Chapter Seven. Wages and Employment in a Single Labour Market. Main Questions

Transcription

1 202 PART 3: Labour Supply and Demand Together Chapter Seven Wages and Employment in a Single Labour Market Main Questions How is the equilibrium wage and employment level determined in a single labour market? How does imperfect competition affect the way in which we use the supply and demand model to analyze wages and employment determination? Are payroll taxes job killers? Is employment lower in Europe than in North America because of higher taxes? What is monopsony? How do wage and employment outcomes differ in labour markets where firms have market power in the hiring of labour? How much worse off might workers be if they have few alternative places of employment? Do minimum wages do more harm than good? In this chapter we complete the neoclassical model by analyzing the interaction of labour supply and demand in a single market. The firm may be operating in a competitive or noncompetitive (monopoly, oligopoly, monopolistic competition) product market. Alternatively, the firm may be competitive or not competitive (i.e., possess some market power, or monopsony) in the labour market. Throughout the analysis we assume that workers are selling their labour on an individual basis; in later chapters we analyze the situation of collective bargaining via unionization. In dealing with the interaction of supply and demand in various market structures, it is important to be specific about the level of aggregation that is being analyzed. In this section we begin at the level of aggregation of the individual firm, and consequently focus on the firm as the decision-making unit. Subsequently we move to the market, dealing with higher levels of aggregation such as the occupation, industry, region, and economy as a whole the levels at which the market wage is determined in competitive labour markets. We then relax the assumption of perfect competition in the product market. As will be seen, this does not substantially alter the supply and demand analysis. We describe the use of the supply and demand framework in policy analysis, with an extended investigation into payroll taxes. Relaxing the assumption of perfect competition in the labour market has a greater impact on the analysis of wage and employment determination. We move from this discussion of monopsony to an in-depth examination of the impact of minimum wages on the labour market, where the monopsony and perfectly competitive models of the labour market have been brought to bear in the interpretation of recent, potentially anomalous, empirical evidence. 202

2 Chapter 7: Wages and Employment in a Single Labour Market 203 A GUIDE TO CHAPTER 7 Chapter 7 pulls together the components of the supply and demand framework developed in Chapters 1 to 6, exploring how firms and workers interact in the market in determining the level of employment and wages. The key model is the competitive, neoclassical, market-clearing, supply and demand model. This model underlies the topics in the remainder of the text. The main objective of this chapter, then, is to see how the model works, and to consider some of the ways in which it may be wrong. One obvious set of ways in which the model may be wrong is in the assumption of perfect competition, that is, the assumption that no one firm or individual can affect the prevailing wage: One possible route that imperfect competition may take is via the product market, where employers may have some degree of monopoly power. We saw in Chapter 5, and will see confirmed here again, that while the slope of the labour demand function may be affected, the basic labour supply and demand framework is unaffected by imperfect competition in the product market. Alternatively, there may be imperfect competition within the labour market itself. For example, there may be only a few (or one) employers, in which case workers have few options of where to work. The case of a single employer is called monopsony, and it significantly affects the way in which employment and wages are determined in a labour market. Imperfect competition could run the other way as well. Labour may be sold to firms on a monopolistic basis by labour unions. Because unions play such an important role in the labour market, we defer their study to separate chapters (Chapters 14 16). The outline of Chapter 7 is thus: 1. Imperfect competition in the product market, and its impact on the labour demand function. 2. The competitive supply and demand model: solving for and interpreting the equilibrium, and applying the model to an evaluation of payroll taxes. 3. Monopsony in the labour market, or employment determination when firms (but not workers) can affect the wage. 4. Measuring the impact of the minimum wage on employment. Traditionally, this topic would be discussed in a labour demand chapter. However, recent empirical evidence requires an understanding of the impact of minimum wages on monopsonistic markets, before it can be presented. THE COMPETITIVE FIRM S INTERACTION WITH THE MARKET We first examine the case in which the firm is a competitive seller of its output in the product market and a competitive buyer of labour in the labour market. In essence, the firm is so small relative to both markets that it can sell all of the output it wants at the going price of the product, and it can buy all of the labour it wants at the going wage rate. The firm is both a price- and a wage-taker:it cannot influence either the product price or the wage rate. This situation is depicted in Figure 7.1(a) and (b) for two competitive firms. In both cases, their supply schedules for a given homogeneous type of labour are perfectly elastic (horizontal) at the market wage rate, W c. The firms are wage-takers, not wage-setters, and consequently can employ all of the labour they want at this market wage rate. The market wage rate for this specific, homogeneous type of labour is determined by the interaction of supply and demand in the aggregate labour market as depicted in Figure 7.1(c). This aggregate labour market could be a regional labour market for a particular

3 204 PART 3: Labour Supply and Demand Together Figure 7.1 Market demand for labour is the sum of the demand for labour by each firm at each wage. At a wage W 0, market demand is N N0 2, plus remaining firms demand at W 0. In a competitive market, each firm faces perfectly elastic labour supply at the prevailing wage. For the whole market, the market supply of labour is relevant, and in panel (c), W C is the equilibrium wage. W W 0 W c N 0 1 N 1 Competitive Product and Labour Markets N W W 0 S 1 W c S 2 N 0 2 N 2 N W W c ΣN i S D = ΣD i N (a) Firm 1 (b) Firm 2 (c) Aggregate labour market (for given occupation, region, industry) occupational category of labour. For example, it could be the Halifax labour market for junior accounting clerks. By assuming that the firms are in the same region and are hiring the same homogeneous occupational type of labour, we are able to minimize the intervening influence of these factors on the wage determination process, and to thereby focus on the issue of wage and employment determination at the disaggregate or microeconomic level of the firm. In the subsequent chapters on wage structures, these assumptions are relaxed sequentially and the resultant wage structures analyzed. The demand schedules for labour in the two firms are the schedules of the value of the marginal products of labour, defined as the marginal physical product of labour times the price at which the firms can sell their products. (These were derived formally in Chapter 5.) Since the firms are assumed to be competitive sellers of their products, their product prices are fixed at p 1 * and p 2 *. Only if the firms are selling the same output would their product prices have to be the same; otherwise p 1 * need not equal p 2 *. The magnitude and the elasticity of the demand for labour also are depicted as being different simply to emphasize that the market wage is the same irrespective of these factors. The demand schedules determine the level of employment in each firm: in this case N 1 and N 2 units of labour, respectively in firms 1 and 2. The market demand curve, as depicted in Figure 7.1(c), is the summation of the demand curves of the individual firms, such as those shown in Figures 7.1(a) and (b). Conceptually the market demand curve can be obtained as follows. For any specific wage rate, such as W 0 in Figure 7.1, determine the quantity of labour that each firm in the market would wish to employ. For the two firms depicted in Figures 7.1(a) and (b), these quantities are N 0 1 and N 0 2 respectively. Adding these quantities gives the total market demand at that wage rate. Repeating this process for all wage rates traces out the market labour demand curve. In summary, when the firm is a competitive buyer of labour, it faces a perfectly elastic supply of labour at the market wage. When the firm is a competitive seller of its output on the product market, it regards the price at which it sells its output on the product market as fixed, and its derived demand for labour schedule is the value of the marginal product of labour, defined as the marginal physical product of labour times the fixed price at which the firm sells its output. Because the labour supply schedule to the firm is perfectly elastic at the market wage rate, the intersection of the firm s labour supply and demand schedules determines the employment level of the firm for that particular type of labour

4 Chapter 7: Wages and Employment in a Single Labour Market 205 wages are determined elsewhere, specifically in the aggregate labour market for that particular type of labour. The previous analysis was based on the long-run assumption that the firm could get all of the labour it needed at the market-determined wage rate. That is, in order to expand its work force it need hire additional workers only at the going wage rate: there is no need to increase wages to attract additional workers. In the short run, however, even a firm that is competitive in the labour market may have to raise its wages in order to attract additional workers. This situation is often referred to in the literature as dynamic monopsony (Baily, 1975). In such circumstances the firm s short-run labour supply schedule could be upward sloping, as depicted by the schedule S s in Figure 7.2. In the short run, in order to expand its work force so as to meet an increase in the demand for labour from D to D', the firm may have to pay higher wages, perhaps by paying an overtime premium to its existing work force or by paying higher wages to attract local workers within the community. The resultant expansion of the work force can be depicted as a movement up the short-run supply curve in response to the new higher wage of W s, occasioned by the increase in the demand for labour from D to D'. In the longer run, however, a supply influx of other workers will be forthcoming because the firm is paying an above-market wage (i.e., W s > W c ) for that particular type of labour. The supply influx may not be instantaneous, but may occur in the long run because it may come from other firms or perhaps from outside of the labour force, and such adjustments take time. The new supply influx in response to the higher wage could be depicted by the S s ' supply schedule of labour. The supply influx would depress the temporarily high, short-run wage of W s back to its long-run level of W c. Thus the long-run supply of labour schedule to the firm, S 1, can be thought of as a locus of long-run equilibrium points, traced out by various shifting short-run supply schedules of the firm, as the firm tries to expand its work force. In essence, temporary wage increases above the competitive norm are consistent with the firm being a competitive buyer of labour in the long run. In fact, short-run wage increases can be a market signal for the supply response that ensures that market forces operate in the longer run. Figure 7.2 Long-run labour supply to the firm, S 1, is perfectly elastic at W C. In the short run, mobility restrictions may yield upwardsloping labour supply, such as S S. At higher wages, more workers will want to work at the firm, and at lower wages, the firm will lose some (but not all) of its workers. If demand shifts from D to D', the firm can hire more workers at W S. In the long run, workers will be attracted to this firm, shifting labour supply to S' S, until the firm is paying W C. Short-Run and Long-Run Labour Supply Schedules to Firm W S s S s W s W c S 1 D D N 0

5 206 PART 3: Labour Supply and Demand Together IMPERFECT COMPETITION IN THE PRODUCT MARKET Monopoly When an industry is competitive in the product market, the industry demand for labour is obtained by aggregating the labour demand of each of the firms in the industry, as illustrated in Figure 7.1. In the case of monopoly in the product market, the firm is the industry, and therefore there is no need to distinguish between the firm s labour demand and that of the industry as a whole. As discussed in Chapter 5, the labour demand schedule for the competitive firm on the product market is given by: w* = MPP N p Q * = VMP N [competitor] (7.1) where the value of the marginal product of labour (VMP N ) is used to distinguish: w* = MPP N MR Q = MRP N [monopolist] (7.2) which applies to the monopolist since its output price is not fixed. Rather, for the monopolist, marginal revenue is the relevant factor. The difference between equations 7.1 and 7.2 highlights the fact that when the monopolist hires more labour to produce more output, not only does the marginal physical product of labour fall (as is the case with the competitor), but also the marginal revenue from an additional unit of output, MR Q, falls. This latter effect occurs because the monopolist, unlike the competitive firm, can sell more output only by lowering the product price and this, in turn, lowers marginal revenue. Because both MPP N and MR Q fall when N increases in equation 7.1, then the monopolist s demand for labour falls faster than it would if it behaved as a competitive firm in the product market, in which case only MPP N would fall, as given in equation 7.2. The difference between the labour demand schedule for a monopolist and the schedule that would prevail if the industry were competitive in the product market is illustrated in Figure 7.3. This comparison is most meaningful if it involves two situations that are identical except for the difference in market structure. To carry out the comparison, begin with a large number of price-taking firms. Aggregating the labour demand of each of the firms in this competitive industry gives the industry demand curve D C = VMP N in Figure 7.3. Now suppose these firms form a cartel and set the product price to maximize total industry profit, as would a monopolist that owned all of the firms in the industry. The labour demand schedule for the monopolist is D M = MPP N MR Q = MRP N. Since MR Q < P Q then D M lies below D c. The monopolist (or a cartel of firms acting like a single monopolist) raises the industry price relative to the competitive equilibrium level, which reduces output and employment. Thus, at any particular wage (e.g., W* in Figure 7.3), employment will be lower if the industry is monopolized than in an otherwise identical competitive industry. As long as the number of workers employed by the monopolist is small relative to the size of the relevant labour market, the fact that the monopolist has market power in the product market does not translate into market power in the labour market. A standard monopolist that simply tries to maximize profits should just pay the going competitive wage. As we will see below, however, there are other reasons a monopolist may end up paying more than the equilibrium market wage. Monopolistic Competition and Oligopoly Between the polar cases of competition and monopoly in the product market are a variety of intermediate cases. Firms can be monopolistically competitive, a situation characterized

6 Chapter 7: Wages and Employment in a Single Labour Market 207 Figure 7.3 In a competitive product market, demand for labour is given by D C = MPP N P Q = VMP N. If a monopolist purchased these firms, demand for labour would be D M = MPP N MR Q = MRP N. At W*, the demand for labour is lower for the monopolist, that is, N* M < N* C, because the monopolist produces lower output. Furthermore, the demand curve is steeper for the monopolist, since MR Q falls as employment and output increases, whereas P Q remains the same. Monopolist Versus Competitive Demand for Labour W W* D C = ΣMPP N P Q = ΣVMP N D M = MPP N MR Q = MRP N 0 N M * N C * N by many firms that are small relative to the total market, but with products that are differentiated in some way, giving the firm some discretion in its price-setting. In such circumstances the demand for the firm s product is not perfectly elastic, as in the competitive case, but rather has some degree of inelasticity reflecting the fact that if the firm raises its price it will not lose all of its market, and if it lowers its price it will not gain all of the market. Under monopolistic competition, as is the case in perfect competition, there are no barriers to entry by new firms. This free entry property implies that firms cannot earn above-normal or monopoly profits in the long run. Oligopoly industries are characterized by few firms that produce sufficiently similar products that the actions of one firm will affect the other firms. Consequently the firms will react to the actions of the other firms, and will take into account the possible reactions of their rivals in making their own decisions. There are many ways in which the firms can react; consequently, there is a large number of possible ways to categorize oligopoly situations. Oligopoly industries are generally characterized by some barriers to entry by new firms, so that above-normal profits may be earned by oligopolists in the long run. The general conclusions reached in the previous analysis of perfect competition and monopoly continue to apply to the intermediate cases of monopolistic competition and oligopoly. In particular, market power in the product market is consistent with the firm s being competitive in the labour market (or labour markets) in which it operates. Thus firms that exercise some discretion, or even exert considerable control, over the product price may be wage-takers in the labour market. In these circumstances, they would pay the market wage for the types of labour they employ, and could increase or decrease their employment without affecting the prevailing market wage. Product Market Structure and Departures from Market Wages The previous analysis highlights that, as long as market power in the product market does not translate into market power in the labour market, there is no particular reason to expect a profit-maximizing firm not to simply pay the going market wage. It is important to realize,

7 208 PART 3: Labour Supply and Demand Together however, that when a monopolist pays the going market wage, all of the monopoly profit or rent goes to the owners of the firm. Workers employed by the monopolist earn the same wage as they would if they worked for a competitive firm making zero profits. As we will see in Chapter 15, however, this scenario is unlikely to hold when workers employed by the monopolist are members of a union that collectively bargains with the employer. In such a setting, there is clear empirical evidence that unions are able to negotiate a wage above the going market wage, thereby redistributing some of the monopoly profits to workers in terms of higher wages. Existing empirical evidence suggests that even when workers are not unionized, they still tend to earn higher wages in industries where firms earn higher profits because of market power in the product market or other considerations. 1 Similarly, organized workers employed by oligopolistic firms will also likely be able to negotiate a wage above the going market wage. A leading example of this phenomena is the North American automobile sector that was traditionally dominated by the Big Three automakers (General Motors, Ford, and Chrysler). While the sector has been under intense foreign competition since at least the 1980s, it was widely believed to be a typical case of an oligopoly making large excess profits during the decades following World War II. Unions were also traditionally very strong in this sector and managed to turn part of the oligopolistic profits into better than average wages and work conditions. More generally, we will see in Chapter 10 that large firms systematically pay higher wages than smaller firms. There are several reasons this may be the case. For example, large firms may have to pay a wage premium to compensate for the rigid work schedules and mass-production techniques associated with larger size or to substitute for costly monitoring. By virtue of their size, large firms may also have to follow administratively determined wage policies: unable to pay each worker its marginal product, such firms may pay wages approximating the productivity of the more productive workers within the group. For these reasons and others, we may expect monopolists and oligopolists that are, by definition, large firms, to pay higher wages than smaller competitive firms. This situation is quite different for firms that operate under monopolistic competition. Under monopolistic competition, free entry implies that firms should not earn above-normal profits in the long run (though they may in the short run, as would occur if there were an unexpected increase in demand for the product). Such firms are also generally small in size, such as retail outlets that are differentiated by location and possibly also by the merchandise carried. Thus, the two characteristics that might cause otherwise wage-taking firms to pay above-market wages economic rents and large size are absent in the case of monopolistic competition, as they are in the case of perfect competition. For these reasons we would expect firms that are monopolistically competitive in the product market but perfectly competitive in the labour market to pay the prevailing market wage, over which changes in their employment levels will exert no influence. WORKING WITH SUPPLY AND DEMAND One of the most common types of applied policy analysis is to simulate the effects of a policy change on the equilibrium level of employment and wages. In order to do this, it is necessary to be able to solve explicitly for the market equilibrium. Consider the most general form of the model, with labour supply and demand functions given by N S = f(w;x) N D = g(w;z) 1 See, for example, Blanchflower, Oswald, and Sanfey (1996).

8 Chapter 7: Wages and Employment in a Single Labour Market 209 W, N S, N D are endogenous variables in this system;while Z, X are exogenous. Economic theory guides us in sorting out what are the various shifters Z, X that will affect labour supply and demand. To solve the system, we invoke market-clearing, and set N S = N D. Then we solve for W* and N* (two equations with two unknowns). The results, which express N, W as functions of the parameters and the exogenous variables, are called the reduced form. Once we have solved for the reduced form, we can easily simulate the effect of changing X, Z on the equilibrium. Linear Supply and Demand Functions A useful example is to consider the simplest functional form for the supply and demand functions: straight lines. Equations for these functions are N D = a + bw; b < 0 N S = c + fw; f > 0 Setting N S = N D, we can solve for the equilibrium or reduced form: W* = Substituting W* into either the supply or demand equation yields the equilibrium employment N* = It is easy to add other variables (shifters) like the X or Z, to our models to make them more realistic: a + bw + X = c + fw + Z yielding a c f b af bc f b a c W* = f b f b + where Z, X represent other supply and demand factors that shift the position of the supply and demand functions. If we know how policy changes affect the position of the demand or supply function (i.e., how they affect a or c), then given estimates of the other parameters we can simulate the effect of the policy change on the equilibrium. The equilibrium wage and employment can also be illustrated graphically using a standard supply and demand graph illustrated in Figure 7.4. In order to match the form of the above supply and demand equations with the unconventional economists reverse representation of functions, it is worth re-expressing these equations in terms of W as a 1 W = + N b b c 1 W = + N f f f b These are equations for straight lines. For example, the slope of the demand curve is given by 1/b, while the intercept is a/b. Many thought experiments take the form of chang-

9 210 PART 3: Labour Supply and Demand Together Figure 7.4 The supply function is given by N S = c + fw, transformed so that the vertical axis is the wage and the horizontal axis is employment, the slope is 1/f, and the intercept is c/f. The demand function is given by N D = a + bw, but after its axes are switched the slope is 1/b and the intercept is a/b. The equilibrium N*,W* is given by the intersection of the functions. W a/b W* c/f Linear Supply and Demand Functions N S = c + fw Slope = l/f Slope = l/b N D = a + bw N* N ing the intercept, that is shifting the position of the supply or demand function, keeping the slopes constant. Note that for these linear demand and supply functions, the absolute N change or the slope is constant, so the elasticity varies. Since the elasticity is given by: W W, if the slope is constant, the elasticity will vary with N W. N The equilibrium values of W* and N* in Figure 7.4 correspond to the analytical expressions derived above. It is sometimes useful to work with log-linear instead of simple linear functions. In this case, the functional form for the supply and demand functions is given by log N D = + logw, < 0 log N S = + logw, > 0 The convenience of the log-linear functional form comes from the fact that the slope parameters and are the demand and supply elasticities. The slopes of the log-linear functions are constant, and so too are the elasticities. Because economists often have estimates of these elasticities, the reduced form turns out to be easier to use for simulating the impact of changes in policy. The solution for the reduced form proceeds identically to the linear case. For simplicity, denote n S = logn S, n D = logn D and w = logw. An equilibrium n*, w* will occur as before, where n S = n D, and implicitly N S = N D. isp/common/ cpptoc_e.shtml Application: Incidence of a Unit Payroll Tax A common use of these equilibrium models is in the evaluation of tax policy. After income taxes, the most contentious tax studied by labour economists is the payroll tax. A payroll tax is a tax levied on employers, based on the level of employment, usually proportional to the firm s payroll. Common examples of payroll taxes in Canada include CPP/QPP premiums, workers compensation, unemployment insurance, and health insurance levies in some provinces. Table 7.1 illustrates these payroll taxes for 1971 and 1997, drawing on Lin (2000). Generally, payroll tax rates have grown in importance, especially for unemployment insurance and the CPP/QPP. In fact, by 2006 the combined tax rate for employer and

10 Chapter 7: Wages and Employment in a Single Labour Market 211 Table 7.1 Payroll Taxes as a Percentage of Labour Income by Type of Payroll Tax, Canada, 1971 and 1997 Type of Payroll Tax Workers compensation Unemployment insurance Canada/Quebec pension Health and/or education Total Source: Authors calculations based on data provided by Lin (2000). employee contributions to the CPP/QPP had reached 9.9 percent. These taxes are often viewed as taxes on employers, rather than on workers. As such, they are often attacked as taxes on jobs, or job killers. By investigating the possible impact of such taxes on wages and employment, we can evaluate the truth to these claims. To keep the notation simple, we will consider a simple per-unit tax, T, applied on a per-employee basis to firms. We will use the linear system of equations developed earlier to further simplify the analysis. Workers are paid the wage W for a unit of work, while firms pay W + T per unit hired, W to the employee, and T to the government. On the face of it, it appears that the commonsense debate has substance: the employers pay the tax to the government (the workers do not), and this is likely to reduce employment because it raises the cost of hiring labour. Unfortunately, this common sense ignores the impact of the tax on the labour market and the adjustments that might occur. The effect of the tax on labour demand is depicted in Figure 7.5. The initial equilibrium is labelled A, with equilibrium employment N 0 and wage W 0. The imposition of the tax shifts the demand schedule N D down by T units for every employment level. The initial demand curve gives the optimal labour demanded at a cost of W per unit, whereas the shifted curve accounts for the fact that the price of labour is now W (to the workers) plus T (to the government), for a total of W + T. If there were no other changes in the labour market (i.e., the wage stayed the same) demand would fall from N 0 to N'. As we can see, however, as long as the supply curve is not horizontal, the equilibrium wage drops to W 1, and employment falls only to N 1. The lower wage effectively means that the workers pay some portion of the tax. In fact, at the new level of employment, N 1, we can account for who pays the tax. Workers wages are lowered by W 0 W 1 (given by BC in Figure 7.5), and this represents their share. The remainder of T, CD, is paid by the firm. Therefore, the incidence of the tax does not necessarily fall only on the party that physically pays the tax or fills in the forms. Algebraically, we can see this by substituting the appropriate post-tax prices into the demand and supply equations: So that N D = a + b(w + T) = a + bw + bt N S = c + fw a c f b a + bw + bt = c + fw, and W 1 = T < W 0 where W 0 is the equilibrium wage in the absence of the tax. b b f

11 212 PART 3: Labour Supply and Demand Together Figure 7.5 The initial equilibrium is A, at a wage of W 0. The payroll tax, T per worker, raises the price of labour to W + T. This shifts labour demand down by T units at each employment level. The new equilibrium is denoted by B, at the lower wage W 1. Some of the tax is paid by workers through lower wages. The workers share of the tax is given by the vertical distance BC, while the incidence of the tax on employers is the remainder of T, which is given by CD. The Effect of a Payroll Tax on Employment and Wages W 0 W 1 N C D B A T N S N D (W) ND (W + T) N 1 N 0 Thus, of the T tax dollars, the worker s wage is reduced from the original level by b b f T. This is the worker s share of the taxes. Clearly, the worker s share will depend on the relative slopes of the supply and demand functions. For example, if b = 3 and f = 1, the workers share is 3 1 = 0.75, so that most of the tax is shifted to workers in terms of lower 3 wages. Note that the effect of a payroll tax may differ in the short and long run. For example, if labour supply is perfectly inelastic in the long run (f = 0),workers will end up paying the entire tax and the employment level will be unaffected. In the short run, with more elastic labour supply, firms will pay part of the tax, and employment will be lower than before the tax (but again, not as much as would be the case if the market wage did not adjust). It is an empirical question what the ultimate incidence of the payroll tax is. Kesselman (1996) provides a very useful summary of the state of empirical knowledge on payroll taxes, as well as a critical review of some of the theoretical and practical issues related to the simple model we have just outlined. One of the first important points he raises and this is very important is that economists pride themselves on evaluating the unanticipated consequences of a given policy due to unaccounted-for equilibrium effects. Unfortunately, they often fall into the trap of ignoring the greater economic system while concentrating on the specific market they are studying. Given that the government needs to raise a given amount of revenue, it must use some form of tax, and all taxes have distortionary effects on markets. The question is then which tax distorts least. Kesselman reviews a number of reasons payroll taxes may actually be useful, and relatively efficient tax instruments, especially if the tax is collected for a particular purpose, such as CPP/QPP. For example, compared to income taxes, payroll taxes are generally easier to administer, compliance is easier, and evasion is lower. The growing body of empirical evidence also seems to support the conclusion that in the long run (as suggested by the model above), the incidence of payroll taxes falls largely on workers, and that the disemployment effect is small. There may be some disemployment effects in the short run, but it does not appear that payroll taxes are the job killers they are made out to be. 2 The analysis of payroll taxes 2 See Hamermesh (1993) for an overview of the empirical evidence on payroll taxes. Dahlby (1993), Beach, Lin, and Picot (1996), Beach and Abbott (1997), and Lin (2000) provide Canadian evidence. Also see Exhibit 7.2 for interesting evidence from Chile.

12 Chapter 7: Wages and Employment in a Single Labour Market 213 Exhibit 7.1 Why Do Americans Work So Much More than Europeans? There is an emerging puzzle about the effect of payroll and other taxes on employment. On the one hand, studies like Beach and Abbott (1997) that look at the effect of payroll taxes within countries generally fail to find evidence that payroll taxes are the job killers they are made out to be. On the other hand, a growing number of studies point out that there is a strong negative relationship between taxes and employment across countries. For example, Prescott (2004) argues that differences in taxes are the primary reason more heavily taxed Europeans now work much less than Americans. Davis and Henrekson (2004) reach a similar conclusion looking at a richer set of cross-country data. As we saw in Figure 7.5, an increase in taxes shifts down the labour demand curve along the labour supply curve. The impact of taxes on employment thus critically depends on the elasticity of labour supply. When labour supply is very elastic, a small increase in taxes can result in a large negative employment effect. By contrast, taxes have little effect on employment when the labour supply curve is inelastic. For the entire working age population, Americans adults now work, on average, 25.1 hours a week compared to 18.0 hours a week in France and 18.6 in Germany. In other words, Americans now work percent more hours a week than people in France or Germany. Prescott argues that taxes can explain this large and growing gap provided that the labour supply elasticity is large enough. To see how large the elasticity needs to be, note that tax rates are 15 to 20 percentage points higher in France and Germany than in the United States (depending on how taxes are measured). The elasticity of hours of work with respect to tax rates must, therefore, be around 2 for a percentage point gap in taxes to translate into a percentage point gap in hours of work. This is a much larger elasticity than what is typically found in the empirical studies of labour supply discussed in Chapter 3. Indeed, Alesina, Glaeser, and Sacerdote (2005) argue that taxes cannot explain differences in hours of work between the United States and Europe because the elasticities required are implausibly large. Alesina, Glaeser, and Sacerdote instead argue that European labour market regulations explain the bulk of the difference in employment and hours of work between the United States and Europe. They also conjecture that a social multiplier in leisure may be playing a role in the U.S. Europe difference. The idea is that leisure is more valuable when other people also take time off, as they do in Europe, than when other people take little time off, as they do in the United States. If this reasoning is correct, it implies that the marginal utility of leisure is higher in Europe than in the United States, and that Europeans rationally choose to consume more leisure than Americans. This social multiplier is closely linked to the concept of social capital popularized by Putnam (2000) in his aptly titled book Bowling Alone. The fact that it is not much fun to go bowling when nobody else does so nicely illustrates how the marginal utility of leisure depends on how much leisure other people also consume. This example highlights how social interactions can have important implications for economic behaviour. Indeed, developing new economic models and empirical tests of social interactions is an active area of research in economics. thus seems to be one area where economists have contributed in a positive way to the evaluation of public policy.

13 214 PART 3: Labour Supply and Demand Together MONOPSONY IN THE LABOUR MARKET To this point, we have examined the wage and employment decision when the firm is both a competitive seller of its output in the product market and a competitive buyer of labour in the labour market. We also relaxed one of those assumptions that of a competitive seller of its output and examined the labour market implications when the firm is a noncompetitive seller of its output, but still a competitive buyer of labour in the labour market. In this section the assumption of being a competitive buyer of labour is relaxed. So as to trace out the implications of this single change, termed monopsony, the firm is still assumed to be a competitive seller of its product. 3 The results of relaxing both assumptions simultaneously that of competition in the labour market and the product market follow in a straightforward fashion from the results of each separate case. Simple Monopsony The situation in which a firm is sufficiently large relative to the size of the local labour market that it influences the wage at which it hires labour is referred to as monopsony. The monopsonist is a wage-setter, not a wage-taker. In order to attract additional units of labour, the monopsonist has to raise wages; conversely, if it lowers the wage rate it will not lose all of its work force. Consequently, the monopsonist faces an upward-sloping labour supply schedule rather than a perfectly elastic labour supply schedule at the going wage, as was the case when the firm was a competitive buyer of labour. This labour supply schedule shows the average cost of labour for the monopsonist because it indicates the wage that must be paid for each different size of the firm s work force. Since this same wage must be paid for each homogeneous unit of labour, then the wage paid at the margin becomes the actual wage paid to all of the workers, and this same average wage is paid to all. The firm s labour supply or average-cost-of-labour schedule is not its relevant decisionmaking schedule. Rather, the relevant schedule is its marginal cost of labour, which lies above its average cost. This is so because when the firm has to raise wages to attract additional units of labour, in the interest of maintaining internal equity in the wage structure it also has to pay that higher wage to its existing work force (intramarginal workers). Thus, the marginal cost of adding an additional worker equals the new wage plus the addition to wage costs imposed by the fact that this new higher wage must be paid to the existing work force. Consequently, the marginal cost of adding an additional worker is greater than the average cost, which is simply the wage. This situation can be depicted by a simple hypothetical example. Suppose the monopsonist employed only one worker at a wage of one dollar per hour. Its average cost of labour would be the wage rate of one dollar. This would also be its marginal cost; that is, the extra cost of hiring this worker. If the monopsonist wanted to expand its work force, however, it would have to pay a higher wage of, for example, $1.20 to attract an additional worker. Its average cost of labour is the new wage of $1.20 (i.e., ( )/2); however, the marginal cost of adding the new worker is the new wage of $1.20 plus the additional $0.20 per hour it has to pay the first worker in order to maintain internal equity in 3 See Thornton (2004) for an interesting discussion of how Cambridge economist Joan Robinson coined the term monopsony over tea with Cambridge classical scholar B.L. Hallward. Robinson had asked Hallward to make up a word parallel to monopoly but with the emphasis on buying rather than selling. Even though the greek word opsonein refers specifically to purchases of dried fish, Hallward thought that the word monopsony sounded better than other alternatives he thought of.

14 Chapter 7: Wages and Employment in a Single Labour Market 215 the wage structure. Thus the marginal cost of the second worker is $1.40 per hour, which is greater than the average cost or wage of $1.20. If a third worker costs $1.40, then the average cost or wage of the three workers would be $1.40, while the marginal cost of adding the third worker would be $1.80, composed of $1.40 for the third worker plus the additional $0.20 for each of the other two workers. (This example is further extended in the first three columns of the table in the Worked Example on page 225, which is presented later to show the impact of minimum wages on a monopsony situation.) The monopsony situation is illustrated diagrammatically in Figure 7.6. The essence of monopsony is that the firm faces an upward-sloping supply schedule for labour and hence has a marginal cost-of-labour schedule that lies above the supply or average cost schedule. The firm maximizes profits by hiring labour until the marginal cost of an additional unit of labour just equals the marginal revenue generated by the additional unit of labour. Marginal revenue is given by the VMP schedule in this assumed case of the firm being a competitive seller of its product. This equality of marginal cost and VMP occurs at the employment level N M. The VMP curve for the monopsonist is not its demand curve for labour in the sense of showing the various quantities of labour that will be demanded at various wage rates. This is so because the monopsonist does not pay a wage equal to the VMP of labour. For example, in Figure 7.6 at the wage W M, the quantity of labour demanded is N M, not W M V M, as would be the case if VMP were a demand schedule equating W with VMP. In essence, the demand for labour is determined by the interaction of the MC and VMP schedules and this depends on the shape of the supply schedule (and hence the MC schedule) as well as the VMP schedule. This line of reasoning is analogous to that underlying the fact that a product market monopolist has no supply curve for its product. Figure 7.6 The marginal benefit of hiring an additional worker is VMP N = MPP N P Q. The supply curve, S, indicates the wage needed to induce a given number of workers to work for the firm. That wage is also the average cost of labour, so S = AC. The marginal cost of labour, MC, is higher than the average cost, because the firm must raise the wage for all of its existing employees, not just the marginal employee hired. The profit-maximizing employment level, N M, occurs at MC = VMP N. In order to attract N M workers, the firm must pay W M. Monopsony W VMP M W C. W M V M S M S 0 0 N M N C N MC S = AC VMP N = MPP N P Q

15 216 PART 3: Labour Supply and Demand Together Implications of Monopsony The level of employment, N M, associated with monopsony is lower than the level, N c, that would prevail if the monopsonist behaved as a competitive buyer of labour, equating the supply of labour with the demand. The monopsonist restricts its employment somewhat because hiring additional labour is costly since the higher wages have to be paid to the intramarginal units of labour. For N M units of labour, the monopsonist pays a wage of W M as given by the supply schedule of labour. The supply schedule shows the amount of labour that will be forthcoming at each wage and, for a wage of W M, then N M units of labour will be forthcoming. This wage is less than the wage, W c, that the monopsonist would pay if it employed the competitive amount of labour N c. The monopsonist wage is also less than the value of the marginal product of N M units of labour. That is, the monopsonist pays a wage rate that is less than the value of the output produced by the additional unit of labour (although the value of that output does equal the marginal cost of producing it). This monopsony profit or difference between wages and the value of marginal product of labour has been termed a measure of the monopsonistic exploitation of labour, equal to VMP M W M per worker or (VMP M W M ) N M for the monopsonist s work force. This monopsony profit accrues to the firm because its wage bill, W M N M, is less than the market value of the marginal output contributed by the firm s labour force, VMP M N M. Because the value of the marginal product of labour for the monopsonist is greater than the value of the marginal product if that firm hired competitively (VMP M > VMP c, the latter of which equals W c ), welfare to society could be increased by transferring labour from competitive to monopsonistic labour markets. This seemingly paradoxical result an increase in welfare by expanding noncompetitive markets relative to competitive markets occurs simply because the value of the marginal product of labour is higher in the monopsonistic market. In effect, transferring labour is akin to breaking down the barriers giving rise to monopsony: it is allocating labour to its most productive use. Although monopsony leads to a wage less than the value of the marginal product of labour, it is also true that intramarginal workers are receiving a seller s surplus; that is, the wage they are paid, W M, is greater than their reservation wage as indicated by the supply schedule. The existing work force up until N M units of labour is willing to work for the monopsonist for a wage that is less than the wage, W M, they are paid. The wage that they are willing to work for is illustrated by the height of the labour supply schedule, which reflects their preferences for this firm and their opportunities elsewhere. However, to the extent that they are all paid the same wage, W M, then the existing work force receives a seller s surplus or economic rent equal to the triangle S 0 S M W M. This is why the term monopsonist exploitation of labour should be used with care. It is true that the monopsonist pays a wage less than the value of the marginal product of labour. However, it is also true that the monopsonist pays a wage greater than the reservation wage (opportunity cost, supply price) that intramarginal employees could get elsewhere for their labour. A final implication of monopsony is that there will be equilibrium vacancies (equal to V M S M in Figure 7.6) at the wage paid by the monopsonist. In other words, the monopsonist would report vacancies at the wage it pays, but it will not raise wages to attract additional labour to fill these vacancies. In this sense the vacancies are an equilibrium, since there are no automatic forces that will reduce the vacancies. The monopsonist would like to hire additional labour at the going wage since the value of the marginal product exceeds the wage cost of an additional unit. However, because the higher wage would have to be paid to intramarginal units of labour, the value of the marginal product just equals the marginal cost, and that is why there are no forces to reduce the vacancies. The monopsonist is maximizing profits by having vacancies. It does not reduce the vacancies because

16 Chapter 7: Wages and Employment in a Single Labour Market 217 it would have to raise wages to do so and the marginal cost of doing so is more than the marginal revenue. Characteristics of Monopsonists As long as there is some inelasticity to the supply schedule of labour faced by a firm, then that firm has elements of monopsony power. To a certain extent most firms may have an element of monopsony power in the short run, in the sense that they could lower their wages somewhat without losing all of their work force. However, it is unlikely that they would exercise this power in the long run because it would lead to costly problems of recruitment, turnover, and morale. Facing an irreversible decline in labour demand, however, firms may well allow their wages to deteriorate as a way of reducing their work force. In the long run, monopsony clearly will be less prevalent. It would occur when the firm is so large relative to the size of the local labour market that it influences wages. This could be the case, for example, in the classic one-industry towns in isolated regions. Such firms need not be large in an absolute sense: they are simply large relative to the size of the small local labour market, and this makes them the dominant employer. Monopsony may also be associated with workers who have particular preferences to remain employed with the monopsonist. In such circumstances, wages could be lowered and they would stay with the monopsonist because their skills or preferences are not transferable. Thus, a mining company could be a monopsonist, even if it were reasonably small, if it were located in an isolated region with no other firms competing for the types of labour it employed. The same company could even be a monopsonist for one type of labour, for example, miners, while having to compete for other types of labour, for example, clerical workers. Conversely, an even larger mining company might be a competitive employer of miners if it were situated in a less isolated labour market and had to compete with other firms. It is size relative to the local labour market that matters, not absolute size. Monopsony may also be associated with workers who have specialized skills (specific human capital) that are useful mainly in a specific firm, or with workers who have particular preferences to remain employed with a specific firm. Because their skills or preferences are not completely transferable, such workers are tied to a single employer which therefore possesses a degree of monopsony power. Contractual arrangements in professional sports, for example, often effectively tie the professional athlete to a specific employer (team), giving the employer a degree of monopsony power. At a minimum, the employer need pay a salary only slightly higher than the player s next-best-alternative salary (e.g., in minor leagues or in another line of work), which would be considerably less than the value of his skills in a competitive market. In practice, a much higher salary is usually paid in order to extract maximum performance. When players are free agents, as for example when different leagues compete, the resultant salary explosions attest to the fact that the competitive salary is much higher than the monopsonist salary. The professional sports example illustrates the important fact that, although monopsonists pay less than they would if they had to compete for labour, they need not pay low wages. In fact, unique specialized skills are often associated with high salaries, albeit they might even be higher if there were more competition for their rare services. Monopsony does not imply low wages: it implies only wages that are lower than they would be if there were competition for the particular skills. Perfect Monopsonistic Wage Differentiation The previous discussion focused on what could be labelled simple monopsony a situation where the monopsonist did not differentiate its work force but rather paid the same wages to all workers of the same skill, both marginal and intramarginal workers. In terms of Figure 7.6, all workers were paid the wage, W M, even though intramarginal workers

17 218 PART 3: Labour Supply and Demand Together would have been willing to work for a lower wage as depicted by their supply price. The resultant seller s surplus or economic rent (wage greater than the next-best-alternative) is appropriated by the intramarginal workers. This highlights another implication of monopsony. The monopsonist may try to appropriate this seller s surplus by differentiating its otherwise homogeneous work force so as to pay each worker only her reservation wage. If the monopsonist were able to do this for each and every worker, the result could be labelled perfect monopsonistic wage differentiation. In this case the supply schedule would be its average cost of labour and its marginal cost, because it would not have to pay the higher wage to the intramarginal workers. In such circumstances the discriminating monopsonist would hire up to the point N c in Figure 7.6. In fact there is an incentive for the monopsonist to try to expand its work force by means that would make the cost of expansion peculiar only to the additional workers. In this fashion the monopsonist could avoid the rising marginal cost associated with having to pay higher wages to intramarginal workers. Thus monopsonists may try to conceal the higher wages paid to attract additional labour so as not to have to pay their existing work force the higher wage. Or they may try to use nonwage mechanisms, the costs of which are specific to only the new workers. Moving allowances, advertising, or other more expensive job search procedures, and paying workers for paper qualifications that are largely irrelevant for the job, are all ways in which monopsonists may try to expand their work force without raising wages for all workers. EVIDENCE OF MONOPSONY Boal and Ransom (1997) provide an excellent overview of theory and evidence pertaining to monopsony in the labour market. Although the empirical evidence is by no means conclusive, there does appear to be evidence of monopsony in at least some particular labour markets. Scully (1974) finds evidence of monopsony in professional baseball, especially among the star players. The dramatic increase in player salaries following the introduction in 1977 of the free agent system (which significantly increased competition among teams for players, who were no longer tied to teams) also indicates that monopsony power was important in this labour market (Hill and Spellman, 1983). However, in their study of salary determination in the National Hockey League, Jones and Walsh (1988) find only modest evidence of monopsony effects on player salaries. Thus, the evidence relating to the importance of monopsony in markets for professional athletes is not entirely conclusive. The professional sports example is also interesting in that it highlights the point that monopsony power need not be associated with low salaries just salaries that are lower than they would be in the presence of competition in the labour market. Empirical studies carried out in the United States, the United Kingdom, and Canada have also found evidence of monopsony in the labour markets for teachers (e.g., Landon and Baird, 1971; Dahlby, 1981; Currie, 1991; Merrifield, 1999), professors (e.g., Ransom, 1993), and nurses (see Exhibit 7.3 on page 224). In Canada, these labour markets are now highly unionized;thus the employer-union bargaining models developed later in this book may be more appropriate than the simple monopsony model of this chapter. In her study of the labour market for Ontario school teachers, Currie (1991) noted that both a bargaining model and a demand-supply framework are consistent with the data. The estimates associated with the demand-supply framework indicate that labour demand is very inelastic and that labour supply is slightly upward sloping, that is, that school boards possess a small amount of monopsony power in wage-setting. Further evidence for newspaper printing employees and construction workers is found in Landon (1970) and Landon and Peirce (1971). The extent to which these results can be generalized, even within the occupations where some monopsony was found, remains an open question. In addition, the extent to

18 Chapter 7: Wages and Employment in a Single Labour Market 219 which it is monopsony power, rather than other factors, that is associated with lower wages could be open to debate. It is unlikely that monopsony can be an extremely important factor in the long run. Improved communications, labour market information, and labour mobility make the isolated labour market syndrome, necessary for monopsony, unlikely at least for large numbers of workers. As these factors improve over time, monopsony should diminish. By contrast, monopsony may be quite common in the short run. Most firms can lower their offered wage and still recruit new employees, albeit perhaps at a slower pace than at a higher offered wage. Similarly, firms that offer a higher wage are likely to experience both more applicants and a higher acceptance rate of job offers. In these circumstances, firms face an upward-sloping labour supply curve in the short run, even though they may face a perfectly elastic labour supply curve over a longer horizon. This situation of dynamic monopsony is especially likely to occur in an environment of imperfect information in which workers are searching for jobs and employers are searching for employees. These and other implications of imperfect information are examined further in Chapter 18. Manning (2003) goes one step further and argues that monopsony is in fact a much more natural way of understanding how labour markets work than the traditional competitive model. Manning s view is based on two key observations. First, there are frictions in the labour market. Second, employers set wages. The first observation is clearly not controversial. Changing one s job or leaving the labour force is a costly decision that people do not undertake unless the benefits of doing so are large enough. And as we just said, it is clear that firm can set wages, at least in the short run. The difficult question is not whether these imperfections exist, but how significant they are and whether even small frictions can lead to very different policy prescriptions. Manning argues that monopsony provides a better explanation than the competitive model for a whole range of labour market phenomena such as unemployment, employers wage policies, and employment effects of the minimum wage, to mention a few examples. As we will shortly see in the case of the minimum wage, there is still much debate about whether monopsony or the competitive model best describe the way the labour market actually works. These debates aside, Manning makes an important point by showing how even small frictions can result in monopsony power, and how this can lead to very different conclusions about the desirability of different labour market policies. MINIMUM-WAGE LEGISLATION common/workplace. shtml Minimum-wage laws provide a useful illustration of the practical relevance of our theoretical knowledge of neoclassical labour markets. The estimation of the impact of minimum wages has also been an important area of research and considerable disagreement in recent years. In Canada, labour matters are usually under provincial jurisdiction; hence, each province has its own minimum-wage law. Federal labour laws cover approximately 10 percent of Canadian workers. Industries of an interprovincial or international nature for example, interprovincial transportation and telephone communication, air transport, broadcasting, shipping, and banks are under federal jurisdiction. The influence of federal laws may be larger to the extent that they serve as a model for comparable provincial legislation. The recent evolution of minimum wages in Canada and the United States is presented in Figure 7.7. For the Canadian wage we show a population-weighted average of the provincial minima, while for the U.S. we show the federal minimum. The minimum wage is shown relative to the average manufacturing wage, an important benchmark for the relative price of low-skilled labour. The graph shows that the Canadian and U.S. minimum wages track each other quite closely, and that after declining sharply between 1975 and 1985, the real value of the minimum wage has remained more or less stable over the past 20 years.

19 220 PART 3: Labour Supply and Demand Together Exhibit 7.2 Estimating the Incidence of Payroll Taxes It is less than ideal to use the theoretical formulae in order to estimate the incidence of a payroll tax: an important distinction must always be made between even well-informed simulation and estimation. In order to estimate the impact of a payroll tax, a researcher would need to provide direct evidence of a link between a change in a payroll tax and a change in employment and wages. This exercise can be difficult for a number of reasons: The tax changes may be very small, for example moving from 4 percent to 5 percent. Given the noise and underlying variation in employment and wages, it may be difficult to detect the impact of such a small change. Causality may run in both directions. The tax changes may themselves depend on the state of the labour market, being increased when wages and employment are relatively high. In this case, we might find a positive relationship between payroll taxes and employment and wages, yet it would be inappropriate to conclude that the taxes raised employment and wages. Especially combined with the previous two problems, it may be difficult to hold constant the many other factors that affect employment and wages. Jonathan Gruber (1997) explores the impact of a dramatic change in payroll taxes in Chile that avoids these problems. In 1981, Chile privatized its social security (public pension) system, moving from a payroll-tax-based, pay-as-you-go system (like the CPP/QPP) to an individual savings-based system. In 1980, the payroll tax to support the old pension system was 30 percent for employers and 12 percent for employees. In May 1981 the new system was implemented, and payroll taxes were slashed, so that they were down to 8.5 percent (for employers) by The new social security system was financed by contributions from employees. The obvious question concerning payroll taxes is whether the reduction in payroll taxes led to higher wages, higher employment, or both. To some extent, the tax reduction had to lead to higher wages, because the government legislated firms to increase nominal wages by 18 percent as a consequence of the drop in payroll taxes. However, inflation was running at 25 percent at the time, so it is not obvious that workers enjoyed any real wage increase corresponding to the payroll tax reduction. Using detailed firm-level data on employment and wages, Gruber explores the impact of the shift in tax regimes. He exploits differences in wages and employment over time (before and after 1981) and in the likely impact of the tax change across firms. His results show that virtually all of the tax reduction led to higher wages, with little effect on employment. His results are thus consistent with the "back of the envelope" calculations that imply that workers bear most of the incidence of payroll taxes and that these taxes have negligible effects on employment. The rationale behind minimum-wage laws has not always been explicit. Curbing poverty among the working poor, preventing exploitation of the unorganized nonunion sector, preventing unfair low-wage competition, and even discouraging the development of lowwage sectors have all been suggested as possible rationales. In the early days of union organizing, it is alleged, minimum-wage laws were also instituted to curb unionization: if the wage of unorganized labour could be raised through government legislation, there

20 Chapter 7: Wages and Employment in a Single Labour Market 221 Figure 7.7 This figure shows the relative value of the minimum wage to the average wage, from 1975 to 2004, for both Canada and the United States. For Canada (dashed line), the graph shows the ratio of a weighted average of provincial minimum wages to the average manufacturing wage in each year. For the United States (solid line), the graph shows the ratio of the federal minimum wage to the average manufacturing wage The Ratio of Minimum Wages to Average Wages, Canada and the United States, Canada United States Notes: For Canada, minimum wage is a population-weighted average of provincial minimum wages. The minimum-wage data are reported in Labour Canada, Labour Standards in Canada, ; and Labour Canada, Employment Standards Legislation in Canada, The average wage is the average manufacturing wage, retrieved from CANSIM. The Ratio of Minimum Wages to Average Wages, Canada and the United States, , adapted in part from CANSIM database, For the U.S., minimum wage is the mandated federal amount, while the average manufacturing wage is retrieved from CITIBASE. would be less need for unions. As is so often the case with legislation, its actual impact may be different from its intended impact, or at least it may have unintended side effects. Economic theory may be of some help in shedding light on this issue. Expected Impact: Competitive Labour Market The primary model for evaluating the impact of minimum wages is the neoclassical supply and demand model. Indeed, it is the evaluation of minimum wages that provides one of the most common illustrations of the insights provided by this simple model. Economic theory predicts that a minimum wage in a competitive labour market will have an adverse employment effect; that is, employment will be reduced relative to what it would have been in the absence of the minimum wage. This is illustrated in Figure 7.8 where W c and N c are the equilibrium wage and level of employment, respectively, in a particular competitive labour market. After the imposition of the minimum wage, W m, employers will reduce their demand for labour to N m. Thus (N c N m ) is the adverse employment effect associated with the minimum wage. In addition to the (N c N m ) workers who would not be employed because of the minimum wage, an additional (N s N c ) workers would be willing to work in this sector because the minimum wage is higher than the previous prevailing wage. Thus the queue of applicants for the reduced number of jobs is (N s N m ), with (N c N m ) representing workers laid off because of the minimum wage and (N s N c ) representing new potential recruits who are attracted to the minimum-wage jobs.