___________________________________________________________________________
Monopsony Power in the German low-wage Labour Market of
the Call-centre Industry
______________________________________________________________________________________________ ___________________
Bachelor Thesis Economics and Business Student: Fabian Mankat
Student ID: 10418156 Supervisor: Aaron Kamm
2
Table of contents
Introduction ... 3
Economics of Minimum wages ... 4
Measurement of Monopsony Power ... 8
Past Research on Monopsony Power ... 9
Hypothesis... 10
Data ... 11
General Theoretical Framework... 12
Analysis ... 14 Results... 18 Limitations ... 19 Conclusion ... 20 Bibliography... 22 Appendix ... 23
3
Introduction
In January 2015 the first industry wide minimum wage will be implemented within Germany at a wage rate of 8.50€. Large discussions have been apparent, with most critics opposing that the minimum wage will hinder economic growth and drastically decrease employment amongst those who it is originally supposed to benefit the most; the low skilled and low wage labour participants. When examining minimum wages, the employment effect usually seems to play a pivotal role in any discussion (Goldfarb, 1974; Zell, 1978; Levitan and Belous, 1979; in Brown, Gilory, & Kohen, 1982, pp. 487). However, Classic economic theory does not go in unanimous accordance about the effect of minimum wages on
employment. Besides the neoclassical model with perfect competition where minimum wages unambiguously decrease employment, several economists, such as Brown, Gilory and Kohen (1982) and Bhaskar, Manning and To (2002) have proposed models where wage floors could in fact have no effect or even increase employment in case of present monopsony power within the labour market in mind. These models abandoned the assumption of perfect competition and usually relate to monopsonistic or oligopsonistic models where individual firms face a non-perfectly elastic supply curve and posses some wage setting power. Mixed empirical results such as presented by Dickens, Machin and Manning (1999, p. 2) further strengthen the idea of variations in the effect of minimum wages as suggested by the theory.
Thus, to make an accurate statement about the possible effects of a minimum wage policy on the German low-wage industries, it seems of crucial importance to determine the market structure and possible market power of present firms. In order to bring light to the mentioned above, this paper will examine the question, to what extend monopsonistic market power is present within the German call-centre industry with regard to the labour labour
market .
The call centre industry has been chosen due to its assumed properties of low-wage and low-skilled participants and thus its representativeness as well as its considerable size indicating some importance to the whole of the German economy. Furthermore, the larger size will hopefully ease the data collection process which will be helpful in order to pursue an empirical analysis later on. In addition, the low-wage labour market is being inspected, as it is the market to be affected by the minimum wage implementation.
In order to solute the research question of this paper, a literature review will be presented first. In this section an introduction into the economics of minimum wages will be
4
provided in order to enhance the reader’s economic understanding throughout the text and clarify the importance of the research pursued. Then, sources of monopsony power, such as imperfect information, switching costs and heterogeneous work preferences as pointed out by Bhaskar, Manning and To (2002, p. 160), are linked to corresponding models and adequate measurements of monopsony power are explained, with labour supply elasticity being the most vital one. This will be followed by an overview of past findings and approaches to examine monopsony power in labour markets. Subsequently, a hypothesis about the beliefs with regard to the research question will be established. The following parts include a description of the data as well as a one of the theoretical framework. Next, the procedure of estimating the variables relevant to the research question by the use of the data gathered will be explained in the Analysis part. The obtained results will then be presented in line with several potential threats to the estimation procedure before concluding the research.
Economics of Minimum wages
The neoclassical model is generally referred to, when analysing a minimum wage implementation and it predicts a negative effect on employment. This model corresponds to a basic demand supply model regarding one perfect competitive labour market with
homogeneous workers earning a wage below the minimum wage in question (Brown, Gilory, & Kohen, 1982, pp. 488-489). Furthermore, each firm faces an individual supply curve which is perfectly elastic. This in turn implies that the firms in the market take wages as given. Each firm hires workers until its marginal product of labour equals the wage rate. The market itself will adjust the wage level until aggregate supply and demand equal one and another (Brown, Gilory, & Kohen, 1982, pp. 488-489). Introducing a minimum wage above the markets competitive level will create distortions and in turn decrease labour demand, leading to a decline in employment and thus creating a welfare loss (Card & Krueger, 1994, p. 772).
Several studies have found supporting evidence in line with this theory. Deere, Murphy and Welch (1995) investigated an overall minimum wage increase from 3.35$ to 4.25$ and its effect on the employment to population ratio in 1991. They concluded that employment has declined twelve month after the introduction of the new minimum wage policy. In addition, Kim and Taylor (1995) inspected the employment effect in retail trade of California’s minimum wage increase from 3.35$ to 4.25$ in 1988. According to their study, minimum wages did have a negative impact on employment.
5
In contrary, several empirical papers, have detected non-existing or even positive effects of minimum wages on employment. Card (1992) also examined the effect of a minimum wage raise from 3.35$ to 4.25$ in the state of California in 1988. His findings cannot provide empirical support to the conventional prediction that employment decreases when raising minimum wages. In a further study, Card and Krueger (1994) inspect the effect of a minimum wage increase from 4.25$ to 5.05$ with regard to Fast food stores across New Jersey. In accordance to Cards earlier findings about employment effects in California, the hypothesis of employment reducing due to the raise of minimum wages could not be
supported either. Dickens, Machin and Manning (1999) examined the British labour market in the time period between 1975 and 1992. They find strong evidence for the compression of earnings distribution; however, a negative employment effect due to the minimum wage policy at that time could not be found. Further research derives similar conclusions, such as Katz and Krueger, Machin and Mannig as well as Bernstein and Schmitt (Dickens, Machin, & Manning, 1999, p. 20).
An explanation for these empirical findings can be given by several models assuming some monopsony power present for the firms in the market and thus abandoning the assumption of perfect competition.
Mostly renown is the model of monopsonistic labour markets, which was firstly introduced by George Stigler in 1946 (Rebitzer & Taylor, 1995, p. 246). In the monopsonistic labour market model, only one firm is present and it possess wage setting power due to non-existent of competitors in the market. The firm in question maximizes profit by choosing labour supply with consideration of its effect on the wage level (Brown, Gilory, & Kohen, 1982, p. 489). In addition, the firm will set the marginal cost of labour equal to its demand to maximize profits. The marginal cost curve lays above the labour supply curve for any given wage, because in order to hire more workers, the firm in mind also has to increase wages for the existing workforce (Hindriks & Myles, 2013, pp. 286-288). Due to this property, the profit maximization process will lead to a lower employment level than in a competitive market. At the prevailing wage level, excess demand for labour will be present. This excess demand of labour creates the possibility of a minimum wage increasing employment.
Introducing a minimum wage takes some wage setting power away from the monopsony and when set moderately, can increase labour supply without lessens labour demand to
intensively. This in turn leads to an increase in employment (Brown, Gilory, & Kohen, 1982, p. 489).
6
Nevertheless, according to Rebitzer and Taylor (1995, p. 246) and Bhaskar, Manning and To (2002, p. 156), this model seems largely unrealistic due to its property of only one firm demanding labour in the market.
Models allowing for monopsony power; however, incorporating the possibility of several firms being active in the market, are introduced by Bhaskar, Manning and To (2002, pp. 159-170) and termed monopsonistic competition model and oligopsony model. The model of monopsonistic competition refers to an oligopsony with free entry, so that firm’s profits are driven to zero (Bhaskar, Manning, & To, 2002, p. 156). These models focus on the feature of every firm facing an individual supply curve which is upward sloping due to several potential sources of monopsony power which will be presented later on. This feature allows for the possibility of attaching some market power to each of the firms in the labour market, for any number of active ones present. The larger the monopsony market power of the firm, the smaller is the substitution of individuals in response to a decrease in wages and thus the steeper is the supply curve faced by the employer.
Manning (2003) presents’ two models which incorporate possible sources for monopsonistic market power of employers and which can be linked to the monopsonistic competition and the oligopsony model presented above.
First, a model is introduced where workers possess full information and mobility costs of job switching are abandoned. However, the firms in the market differ in respect to
dimensions of nonwage characteristics (Manning, 2003, p. 107). Boal and Ransom (1997, p. 92) identify possible nonwage characteristics, such as working conditions and location, which effects the workers daily travelling costs. In addition, Bhaskar, Manning and To (2002, p. 160) describe further potential characteristics such as the social environment at the workplace and the activity of the work itself. In case of workers possessing heterogeneous preference with regard to these nonwage characteristics, it can be argued that these characteristics will also enter a worker’s decision process (Bhaskar & To, 1999, pp. 190-192). This indicates that workers are willing to precede a trade of between their wage level and the preferred job characteristics as both factors enter ones Utility function. Thus, any firm, with preferred characteristics for a number of individuals, can be argued to be equipped with a wage setting power regarding the individuals in mind. Moreover, Bhaskar, Manning and To (2002, p. 163) point out, that the larger the degree of preferences amongst workers and the differentiation of jobs by employers, the larger is the monopsony power of the individual firms over the workers with the corresponding preferences.
7
The second type of model presented by Manning is termed searching model and considers job switching costs for workers. Furthermore, one can assume identical jobs due to simplicity without altering the conclusion (Manning, 2003). Dickens, Machin and Manning (1999, pp. 2-3) exploit these switching costs in the labour market as ‘frictions’, which make it unbeneficial for workers to change their occupation in case of a low decrease in wages. An intuitive example is presented by Manning (2003, p. 107) which illustrates that a one percent decrease in wages will not instantly lead to a loss of the entire workforce. Boal and Ransom (1997, p. 92) point out that those searching models correspond to classic differentiation models including a dynamic element. It is explained that firms for which a worker is indifferent before being hired become differentiated once the worker is employed.
Furthermore, it is stated that the importance of the established differentiation depends on the length of the wage contract. Boal and Ransom (1997, p. 93) establish additional assumptions, such as unimportant turnover costs for firms and the inability of committing to wages for workers. However, they query some assumptions established with regard to the low-wage labour market. Boal and Ransom argue that low-skilled workers wages can roughly be specified at low-cost which presumably discharges the assumption of non-commitment to wages. Furthermore, it is argued that switching costs are low considering alternative
employers are likely to be in close proximity. Thus, relating to the argumentation of Boal and Ransom, one might question the relevance of switching costs with regard to the low-wage labour market. However, the first classic differentiation model presented by Manning seems to provide valid argumentation for possible monopsony power even in the low-wage labour market this paper is investigating.
In case, monopsonistic competition or an oligopsony is present in the market, Bhaskar, Manning and To (2002, p. 162) illustrate that profit maximisation takes place in a similar matter as in Stigler monopsony model. However, the labour supply facing a particular firm further depends on the wage level set by its competitors. In particular, an increase in wages by competitors will decrease labour supply for any wage set by the firm. Similar to the
monopsony model, wages deceed the level of a competitive market outcome and excess labour demand is created. This creates the possibility of minimum wages increasing labour supply without lowering labour demand to intensively and in turn increasing employment. Bhaskar, Manning and To (2002, p. 168) term this effect as the ‘oligopsony’ effect and it is clearly employment increasing. However, due to the existence of competitors in the market, an employment decreasing effect is present. As a minimum wage is introduced, competitors also have to implement this wage floor, leading to an increase in competitor’s wages and thus
8
an inwards shift of the firms labour supply curve. This overall increase in wages may lead to some firms being forced to leave the market and thus less jobs being available. Bhaskar, Manning and To (2002, pp. 168-169) term this employment decreasing effect as the ‘exit’ effect. Considering the mentioned, the overall effect on employment is ambiguous due to the two relevant effects being opposing. Furthermore, it is argued that in such a setting the overall effect is likely to be small due to the two counterworking forces effecting employment.
Regarding the models presented, one can argue that the monopsony power and thus the wage setting power of individual firms in the market is highly linked to the employment effects of a minimum wage implementation. It was presented that in case of no monopsony power present; a minimum wage unambiguously decreases employment whereas in the presence of monopsony power, an adequately set minimum wage could possibly increase employment. Furthermore, if a minimum wage policy is implemented, it was described that its employment effect, if negative, must be fairly small in the presence of monopsony power. Thus it seems that if larger monopsony power for the firms in the market is presence, the argument of negative employment effects due to minimum wages is debilitated to some extent.
Measurement of Monopsony Power
Monopsony power itself, has been termed as the gap between marginal product of labour and wage, as this gap amounts to zero in case of perfect competition (Bhaskar, Manning, & To, 2002, pp. 162-165). The larger a firms monopsony power, the further it can set wages below the hourly productivity of its workers. As indicated earlier, this wage setting power is also largely linked to the supply elasticity the firm is facing. Hindriks and Myles (2013, pp. 286-288) illustrate, that rearranging the elasticity of labour supply and the firms profit maximization conditions yields;
This can further be manipulated to show;
. The second equation provided, yields the wage as a percentage of Marginal Product. These equations demonstrate that as the wage elasticity of labour supply raises, the gap between marginal product of labour and the wages offered decreases. It is arguable that the elasticity of supply is an easier measurable variable than marginal productivity of labour which explains why in order to measure Monopsony power in a market, most past research has used the approach of estimating labour supply elasticity’s.
9
Past Research on Monopsony Power
Most renowned research approaches examining Monopsony power estimated labour supply elasticity’s at the establishment level. The first study having done so in a static model was pursued by Sullivan in 1989 (Falch, 2010, p. 237). Wage elasticity’s to individual hospitals over a 1-year and 3-year period were estimated at 1.26 and 3.85 accordingly by the use of a two stage least square approach. Panel data from 1980 to 1985 was used, which corresponded to a national sample of hospitals in the United States (Staiger, Spetz, & Phibbs, 2010, pp. 217-219). In addition, exogenous variables, measuring the caseload of the hospitals, were used as an instrument for the level of labour demand (Bhaskar, Manning, & To, 2002, p. 171). The findings proposed the wage amounting to 79 percent of each workers marginal product and thus rather small monopsony power to have been present.
Staiger, Spetz and Phibbs (2010, pp. 214-215) criticed the paper by Sullivan and argued for two potential sources of elasticity overstating. First, it is pointed out that in the paper of Sullivan, hospital days are exogenous demand shifters. Staiger, Spetz and Phibbs questioned the instrument to be exogenous, referring to the introduction of the Medicare’s Prospective Payment System in 1984, which in turn led hospitals to reduce the days spent in hospitals. This suggests that the variation in days in hospital was actually chosen endogenously, which would have introduced a positive bias into the 2SLS method pursued by Sullivan. Secondly, Staiger, Spetz and Phibbs (2010, pp. 214-215) explained that Sullivan used the average wage of registered nurses which again could have introduced a positive bias into the 2SLS
regression. It was argued that this did not account for disproportionate hiring’s at entry levels which again led to an upwards bias due to the change in average wages understating the change in the actual wage levels. Furthermore, Staiger, Spetz and Phibbs then pursued a 2SLS regression themselves, using a legislated change in wages for registered nurses by the
Veterans administration in 1991, arguing that this change is fully exogenous. They found an elasticity of 0.1 concluding that hospitals have a larger monopsony power over their nurses than proposed by Sullivan. In addition, their findings imply that hospitals set their wages at about 9 percent of the nurse’s marginal product.
In order to avoid the problem of finding accurate instruments, Falch (2010) estimated the supply elasticity in the Norwegian teacher labour market. It is stated that the wages across schools were solely set by the government. Moreover, Falch emphasised that in contrast to earlier literature, he was able to determine the establishment with excess demand. In case of
10
fixed-effects models, the elasticity of labour supply is estimated to have been in the range of 1.0-1.9 indicating moderate to high monopsony power present.
Manning (2003) presented one of the first dynamic employment models. The main idea of Manning’s model is that a firm’s employment level represents equilibrium between the flow of workers who leave and the flow of workers who join the firm. Both these flows seem to be dependent on the wage level set by the firm in question (Manning, 2003). Manning derived that according to his model the labour elasticity to a firm is simply twice the elasticity of the quit rate with respect to wages. According to Ashenfelter, Farber and Ransom (2010, p. 206), these estimates represent long-run elasticity’s as they are founded on equilibrium
conditions. Several economists have adopted the proposed model of Manning in order to determine an expressing elasticity going beyond the static analysis as described earlier. Hirsch, Schank and Schnabel have estimated elasticity’s from a large data set corresponding to the German labour market in 2010 (Ashenfelter, Farber, & Ransom, 2010, p. 207). They examined a wide variety of jobs and workers and found average elasticity’s in the range of 2 to 4 indicating significant present monopsony power. Furthermore, Ransom and Sims (2010) investigated the monopsony power in the Missouri school districts by having used the
dynamic approach introduced by Manning. They estimate elasticity’s of about 3.7 suggesting that the regarded school districts possessed some market power with regard to teacher
supplying their work.
Hypothesis
Regarding the literature presented, the presence of nonwage characteristics and
preferences across jobs and workers accordingly seems plausible with regard to the call centre industry. However, the presence of ‘frictions’ in the labour market, as presented by Dickens, Machin and Manning (1999, pp. 2-3), is believed to be irrelevant considering the German call centre industry by reason of Boal and Ransom’s argumentation (1997, p. 93). The past
empirical findings in mind, in combination with the theoretical indications, the presence of monopsony power in the German call centre industry seems probable; however, whether it is of considerable size is questionable. The beliefs about the empirical findings will not be set into a formal hypothesis, as no clear-cut value for elasticity indicates the presence of a monopsony. The elasticity’s estimated rather serve as a measure of the corresponding degree of monopsony power in the market.
11
Data
Unfortunately, the data to conduct elasticity estimation at the establishment level, as most past research did and as it was planned originally in this paper, was not available in sufficient form. Therefore, as an alternative to estimate prestigious wage elasticity’s each call centre is facing, a German wide survey, aimed at workers employed in those institutions at a low wage leveland full time employment, was conducted. The data obtained from the survey, in combination with a later derived model, will serve to estimate the average supply elasticity facing a call centre. Furthermore, the established survey was send to human resource
departments of all call centres which contact information were accessible, to reach the aimed at workers. Clear instructions about who needed to fill in the information were provided. The precise criteria for the employees investigated are described above; however whether only the targeted group did participate at the survey cannot be secured, considering that much of the supervision responsibility was handed towards the firms themselves. The survey was designed to obtain a self-assessment of each workers trade-off between wages and working hours as well as their substitution away from employees, given a change in the job conditions. These self-assessments in combination with the general information about each worker, which were also demanded in the survey, will serve to estimate wage elasticity’s in order to answer the research question. General information demanded in order to pursue the empirical analysis included weekly working hours and yearly paid holidays. The first self-assessing question introduced asked, how much yearly holidays a worker would be willing to give up in exchange for a ten per cent increase in their wage level. The trade of between wages and holidays in contrary to working hours was consciously chosen in order to create a somewhat more imaginable and realistic setting for the workers surveyed. Secondly, it was asked on how much of their current wage level in terms of per cent workers would be willing to
relinquish before considering a termination of their working contract at their current firm. The third question introduced, is similar to the second, however now it was asked on how much yearly holidays workers would be willing to renounce before considering a termination of their current working contract. However, this question will be disregarded in the later
empirical analysis. In total, 48 different individuals where surveyed. The corresponding data relevant to the analysis conducted and the survey given out can be regarded in the Appendix part.
12
General Theoretical Framework
In order to estimate the labour supply elasticity’s relevant, a general model will be introduced, which in addition to the information given will serve to estimate the elasticity of supply faced by the call centres. The general framework of the model goes as follows;
The total supply elasticity facing a call centre at any wage level, termed aggregate supply elasticity, relates to a measure of changes in labour supply due to changes in wages. In this framework, two different effects of wages on labour supply will be present. First,
changing wages will lead to the current employed workers changing the amount of working hours they are willing supply. Second, a change in wages will also lead to job switching behaviour by employees across firms. Thus, the effect on labour supply to the individual firm due to a change in wages corresponds to the changes in labour supply by the employed workforce; termed the ‘employed workforce effect’, and secondly, the change in employees willing to work at the corresponding firm; termed ‘entry/exit effect’.
The general form of wage elasticity is given by economic theory and holds on the individual as well as the aggregate level. It is defined as the percentage change of labour supply corresponding to a percentage change in wages. Thus, formally written the following general formula is obtained;
.
In order to examine individuals labour supply elasticity, the following individual labour supply function is assumed for each worker at the current occupation;
and relate to Labour supply and wage level correspondingly. Furthermore, the parameters and can vary across workers to capture the individuality of those. Wages by competitors are assumed to be fixed and are not adjustable throughout the models corresponds to the individual cut-off wage, the current firm can set and below which employees leave the firm due to preferred outer working options. This cut-off wage inhabits the wage rate of competitors as well as non-wage
characteristics across jobs and is further determined by those. In addition, it will be assumed that is independent of all parameters contained within the individuals supply function throughout the analysis and thus the supply function only relates to the trade-off of labour supplied and free time by each individual. The specific values for the labour supply functions do not matter in order to pursue the analysis, rather than the theoretical implications it offers. The proposed supply curve offers two convenient features for the analysis pursued later on.
13
First, when deriving the elasticity for the general form of the supply curve, one can notice the presence of constant supply elasticity’s being equal to the parameter for the individuals in the market;
This feature seems largely helpful in the later calculation process as, allowing for constant elasticity’s, enables one to manipulate the elasticity fractions obtained by the analysis in a simpler matter than otherwise, as the ratio of percentage change in Labour supply
corresponding to a percentage change in wages will remain constant at any level. However, in the established framework there is one exception present. Considering that below the cut-off wage a worker retracts his whole labour supply, the individual’s elasticity in exactly his cut-off wage must be equal to infinity for a reduction in wages. It should be noted that this is due to the introduction of cut-off wages which leads to discontinuous individual supply curves.
The second convenient feature corresponds to the fact that when aggregating the individuals Supply curves, it can be derived that the elasticity of the aggregated labour supply is simply the weighted average of the individual elasticity’s;
This feature significantly reduces complexity when estimating the elasticity of the existing workforce as well as the aggregate supply elasticity.
To determine an aggregate supply for a firm given any wage in the established model, the individuals labour supplies must simply be summed up for all workers with a cut-off wage equalling or exceeding the prevailing wage rate . Thus;
The expected aggregate Labour supply could possibly further be approximated by;
with relating to the to the average individual’s supply curve and describing the exit and entry behaviour of individuals by incorporating the distribution of the cut-off wages across the population of workers. However, in order to pursue the later analysis this will not be of importance.
14
It is worth noticing, that even though constant individual elasticity’s are given, the elasticity of the aggregate supply facing a firm must not be. This can be simply argued due to the entry and exit behaviour of employees. It was stated that each individual’s elasticity is constant. Considering that each individual’s cut-off wage is independent of the workers supply curves and the containing parameters; the expected average elasticity due to the ‘employed workforce effect’ is independent in wages and thus constant in this framework. When introducing entry and exit behaviour, the aggregate elasticity further changes due to the change in the available workforce. It can simply be illustrated that entry and exit can occur in such a way that aggregate elasticity is not constant, by assuming for example a uniform distribution of the cut-off wage across the population.
In addition to the described above, the framework under which the analysis will be conducted is subject to several further assumptions. Firstly, a fixed yearly income; , for the surveyed workers is believed. As mentioned in the survey and data section, the workers questioned are also believed to be full time employed. In addition, it is assumed that workers split their weekly working hours evenly over five days. Assuming five working days per week seems plausible as the survey was conducted for full time employed workers and this seems to be the usual implemented set up in reality. Furthermore, this will ease the calculations later on. Moreover, workers must be able to choose their yearly supplied working hours somewhat freely for example via the weekly working hours to be supplied or the yearly holidays to be taken.
Analysis
The analysis following is constructed on the basis of the established theoretical framework, using the data obtained from the survey. The two employment changing effects introduced earlier, ‘entry/exit effect’ and ‘employed workforce effect’, will be regarded individually until they are set into one framework in order to estimate the relevant elasticity.
The first question introduced to the low-wage labour participants in the call centre industry asked how much yearly holidays they would be willing to give up in exchange for a ten per cent increase in their wage level. Regarding this question, the individuals wage
elasticity’s can be determined. Having obtained all individuals elasticity’s, the wage effect on aggregate supply facing the firm due to the ‘employed workforce effect’ can be established.
The daily working hours per individual; correspond to the weekly hours given by survey, divided by five; This measure, than further matches the amount of working
15
hours corresponding to one paid holiday of the worker. Yearly supplied working hours of the individual, , than amount to the weeks per year times the weekly working hours. However, the working hours per holiday times the amounts of holidays available to the individual in one year have to be deducted. Following formula is obtained;
.
A percentage change of yearly Labour hours supplied by the individual then corresponds to;
.
In order to measure the elasticity using the data provided, the elasticity function is
adjusted as follows;
Due to the structure of the question in the survey, the percentage change in Yearly income per individual worker is given at ten percent; .
The percentage change in yearly supplied working hours by the workers will be given by the formula established earlier in addition to the submitted answer for question one.
Given the obtained values and the formula for the elasticity; elasticity’s for each
individual, serving to determine the aggregate elasticity; and other statistical variables using statistical software’s, will be calculated. To estimate the aggregate elasticity the individuals elasticity’s will be weighted by the corresponding labour supply, rather than just calculating the normal average elasticity. Considering the link between elasticity and Labour supply as well as the relation of aggregate elasticity and individual’s elasticity’s established in the theoretical framework, this seems to provide a more accurate measure when estimating the aggregate supply elasticity arising from the ‘employed workforce effect’.
As mentioned earlier, this estimate of elasticity alone does not incorporate the job entry behaviour by employees. The analysis solely focuses on the existing workforce, rather than the whole potential labour supply including workers outside the firm. Thus when regarding the supply the firm is facing, the elasticity obtained is potentially underestimated. However, it should be considered that in case of a market faced by a monopsony one can argue that the elasticity estimated could be valid in order to determine the size of monopsony power. However, considering the interest rather lies in determining the presence of
monopsony power, this measure does not yield any significant conclusions yet.
To include entry and exit behaviour by workers, the analysis has to be extended. The data obtained in the second question, which asked on how much of their current wage level in
16
terms of per cent workers would be willing to relinquish before considering a termination of their working contract at their current firm, can be used to resolve the ‘entry/exit effect’ at the prevailing wage rate. Workers remaining employed at the firm in case of an infinitely small decrease in wages, and thus corresponding to a submitted value larger than zero, have a non-infinite elasticity at that wage level and consequently their decrease in labour supply
corresponds to the ‘employed workforce effect’. However, those workers having submitted a value of zero will leave the firm instantly when wages will decrease. Having the theoretical framework in mind, those workers must have a cut-off wage equalling and an elasticity being infinite in the prevailing wage rate. Therefore, the individuals in mind determine the labour supply decreases due to the ‘entry/exit effect’. As mentioned earlier, it is believed that those workers having submitted a value of zero, posses a smaller elasticity than infinity for larger wage rates. However, because they are employed at the wage level of their cut-off
point; their elasticity’s correspond to infinity at this wage level when decreasing wages.
In general, the decrease in labour supply due to the ‘entry/exit effect’ would amount to the sum of the hours by those workers whose Cut-off wage equals the current employed wage rate; To obtain a population estimate of the elasticity relating to the above, the proportion of workers being employed at their cut-off wage; will be estimated. This proportion will then be multiplied with the aggregate supplied hours to determine an estimate of the labour supply decrease due individuals leaving the firm; . Considering that the individuals cut off wages are assumed to be independent of their corresponding Labour supply curves, using just the proportion of individuals leaving their occupation, rather than accounting for the specific labour hour decrease corresponding to each exiting worker, seems to provide a more accurate
approximation of the general exit effect due to wages. The multiplication above, divided by the aggregate supplied labour hours corresponds to the estimate of the percentage change in
labour supplied; . In order to obtain the elasticity due to the ‘entry/exit effect’, it has to be corrected for percentages; . To
estimate the proportion of individuals; the analysis will be conducted in accordance with the Sample proportion approximation approach introduced by Keller (pp. 416-419). The population estimate for the proportion simply relates to the number of individuals having submitted the value of zero divided by the sum of individuals having participated at the
17
survey; The corresponding variance is determined as follows; . The parameter resembles the population proportion. Considering, that is unknown; the sample proportion will be used to receive an estimate of the variance; . The
parameters estimated will further serve to establish a 95% Confidence Interval for . To determine the relevant elasticity when regarding the supply curve each call centre is facing, the two earlier estimated measures will be combined. In order to pursue the analysis, the total labour supplied in the starting point will be split up into two components, being first the labour supplied by those workers leaving the firm in case of any wage decrease;
and second the labour supplied by those workers staying at the firm in case of a sufficiently small wage decrease; . The estimated change in Labour supply to the individual firm due to a wage decrease corresponds to the decrease in both those components;
Thus overall elasticity yields;
As all workers corresponding to will instantly leave the firm in case of a wage decrease the change in correspondence yields as well. Thus, the change in as a percentage of the overall Labour supply gives;
The change in relates to the change in Labour supply by the individuals remaining at the firm. Rearranging hows;
The term,
, now corresponds to the percentage change in labour supply due to the ‘employed
workforce effect’. Thus, the change of labour supply by those workers remaining at the firm as a percentage of the labour in the starting point yields;
Plugging the findings into the original function of overall elasticity in terms of and presents;
18
This formula seems to go along with intuitive thinking, as the overall labour supply clearly decreases by the people exiting the firm. In addition, labour will be decreased by each
individual remaining at the firm and thus the ‘employed workforce effect’ must be somewhat related to the ‘exit/entry effect’ when determining the aggregated change in Labour supply. The estimate is determined by substituting the approximations of and into the established formula. Considering the complexity of constructing an exact 95% Confidence Interval for , due to the mixed properties of its estimated parameters, suchlike will not be established. Nevertheless, a highly Probable Interval for will be introduced, by simply substituting the lower and upper bound limits of the 95% Confidence Intervals for and into The value obtained by substituting in the lower bound limits of and evidently correlates to the lower bound limit of the Probable Interval, while the value obtained by upper bound limits relates to the upper bound. Bearing in mind that no formal hypothesis was introduced, the exact values for a 95% Confidence Interval do not seem to alter the conclusions significantly when compared with any probable Interval for .
Results
Following results are obtained when pursuing the analysis in order to estimate the elasticity arising due to the employed workforce effect;
Constructing a 95% Confidence Interval for the estimated effect yields;
The value obtained for indicates that on average individuals being employed at a firm are willing to supply 0.83 percent more labour, if wages increase by one percent. The results further suggest that workers do not drastically correspond to changes in wages as the elasticity for the workforce being employed lies between 0.54 and 1.29 and thus appears to be modest.
Ei 48 1811 .8362515 1.011817 0 4.638321 Variable Obs Weight Mean Std. Dev. Min Max
19
Estimating the proportion as described in the analysis section yields following;
The results propose a large fraction of individuals leaving the firm in case of a wage decrease, being 25%. The corresponding elasticity arising due to entry exit behaviour of individuals relates to 25. This estimate appears to be notably large. However, regarding the Confidence Interval, a large dispersion is apparent, yielding the possibility of both noteworthy larger or smaller elasticity and thus corresponding entry exit behaviour by employees.
Given the values and , the estimation of yields the following;
When proceeding with the analysis in order to achieve the Probable Interval for following results are obtained;
The elasticity results relating to the aggregated supply curve suggest monopsony power present, but being rather small. The estimated average, predicts wages amounting to about 96% of the workers marginal product, with a probable Interval for being between 93% and 98%. These measures suggest a highly competitive environment as compared with the estimated elasticity’s by past researchers, but nevertheless some present monopsony power introducing the possibility of minimum wages increasing efficiency in the market. However, considering wages in the market being close to those arising in a competitive market, efficiency could be argued to possibly only be increased little when introducing an optimal minimum wage.
Limitations
The validity of the obtained results is possibly exposed to several threats. First, due to several reasons, it is uncertain whether the conducted survey yields legitimated data to pursue the analysis. The survey was created by an unschooled person and thus it is likely not to be designed in an ideal way. Besides possible apparent ankers creating some biasness concerning the surveyed individuals, unclarity about the questions in the survey might be present. One possible unclarity threatening the validity is the phrasing of the second question in the survey. It should have been explicitly stated that labour supply by surveyed individual is assumed to always be adjustable in correspondence to any wage change by the employer. Furthermore, as
48 .25 .0625 .1363723 .3959593 Variable Obs Mean Std. Err. [95% Conf. Interval] Binomial Exact
20
described in the data section, it is uncertain whether only the workers relevant to the research question did participate at the survey, considering that much of the supervision responsibility was handed towards the Call-Centres. In case elasticity’s differ across workers depending on their wage level, including workers into the survey who are not affected by the minimum wage implementation could also distort the results. In addition, whether the elasticity’s approximated via the self assessment by the individuals relate to the elasticity’s in reality, is questionable. It could be argued that individuals are overconfident about their outer working possibilities and thus overstate their exit behaviour in correspondence to a wage decrease. This in turn would lead to an overestimation of the aggregated elasticity.
Besides possible threats concerning the data obtained, the theoretical framework might be exposed to further limitations. The analysis conducted only seems valid in case workers can individually choose their labour supply, which in turn implies firms not to have a standardized contract across employees. Moreover, in case the assumption of individual’s constant elasticity’s of labour supply is invalid the estimated elasticity relating to the ‘employed workforce effect’ only serves as an approximation rather than an accurate
calculation. In addition, the assumption of the independence of individuals supply curves and their corresponding cut-off wage is possibly invalid. For example, one might argue the labour supply elasticity for an individual worker to be closely related to the level of the workers pleasantness with regard to his occupation. Thus, a small individual elasticity for workers might indicate the worker to relatively like his occupation. Furthermore, small values of the cut-off wage seem to indicate similar, as the workers accept lower wages until they move to competitors. Considering the above, the aggregated elasticity could be argued to be
overestimated, as the workers with the highest elasticity are likely to leave their occupation the earliest.
Conclusion
This paper examined the monopsony power within the German labour market of the call-centre industry in order to determine the possible effects of the industry wide minimum wage implementation. In order to resolute the research question, the average elasticity facing the individuals firms was estimated using data from a conducted survey in combination with an introduced theoretical framework. The probable Interval for such elasticity was estimated to be between 14.11 and 40.28 corresponding to wages being between 93% and 98% of the workers marginal product of labour. These results indicate a highly competitive environment,
21
but nevertheless some existing monopsony power introducing the possibility of an optimal minimum wage improving efficiency slightly. Thus, the argument of a negative employment effect due to minimum wage in the german Call-Center industry seems debilitated to some extent.
However, several threats described in the limitations section concerning the estimation procedure possibly introduce some positive bias into the elasticity approximation.
Furthermore, a large dispersion is apparent when regarding the probable interval for the elasticity estimated. Thus, in order to get a more expressive result, the analysis conducted should be repeated with sufficient data. Furthermore, to obtain valid and unbiased data, the survey conducted should be redesigned by someone with expertise in this field. As an alternative, the elasticity in question could be re-estimated at the establishment level to exclude any problems arising due to the introduction of surveys.
It is highly notable that this paper solely focused on the determination of monopsony power within the labour market for call-centres and thus the possibility of minimum wages increasing employment and efficiency. Considering the theoretical framework, even though some degree of monopsony power was found in the market examined, it is not determined whether minimum wages in fact can increase employment but rather that the possibility for doing so exists. This question should be resolved within further research regarding the minimum wage implementation. In addition, further research should examine whether the minimum wage to be implemented is likely to be at an optimal rate in order to make further judgements about its effects on employment.
22
Bibliography
Ashenfelter, O. C., Farber, H., & Ransom, M. R. (2010). Labor Market Monopsony. Journal of Labor
Economics , 28 (2), pp. 203-210.
Bhaskar, V., & To, T. (1999). Wages for Ronald McDonald Monopsonies: A Theory of Monopsonistic Competition. The Economic Journal , 109 (455), pp. 190-203.
Bhaskar, V., Manning, A., & To, T. (2002). Oligopsony and Monopsonistic Competition in Labor Markets. Journal of Economic Perspectives , 16 (2), pp. 155-174.
Boal, W. M., & Ransom, M. R. (1997). Monopsony in the Labor Market. Journal of Economic
Literature , 35 (1), pp. 86-112.
Brown, C., Gilory, C., & Kohen, A. (1982). The effect of the minimum wage on Employment and Unemployment. Journal of Economic Literature , 20 (2), pp. 487-528.
Brueckner, J. K., Thisse, J.-F., & Zenou, Y. (2002). Local Labor Markets, Job Matching, and Urban Location. International Economic Review , 43 (1), pp. 155-171.
Card, D. (1992). Do Minimum Wages Reduce Employment? A Case Study of California, 1987-89.
Industrial and Labor Relations Review , 46 (1), pp. 38-54.
Card, D., & Krueger, A. B. (1994). Minimum Wages and Employment: A Case Study of the Fast-Food Industry in New Jersey and Pennsylvania. The American Economic Review , 84 (4), pp. 772-793. Deere, D., Murphy, K. M., & Welch, F. (1995). Employment and the 1990-1991 Minimum-Wage Hike.
The American Economic Review , 85 (2), pp. 232-237.
Dickens, R., Machin, S., & Manning, A. (1999). The Effects of Minimum Wages on Empoyment: Theory and Evidence from Britain. Journal of Labor Economics , 17 (1 ), pp. 1-22.
Falch, T. (2010). The Elasticity of Labor Supply at the Establishment Level. Journal of Labor Economics
, 28 (2), pp. 237-266.
Hindriks, J., & Myles, G. D. (2013). Intermediate Public Economics (2nd ed.). Cambridge, Massachusetts: The MIT Press.
Kim, T., & Taylor, L. J. (1995). The employment Effect in Retail Trade of California`s 1988 Minimum Wage Increase. Journal of Business & Economic Statistics , 13 (2), pp. 175-182.
Manning, A. (2003). The real thin theory: monopsony in modern labour markets. Labour Economics ,
10 (2), S. 105-131.
Ransom, M. R., & Sims, D. P. (2010). Estimating the Firm´s Labor Supply Curve in a "New Monopsony" Framework: Schoolteachers in Missouri. Journal of Labor Economics , 28 (2), pp. 331-355.
Rebitzer, J. B., & Taylor, L. J. (1995). The consequences of minimum wage laws Some new theoretical ideas. Journal of Public Economics , 56 (2), pp. 245-255.
23
Staiger, D. O., Spetz, J. S., & Phibbs, C. S. (2010). Is There Monopsony in the Labor Market? Evidence from a Natural Experiment. Journal of Labor Economics , 28 (2), pp. 211-236.
Appendix Data
weekly working hours
Yearly
holidays AQ1 AQ2
40 26 2 10 36 24 10 10 40 24 5 9 40 24 5 10 40 23 1 10 38 24 0 10 40 25 0 10 40 24 0 10 40 25 0 1,5 30 27 1 10 40 26 3 10 40 26 0 5 40 35 0 0 40 28 0 10 40 25 5 10 37,5 30 3 5 45 27 0 10 40 30 5 10 25 28 0 0 1 1 1 1 30 26 1 1 40 23 0 3 40 28 1 0 40 25 0 5 40 35 5 0 40 30 2 5 40 27 0 0 20 30 0 8 40 30 5 10 60 30 0 0 40 31 1 5 40 33 0 0 40 28 0 0 60 30 0 0 40 30 0 0 40 28 1 5 40 24 4 0 4 2 12 12 40 22 2 2 40 26 2 1 39 30 3 20 40 15 5 20 38,5 28 2 15 40 30 0 0 39,5 30 7 25 35 15 5 5
24
39,5 27 2 10
33 10 3 20
Survey
Umfrage: Lohnelastizität in der Call-Center Branche
Dieser Fragebogen ist Teil meiner Bachelorarbeit verfasst an der Universität van Amsterdam und dient als Grundlage für den Praxis-Teil der Arbeit. Die Bearbeitungsdauer sollte nicht länger als 3 Minuten dauern. Vielen Dank für Ihre Unterstützung!
Fabian Mankat
Bitte beantworten jede der folgenden Fragen mit einer konkreten Angabe. Des Weiteren möchte ich Sie darauf hinweisen, dass diese Umfrage streng vertraulich und anonym behandelt wird.
Geschlecht Alter Wöchentliche Arbeitsstunden Jährlicher Urlaubstage
1. Auf wie viele jährliche Urlaubstage würden Sie verzichten, im Gegenzug zu einer 10% Lohnsatzerhöhung?
2. Auf wie viel Prozent Ihres jetzigen Lohnsatzes würden Sie verzichten, bevor Sie eine Kündigung Ihres jetzigen Arbeitsvertrags in Betracht ziehen würden?
3. Auf wie viele jährliche Urlaubstage würden Sie verzichten, bevor Sie eine Kündigung Ihres jetzigen Arbeitsvertrags in Betracht ziehen würden?
