wages and salaries of wholesale trade in the period from 1995 to 2003 was 5.9 p.p.a. (faster than the average 5.3 p.p.a. for the whole economy). As the number of employees in wholesale trade increased during the same time by 3.2 p.p.a. therefore average earnings rose by 2.7 p.p.a. At the level of the whole economy the number of employees increased moderately in 1975-90 (by 0.9 p.p.a.) and wages and earnings rapidly (10.3 p.p.a. and 9.4 p.p.a. respectively). During the early 1990s the wage sum decreased (-1.1 p.p.a.), how-ever, as the number of employees diminished even more (-3.8 p.p.a.) average earnings increased at a pace of 2.8 p.p.a. After 1995 earnings grew at a rate of 3.3 p.p.a. as employees increased by 2.0 and the aggregate wage sum by 5.3 p.p.a. The only two industries that have consistently managed to main-tain a faster-than-average growth in average earnings is total manufacturing and financial intermedia-tion & insurance. Total manufacturing did the trick by having a slower than average growth in the number of employees in 1975-90 and 1995-2003 and a higher than average growth in the wage sum during the recession. In financial intermediation the wage sum rose faster than average in 1975-90 and the sectors’ employees decreased in 1990-2003. In retail trade earnings grew at a faster rate than the national average in 1975-95, due to lower than average growth in the number of employees, and after 1995 the growth was slower than the national average (2.9 p.p.a.). The 1990s recession was the only time that wholesale trade could reach an average growth rate in earnings; whereas the sale of motor vehicles did it from 1975-1995. Only during the post-1995 era was it slightly lagging behind the nation’s average.

21

Table 11 Average earnings (wages and salaries) per employee, Euro


Table 12 Household consumption by purpose and durability, %-share and value in million Euro


22

Table 13 Average growth of wages and salaries, employees and average earnings by industry, ln-%

Wages and salaries Employees Average earnings

1975-1990 1990-1995 1995-2003* 1975-1990 1990-1995 1995-2003* 1975-1990 1990-1995 1995-2003*

A Agriculture, forestry & hunt. 6.9 -5.4 2.2 -1.1 -5.4 -1.3 8.0 0.1 3.5

B Fishing 10.7 15.8 0.0 3.4 6.7 -4.2 7.3 9.0 4.2

CDE Total manufacturing 8.7 0.4 4.4 -1.0 -4.0 0.6 9.7 4.4 3.8

F Construction 8.4 -11.1 7.2 -0.7 -12.5 3.5 9.1 1.4 3.7

G Trade 9.7 -2.7 5.8 0.3 -6.1 2.8 9.3 3.3 3.0

Sale of motor veh.; serv. stat. 11.4 -4.9 6.8 1.9 -7.7 3.7 9.5 2.8 3.1

Wholesale trade 8.8 -1.2 5.9 -0.2 -4.1 3.2 9.0 2.9 2.7

Retail trade 10.1 -3.8 5.1 0.3 -6.8 2.2 9.8 3.0 2.9

H Hotels & restaurants 10.9 -3.0 5.3 0.7 -5.7 2.5 10.1 2.7 2.8

I Transport, storage & comm. 10.0 -0.4 4.6 1.1 -3.5 1.0 8.9 3.2 3.7

J Financial interm. & insurance 11.8 -1.1 0.8 1.8 -6.3 -2.6 9.9 5.2 3.4

K Real estate & business act. 13.6 -0.6 10.1 4.5 -2.4 6.4 9.1 1.8 3.8

L Admin., comp. soc. serv. 11.1 0.5 4.4 1.8 -1.0 1.3 9.3 1.5 3.2

M Education 11.0 1.7 4.7 1.8 0.1 2.0 9.2 1.5 2.7

N Health & social work 13.4 1.4 5.0 4.5 -1.5 1.8 8.9 2.8 3.2

O Other comm., soc. & pers. s. 12.7 0.2 5.0 3.5 -1.5 2.7 9.1 1.8 2.3

P Household service activities 7.3 9.0 11.8 -4.9 8.7 8.5 12.2 0.3 3.3

FISIM

GDP at bp 10.3 -1.1 5.3 0.9 -3.8 2.0 9.4 2.8 3.3

*=Preliminary estimate.

Source: Own calculations, data from Statistics Finland.

23

4. … and Productivity

4.1 Labour productivity Finnish economic growth switched into a faster gear as the earlier average growth rate of GDP at ba-sic prices of 3.1 p.p.a. in 1975-90 changed to 3.6 p.p.a. in 1995-2003 (table 14). This more rapid growth in aggregate value added was not matched by a corresponding step-up in the ratio of the vol-ume of value added to hours worked, i.e. labour productivity (LP) growth. On the contrary, aggregate LP increased by 3.0-3.1 p.p.a. 1975-95 and by 2.3 p.p.a. after the recession. Three industries managed to keep the growth rate of value added higher than the mean throughout the observation period: total manufacturing, transport & communication and real estate & business activities. In 1995-2003 they were joined by the trade industries that all demonstrated robust growth. The sale of motor vehicles even took the lead with 5.1 p.p.a. The 1990s were pro-cyclical for the trade industries as their value added plummeted with the national average during the recession and picked-up with the post-1995 boom. The cyclicality has been more pronounced in trade than at the level of the aggregate economy, which is understandable since trade is dependent both on the production and import of tradeable goods and the demand for these goods. Also vis-à-vis hours worked the development in the trade in-dustries was pro-cyclical in the 1990s. During the recession labour input in trade fell by 5.5 p.p.a. (-3.7 p.p.a. in the whole economy) and increased in 1995-2003 by 1.9 p.p.a. (1.3 p.p.a. in the whole economy). After 1995 retail trade increased its labour input only by 1.2 p.p.a., which resulted in the strongest LP growth of all trade industries (2.6 p.p.a.). Alongside total manufacturing only primary production kept its LP change above mean from 1975 to 2003. Manufacturing achieved this mainly by strong value added growth and primary production by big cuts in labour input. LP in transport & communication grew strongly in the 1990s and in the financial sector in 1975-90 and 1995-2003. The trade industry’s LP grew at the same speed as the national average in 1975-1990 (3.1 p.p.a.), with wholesale trade standing out to its advantage with 4.1 p.p.a. After the recession LP in trade grew at the rate of 2.5 p.p.a. which is slower than prior to the recession but faster than the national mean. Trade’s strong post-recession LP growth is thanks to retail trade’s 2.6 p.p.a. and the sale of motor vehicles’ 2.4 p.p.a. Wholesaleing showed an increase in LP growth of 2.1 p.p.a.

In table 14 we observed the growth of LP. To see how the levels of labour productivity have evolved we divide value added by hours worked. We use value added at current prices and normalize the total economy to equal 100, which means that the overall improvement in LP from 1975 to 2003 is not taken into account. We look at yearly cross-sections to ascertain the industries relative positions with respect to the national average. In 1975 eight industries equaled or surpassed the nation’s average LP level: total manufacturing, construction, wholesale trade, transport & communication, financial inter-mediation & insurance, real estate & business activities, education and other community, social and personal services (table 15). By 1990 only one of these eight, construction, slipped below average. But also the LP level of real estate & business activities, education and other community, social and personal services worsened relative to the national mean. Of the eight original high productivity in-dustries only five still maintained their stances as high productivity industries in 2003: total manufac-turing, wholesale trade, transport & communication, financial intermediation & insurance and real estate & business activities. Of these five the first four industries had in 2003 a level of LP that was higher than it was in 1975 relative to the total economy. Of these four industries total manufacturing and transport & communication both have a very dynamic ICT-producing sub-industry (radio, televi-

24

sion & communication equipment and telecommunications services) that performed double-digit LP growth in 1995-2002 (Statistics Finland, 2004c). Wholesale trade and financial intermediation & in-surance on the other hand are two major ICT-using industries (see van Ark, 2001; Jalava, 2003).

To find out what the effect of labour shifting to industries with either a higher level of or higher growth rate of LP on LP growth is, we performed a shift-share analysis (see van Ark, 2001). The rela-tive change in labour productivity can be expressed as:

( ) ( ) ( )(, , 1 , 1 , , 1 , 1 , , 1 , , 11 1 1 1

1 1

n n n

i t i t i t i t i t i t i t i t i t i tt t i i i

t t

LP LP S S S LP S S LP LPLP LPLP LP

− − − − − −− = = =

− −

− + − + − −−

=∑ ∑ ∑ )

(4)

where LP is the level of labour productivity, Si is industry i's share of all hours worked and t is time. The first term on the right side of the equation is the industries' internal (within) productivity effect, i.e., sub-industries impact on aggregate productivity change. The second term on the right is the static shift effect of labour, that is, the contribution of a shift of labour to industries with a higher level of LP. The third term on the right captures the dynamic shift effect of labour, i.e., the contribution of labour shifting to industries with a higher than average LP growth rate. Thus the second and third terms quantify the above mentioned Baumol’s effect.

The results of the shift-share analysis can be seen in table 16. The within effect was the most signifi-cant factor explaining labour productivity growth in the Finnish economy. The contribution of the within effect increased in the 1990s compared with 1975-90. One quarter of LP growth was to be attributed to labour shifting to industries with a higher productivity level prior to the recession but only one fifth in the early 1990s and 13 per cent after 1995. The contribution of the dynamic shift was negative the whole time, but increased from –0.9 via –1.6 to –2.6 per cent.

Table 17 reinforces the view of productivity growth being more concentrated after 1990 than before the recession. When summing the three most influential industries (total manufacturing, real estate & business activities and transport & communications) combined shares in within, dynamic and static components in 1995-2003 the sum is 71.4 per cent. In contrast the three most influential industries (total manufacturing, real estate & business activities and health & social work) in 1975-90 only amounted to 54.1 per cent. Were we to add the fourth most important industry to LP growth in both periods, trade, the sums would change to 84.3 per cent in the post-1995 period and 64.6 per cent prior to 1990. So LP growth became very concentrated indeed. Trade’s share of the aggregate within com-ponent was 14.9 per cent in 1975-90 and 11.8 per cent in 1995-2003. Industries that significantly in-creased their share of the within component from pre-1990 to post-1995 were: total manufacturing (from 45.4 to 52.8 per cent), transport & communication (from 9.1 to 18.7 per cent) and financial intermediation & insurance (from 6.6 to 12.6 per cent).

25

Table 14 Average growth of value added, hours worked and labour productivity (output per hours worked) by industry, ln-%

Value added Hours worked Labour productivity

1975-1990 1990-1995 1995-2003* 1975-1990 1990-1995 1995-2003* 1975-1990 1990-1995 1995-2003*

A Agriculture, forestry & hunt. 0.4 -0.6 2.0 -3.5 -5.0 -4.0 3.9 4.4 5.9

B Fishing 1.8 6.5 -1.4 -1.4 -2.7 -4.6 3.2 9.1 3.3

CDE Total manufacturing 3.7 2.2 4.6 -1.3 -4.1 0.5 5.0 6.3 4.2

F Construction 1.5 -9.9 2.9 -0.1 -11.1 3.2 1.6 1.1 -0.2

G Trade 3.0 -4.4 4.4 -0.2 -5.5 1.9 3.1 1.1 2.5

Sale of motor veh.; serv. stat. 2.6 -2.9 5.1 1.2 -6.5 2.8 1.3 3.5 2.4

Wholesale trade 3.5 -5.9 4.5 -0.6 -3.5 2.4 4.1 -2.5 2.1

Retail trade 2.3 -2.2 3.8 -0.3 -6.4 1.2 2.6 4.3 2.6

H Hotels & restaurants 2.7 -2.1 2.8 1.1 -5.9 3.4 1.6 3.9 -0.6

I Transport, storage & comm. 3.4 1.6 4.9 0.7 -2.5 1.0 2.7 4.1 3.9

J Financial interm. & insurance 5.7 -7.0 4.2 1.8 -6.4 -2.9 3.9 -0.5 7.1

K Real estate & business act. 4.2 2.1 4.0 4.3 0.0 4.8 -0.1 2.0 -0.9

L Admin., comp. soc. serv. 2.4 -0.8 1.1 1.3 -0.7 1.1 1.2 -0.1 0.0

M Education 2.1 0.2 1.8 2.3 1.3 0.8 -0.2 -1.1 1.0

N Health & social work 4.0 -1.6 2.0 3.2 -1.3 2.4 0.8 -0.3 -0.4

O Other comm., soc. & pers. s. 4.0 -1.0 2.5 3.0 -1.0 2.9 1.0 0.0 -0.4

P Household service activities -5.3 9.4 10.1 -5.1 8.8 8.6 -0.2 0.6 1.5

FISIM

GDP at bp 3.1 -0.7 3.6 0.0 -3.7 1.3 3.1 3.0 2.3



26

Table 15 Levels of labour productivity (current price value added per hours worked), total economy=100

1975 1980 1985 1990 1995 2000 2003*

A Agriculture, forestry & hunt. 52 54 49 56 42 41 48

B Fishing 32 28 26 22 22 20 19

CDE Total manufacturing 123 129 129 128 147 147 140

F Construction 100 83 80 85 66 68 67

G Trade 87 89 90 86 83 80 83

Sale of motor veh.; serv. stat. 91 91 86 79 72 65 70

Wholesale trade 126 133 144 139 125 126 127

Retail trade 61 62 59 56 59 51 54

H Hotels & restaurants 60 59 63 60 56 43 46

I Transport, storage & comm. 120 129 128 123 123 138 142

J Financial interm. & insurance 159 171 156 197 190 244 210

K Real estate & business act. 322 264 243 203 208 190 188

L Admin., comp. soc. serv. 86 79 83 84 73 69 69

M Education 143 121 122 113 92 89 92

N Health & social work 83 76 82 80 73 66 66

O Oth. comm., soc. & pers. s. 118 116 112 114 99 87 85

P Household service activities 32 40 45 48 36 29 34

FISIM

GDP at bp 100 100 100 100 100 100 100



27

Table 16 The impact of structural change on labour productivity growth, %

1975-1990 1990-1995 1995-

2003*

Within 73.6 82.9 89.8

Static 27.3 18.8 12.8

Dynamic -0.9 -1.6 -2.6

Total 100.0 100.0 100.0*=Preliminary estimate.

28

Table 17 Shares of aggregate within component by industry and sum of within, static and dynamic components by industry, %

Within Sum of within, static and dynamic shares

1975-1990 1990-1995 1995-2003* 1975-1990 1990-1995 1995-2003*

A Agriculture, forestry & hunt. 10.1 7.6 10.8 1.4 4.4 1.1

B Fishing 0.2 0.4 0.2 0.1 0.4 -0.1

CDE Total manufacturing 45.4 58.1 52.8 24.7 46.3 37.2

F Construction 6.5 5.1 -0.7 4.0 -14.9 4.2

G Trade 14.9 3.3 11.8 10.5 -3.5 12.9

H Hotels & restaurants 1.0 2.3 -0.4 1.3 0.8 0.9

I Transport, storage & comm. 9.1 15.1 18.7 8.5 16.5 15.4

J Financial interm. & insurance 6.6 -1.1 12.6 7.0 -4.7 4.5

K Real estate & business act. -1.0 12.9 -6.6 17.9 31.0 18.8

L Admin., comp. soc. serv. 3.2 -0.2 0.2 4.9 6.0 -0.5

M Education -0.5 -2.4 2.5 3.8 7.2 1.1

N Health & social work 3.1 -1.3 -1.4 11.5 6.8 2.3

O Other comm., soc. & pers. s. 1.5 0.0 -0.6 4.5 3.5 1.8

P Household service activities 0.0 0.0 0.1 -0.2 0.3 0.4

FISIM

TOTAL 100.0 100.0 100.0 100.0 100.0 100.0

May not sum to totals due to rounding.


29

4.2 Unit labor cost The norm for maximum wage increases in the Finnish centralized wage bargaining system was during the period we are analyzing aggregate LP growth corrected for inflation. However, the new post-recession state-of-play with fewer industries contributing more to aggregate LP than before is prob-lematic for the old paradigm. How can low productivity industries afford the wage increases that high productivity industries can? A useful way of looking at how a change in hourly compensation is fi-nanced is by decomposing it into the change in unit labour cost (ULC) and labour productivity. Thus ULC equals hourly compensation less LP. Alternatively ULC can be calculated as the total compensa-tion9 of employment divided by real value added, which means that ULC is total labour cost per unit of output.

In the period 1975-90 the development in the trade industry was remarkably similar to that of the total economy (table 18). Total compensation grew at an average rate of 10.2 per cent both in trade and overall. The sale of motor vehicles increased its total compensation by 11.7 p.p.a., retail trade by 10.6 p.p.a. and wholesale trade by 9.4 p.p.a. Hours worked stayed level in the whole economy and dimin-ished somewhat (-0.2 p.p.a.) in trade. The hourly compensation in trade was only growing a bit faster than the national average (10.4 vs. 10.2 p.p.a.), with retail trade’s 10.9 p.p.a. and sale of motor vehi-cles’ 10.5 p.p.a. above average and wholesale trade at 10.1 p.p.a. Trade and the aggregate economy both had a LP growth of 3.1 p.p.a. (the sale of motor vehicle’s LP grew only by 1.3 p.p.a., the corre-sponding figures for retail and wholesale trade were 2.6 and 4.1 p.p.a.). Trade’s real value added in-creased by 3.0 p.p.a. and the nation’s average by 3.1 p.p.a. Of the trade industries wholesale trade demonstrated the strongest growth (3.5 p.p.a.) with the sale of motor vehicles and retailing performing much worse (2.6 and 2.3 p.p.a., respectively). Therefore ULC in trade grew by 7.2 p.p.a. in compari-son with 7.0 p.p.a. for the whole economy. In the trade industries wholesale trade showed the slowest ULC growth (6.0 p.p.a.) as retail trade’s and the sale of motor vehicles’ ULC grew at rates of 8.3 and 9.2 p.p.a., respectively. The recession hit trade harder than the whole economy as its total compensa-tion, hours worked and real value added all plummeted. The result was that ULC grew much faster in trade than the national average. In terms of unit labour cost the previous best performer among the trade industries, wholesale trade, did not manage to slow down its ULC growth during the recession. After the recession ULC change in trade experienced a change for the better. Due to a strong 4.4 p.p.a. growth in real value added in trade (3.6 p.p.a. in the whole economy) LP increased at a 2.5 p.p.a. pace with 2.3 p.p.a. as the national mean. Trade’s ULC growth slowed down to only 0.8 p.p.a. which is better than the 1.1 p.p.a. the whole economy performs. Trade’s low ULC growth stemmed from the low 0.7 p.p.a. ULC change in retail trade. Motor vehicle sales and retailing were close to the national mean. at 1.0 p.p.a., respectively. On the whole, ULC growth slowed down spectacularly both in the whole economy and trade from what it was prior to the recession.

9 Wages, salaries and employers’ social contributions for employees plus the imputed wage share for the self-employed. The imputation is performed by multiplying the hours worked by the self-employed with the em-ployees’ average earnings per hour.

30

Table 18 Average growth of unit labour cost and related variables in trade, its sub-industries and the total economy,

ln-%

G Trade: 1975-

1990

1990-

1995

1995-

2003*

50, 51, 52: 1975-1990 1990-1995 1995-2003*

Total compensation 10.2 -2.6 5.2 Total compensation 11.7, 9.4,

10.6

-4.0, -1.2, -3.6 6.2, 5.5, 4.5

Hours worked -0.2 -5.5 1.9 Hours worked 1.2, -0.6, -

0.3

-6.5, -3.5, -6.4 2.8, 2.4, 1.2

Hourly compensation 10.4 2.9 3.3 Hourly compensation 10.5, 10.1,

10.9

2.5, 2.3, 2.8 3.4, 3.1, 3.3

Real value added 3.0 -4.4 4.4 Real value added 2.6, 3.5, 2.3 -2.9, -5.9, -2.2 5.1, 4.5, 3.8

Labour productivity 3.1 1.1 2.5 Labour productivity 1.3, 4.1, 2.6 3.5, -2.5, 4.3 2.4, 2.1, 2.6

Unit labour cost 7.2 1.8 0.8 Unit labour cost 9.2, 6.0, 8.3 -1.0, 4.7,-1.5 1.0, 1.0, 0.7

Total economy:

Total compensation 10.2 -0.6 4.7

Hours worked 0.0 -3.7 1.3

Hourly compensation 10.2 3.1 3.3

Real value added 3.1 -0.7 3.6

Labour productivity 3.1 3.0 2.3

Unit labour cost 7.0 0.0 1.1



4.3 Multi-factor productivity The natural next step in our venture to account for the growth and productivity in trade and in the rest of the economy is to look at which proximate factors determine LP growth. In a neoclassical growth accounting framework labour productivity growth can be decomposed into the contributions of capital deepening and multi-factor productivity (MFP):

(5) , AHKvHY Kˆ)ˆˆ(ˆˆ +−=−

where the ^-symbol indicates the rate of change, Y is real value added, H is hours worked, v is the nominal income share, K is capital stock and A is MFP. It shows that there are two sources of labour productivity growth. The first one is capital deepening, i.e. an increase in capital per hour worked weighted by capital’s income share. Increases in capital deepening are usually sustained by a high investment ratio that ensures that the stock of fixed capital grows faster than the labour input. Alterna-tively capital deepening can increase by decreases in the labour input realitive to capital. The second

31

source is a general advance in multi-factor productivity. MFP growth is calculated as the geometric average of LP and capital productivity (CP) change. The weights used are the arithmetic averages of capital’s and labour’s respective income shares in period t and t-1. Labour’s share has been con-strained to a maximum of 1. The rental price of capital is calculated as the ratio of capital income to the real capital stock. As a measure of capital we use the gross capital stock (GCS) which is not the ideal measure but is in keeping with our effort to stay as closely as possible to official national ac-counts data.10 Aulin-Ahmavaara and Jalava (2003) showed for the total Finnish nonresidential econ-omy that MFP measures calculated using the GCS did not differ much from the other capital measures. The exception was the early 1990s recession when GCS based MFP measures were lower than the alternative ones.

From 1975 to 1990 the contributions from capital deepening and MFP – as well as the income shares of labour and capital – were virtually identical in the trade industry and the whole economy (tables 19 and 20, figures 2 and 3). During the 1990s recession the contribution from MFP vanished in trade thus shrinking the LP change to a third from before. As can be seen in figure 3 trade’s low MFP growth stemmed from a massive decrease in CP change. After 1995 MFP growth in trade rebounded to 3.5 p.p.a., thanks to a major step-up in CP growth, which offset the negative contribution from capital deepening. In the total economy capital deepening’s contribution shrank to zero but MFP change stayed near its pre-recession track record even after 1995. The overall productivity change is less cy-clical in the total economy than in trade and its sub-industries (figures 4, 5 and 6). Primary production and total manufacturing show above average MFP growth throughout the 1975 to 2003 period. The low contribution from capital deepening post-1995 stems from lower investment ratios than before. The investment ratio (current price gross fixed capital formation per current price gross value added) in trade declined from an arithmetic average of 16.1 per cent in 1975-90 to 10.2 in 1995-2003. In the same periods the aggregate investment ratio declined from 30.9 per cent to 21.7 per cent.11

10 See Statistics Finland (2004a) for definitions on how Gross and Net Capital Stocks are calculated in Finnish national accounts. 11 Here the investment ratio for the total economy is gfcf per value added at basic prices. If the more commonly used value added at market prices (which also takes into account the taxes on production and imports as well as subsidies) is used the investment ratios are approximately 3 percentage points lower.

32

Table 19 Decomposition of compound annual average labour productivity growth by industry, ln-%

Labour productivity Capital deepening MFP

1975-1990 1990-1995 1995-2003* 1975-1990 1990-1995 1995-2003* 1975-1990 1990-1995 1995-2003*

A Agriculture, forestry & hunt. 3.9 4.4 5.9 0.5 0.6 0.9 3.4 3.8 5.1

B Fishing 3.2 9.1 3.3 0.0 0.0 0.0 3.2 9.1 3.3

CDE Total manufacturing 5.0 6.3 4.2 1.6 1.5 -0.1 3.4 4.8 4.2

F Construction 1.6 1.1 -0.2 0.2 0.4 -0.2 1.4 0.7 -0.1

G Trade 3.1 1.1 2.5 0.7 1.2 -0.8 2.4 -0.1 3.5

Sale of motor veh.; serv. stat. 1.3 3.5 2.4 0.5 0.9 -1.0 0.9 2.6 3.4

Wholesale trade 4.1 -2.5 2.1 1.1 1.1 -1.3 3.0 -3.6 3.3

Retail trade 2.6 4.3 2.6 0.6 0.8 -0.2 2.0 3.5 2.9

H Hotels & restaurants 1.6 3.9 -0.6 0.2 0.3 -0.1 1.3 3.6 -0.5

I Transport, storage & comm. 2.7 4.1 3.9 0.7 1.6 0.3 2.0 2.5 3.6

J Financial interm. & insurance 3.9 -0.5 7.1 0.9 2.0 -0.2 3.0 -2.6 7.3

K Real estate & business act. -0.1 2.0 -0.9 0.1 1.3 -1.6 -0.2 0.8 0.7

L Admin., comp. soc. serv. 1.2 -0.1 0.0 0.3 0.3 0.1 0.9 -0.4 -0.1

M Education -0.2 -1.1 1.0 0.1 0.1 0.1 -0.3 -1.2 0.9

N Health & social work 0.8 -0.3 -0.4 0.0 0.2 0.0 0.8 -0.5 -0.4

O Other comm., soc. & pers. s. 1.0 0.0 -0.4 0.4 0.9 -0.3 0.6 -0.9 -0.1

P Household service activities

FISIM

GDP at bp 3.1 3.0 2.3 0.8 1.2 0.1 2.3 1.9 2.2

*=Preliminary estimate. Source: Own calculations, data from Statistics Finland [Table 19]

33

Table 20 Arithmetic average of income shares, investment ratios and profitability in trade and total economy, %

G Trade: 1975-1990 1990-1995 1995-2003*

Labour’s income share 76.0 77.2 67.7

Capital’s income share 24.0 22.8 32.3

Investment ratio 16.1 14.8 10.2

Profitability 16.7 13.5 28.8

Total economy:

Labour’s income share 76.3 74.3 66.5

Capital’s income share 23.7 25.7 33.5

Investment ratio 30.9 23.5 21.7

Profitability 6.5 6.7 10.3

*=Preliminary estimate. Source: Own calculations, data from Statistics Finland

34

Figure 2 Productivity change in total economy, 1975-2003*, 1975=LN(100)

4,04,24,44,64,85,05,25,45,65,86,0

1975

1977

1979

1981

1983

1985

1987

1989

1991

1993

1995

1997

1999

2001

LP CP MFP



35

Figure 3 Productivity change in trade, 1975-2003*, 1975=LN(100) Figure 4 Productivity change in sale of mot. veh. etc., 1975-2003*, 1975=LN(100)

4,04,24,44,64,85,05,25,45,65,86,0

1975

1977

1979

1981

1983

1985

1987

1989

1991

1993

1995

1997

1999

2001

LP CP MFP

4,04,24,44,64,85,05,25,45,65,86,0

1975

1977

1979

1981

1983

1985

1987

1989

1991

1993

1995

1997

1999

2001

LP CP MFP

*=Preliminary estimate. *=Preliminary estimate. Source: Own calculations, data from Statistics Finland. Source: Own calculations, data from Statistics Finland.

Figure 5 Productivity change in wholesale trade, 1975-2003*, 1975=LN(100) Figure 6 Productivity change in retail trade, 1975-2003*, 1975=LN(100)

4,04,24,44,64,85,05,25,45,65,86,0

1975

1977

1979

1981

1983

1985

1987

1989

1991

1993

1995

1997

1999

2001

LP CP MFP

4,04,24,44,64,85,05,25,45,65,86,0

1975

1977

1979

1981

1983

1985

1987

1989

1991

1993

1995

1997

1999

2001

LP CP MFP

*=Preliminary estimate. *=Preliminary estimate. Source: Own calculations, data from Statistics Finland. Source: Own calculations, data from Statistics Finland.

36

The functional income distribution shifted by close to 10 percentage points in capital’s favour in both trade and the total economy from 1975-90 to 1995-2003. Sauramo (2004) calculated so called equilib-rium labour income shares and concluded that the functional income distribution was not in equilib-rium at the turn of the millennium. As an explanation for this he offered the exceptionally high profitability in Finnish firms in the 1990s. When looking at the non-financial corporations’ financial status in the national accounts’ institutional sector accounts it is easy to concur with Sauramo’s con-clusion. Throughout the 1975 to 1992 period (with the sole exception of the year 1978) non-financial corporations were forced to borrow funds as their net lending was negative. After 1993 the net lending of non-financial corporations was positive. One way of calculating profitability in the macroeconomic framework is to divide the capital income, which equals nominal value added less total labour com-pensation, with the value of the net capital stock. We found that profitability indeed rose by 3.8 per-centage points in the whole economy (from 6.5 per cent in 1975-90 to 10.3 per cent in 1995-2003). Profitability in trade has also increased by double-digit figures thanks to favorable developments in wholesale trade (from 16.7 per cent in 1975-90 to 28.8 per cent in 1995-2003).

5. Conclusions

In this paper we took a quantitative look at the growth and productivity of trade and the rest of the economy in 1975-2003. During our observation period the share of primary and secondary production in the total economy dwindled from half to a third. Overall economic growth shifted into a faster gear after 1995 as it was 3.6 p.p.a. compared with the pre-recession 3.1 p.p.a. The trade sector’s value added joined post-1995 the growth club of total manufacturing, transport & communication and real estate & business services that all performed better than the national average. Average earnings dif-ferentials decreased from 1975 to 1990 but increased again from 1990 to 2003. In trade the average earnings grew more slowly in the 1995 to 2003 period than the economy’s mean. Increases in hourly compensation were post-1995 financed by lower unit labour costs than previously. The trade sector’s ULC growth was – thanks to retailing – less than what it was in the rest of the economy.

Overall LP growth slowed from 3.1 p.p.a. in 1975-90 to 2.3 p.p.a. in 1995-2003. In the latter period primary production, total manufacturing, trade, transport & communication and the financial sector were strong performers. Primary production achieved LP growth mainly by strongly cutting back on the labour input. Among the rest of the rapid growers were found either dynamic ICT-producing sub-industries or use of ICT. High LP levels were achieved by fewer industries in 2003 than previously, and wholesale trade was one of the few. When decomposing LP change into the contributions of capi-tal deepening and multi-factor productivity, we found the post-recession slow aggregate LP growth to stem from a nonexistent contribution from capital deepening. The trade industry, along with the strong performers in LP, experienced a significant boost in MFP change that clearly surpassed the national mean.

So what inferences if any can we draw on Baumol’s disease for Finland? At first sight the evidence looks inconclusive. When comparing the LP growth in 1975-90 vs. 1995-2003 slightly more service industries experienced a slowing down of LP growth than a speeding up. For MFP it was vice versa: more service industries experience a speeding up than a slowing down. Were we to leave the public sector out of the analysis (as in Statistics Finland, 2004b) the verdict concerning LP growth would still be the same. However, for MFP growth the picture changes dramatically: all service industries shift into a faster gear from the 1975-90 period to the 1995-2002 period. Furthermore, the results of our shiftshare analysis clearly indicate that in the post-1995 era LP growth was much more a within industry story than before. Based on the shiftshare analysis and the evidence on MFP growth in Fin-nish service industries our verdict on Baumol’s disease then becomes a definitive not guilty.

Be that as it may, for the trade industries the post-recession period looked good. Trade joined the growth clubs in value added, LP and MFP. ULC growth was moderate and profits were high. How-ever, vis-à-vis the level of LP only wholesale trade was above the national average whereas the sale of motor vehicles and retailing slipped further behind the high productivity industries. Times were com-petitive as strong productivity growth was displayed by fewer industries post-1995 than in earlier periods.

38

References

van Ark, B. (2001): ‘The Renewal of the Old Economy: An International Comparative Perspective’, STI Working Papers 2001/5, Paris:OECD.

van Ark, B. and Timmer, M. (2001): ‘PPPs and International Productivity Comparisons: Bottlenecks and New Directions’, paper presented at: Joint World Bank – OECD Seminar on Purchasing Power Parities: Recent Advances in Methods and Applications, 30.1.–2.2.2001, Washington, D.C.

Aulin-Ahmavaara, P. and Jalava, J. (2003): ‘Capital and its Productivity in Finland, 1975-2001’, Scandinavian Economic History Review, Vol. 50 (3), pp. 62-79.

Baily, M. N. and Lawrence, R. Z. (2001): ‘Do We Have a New E-conomy?’, American Economic Review, Vol. 91 (2), pp. 308-312.

Baily, M. N. and Zitzewitz, E. (2001): ‘Service Sector Productivity Comparisons: Lessons for Measurement’, in C.R. Hulten, E.R. Dean and M.J. Harper, New Developments in Productivity Analysis, NBER Studies in Income and Wealth, Volume 63, Chicago/London: University of Chicago Press.

Baumol, W. J. (1967): ‘Macroeconomics of Unbalanced Growth: The Anatomy of Urban Crisis’, American Economic Review, Vol. 57 (3), pp. 415-426.

Broadberry, S.N. (1998): ‘How did the United States and Germany Overtake Britain? A Sectoral Analysis of Comparative Productivity Levels, 1870-1990’, Journal of Economic History, Vol. 58 (2), pp. 375-407.

Commission Regulation (EC) No. 1889/2002.

Feinstein, C. H. and Thomas, M. (2002): Making History Count: A Primer in Quantitative Methods for Historians, Cambridge, UK: Cambridge University Press.

Forssell, O. (1979): Finland’s Domestic Trade, 1860-1960 (in Finnish with English Summary), Studies on Finland’s Economic Growth X, Helsinki: Bank of Finland Publications.

Gordon, R. (2002): ‘Recent Productivity Puzzles in the Context of Zvi Griliches' Research’, mimeo, Northwestern University/NBER.

Griliches, Z. (1992): ‘Introduction’, in Z. Griliches (ed.), Output Measurement in the Service Sectors, NBER Studies in Income and Wealth, Volume 56, Chicago/London: University of Chicago Press.

Hill, T.P. (1977): ‘On Goods and Services’, Review of Income and Wealth, Vol. 123, pp. 315-338.

Hill, T.P. (1999): ‘Tangibles, Intangibles and Services: A New Taxonomy for the Classification of Output’, Canadian Journal of Economics, Vol. 32 (2), pp. 426-446.

Hjerppe, R. (1988): Suomen talous 1860-1985, kasvu ja rakennemuutos, Studies on Finland’s Economic Growth XIII, Helsinki: Bank of Finland Publications.

Jalava, J. (2003): ‘Den nya ekonomin i Finland: produktion och användning av IKT’, Ekonomiska

39

Samfundets Tidskrift, Vol. 56 (1), pp. 17-24.

Jalava, J. and Pohjola, M. (2002): ‘Economic Growth in the New Economy: Evidence from Advanced Economies’, Information Economics and Policy, Vol. 14 (2), pp. 189-210.

Maliranta, M. (2003): Micro Level Dynamics of Productivity Growth, An Empirical Analysis of the Great Leap in the Finnish Manufacturing Productivity in 1975-2000, The Research Institute of the Finnish Economy (ETLA), A 38, Helsinki: Taloustieto Oy.

Mankinen, R., Rouvinen, P. and Ylä-Anttila, P. (2002): ‘Palveluiden tuottavuus – kilpailu ja teknologia muuttavat rakenteita’, The Research Institute of the Finnish Economy (ETLA), Discussion papers No. 829, October.

McKinsey Global Institute (1992): Service Sector Productivity, Washington, D.C: McKinsey & Co.

OECD (2003): National Accounts of OECD Countries, Main Aggregates, Volume I, 1990-2001, Paris: OECD.

Oulton, N. (2001): ‘Must the Unbalanced Growth Rate Decline? Baumol's Unbalanced Growth Rate Revisited’, Oxford Economic Papers, Vol. 53 (4), pp. 605-627.

Rouvinen, P. and Ylä-Anttila, P. (2004): ‘Palveluyrittäjyys kasvaa nopeimmin ja työllistää eniten’, The Research Institute of the Finnish Economy (ETLA), Discussion papers No. 913, May.

Sauramo, P. (2004): ‘Funktionaalinen tulonjako Suomessa: ollaanko tasapainossa?’, Kansantaloudellinen aikakauskirja, Vol. 100 (1), pp. 33-49.

Statistics Finland (1983): Panos-tuotostutkimus 1980, alustavia tietoja, KT 1983:9, Helsinki: Government Printing Centre.

Statistics Finland (1985): Panos-tuotostutkimus 1982, alustavia tietoja, KT 1985:4, Helsinki: Government Printing Centre.

Statistics Finland (1988): Panos-tuotos 1985, Kansantalous 1988:1, Helsinki: Government Printing Centre.

Statistics Finland (1992): Panos-tuotos 1989, taulukot, SVT Kansantalous 1992:17, Helsinki.

Statistics Finland (1996): Panos-tuotos 1992–1993, SVT Kansantalous 1996:19, Helsinki: Edita Oy.

Statistics Finland (2003): Talouden rakenne 1995–2000, tarjonta- ja käyttötaulukot, panos-tuotos, SVT Kansantalous 2003:4, Helsinki: Yliopistopaino Oy.

Statistics Finland (2004a): Kiinteä pääomakanta 1975-2002, Aikasarjat, SVT Kansantalous 2004:1, Helsinki: Multiprint Oy.

Statistics Finland (2004b): Tuottavuuskatsaus 2003, Katsauksia 2004/3, Helsinki: Multiprint Oy.

SNA93 (1993): System of National Accounts 1993, UN, OECD, EU, IMF, World Bank, Series F, No. 2, Rev. 4, New York: United Nations.

40

Appendix

The Leontief inverse matrix is calculated in practice as shown in this fictitious example. The identity matrix I is:

(A.1)

1001

Assuming that the input coefficient matrix IC is:

(A.2) ,

55.007.015.04.0

then the (I-IC) matrix is:

(A.3) .

−

−45.007.015.06.0

Its inverse L=(I-IC)-1 is:

(A.4) .

312.2270.0578.0734.1

As we know that the product of a matrix and its inverse is the identity matrix we can validate the re-sults:

(A.5) L11 = 0.6*1.734 + (-0.15)*0.270=1

(A.6) L12 = 0.6*0.578 + (-0.15)*2.312=0

(A.7) L21 = (-0.07)*1.734 + 0.45*0.270=0

(A.8) L22 = (-0.07)*0.578 + 0.45*2.312=1.

41

Papers issued in the series of the EU KLEMS project

All papers are available in pdf-format on the internet: http://www.euklems.net/

Nr.1 Mas, Matilde and Javier Quesada, ICT and Economic Growth in Spain 1985-2002 (2005)

Nr.2 Jalava, Jukka, Growth and Productivity in the Finnish Trade Industry, 1975-2003: A National Comparative Perspective

42

growth and productivity in the finnish trade industry ...euklems.net/pub/no2.pdf · growth and...

Documents