Relative labor demand in an open economy.
This paper contributes to the discussion concerning the nature of the well-documented worsening of wage and employment inequality in western economies during the past three decades. It critically discusses the use of the traditional Heckscher and Ohlin approach to analyze the distributional effects of international competition. The paper also discusses an innovative theoretical scenario in order to effectively explain the empirical observations. The model overcomes the problem of a dichotomized labor market, which is an unfavorable result of the traditional approach. Furthermore, the factor-biased character of the technological change becomes endogenous as the strength of foreign competition and the induced incentives for technical innovations are taken into consideration. (JEL F10, F41, O33)
The worsening of inequality between workers of different skill levels for the past 25 years is an unquestionable fact. Slaughter  reports that the premium for college-educated male workers in the U.S. rose from 30 percent in 1979 to about 70 percent in 1995. These sharply widening gaps would be less noteworthy if a sustained overall real wage increase accompanied them. Unfortunately, real wages have fallen in the U.S. by a yearly average of about 0.4 percent since 1973. Analogously, the employment of production workers in the British manufacturing sector fell between 1979 and 1992 by 41 percent, while the decrease for the non-production workers was only 26 percent. According to the existing evidence (worsening of wage and employment perspectives), the relative demand for unskilled labor fell over the last two decades. Labor supply has partly but not sufficiently adapted to prevent the increase of inequality [Hine and Wright, 1998].
Since the beginning of the 1990s, discussions about the steady worsening of employment perspectives for the unskilled workers intensified. Supporters of trade liberalization policies based their arguments on the relatively small volume of trade. Additionally, if the rising relative wages for skilled workers were to be explained by the factor price equalization theorem, one should expect an accompanied decrease in the ratio of skilled to unskilled labor in all industries. Yet, while the wage premium associated with education rose sharply since the 1970s in the U.S., the share of college-educated workers rose as well [Krugman and Lawrence, 1993; Lawrence and Slaughter, 1993]. Consequently, the main source of this development must be a skill-biased technical change. [Bound and Johnson, 1992; Berman, Bound, and Griliches, 1994; Machin, Ryan, and van Reenen, 1996].
Nevertheless, it is crucial to identify the source of this technology-skill complementarity. In fact, the character of technological development was not always a skill-biased one. The evolution through the 18th century, from artisan shops to the earliest factories, was characterized by a substitution of highly skilled individuals with physical capital and less-skilled labor [Goldin and Katz, 1998, p. 694]. Acemoglu  shares the same belief by saying, "... new technologies are not complementary to skills by nature, but by design." Even current innovations like simplifying complex tasks with the use of computers, can cause the demand for skilled labor to fall. Wood  regards defensive labor-saving innovation including technological progress in general, as a partial response of domestic producers to increasing foreign competition.
Failures of the Heckscher-Ohlin Approach
The Heckscher-Ohlin model is a valuable approach for defining gains from trade and specialization tendencies for each country in a world of completely open economies and with identical production technologies. However, the main problem arises as the same theoretical framework is used to examine the effects on the domestic distribution of income.
The Labor Market Dichotomy in the Heckscher-Ohlin Framework
The Heckscher-Ohlin model implies a particular picture of the labor market in relative terms [Wood, 1995, pp. 59-62]. If a country is completely open to trade, relative labor demand consists of two downward-sloping curves lying within a horizontal segment covering the range of labor endowments in which the country will not be specialized. (1) With such a crooked relative labor demand, the labor supply may only be significant for relative wages in specialized countries. The opposite is true for trade-induced changes of relative labor demand. They will affect relative wages, only for non-specialized countries with the labor supply lying within the horizontal segment of demand. According to the implications of the standard non-profit equation, relative wages can be affected either by the changes in labor supply or by trade induced shifts of labor demand. However, it can not be affected by both of them simultaneously.
Reverse Adjustments of Skill Intensity
To further evaluate the ability of standard theoretical tools in explaining distributional effects, recent developments, such as the paradigm of new competitors emerging in international markets from regions with a relative abundance of unskilled employees should be examined. The price of the unskilled, labor-intensive product will fall in relative terms, generating an analogous rise of the skill premium and thereby, a specialization tendency towards skill-intensive branches. At the same time, skill intensity decreases in both productions in order to make the skilled workers available to expand the skill-intensive sector. Nevertheless, this is not exactly what happened in previous decades. Recent developments in western economies imply the following: first, a rising skill-premium; next, specialization tendencies toward skill-intensive productions; and finally, a generalized tendency of increasing skill intensity in all branches [Francois and Nelson, 1998, p. 1491; Berman, Bound, and Griliches, 1994, p. 380].
Many authors claim that the only reason for a rising skill premium and an accompanied increase in skill intensity is a skill-biased technical change [Krugman and Lawrence, 1993; Baldwin, 1995]. Figure 1 depicts the effects of such a technical evolution that occurs simultaneously in both industries with xx and yy moving down and to the left. As technical improvement affects both branches in the same way, the common tangent of the new isoquants is simply a parallel line to the previous (through points A' and B'). For the moment, relative factor returns remain unaffected and both industries produce relatively more skilled employees than before. The cone of diversification moves to the left, which is equivalent to an imaginary increase of the country's relative abundance in unskilled labor, so that production shifts to the unskilled, labor-intensive product y.
[FIGURE 1 OMITTED]
According to the basic assumption of the international common technologies, technologiical change will spread throughout the world, increasing the worldwide supply of y. With homothetic tastes remaining unchanged, the relative price of y will fall, yy' moves up to yy[??]. The new common tangent through the points A" and B" has a smaller slope than the previous one, implying a falling relative wage for the unskilled worker. Suppose that the price adjustment is not as strong and the new cone of diversification (qx[??] - qy[??]) is still to the left compared to the original situation (qx - qy), then there will be a rising skill premium and an increasing skill intensity for each different production. However, the tendency to specialize skill-intensive production is missing from the previous scenario. The skill-biased technological change induced a boost in the relative demand for skilled employees, which was balanced by an increase in the skill premium and a relative vivification of the less skill-intensive branches.
The Heckscher-Ohlin model has significant difficulties in explaining the simultaneous appearance of the three key factors previously mentioned. It is restricted by the full-employment assumption. The model denies the possibility of a parallel within-industry and between-industry adjustments. These changes should have an opposite sign, regardless of whether international trade or factor-biased techological changes is the underlying reason.
Empirical observations [Zarotiadis, 2001] show that there is a positive correlation between within-industry and between-industry adjustments of relative labor demand. Specifically, there is a significant positive correlation between annual within-industry and between-industry adjustments in the skill intensity in Austria, Greece, and the U.K. Additional comparisons indicate that skill-biased technological changes are usually applied in industries with a relatively lower skill intensity. Within-industry adjustments in skill intensity seem to be the alternative reaction to changes in the share of total employment for each industry.
Far from being an accidental occurrence, the tendency to specialize among the wider branches and adjustments within each specific industry could be caused by emerging globalization and increased international competition. The model presented in the following pages considers this possibility and depicts the overall impact of global economic circumstances on domestic labor demand and thereby, the local distribution of income.
An Innovative Theoretical Approach
There are two main issues that should be taken into account. First, the effects of relative labor demand and supply should be dealt with simultaneously, in contrast to the picture of the labor market derived by the Heckscher-Ohlin approach. Wages will be affected by the changes in relative demand (derived from international competition), as well as by the changes in a factor's relative abundance. Second, technology will be viewed as affected by competition from abroad. Trade with less developed countries will generate skill-biased technical changes and modify relative wages, thereby increasing the skill intensity of the economy even if worldwide product prices remain unaffected.
For simplicity, attention will be paid to a small economy, examining a limited case with only two separate industries and three production factors (capital, skilled labor, and unskilled labor). Each industry is formed from a total of various firms. It is assumed that the entire output is supplied in a unique world market. Because of its restricted size, the economy cannot affect international prices by its supply decisions.
Assuming there is a world demand strong enough to absorb the fluctuations of home supply, its solitary strategic option is the division of its production endowments between the two industries. Suppose further that both industries produce using a linear limitational technology. (2) In particular, assume that the output is fully determined by capital. Its autonomous variation causes an analogue variation in the use of the other two factors until the indispensable proportion in the input mix is achieved. Both the linear limitational character and capital's leading position are responsible for the long-run dimension of the production function and enable the examination of structural changes:
[Y.sub.i] = [F.sub.i][[K.sub.i], [S.sub.i]([K.sub.i]), [U.sub.i]([K.sub.i])], i = 1, 2. (1)
The output of each industry, [Y.sub.i], depends on the invested capital, [K.sub.i], which also defines the use of skilled and unskilled employees ([S.sub.i] and [U.sub.i]) according to the following linear relationships:
[S.sub.i] = [[xi].sub.i][K.sub.i] and [U.sub.i] = [[zeta].sub.i][K.sub.i], where [[xi].sub.i], [[zeta].sub.i] > 0 (2)
Suppose that the two industries differ in the intensity of using skilled and unskilled labor. For instance, think of industry 1 as relatively skill-intensive:
[d[S.sub.1]/d[K.sub.1]]/[d[U.sub.1]/d[K.sub.1]] > [d[S.sub.2]/d[K.sub.2]]/[d[U.sub.2]/d[K.sub.2] [right arrow] [[xi].sub.1]/[[zeta].sub.1] > [[xi].sub.2]/[[zeta].sub.2]. (3)
The effect of a marginal change of invested capital on the demand for skilled labor in comparison to the demand for unskilled labor is greater in industry 1 than in industry 2.
Analyzing the Shift-Effect
Because of the linear limitational technology and the assumption of an infinite demand, the described economy would be fully specialized in one or the other industry. In order to simulate the gradual adjustment of specialization structures, assume that each branch consists of several firms with non-homogeneous supply products. The distinction between them is related to the differences in the quality and other special features of their products. Also, the decision to invest in an unknown branch embodies many risks concerning the magnitude of transfer costs and the extent of extra achievable profits.
Assume the country's total capital is divided between the two branches ([K.sub.1] and [K.sub.2]). Imagine further that in each period, a fraction of [K.sub.i]([a.sub.i]) will be released. (3) Some of the weaker, less competitive firms in each industry will close. From total released capital, only a portion will be transferred to the other production ([[phi].sub.i]), while the rest will be reinvested in the original industry. Thereafter: (4)
([K.sub.1])[.sub.t] = ([K.sub.1])[.sub.t-1] - [a.sub.1][[phi].sub.1]([K.sub.1])[.sub.t-1] + [a.sub.2][[phi].sub.2]([K.sub.2])[.sub.t-1],
([K.sub.2])[.sub.t] = ([K.sub.2])[.sub.t-1] - [a.sub.2][[phi].sub.2]([K.sub.2])[.sub.t-1] + [a.sub.1][[phi].sub.1]([K.sub.1])[.sub.t-1]. (4)
To find a condition for a stable [K.sub.1]/[K.sub.2] (meaning that ([K.sub.1])[.sub.t]/([K.sub.2])[.sub.t] = ([K.sub.1])[.sub.t-1]/([K.sub.2])[.sub.t-1]), divide the two equations of (4) and solve for ([K.sub.1])[.sub.t]/([K.sub.2])[.sub.t]. The result is:
[a.sub.1][[phi].sub.1]/[a.sub.2][[phi].sub.2] = ([K.sub.2])[.sub.t-1]/([K.sub.1])[.sub.t-1]. (5)
Equation (5) defines the steady state specialization structure. As the left side increases, capital will flow in the following periods from industry 1 to industry 2. The shrinking industry will experience a selection of the best firms, while competition will increase in the expanding industry. Specialization tendencies will remain until condition (5) is restored.
To further clarify the meaning of equation (5), one can interpret [[phi].sub.i] as an expression of attractiveness, defined by the relative profitability of each branch, which in turn depends on the relative unit labor costs (in case of constant, exogenously defined prices). Keeping in mind the assumption about product non-homogeneity and the risks of a successful transfer from one industry to the other, ratio [[phi].sub.1]/[[phi].sub.2] can be defined as a monotone increasing function of skill premium [w.sub.S]/[w.sub.U], (5) for [SIGMA] > 1: (6)
[[phi].sub.1]/[[phi].sub.2] = (p)[.sup.-1]([[zeta].sub.1] + [[xi].sub.1][omega])/([[zeta].sub.2] + [[xi].sub.2][omega]), (6)
where [omega] = [w.sub.S]/[w.sub.U] is the domestic relative wage and p = [p.sub.1]/[p.sub.2], with [p.sub.i] expressing international average prices for the products in industry, i.
Now consider international economic conditions by defining international prices as a simple function of a relative world demand and international relative, unit labor costs:
p = ([d.sub.1]/[d.sub.2])([[zeta].sub.1] + [[xi].sub.1][omega]')/([[zeta].sub.2] + [[xi].sub.2][omega]'), (7)
where [d.sub.1]/[d.sub.2] is the international demand structure and [omega]' = [w.sub.S']/[w.sub.U'] is the international relative wage. (7) By substituting (7) to (6) and finally in equation (4), the following expression is obtained, given the decomposition of capital between the two industries. In other words, the specialization structure of the small economy, due to domestic and international relative wages, is:
[K.sub.2]/[K.sub.1] = ([a.sub.1]/[a.sub.2])([d.sub.2]/[d.sub.1])([[zeta].sub.1] + [[xi].sub.1][omega])([[zeta].sub.2] + [[xi].sub.2][omega]')/([[zeta].sub.1] + [[xi].sub.1][omega]')([[zeta].sub.2] + [[xi].sub.2][omega]). (8)
Assuming linear limitational technologies and non-homogeneous supply of products, any variation of [omega] or [omega]' alters the relative profitability of the two industries and generates equivalent specialization tendencies.
The Definition of Relative Wages
First, think of L representing the country's total labor force, which is divided into skilled and unskilled employees. Also, assume that the country's relative abundance is exogenously defined so that L = S + U (S, U is exogenous.) In addition, labor demand has to be considered. As mentioned earlier, the industry-specific demand for skilled and unskilled workers is a function of the invested capital. Therefore, defining total demand for skilled and unskilled employees ([L.sub.S] and [L.sub.U]) is an easy job:
[L.sub.S] = [[xi].sub.1][K.sub.1] + [[xi].sub.2][K.sub.2] and [L.sub.U] = [[zeta].sub.1][K.sub.1] + [[zeta].sub.2][K.sub.2]. (9)
Notice that [L.sub.S,U] is the actual amount of employed labor of each category. Furthermore, one can define relative labor demand for different skill groups in terms of capital distribution by dividing the above two equations:
[L.sub.S]/[L.sub.U] = ([[xi].sub.1] + [[xi].sub.2][K.sub.2]/[K.sub.1])/([[zeta].sub.1] + [[zeta].sub.2][K.sub.2]/[K.sub.1]),
d([L.sub.S]/[L.sub.U])/d([K.sub.2]/[K.sub.1]) < 0 for [SIGMA] > 1. (10)
Wages will be regarded as the outcome of negotiations. The position of labor groups in this bargaining depends on their relative importance for the production process. Therefore, one can suppose that wages are positively correlated with specific labor group demand and negatively with the abundance of each group:
[w.sub.S] = [[eta].sub.S]([L.sub.S]/S)[.sup.v] and [w.sub.U] = [[eta].sub.U]([L.sub.U]/U)[.sup.v], (11)
where v > 0 represents the significance of employment rates for wages. (8) [[eta].sub.S,U] expresses all other factors that can be decisive for the level of wages.
Equation (11) gives a very simple explanation of wages. The group-specific unemployment rate reduces the wages determined by all other group-specific factors ([[eta].sub.S,U]). The lower the level of unemployment, or the closer the employment rate is to unity, the higher the [w.sub.S,U]. The outcome of income distribution will be obtained by dividing the last two equations. It is defined by the relative demand for the two labor groups:
[omega] = ([[eta].sub.S]/[[eta].sub.U])[([L.sub.S]/S)/([L.sub.U]/U)][.sup.v]. (12)
Substituting (10) to (12) yields (13), the second relationship between specialization structure and relative wages:
[omega] = ([[eta].sub.S]/[[eta].sub.U])[([[xi].sub.1] + [[xi].sub.2][K.sub.2]/[K.sub.1])U/([[zeta].sub.1] + [[zeta].sub.2][K.sub.2]/[K.sub.1])S][.sup.v]. (13)
According to the last equation, the relative wage of skilled (unskilled) labor is a monotone decreasing (increasing) function of the ratio [K.sub.2]/[K.sub.1], for [SIGMA] > 1.
Analyzing the Effect of Technical Change
The above analysis provides the appropriate theoretical tools to depict the direct effect of international competition in the domestic labor market through the adjustment of the specialization tendencies. Now, the paper introduces the technical change in the developed framework. For this purpose, [[xi].sub.1,2] and [[zeta].sub.1,2] can be expressed by the following identities:
[[xi].sub.i] [equivalent to] [v.sub.S][v.sub.i][x.sub.i] and [[zeta].sub.i] [equivalent to] [v.sub.U][v.sub.i][z.sub.i]. (14)
Use intensity of each labor group in each industry can be expressed as the product of a core-term ([x.sub.i] and [z.sub.i]), an industry-specific ([v.sub.i]), and a factor-specific term ([v.sub.S,U]). This definition helps to distinguish between the different kinds of technical change. For instance:
1) Industry-specific factor-biased technical change: changes of x or z occurring in a single industry. [SIGMA] changes accordingly. 2) Industry-specific Hicks neutral technical improvement: changes of [v.sub.1]/[v.sub.2], altering the relative unit costs. [SIGMA] remains stable. 3) Economy wide factor-biased technical change: changes of [v.sub.S]/ [v.sub.U]. [SIGMA] remains stable. (9)
Further, three crucial assumptions are needed. First, given international identical technologies, assume that innovations, concerning the production process, spread automatically around the world. Furthermore, the firms are completely aware of this fact. The importance of this assumption is straightforward. Because all the domestic investors know that foreign competitors will adopt any new technical development immediately, they have a strong incentive to introduce new technologies. They are at least as efficient as the latest technologies but at the same time, change the input mix in favor of the factors, where the domestic firms have a comparative advantage. (10) Second, assume that innovations can be created only in the highly developed and relatively skill-abundant regions. The firms of the advanced northern countries are the technology leaders, while the southern competitors are the imitators and have no influence on the direction of relative and demand (R & D) activity. It is a realistic assumption, "... since most of the world's research capability, most of its advanced equipment manufacturers and most of their customers are located in the North ..." [Wood, 1994, p. 159]. Finally, assuming perfect interindustry spillovers, a variation of [z.sub.i]/[x.sub.i] will automatically affect the relative factor intensity of the other industry and a change will occur in [v.sub.U]/[v.sub.S].
Now imagine a decrease in [w.sub.U'] (or a rise of [omega]'). Through the effect on international prices, [p.sub.i], firms in both industries experience a decrease in the attainable profits ([[phi].sub.i])[.sup.-1]. This decrease will worsen for the less skill-intensive sector, inducing capital moving to industry 1. So far, the analysis of the shift-effect was simply recapitulated. However, domestic investors have an additional reaction possibility for improving their attainable revenues. Firms in industry 1 and 2 can concentrate some of their resources in order to improve production efficiency and regain a part of their missing profits. Given that the domestic firms anticipate that technical changes will be applied sooner or later by all competitors, they will attempt to reduce the relative use of the unskilled labor, besides the improvement of production's efficiency. Because of interindustry spillovers, all skill-biased technical changes can be summarized in an aggregate reduction of economy-wide [v.sub.U]/[v.sub.S]. (11) In conclusion, the relationship between [v.sub.U]/[v.sub.S] and [omega]' must be such that d([v.sub.U]/[v.sub.S])/d[omega]' < 0. (12)
Economy-Wide, Skill-Biased Technical Change
In both (8) and (13), [[xi].sub.1,2] and [[zeta].sub.1,2] are meaningful parameters. The effects of technical changes on (9) by substituting the identities (14) should first be considered. Then:
[K.sub.2]/[K.sub.1] = ([[alpha].sub.1]/[[alpha].sub.2])([d.sub.2]/[d.sub.1])([z.sub.1][v.sub.U]/[v.sub.S] + [x.sub.1][omega])([z.sub.2][v.sub.U]/[v.sub.S] + [x.sub.2][omega]')/([z.sub.1][v.sub.U]/[v.sub.S] + [x.sub.1][omega]')([z.sub.2][v.sub.U]/[v.sub.S] + [x.sub.2][omega]). (8')
If the first derivative of [K.sub.2]/[K.sub.1] is taken, with respect to [v.sub.U]/[v.sub.S], then the sign of the effect found depends on the following three terms: ([omega] - [omega]'), ([z.sub.1][x.sub.2] - [x.sub.1][z.sub.2]), and [[z.sub.1][z.sub.2]([v.sub.U]/[v.sub.S])[.sup.2] - [omega][omega]'[x.sub.1][x.sub.2]].
Recall the second assumption, that innovations will be generated in the advanced, skill-abundant countries. Consequently, the first term will be less than zero. The second term will also be negative because of [SIGMA] > 1. Hence, the sign of the derivative depends on the third term. By readjusting [[z.sub.1][z.sub.2]([v.sub.U]/[v.sub.S])[.sup.2] - [omega][omega]'[x.sub.1][x.sub.2]], the following expression can be obtained. It is called inequality X:
[[zeta].sub.1][[zeta].sub.2]/[[xi].sub.1][[xi].sub.2] > or < [w.sub.S][w.sub.S']/[w.sub.U][w.sub.U'],
[right arrow] d([K.sub.2]/[K.sub.1])/d([v.sub.U]/[v.sub.S]) > or < 0 respectively. (inequality X)
This relationship compares the universal level of wage inequality between the two labor groups, with the reciprocal ratio of their overall use intensities. If the skilled workers are overpaid in a grade that is above (below) the economy-wide, use-intensity of the unskilled worker, then a change of [v.sub.S]/[v.sub.U] will have the opposite (same) effect for the ratio [K.sub.2]/[K.sub.1] in the northern countries. (13)
The effects on the definition of wages by substituting the identities (14) to (10) and thereby, to (13) have to be thought of:
[omega] = ([[eta].sub.S]/[[eta].sub.U])([v.sub.U]/[v.sub.S])[.sup.-v][([x.sub.1][v.sub.1]/[v.sub.2] + [x.sub.2][K.sub.2]/[K.sub.1])U/([z.sub.1][v.sub.1]/[v.sub.2] + [z.sub.2][K.sub.2]/[K.sub.1])S][.sup.v]. (13')
If the first derivative is calculated with respect to [v.sub.U]/[v.sub.S], then it is smaller than zero. (14)
Definition of [v.sub.U]/[v.sub.S] Due to International Relative Wages
As already mentioned, [v.sub.U]/[v.sub.S] is thought of as a monotone decreasing function of the foreign relative wage, given the domestic relative wage. In a broader definition, the gap between [omega] and [omega]' defines the evolution of the technology's skill-biased character. This notion can be expressed by the following: (15)
[v.sub.U]/[v.sub.S] = ([[omega].sup.N]/[omega]')[.sup.(1-q)], (15)
where q [member of] [0, 1] is a parameter representing the research and development costs or in a more general description, the incentive of a country to innovate. (16) Substitution of (15) to (9') yields:
[K.sub.2]/[K.sub.1] = ([[alpha].sub.1]/[[alpha].sub.2])([d.sub.2]/[d.sub.1])([z.sub.1][bar.[omega].sup.(1-q)] + [x.sub.1][omega])([z.sub.2][bar.[omega].sup.(1-q)] + [x.sub.2][omega]') / ([z.sub.1][bar.[omega].sup.(1-q)] + [x.sub.1][omega]')([z.sub.2][bar.[omega].sup.(1-q)] + [x.sub.2][omega]). (8")
The first derivative of [K.sub.2]/[K.sub.1] with respect to [omega] is still greater than zero, as long as [SIGMA] > 1.
Nevertheless, an interesting and strange result arises concerning the effect of foreign relative wages. In fact, it would be the same as before the internalizing, skill-biased technical changes if the following condition is satisfied: [omega]'(1 - q) / (2 - q) [less than or equal to] [omega] [less than or equal to] [omega]'(2 - q) / (1 - q). This inequality shows that the differences between the domestic and the international relative wages must not be very large. For the special case of q = 0 or when a country has the maximum incentive to innovate, [omega] must be greater than half and smaller than double of [omega]' in order to ensure the observation of shifts in the relationship (8"). Finally, substituting (15) to (13') shows the following definition of domestic relative wages:
[omega] = ([[eta].sub.S]/[[eta].sub.U])[.sup.k] ([omega]')[.sup.v(1-q)k] [([x.sub.1][v.sub.1]/[v.sub.2] + [x.sub.2][K.sub.2]/[K.sub.1])U/([z.sub.1][v.sub.1]/[v.sub.2] + [z.sub.2][K.sub.2]/[K.sub.1])S][.sup.vk], (13")
k [equivalent to] 1/[[v(1 - q) + 1].
The definition of relative wages remains negatively sloped, as it was without the consideration of the substitution effect (since k and therefore, vk is always positive). However, the entire effect of [K.sub.2]/[K.sub.1] on [omega] is weaker than before because of further reactions through the adjustment of [v.sub.U]/[v.sub.S]. Furthermore, there is a new term in the equation, namely [omega]', which shows a positive, direct substitution effect of foreign relative wage on the domestic, through the generation of skill-biased technical changes.
Depicting the Shift and Substitution Effect (17 18)
Figure 2 depicts the shift and substitution effect in the same diagram in which the vertical axis describes the domestic specialization structures (ratio [K.sub.2]/[K.sub.1]) and the horizontal axis shows the skill premium. The ID-curve is equation (13"). It shows the income distribution according to the relative labor demand. The ST-curve shows the specialization structures for different levels of the relative wage equation (8"). At the point where the two curves meet, the pressures in the labor market and the production structure are balanced and the economy would not show any further dynamic for adjustments ([A.sub.0]). If there was a rise in the foreign relative wages [omega]' and assuming q = 1, then technology would stay unaffected and the downward shift of ST to S[T.sub.1] would be the only effect from outside (equation 8 and 13 would be relevant). The economy would move from [A.sub.0] to [A.sub.1] and the fall of ([K.sub.2]/[K.sub.1])[.sub.0] to ([K.sub.2]/[K.sub.1])[.sub.1] induces a rising domestic relative wage.
[FIGURE 2 OMITTED]
Further assume that there is also an effect on [v.sub.U]/[v.sub.S], with q being smaller than one and that the effect of this change on the specialization tendencies is insignificant, (19) then ST would remain in its new position. However, if ID moves to the right and the economy's new equilibrium is [A.sub.2], then an additional deterioration of the relative wage for the unskilled workers is implied ([omega] rises to [[omega].sub.2]). At the same time, the country's new specialization structure is ([K.sub.2]/[K.sub.1])[.sub.2]. Although the difference between the two industries in the skill intensity remains the same as before ([SIGMA] is constant) and the effect of [omega]' on their attractiveness is unchanged, capital will show a modest tendency to move toward industry 1 because there is an overall rise in the domestic relative wage of the skilled workers, according to the direct influence from [omega]' (I[D.sub.1]). What does it mean? Focusing on the actual specialization structures as the only sign of the effect from international trade on domestic distribution patterns is underestimating the significance of global economic developments. Not only because it fails to show the substitution-effect on [omega] (from [[omega].sub.1] to [[omega].sub.2]), but also because it ignores a notable part of the entire impact on the dynamics of specialization tendencies (the vertical distance between S[T.sub.0] and S[T.sub.1]). (20)
Figure 3 describes the functioning of labor market as previously mentioned. In contrast to the standard approach, the presented framework implies no flat segment for the relative labor demand curve. Here LD is continuously negatively sloped as in the autarky case. However, there is a different reason for that. With linear limitational production technologies, the marginl product of labor remains stable with increasing levels of employment. Therefore, the negative slope of LD cannot be justified as in the traditional classical view. Aggregate relative labor demand for unskilled labor declines with rising [w.sub.U]/[w.sub.S] because of the generated specialization tendencies toward skill-intensive industries and because of the skill-biased technical changes. As previously mentioned, the relative labor demand curve is equation (10) after [K.sub.2]/[K.sub.1] is substituted at the right hand side of equation (13"). Notice, there is a line again representing the relationship of ID. However, this time it is equation (12). It consists of all the feasible combinations of relative employment and relative wages, according to the domestic process of wage determination and the international conditions.
[FIGURE 3 OMITTED]
For domestic circumstances (expressed by LS) and international conditions (expressed by [omega]'), the economy will stabilize at point [A.sub.0], which is equivalent to [A.sub.0] in Figure 2. (21) An exogenous increase in [omega]' shifts LD rightwards and the ID upwards, leading the economy to the new equilibrium [A.sub.2] (equivalent to [A.sub.2] in Figure 2). Notice that [A.sub.2] is not equilibrium with full employment. In this case, unemployment will be relatively higher for the unskilled employees.
Based on the standard assumptions of identical technologies and differences in relative factor abundance, the importance of global economic circumstances for the inequalities among skilled and unskilled employees was restated. Relative labor demand is continuously negatively sloped in contrast to the strange dichotomized picture implied by the Heckscher-Ohlin approach. Accordingly, domestic skill premium affects the relative demanded amount of unskilled workers. Responding to any decrease in [omega], the investors find it profitable to either move their funds to more skill-intensive products or to increase the skill intensity of the production. Simultaneously, international relative wages are also significant. For instance, given an exogenous rise in the foreign skill premium, the reaction of domestic investors generates pressure towards the relative labor demand. Domestic [omega] will grow and the relative employment of skilled employees will rise. The country will specialize in skill-intensive products, while at the same time, the overall skill intensity will increase (moving from point [A.sub.0] to [A.sub.2] in Figure 3). The framework presented overcomes the previously described difficulties of explaining the simultaneous appearance of the three key observations.
International trade is indeed a powerful engine of wealth creation. At the same time, theory denotes that even if trade liberalization improves national welfare, it does not necessarily imply that all economic actors will gain. Supporters of liberalization policies usually refuse the thought that commercial relationships with developing countries could have a significant negative impact on the jobs in the advanced countries because they see the danger of protectionism. Certainly, protectionism cannot be the answer for the discussed inefficiencies. At the same time, the severe developments of wider groups of population in the western economies are an unquestionable fact. Ignoring them or disregarding the significance of international competition does not help find the appropriate answers. According to the framework presented, international conditions affect the decisions of domestic investors, thereby changing the picture of relative labor demand and generating adjustments in relative wages and employment perspectives. As much as the lagging adjustment of labor supply to the new requirements is one of the key features in this analysis, active labor market policies and all similar activities that build up the skills of the labor force would be much more effective than any protectionist strategy. Recognizing the international comparative advantages for each region and applying appropriate education and training programs in the frame of a well-suited, regionally-adjusted active labor market policy is the answer to the emerging problems for a part of the population and also strengthens the country's comparative advantage itself.
(1) Analogous to the cone of diversification: as long as the line presenting the country's relative abundance lies within the cone, relative factor returns remain constant, regardless of the relative endowments (Rybzinski's theorem).
(2) Linear limitational technology implies, on the one hand, a limitation due to the constant proportions in the input mix. The reason for adopting this type of production function is to distinguish clearly among the variations of relative labor demand due to between-industry and within-industry shifts. Ignoring strict concave--substitutional technologies and thereby, the short-term variations in the input mix helps to tell apart the direct effect of trade through specialization tendencies from the indirect one, via affecting the skill intensity of technology. On the other hand, linearity means constant returns to scale. Limitational functions can also be non-linear in that sense. Yet, considering increasing or decreasing returns to scale in the previously described context will generate countries which are specialized in one or the other industry and removes from our theoretical framework the ability of simulating the smooth developments in the international specialization structure.
(3) In other words, [a.sub.i] can be understood as the regeneration rate of capital or of production function in each industry, which can be affected by the average size of the firms in an industry, the speed of technical evolution in this branch, and finally the intensity of competition and the competitiveness of the industry as a whole.
(4) The reason for distinguishing the interindustrial movement of capital in two different effects ([a.sub.i] and [[phi].sub.i]) is to be able to discuss separately the underlying motives causing the specialization tendencies.
(5) Although it is not introduced in this paper, it is easy to imagine the modeling of branches with products differing in the qualitative characteristics.
(6) Where [SIGMA] = [[xi].sub.1][[zeta].sub.2]/[[xi].sub.2][[zeta].sub.1] expresses the gap between the two industries in the use intensities for each labor group.
(7) At this point, it is assumed implicitly that there are internationally identical technologies. The notation for the international use intensities is the same as for the domestic intensities.
(8) Parameter v can also be interpreted as the degree of sensitivity of the labor market or of the bargaining system, concerning unemployment. For example, the more decentralized the system of wage negotiations, the lower v will be, expressing the lower importance of unemployment and the stronger occurrence of macroeconomic externalities [Layard, Nickell, and Jackman, 1991, pp. 52-9]. Similar arguments have been mentioned in OECD-Jobs Study, Part 2, 1994.
(9) There is also a forth case, namely economy-wide Hicks neutral technical improvement. In this case, all labor use intensities alter in the same direction and the same proportion for both industries. In relative terms, nothing changes. Therefore, these developments are irrelevant for the analysis.
(10) This further simplifies the analysis by overruling an industry-specific, Hicks neutral technical improvement, changing it to [v.sub.1]/[v.sub.2], and leaving the international competitiveness of each industry unaffected.
(11) Notice, in the context of this framework skill-biased innovation can either be defensive or aggressive.
(12) Furthermore, the induced changes of [v.sub.U] and [v.sub.S] must satisfy the Rationality Restriction (RR): -[z.sub.i]([v.sub.U']-[v.sub.U]) [greater than or equal to] [omega][x.sub.i]([v.sub.S']-[v.sub.S]), where [v.sub.U'] and [v.sub.S'] are the new, economy-wide, factor-specific terms.
(13) Today, it is reasonable to assume that the right hand side of (inequality X) is greater than 1 and rises gradually, while the left hand side shows a tendency to fall, making the scenario of d([K.sub.2]/[K.sub.1])/d([v.sub.U]/[v.sub.S]) being less than zero more likely.
(14) Notice, in (13'), parameters [v.sub.i] are exogenous, according to the assumption of international identical, production technologies. However, a Hicks neutral development in favor of industry 1 ([v.sub.1]/[v.sub.2] falls) will shift ID to the left as long as [SIGMA] > 1.
(15) Also notice that among the explaining variables, there is not just the domestic relative wage but also the relative wage of the northern world ([[omega].sup.N]). In other words, [v.sub.U]/[v.sub.S] is an exogenous variable in developing countries.
(16) Skill-biased innovation can be determined more precisely according to the extent to which the attainable profits can be increased, the society' time preferences, and the R & D costs. Yet, this definition is chosen for reasons of simplicity.
(17) Meaning the direct impact of international competition on the domestic relative use of different types of labor (relative labor demand) through the effect on specialization structure, [K.sub.2]/[K.sub.1] (between-industry changes of skill intensity).
(18) Meaning the indirect impact of international competition through the effect on [v.sub.U]/[v.sub.S] (in other words within-industry changes of skill intensity).
(19) This could be the case, if for example (inequality X) becomes [[zeta].sub.1][[zeta].sub.2]/[[xi].sub.1][[xi].sub.2] [congruent to] [w.sub.S][w.sub.S']/[w.sub.U][w.sub.U'].
(20) Notice, the framework presented allows an overall, perverse effect of foreign relative wages on specialization structure, even if the shift-effect will be a normal one (d[K.sub.2]/[K.sub.1]/d[omega] < 0). ST can move downwards but if the move of ID is sufficiently large or the slope of ID is small enough (in absolute terms), one could observe a rise in [K.sub.2]/[K.sub.1]. In terms of first derivatives, this means that d[K.sub.2]/[K.sub.1]/d[omega]' < d[omega]/d[omega]' and d[K.sub.2]/[K.sub.1]/d[omega] > d[omega]/d[K.sub.2]/[K.sub.1] to a sufficient extent.
(21) Notice, [A.sub.0] is deliberately drawn as a full employment situation.
Acemoglu, D. "Why Do New Technologies Complement Skills? Directed Technical Change and Wage Inequality," Quarterly Journal of Economics, November, 1998.
Baldwin, R. E. "The Effect of Trade and Foreign Direct Investment on Employment and Relative Wages," NBER Working Paper 5037, 1995.
Berman, E.; Bound, J.; Griliches, Z. "Changes in the Demand for Skilled Labor Within U.S. Manufacturing: Evidence From the Annual Survey of Manufactures," Quarterly Journal of Economics, May, 1994.
Bound, J.; Johnson, G. "Changes in the Structure of Wages During the 80s: An Evaluation of Alternative Explanations," American Economic Review, June, 1992.
Francois, J. F.; Nelson, D. "Trade, Technologies and Wages: General Equilibrium Mechanics," The Economic Journal, 108, September, 1998.
Goldin, C., Katz, L. F. "The Origins of Technology-Skills Complementarity," Quarterly Journal of Economics, August, 1998.
Hine, R. C.; Wright, P. W. "Trade With Low-Wage Economies, Employment, and Productivity in Manufacturing," The Economic Journal, 108, September, 1998.
Krugman, P.; Lawrence, R. Z. "Trade, Jobs, and Wages," NBER Working Paper 4478, 1993.
Layard, R.; Nickell, S.; Jackman, R. Unemployment: Macroeconomic Performance and the Labour Market, Oxford University Press, New York, 1991.
Lawrence, R. Z.; Slaughter, M. J. "International Trade and U.S. Wages: Great Sucking or Small Hiccup," Paper Presented at the MICRO-BPEA Meeting, Washington, D.C., June, 1993.
Machin, S.; Ryan, A.; van Reenen, J. "Technology and Changes in Skill Structure: Evidence from an International Panel of Industry," CEPR Discussion Paper 1434, 1996.
Slaughter, M. J. "International Trade and Labour-Market Outcomes: Results, Questions, and Policy Options," The Economic Journal, 108, September, 1998.
Wood, A. North-South Trade, Employment, & Inequality: Changing Fortunes in a Skill-Driven World, Clarendon Press-Oxford, U.K., 1994.
__. "How Trade Hurt Unskilled Workers," Journal of Economic Perspectives, Vol. 9, No. 3, Summer, 1995.
Zarotiadis, G. "Within-and Between-Industry Changes in the Relative Use of Non-Operative Workers in the Austrian, Hellenic, and British Manufacturing," Paper Presented at the ASSET Euroconference, Rehthymno, Greece, October, 2001.
*University of Ioannina--Greece. This paper was presented at the 55th International Atlantic Economic Conference, Vienna, Austria, March 12-16, 2003.
|Printer friendly Cite/link Email Feedback|
|Publication:||International Advances in Economic Research|
|Date:||May 1, 2004|
|Previous Article:||Modeling the European cycle with factor structure and regime switching.|
|Next Article:||Endogenous employment rate in the efficiency wage shirking approach.|