General Considerations

A useful indicator of competitiveness, i.e., the real exchange rate, would have the property that movements in the indicator are associated with changes in a country balance of trade in goods and nonfactor services. This arises from the fact that in its simplest representation, the trade balance is a function of aggregate income at home Y, aggregate income in the foreign country Y*, and relative prices. In functional form:

and

where X denotes aggregate exports, M denotes aggregate imports, E is the nominal exchange rate (foreign currency units of the domestic currency), P is the price of domestically produced goods and P* is the price of foreign goods. In practice, there are many real exchange rates corresponding to different measures of relative prices. What equations (1) and (2) suggest is that it is useful to find an indicator whose movements are important for explaining (X - M), a country’s trade balance.

Frequently, the prices of exports, imports, or both are used in the construction of real exchange rate indices, but it is unclear exactly what types of prices should be used in such indices. For example, an increasingly large proportion of the volume of international trade takes place in intermediate inputs, raw materials, and capital goods, rather than final goods. If the objective is to explain the balance of trade in goods, then it seems appropriate to obtain information on the prices of these intermediate goods and incorporate them in a real exchange rate index. A real exchange rate based on consumer price indices contains information on the prices of final consumer goods, but does not contain information on the prices of traded intermediate goods directly, so this type of indicator may not be appropriate for explaining an important component of trade flows.

All of the real exchange rate indices discussed in this paper, except for the relative price of traded to nontraded goods, are sensitive to the weighting scheme adopted involving trading partners. Real exchange rate indicators usually involve the comparison of a domestic index with the same index in the partner country, but this “partner” is in reality a weighted average of the index in all competing countries. ^6/ Thus, any computed real exchange rate index will depend on the adopted weighing scheme. ^7/

An important consideration concerning the usefulness of alternative indicators is the amount of information contained in the indicator and its relevance for explaining the balance of trade in goods and nonfactor services. As the objective of using an indicator is to provide information for an assessment of movements in a country’s trade balance, indicators which include information on both exports and imports are preferable to those which contain information on only one component of the trade balance. For example, export unit values of manufacturing products contain information for an assessment of exports, but other indicators such as normalized unit labor costs in the manufacturing sector contain information on both exports and imports to the extent that both of these fall into the category of manufacturing products. In this regard, a general real exchange rate measure would be the relative price of traded goods, defined as:

where P_T denotes the price of the home country’s traded goods and $P_{T^{*}}$ denotes the prices of traded goods in the foreign country. The real exchange rate defined in (3) encompasses information on both exports and imports, so it possesses an advantage over a real exchange rate index based on export unit values. ^8/ Also, the definition in (3) may be particularly useful for the case of differentiated products, which is discussed below.

Indices based on manufacturing products alone cannot fully explain trade flows because they exclude information on a wide class of traded goods, such as agricultural commodities. The two indicators which are the “most comprehensive” are real exchange rates based on consumer price indices and the relative price of traded to nontraded goods, as these encompass all sectors of the economy.

Basic Structure

To provide a unifying framework, it is useful to start with a basic structure that describes the production side of the economy. The approach followed here is similar to the approach used in Marston (1987) and Lipshitz and McDonald (1991).

To begin, the value-added functions for both traded and nontraded goods are given by:

where V_j denotes value added in sector j, L_j is employment in sector j, K_j is the amount of capital used in sector j, and t represents technical progress over time. Gross output of final goods, Z_j, is a function of value added V_j and intermediate inputs I_ij:

where I_NT is the amount of nontraded goods used as an intermediate input in the production of traded goods, and I_TN is the amount of traded goods used as an intermediate input in the production of nontraded goods. The corresponding equations which define the prices of each final good are:

where:

PV_j = value-added price of the jth good;

c_VJ = the share of value added in the output of the jth final good;

c_NT = the share of nontraded intermediate inputs in the output of traded goods;

c_T = the share of traded intermediate goods in the output of traded goods;

c_N = the share of nontraded intermediate goods in the output of nontraded goods; and

c_TN = the share of traded goods used as an intermediate input in the production of nontraded goods.

Value-added prices are related to wages by:

where MPL_j is the marginal product of labor employed in the jth sector and W_j is the wage in the jth sector.

Real exchange rate indicators using consumer prices

One frequently used indicator of competitiveness is a real exchange rate index computed using consumer price indices which takes the form:

where CPI_H is the consumer price index in the home country, and CPI_F is the consumer price index in the home country’s trading partner. By construction, this index is comprised of the prices at the consumer level, rather than prices at the producer or wholesale level. A rise in the consumer price index of the home country relative to the consumer price index in the foreign country, i.e., a real appreciation, would be associated with a loss in competitiveness of the home country and therefore a deterioration in the trade balance of the home country.

As the consumer price index is comprised of both traded and nontraded goods, each consumer price index can be written:

where ${\alpha }_T+{\alpha }_N=1,and{\alpha }_T^{*}+{\alpha }_N^{*}=1.$. In the extreme case where all traded goods are homogeneous, then EP_T will equal $P_T^{*}$. If the weights of each type of good in the domestic and foreign CPIs are equal (i.e., $E{P}_T=P_T^{*},{\alpha }_T={\alpha }_T^{*},and{\alpha }_N={\alpha }_N^{*}$), then:

Under these restrictive assumptions, a real exchange rate index based on the ratio of consumer price indices would be a function only of the price of nontraded goods in the home country relative to the price of nontraded goods in the foreign country. In practice, however, traded goods are generally not homogeneous, therefore $E{P}_T\ne P_T^{*}$ and the ratio of consumer price indices becomes a function of the prices of traded and nontraded goods.

As noted by Turner and Van’t dack (1993) and Wickham (1993), a fundamental problem inherent in using CPIs is that the consumer price index is subject to the influence of price controls and other distortions which may introduce “noise” into the performance of the indicator. If the objective is to measure the relative price of traded goods, as shown in equation (3), then Turner and Van’t dack (1993) argue that the consumer price index may be a poor proxy for the price of traded goods. The reason for this is that the consumer price index includes the prices of services, many of which are nontraded. Furthermore, a large portion of trade is in intermediate goods and therefore much of international trade does not take place at consumer prices. These two facts may diminish the usefulness of relying on a real exchange rate based on consumer prices to explain the balance of trade in goods.

Despite these disadvantages, consumer price indices possess some advantages. As Turner and Van’t dack (1993) note, consumer price indices are calculated based on a basket of goods that is fairly comparable across countries. Consumer price indices are constructed with a high degree of accuracy, are readily available, and are published frequently. As it is important for any indicator of competitiveness to provide information on the profitability of producing traded goods, Enoch (1978) and Turner and Van’t dack (1993) suggest that consumer price indices may accurately reflect factor costs. They argue that many productive inputs, such as labor, are priced in line with consumer prices, so consumer price indices may provide a useful indicator of the costs of production. ^9/

Real exchange rates based on export unit values

A second indicator often used to assess price competitiveness is a real exchange rate index based on export unit values of manufacturing products, as these products account for a large proportion of trade in goods. This index is computed by comparing the home country’s export unit values (UV_X) with the export unit values of competitors in a given market $U{V}_{X^{*}}$, expressing both m the same currency:

It should be noted that export unit values themselves are not the actual prices at which transactions take place, but rather serve as proxies for the prices of exports. ^10/ If export unit values are good proxies for the actual prices of exports, then a comparison of export unit values provides useful information relevant for an assessment of a country’s export performance. If the unit values of the home country’s exports fall relative to the unit values of a competitor’s exports, then exports from the home country would be expected to increase, leading to an improvement in the trade balance.

In the extreme case where all traded goods are homogeneous, the prices of traded goods must be equalized, so the ratio of export unit values equals one if $U{V}_X=U{V}_X^{*},$ and from (15) RER_XUVM = 1.0. Therefore, in this case– for example, a good like wheat– the ratio of export unit values has no Implications for competitiveness. However, as this assumption is far too restrictive, models of international trade have increasingly adopted a structure where traded goods are considered to be imperfect substitutes based on country, of origin, as described by Armington (1969). When traded goods are heterogeneous across suppliers, their prices are no longer identical. In this case, a comparison of the export unit values of competitor products in a given import market is indeed meaningful. In the framework of an Armington-type model, consumers in a given importing region are assumed to engage in a three-stage procedure to determine the allocation of expenditure among nontraded goods, import competing goods, and imports from various sources.

In the first stage, consumers determine the mix of expenditure between an aggregate of all traded goods and home (nontraded) goods based on relative prices:

where C is an aggregate of traded and nontraded goods, T is consumption of aggregate traded goods, and N is consumption of nontradables. If the consumer minimizes the expenditure of consuming both goods subject to a given level of C, then the mix of expenditure between T and N will depend on their relative prices and the elasticity of substitution between nontraded goods and traded goods, σ^NT:

In the second stage, consumers determine the allocation of total expenditure on tradables from stage one between aggregate imports and the domestic import competing good. The tradables aggregate (T) is a function of aggregate imports and the domestic, import-competing good:

Expenditure between aggregate imports and the import competing good is a function of relative prices and the elasticity of substitution between aggregate imports and import competing goods σ_M,MC:

At the final stage, the allocation of expenditure on aggregate imports among sources is also a function of their relative prices and the elasticity of substitution between imports from source A and source B σ_MA,MB:

where PM_A is the price of imports from source A and PM_B is the price of imports from source B. Thus, at stage one, consumers decide on the allocation of expenditure between traded and nontraded goods based on the relative price of nontraded to traded goods, and in stage two, allocate their expenditure on a composite of all traded goods based on the price of aggregate imports relative to the price of the domestic import-competing good. At the final stage, aggregate expenditure on imports is determined based on the prices of imports by source country. This type of structure allows the prices of imports to differ by country of origin, and thus, captures the phenomenon of two-way trade because goods are heterogeneous.

In the context of the imperfect substitutes model, a comparison of the prices of traded goods from different sources captures the incentives which guide consumers in stage three of the decision process, rather than the relevant incentives in stage one. As noted by McGuirk (1986), comparing the unit values of exports in a given import market implicitly adopts a view of competitiveness based on competition among exporters in a given import market, and is not a view based on the relative incentives to produce tradable and nontradable goods. ^11/

There are a number of difficulties inherent in a real exchange rate index based on export unit values. The usefulness of export unit values as an indicator of competitiveness depends on how closely movements in the indicator are associated with changes in the trade balance. Although a comparison of export unit values contains information that is directly relevant for assessing changes in exports, i.e., export prices, the comparison does not contain information relevant for assessing the performance of imports, so in this sense an indicator based on export unit values may not incorporate fully the information on competitiveness needed to explain movements in the trade balance. Moreover, as noted above, export unit values are not true prices; they are average values. As such, a comparison of export unit values is sensitive to the composition of exports across countries. It is possible that fluctuations in export unit values are a consequence of changes in the composition of exports across countries, which is unrelated to competitiveness.

To the extent that the competitive situation in the export market involves “pricing to market,” a comparison of export unit values may not reflect underlying competitiveness, i.e., the relative cost position of the country. Turner and Van’t dack (1993) make the point that international competition among sellers places a limit on the extent to which export prices may deviate from each other. This competition may force some exporters to accept prices that do not cover costs of production, which may be possible in the short run, but not in the long run. As a consequence, all potentially exportable goods will not be included in an export unit value index. This is an important shortcoming of comparing export unit values because the index will exclude some goods whose costs exceed world prices, and thus, are not exported. Comparison of export unit value indices will then not provide information on how changes in competitiveness are related to the export performance of goods that are excluded from the index. To get a comprehensive picture of a country’s “underlying competitiveness,” it would be useful to have an indicator that provided information on goods that are currently tradable as well as goods that are potentially tradable.

Real exchange rate as the relative price of traded to nontraded goods

In an important strand of literature on open economies, many economists have calculated the real exchange rate as the relative price of traded to nontraded goods and noted the importance of this relative price in terms of its relationship to a country’s external position. ^12/ The real exchange rate defined in this manner is:

where P_T is a price index of tradable goods and P_N is a price index of nontradable (home) goods. As it is difficult in practice to delineate traded goods from nontraded goods, many economists, such as Marston (1987) and Milesi-Ferretti (1993), designate agricultural and manufacturing goods as traded goods and define the category of nontraded goods to include services (although many services are traded) and construction activities. Once this division is made, a real exchange rate can be computed based on sectoral value-added deflators or other price indices.

Many economists have analyzed how changes in the real exchange rate (P_N/P_T) are associated with changes in a country’s trade balance. For example, Dornbusch (1974a) uses an economy’s aggregate budget constraint to show that the balance of trade in goods equals excess demand in the market for nontraded goods, and thus if a country is initially running a trade surplus, then this surplus must be matched by an excess demand for nontraded goods. If the market for nontraded goods is to clear, the price of nontraded goods must rise relative to other goods; in this example, the price of nontraded goods must rise relative to the exogenous price of traded goods. A rise in P_N relative to P_T reduces the output of tradables and increases output of nontradables, which corresponds to a movement along the economy’s production possibilities frontier. A rise in P_N/P_T will also increase domestic consumption of tradables and discourage consumption of nontradables. Lower production and greater consumption of tradables will reduce the trade surplus, while larger output and reduced consumption of nontradables will eliminate the excess demand. Thus, the trade surplus is eliminated and equilibrium in the market for the nontraded good is restored by a real exchange rate appreciation– a rise in P_N/P_T--so P_N/P_T plays an important role is bringing about changes in a country’s external position.

While the relative price of traded to nontraded goods is an important relative price, looking at movements in this relative price alone is not sufficient as an indicator of competitiveness over time. As demonstrated by Lipshitz and McDonald (1991) and Marston (1987), changes in the relative price of tradables to nontradables may not be a reliable indicator of changes in competitiveness, especially when growth in labor productivity differs across sectors of the economy. Marston (1987) demonstrates that given the high growth in labor productivity in Japan’s traded goods sector, an increase in P_N/P_T, i.e., a real appreciation, has nonetheless been associated with an improved trade position in Japan. Lipshitz and McDonald (1991) suggest that even if the price of tradables to nontradables is adjusted for differences in productivity,’ a rise in this relative price could be consistent with faster growth in the output of traded goods in the home country relative to partner countries if there are rigidities in factor markets and productivity growth in the traded goods sector is sufficiently large. Furthermore, as real per capita income rises over time, the relative price of nontradables will tend to rise (the relative price of tradables will fall). This shift in relative prices should not by itself suggest that the economy has suffered a loss of competitiveness.

Real exchange rate indicators using unit labor costs

Instead of focussing on indicators of competitiveness based on prices, a third indicator–an index of unit labor costs in a country’s manufacturing sector–is often to used to assess competitiveness based on costs. For each country, the unit labor cost index is defined as the ratio of an index of hourly compensation per worker in the manufacturing sector to an index of output per man-hour. ^13/ A real exchange rate indicator is then computed by dividing the index of unit labor costs in the home country by the index of unit labor costs for the sixteen industrial countries for which data are collected by the IMF. For example, the real exchange rate using unit labor costs is:

where all indices are expressed in a common currency. ^14/ The numerator of equation (22) equals labor costs per unit of value added in the traded sector in the home country and the denominator is the corresponding value in the foreign country.

A real exchange rate based on unit labor costs may serve as an indicator of the profitability of producing tradable goods. Lipschitz and McDonald (1991) note that unit labor costs can provide information on the profitability of producing tradable goods if the prices of traded goods are linked through international competition, no intermediate inputs are used in production, the capital stock is fixed, and technology is homogeneous across countries. Under these circumstances, a rise in unit labor costs in the manufacturing sector in the home country, relative to the foreign country, will be associated with a loss in competitiveness and a deterioration of the trade balance of the home country.

Using unit labor costs for an assessment of competitiveness has a number of advantages. For industrial countries, data on wage costs are widely available on a comparable basis across countries. It is also reasonable to focus on developments in labor costs as many industrial countries have singled out the containment of wage costs as an important component of policies designed to achieve macroeconomic stability. Moreover, unit labor costs provide information on an important component of production costs that is nontraded and this is useful information as labor costs can differ widely across countries. Therefore, unit labor costs reveal important information about underlying costs of production.

There are, however, circumstances in which movements in unit labor costs can give misleading signals concerning changes in a country’s trade balance. For example, production activities typically use inputs other than labor such as capital and intermediate inputs. Changes in the prices of these inputs have implications for competitiveness that would be undetected from just an examination of unit labor costs. As Lipshitz and McDonald (1991) point out, when intermediate inputs are important in the production process and their prices vary significantly across countries, then changes in unit labor costs may not be indicative of a change in competitiveness. For example, if the price of an intermediate input such as oil rises differentially in one country, then profits would be reduced for a given price of final output, P^. In this case, the country in question has suffered a loss in competitiveness, yet this loss may not necessarily be captured by movements in an indicator based on unit labor costs.

According to Turner and Van t’ dack (1993), unit labor costs possess two drawbacks. First, it is possible that a unit labor cost indicator will provide a misleading impression concerning a country’s competitiveness in the case where gains in productivity result from a substitution of capital for labor. In this case, the substitution of capital for labor will result in a lower unit labor cost index, but higher capital costs, so the observed decline in unit labor costs may actually overstate any improvement in competitiveness. In addition, unit labor costs are highly sensitive to cyclical movements in labor productivity over the course of the business cycle. An attempt is made to control for these cyclical effects by basing unit labor costs on trend productivity rather than short-run changes in productivity. ^15/

An indicator based on the profitability of producing traded goods

Given the potential shortcomings of relying exclusively on movements in unit labor costs, Lipshitz and McDonald (1991) propose an alternative indicator–a profit-based indicator–as a preferred measure of competitiveness. This profit-based indicator is constructed by dividing the real exchange rate based on unit labor costs by an indicator based on value-added prices:

where PV_T is the value-added deflator in the tradable sector. Substituting for ULC_T from equation (22), equation (23) can be written:

which shows that the profit-based measure, RER_PRF reduces to the ratio of the share of labor costs in valued added in the tradable sector of the home country to the tradable sector in the foreign country. According to Lipshitz and McDonald (199.1), the inverse of the labor share in value added can be interpreted as an indicator of the profitability or the return to capital of producing tradable goods. ^16/ A rise in RER_PRF. would be associated with a loss in competitiveness and a deterioration in the trade balance because the share of labor costs in value added has risen, relative to competitors, resulting in a squeeze on the profitability of producing traded goods in the home country. Unlike the indicator based on unit labor costs, this profit-based indicator would capture a loss in competitiveness which comes, about as a result of a rise in the price of an intermediate input. In this case, for a given price of the final good, an increase in the price of an intermediate good would reduce the value-added price PV-_p, thus the profit-based measure would show an appreciation and a loss in competitiveness. ^17/

Relationship among indicators of competitiveness

The purpose of this section is to provide a discussion of how three of the most commonly used indicators of competitiveness relate to each other. This section relates the indicators based on consumer prices, export unit values in manufacturing, and unit labor costs. All remaining comparisons among indicators are contained in Appendix I.

Consumer price indices and export unit values

The consumer price index is defined as follows:

where α_j is the weight assigned to the jth good in the index. The price index is comprised of all final, consumer-good, prices, so RER_CPI can be expressed as:

where a* denotes a variable in the foreign country. In the extreme case where the prices and weights of traded goods and are equal across countries, the real exchange rate indicator based on a ratio of CPIs is a function of the prices of nontraded goods in two countries, as discussed above. Under more general conditions, a real exchange rate indicator based on consumer prices is a function of the prices of all consumer goods in the home country, including traded and nontraded goods, relative to all prices in the foreign country as shown in equation (26). By contrast, a real exchange rate index using export unit values is simply a comparison of export unit values among competitors in a given export market; thus a comparison of export unit values uses proxies for one component of the consumer price index, the prices of exportable goods consumed at home. The consumer price index includes the price of exportable goods purchased by domestic consumers, while export unit values are proxies for the actual prices of exports. Comparing these two indicators, a real exchange rate based on consumer price indices is preferable since it provides information on the prices of exports and imports rather than on just one component of the trade balance, and consumer price indices are readily available.

Consumer price indices and unit labor costs

As shown in equation (26), the ratio of the CPIs in two countries can be expressed as a function of all final goods prices. The relationship between the final goods prices and the factor prices can be demonstrated by substituting the price equations (8- 9) into equation (13):

Each value-added price in the traded sector PV_T, can be written as a function of labor and capital costs (profits):

where R_j is the return to capital in sector j and K_j is the capital stock in sector j. Equation (28) can be written:

or:

where profit equals capital income per unit of value added in sector j. Equation (30) can be substituted into equation (27), giving the relationship between CPIs and unit labor costs:

From equation (31), the ratio of CPIs is a function of unit labor costs in the traded and nontraded sectors, which encompasses more information than a simple comparison of unit labor costs in the manufacturing sectors across countries. Thus, a real exchange rate index based on unit labor costs in manufacturing contains a subset of the information contained in a real exchange rate index based on the ratio of CPIs. The additional information obtained from using CPIs may be relevant from an economy-wide point of view because developments in an economy’s nontraded sector may have important implications for the performance of the traded sector, especially if factors of production are mobile across sectors. Furthermore, there may be instances where a real exchange rate based on consumer price indices may provide a signal of a change in competitiveness that would not be detected by unit labor costs. For example, as mentioned above, a rise in the price of an important intermediate input may erode competitiveness, but the indicator based on unit labor costs will not necessarily detect this loss in competitiveness. To the extent that increases in production costs are passed on to consumers in the form of higher prices, the consumer price index will capture this price increase and signal a loss in competitiveness. Also, consumer prices may show movements in the prices of traded goods which do not fall under the category of manufacturing goods.

However, real exchange rates based on consumer prices are not necessarily superior to unit labor costs. Consumer price indices tend to exhibit much more volatility than normalized unit costs because consumer prices are sensitive to position of the economy in the business cycle. Furthermore, normalized unit labor costs are more narrowly focused on the traded goods sector than consumer price indices, so it is possible that fluctuations in consumer prices are unrelated to changes in competitiveness. For these reasons, normalized unit labor costs possess advantages over consumer price indices.

Comparison of export unit values with unit labor costs

As noted above, the only case in which the comparison of export unit values is meaningful is the case where traded goods are heterogeneous. In this case, a comparison of export unit values is equivalent to a comparison of all the prices of inputs used in producing exportable goods. The ratio of export unit values can be written:

which is a function of unit labor costs in the exportable sector, as well as the prices of traded and nontraded intermediate inputs. Thus, when comparing export unit values, there is an implicit comparison of unit labor costs in the exportable sector across countries; however, this comparison does not include information on unit labor costs in the importable sector. Real exchange rates based on unit labor costs compare unit labor costs in the manufacturing sectors across countries, which is more comprehensive in coverage. The additional information (coverage) contained in a comparison of unit labor costs is important for an assessment of a country’s trade balance since RER_ULC contains some information concerning imports, so this wider coverage represents a distinct advantage of using unit costs.

Summary

Overall, each measure of the real exchange rate possesses shortcomings, and no one indicator provides an unambiguous assessment of competitiveness, which is a conclusion shared by Durand and Giorno (1987), Dornbusch (1986), and Turner and Van’t dack (1993). Wickham (1993) concludes that reliance on competitiveness indicators should only form part of any assessment of the appropriateness of a country’s exchange rate, given the many limitations inherent in the construction of these indicators. Some authors, such as Artus and Knight (1984), make the case that a real exchange rate based on normalized unit labor costs in manufacturing is the preferred indicator of competitiveness, while others such as Lipshitz and McDonald (1991) argue that RER_PRF is the preferred indicator. In the end, each indicator discussed above does not provide a complete assessment of competitiveness by itself. If anything, the above discussion suggests that competitiveness indicators should be used in conjunction with other indicators, in order to obtain an assessment of competitiveness that is as complete as possible.

However, if only one of the three commonly used real exchange rate indices can be selected (RERs based export unit values, consumer price indices, and normalized unit labor costs), then unit labor costs seem to be the most useful for assessing competitiveness. Export unit values provide information on only one component of trade (exports), cover only exported goods rather than potentially exportable goods, and the composition of goods contained in an export basket is likely to differ across countries. Consumer price indices are subject to the influence of distortions, e.g., taxes, do not include prices of intermediate goods which make up a large component of the volume of trade, and are subject to volatility over the business cycle. Normalized unit labor costs in manufacturing possess the advantages that they provide information on both exports and imports and are not subject to the volatility inherent in consumer price indices. ^18/ As the business cycle proceeds, consumer price indices may exhibit volatility unrelated to competitiveness since the consumer price index contains nontraded goods. ^19/ Thus, use of normalized unit labor costs in manufacturing overcomes this difficulty. For an assessment of competitiveness across well-diversified industrial countries, normalized unit labor costs are the preferred indicator since they are readily comparable across countries and information on labor costs is generally available. For developing countries, however, data on unit labor costs are often nonexistent, so real exchange rates based on consumer prices are often the only alternative, a point make by Wickham (1993). ^20/ Despite the advantages of unit labor costs, it cannot be overstressed that any single indicator of competitiveness fails to provide a complete and satisfactory assessment of competitiveness in all cases and the basis for selecting normalized unit labor costs is rather weak.