• No results found

On the causes of unemployment

N/A
N/A
Protected

Academic year: 2021

Share "On the causes of unemployment"

Copied!
32
0
0

Bezig met laden.... (Bekijk nu de volledige tekst)

Hele tekst

(1)

On the causes of unemployment

Masterthesis: definitive version

“In the long run we are all dead.”

Erik-Jan Hummel

August 13

th

2010

Tutor: Robbert Maseland

(2)

Content

Introduction 3

I The Keynesian theory 4

1.1 What is unemployment? 4

1.2 How does the labour market affect unemployment? 6 1.3 How do financial markets affect unemployment? 6

1.4 Uncertainty 8

1.5 Subconclusion 9

II The New Classical theory 10

2.1 New Classical theory about the labour market 10

2.2 The NAIRU story 11

2.3 Market rigidities 13

2.4 Subconclusion 14

III Empirical part 15

3.1 Summary theoretical part 15

3.2 Methodology 16

3.3 The variables 18

3.4 Data 19

3.5 Tests 22

IV Conclusions and recommendations 29

4.1 Conclusions 29

4.2 Recommendations 30

(3)

Introduction

The subject of the causes of unemployment has been under a constant debate. Between the Great Depression and the seventies, Keynesian theory offered the main explanation for unemployment. After the oil-price crisis and the stagflation or arguably when Reagan and Tatcher took office, the Keynesian theory lost its dominant position to the New Classical theory.

In the current situation in the world, there is much debate about how to deal with recessions. Should we stimulate the economy, as Keynes advocates and the United States seems to have planned? Or should we base ourselves on Friedman, who claims that stimulation of the economy will not lead to increased consumption, which seems to be the most persistent thought in Europe?

This thesis aims at one part of dealing with recessions: unemployment. It discovers two dominant streams of theory: the Keynesian and the New Classical and tries to both identify and discuss the main differences between the theories, and their assumptions and the claims they make about the causes of unemployment. The concrete aim of this thesis is to arrive at a policy-advice based on theory and empiric results.

(4)

I The Keynesian theory 1.1 What is unemployment?

Unemployment simply means: the existence of people belonging to the workforce who do not have jobs. However, unemployment can be divided into different groups. A main difference between Keynes and the New Classicals is the status of one of these groups: involuntary unemployment.

In the second chapter of his General theory, Keynes had a dialogue with the classical economists such as Professor Pigou about employment. They agreed that there can be frictional unemployment. This means that there are time-lags between jobs or that the entrepreneur has miscalculated the demand. There can also be voluntary unemployment, which is caused by refusal to work, inability, legislation or collective bargaining. However, apart from this agreement, they disagree about the existence of involuntary unemployment. This difference is caused by the difference of opinion about the power of markets. As Stockhammer describes it:

“Unlike what he [Keynes] labelled ‘classical’ theory, he rejected the self adjusting ability of the market economy for several reasons. First, because there is no feedback from unemployment that guarantees that wages will fall. On the labour market money wages, that is nominal wages, are set by employers or in negotiations with labour unions. An increase in unemployment will weaken labour unions and thus reduce money wages, but whether it also decreases real wages depends on prices. If, as is likely in a recession, firms cut prices, real wages need not fall. Second, an important part of effective demand depends on investment decisions. These involve decisions regarding the distant future, about which rational expectations often cannot be made. In a fundamental sense the future is open and therefore uncertain. In such a context investment decisions not only depend on rational factors but also on investors sentiment, or what Keynes famously labelled as ‘animal spirit’ of investors. Third, financial markets, namely stock markets, are prone to mood swings, because investors are trying to anticipate the public’s evaluation of the decisions. Thus financial markets will be, at least at times, a source of destabilization.”

(Stockhammer 2004: 12).

(5)

Let us now return to the notion of involuntary unemployment, what it means and how it is caused.

Keynes defines involuntary unemployment as follows:

“Men are involuntarily unemployed if, in the event of a small rise in price of wage-goods relatively to the

money-wage, both the aggregate supply of labour is willing to work for the current money-wage and the aggregate demand for it at that wage would be greater than the existing volume of employment.”

(Keynes 1946: 15).

This means that the supply of labour will increase if real wages increase, but demand will lag behind and will not catch up with supply. Very concrete this means that there are more people willing to work than jobs offered. This is very distinct from the thought of monetarists that markets will always clear.

The main difference between Keynes and the classical economists is the acceptance of Say’s law. This law claims that supply will create its own demand. (Keynes 1946: 25) It must be said that Say thought of his law in times of scarcity, so that indeed everything that was produced created its own demand. This is to say that economic laws are influenced by other economic factors such as abundance and scarcity of particular goods. When the times wherein Say had lived and wherein Keynes has lived are being compared, then goods are scarce in Say’s time and abundant in Keynes’ time. Therefore it was clear to Keynes that supply would not always create its own demand, and to explain this he uses the concept “the propensity to consume.”

This propensity to consume is the part of the increase in income people spent on consumption and according to Keynes this will decline as income increases:

“Men are disposed, as a rule and on the average, to increase their consumption as their income increases, but not by as much as the increase from their income.”

(Keynes 1946: 96)

(6)

“For consumers will spend less than the increase in aggregate supply price when employment is increased, the increased employment will prove unprofitable unless there is an increase in investment to fill the gap.” (Keynes 1946: 98)

1.2 How does the labour market affect unemployment?

The basis of a labour market is relatively simple. People offer their labour and firms want people to work for them. Firms offer money for this labour and people buy products with this wage. This is again revenue for the firm which can be spent on wages. If revenues are low, people can be fired and if revenues are high, more labour is demanded for. The trouble begins when we make a difference between real and nominal wages. Real wages are the values of the wages compared to the prices of products. So if inflation rises and nominal wages stay the same, the real wages decrease.

Keynes has claimed that since employers and employees negotiate nominal wages and not real wages, there will be a trade-off between inflation and employment. This trade-off gets its form in the Phillips curve.

The Phillips curve shows that inflation and unemployment are negatively related (Islam et al. 2003). The curve was created by A.W. Philips in 1958. The basic idea is that, in a situation where the demand for labour is higher than the supply of labour, aggregate wages are expected to rise. Additional workers are hired and thus the unemployment decreases. The rising wages will lead to inflation as firms have higher costs and employees have more wages to spent (Islam et al. 2003).

1.3 How do financial markets affect unemployment?

Minsky claimed that financial instability is simply part of the capitalistic system. In times of prosperity economic agents will take more risks with riskier products, also known as Ponzi products and thereby putting the total system in danger (Minsky 1992). Note that human behaviour, in this case taking more risks when economic times are well, is the crucial part in this hypothesis.

The human behaviour that is needed to be a good investor is expressed by Keynes in another humorous metaphor:

(7)

the prize being awarded to the competitor whose choice most nearly corresponds to the average preferences of the competitors as a whole; so that each competitor has to pick, not those faces which he himself finds prettiest, but those he thinks likiest to catch the fancy of the other competitors, all of whom are looking at the problem from the same point of view. It is not a case of choosing those which, to the best of one’s judgement, are the prettiest, nor even those which average opinion genuinely thinks the prettiest. We have reached the third degree where we devote out intelligences to anticipating what average opinion expects the average opinion to be. And there are some, I believe, who practise the fourth, fifth and higher degrees.” (Keynes 1946: 156).

So the information investors want to have is what the fancy will be of the aggregate of all investors, which is the summation of all investors’ fancy. If single investors base their choices on caprice, it is at least very hard to make a good prediction. If, however, most choices are based on habit, on market prices, on predictions made by institutions, then predictions can be, for some part, reliable.1 Keynes would claim no accurate predictions can be made, because the future is uncertain.

Nevertheless, there is still a debate going on about the issue. Diamond and Rajan claim that financial instability is not part of the system itself, but rather an externality (Diamond & Rajan 2009). Lourenço claims that the investment volatility put forward by Keynes and Minsky has been overstated (Wray 2008). He says for example about Minsky:

“In this model cyclical reversions remain basically inexplicable expect by shocks – which is not surprising, given the known resistance of the long-term state of expectations to analytical treatment. However, it seems that the excessive dependence of these long-term expectations in past results – especially the excessive sharing of the ‘general state of confidence in business’- contributes to an overstatement of volatility in the behaviour of some economic variables, mainly during debt deflation and/or investment boom periods.” (Wray 2008: 47).

The point of discussion is the possibility to give the long-term state of expectations a ‘analytical treatment.’ It is clear that Lourenço thinks this is possible and the volatility in an investment boom is not so big. Keynes, as mentioned before, thinks long-term expectations are by definition uncertain.

The thought that the future is fundamentally unknown is also expressed by Borio. He claims that since the future cannot be predicted, financial institutions should take less risks, so he wishes for greater transparency and accountability (Borio 2008).2 Borio also

1 To complicate matters: whether a prediction is right can only be tested in the future, which is impossible.

What can be done is base the strength of the current prediction on the strength of foregone predictions.

2

(8)

comments the coming to be of a bubble. Bubbles are in fact asset prices that are over valuated, as for instance the house prices in the US were in 2007. Borio therefore advises to ‘work against the cycle’ (Borio 2008: 19-20). This is to prevent that in good times, when people want to take risks and loan more money and thus expanding the bubble. The question at hand however, is not so much how financial instabilities are caused, but whether financial instabilities have an effect on unemployment. In Keynes’ mathematics, it is clear that in an economic downturn there is a lack of demand and this should be compensated by an increase in investment, to increase the effective demand. So when economic policy is pro-cyclical, the instability will have, in this theory, a great impact on the real economy and on unemployment. The New Classicals do not think financial instability will lead to bigger unemployment, because the labour market is, according to them, self adjusting.

The empirics for now support the idea that an economic downturn, measured in a decline in house prices, will lead to an increase in unemployment (Cleasens et al).

1.4 Uncertainty

Based on the sections above, it seems fair to say that the main point of Keynes is the way we perceive the future. According to Keynes, the future is fundamentally uncertain and therefore all wages can only be nominal wages and not real wages. We are simply not able to predict the inflation accurately. According to Robert Skidelsky, Keynes was right and that the future is indeed fundamentally uncertain (Skidelsky 2009). Or to put it in Keynes own words:

“Is our expectation of rain when we start out for a walk, always more likely than not, or less likely than not, or less likely than not, or as likely as not? I am prepared to agree that on some occasions none of these alternatives hold, and that it will be an arbitrary matter to decide for or against the umbrella. If the barometer is high, but the clouds are black, it is not always necessary that one should prevail over another in our minds, or even that we should balance them – though it would be rational to allow caprice to determine us and waste no time on the debate.”

(Keynes in Skidelsky 2009: 86).

(9)

As we have seen before, the financial market depends on the fancy of investors and when these fancies are based of caprice, instead of on habit and prices, the future will become uncertain indeed.

Skidelsky even stretches his point so far that the idea that the future can be reduced into calculable risks is the core cause for the credit crisis that started in 2007. His claim is that in the risk models extreme outliers are not considered. According to Shiller, this is the reason why the perfect market theory has lost its appeal (Shiller 2003).

Therefore Keynes thought there is a trade-off between inflation and unemployment.

1.5 Subconclusion

According to Keynes unemployment is caused by a lack of demand for labour, which is caused by a lack of effective demand. Effective demand is the sum of consumption and investment. In the situation that the nominal wages increase, the supply for labour increases and more people will be hired. Since firms need more money to pay the higher wages and employees have more money to spent, the situation will lead to inflation. This trade-off between unemployment and inflation is expressed in the Phillips curve.3

A problem however is the propensity to spent. This means that the employees are inclined to spent the increase of their wages, but not the total increase. The problem is then that firms do not receive enough revenue to compensate for the higher wages. This leads to increased unemployment.

It is useful to note that Keynes refers to the aggregate firm and the aggregate consumer. The most important notion of Keynes is that he claims that the future can not be predicted. This is why people are unable to negotiate real wages – they are unaware of the rate of inflation that will come – and that there is a trade-off between inflation and unemployment. For policy this means that governments are to try to keep the investment level high, by means of low interest rates and own investments.

3 Of course, if there is inflation to start with, firms have more revenue and can increase wages. Policy

(10)

II The New Classical theory

2.1 New Classical theory about the labour market

What would be the implications if we could predict the future accurately? This is the core assumption of the New Classical theory and is supported by the rational expectations hypothesis. G.K. Shaw gives an example of the rational expectations hypothesis in practice in his book Rational expectations (Shaw 1984: 3). He introduces Mr Well-to-do and Mr Need-a-Loan and Mr Well-to-do will lent money to Mr Need-a-Loan. The interest rate is set at 5 percent, but the minister of finance will increase the money-stock with 10 percent. The result will be that the new rate of interest will become 15 percent, as both Mr Well-to-do and Mr Need-a-loan will anticipate a ten percent inflation rate. The real interest rate is still 5 percent and Shaw claims investment is a function of real interest rates. Investment does not increase (Shaw 1984: 4).4 The rational expectations hypothesis suggests that economic agents use information of the past and of endogenous shocks to make rational expectations. They are able to predict the future in this way and this is thus contrary to Keynes’ belief. The hypothesis uses the homo economicus (Shaw 1984: 105) and claims all human beings behave in such a way that they maximise their utility (Shaw 1984: 105).5

Do the rational expectations work in practice? The answer seems to be simple. Most people will not know anything about the expected inflation when lending money. Shaw claims that it is not worth some peoples while to get the information and so base their decisions on incomplete information (Shaw 1984: 107).6 The rational expectation theory

4

The same structure can be applied if someone is negotiating his wage and knows about the future inflation and wants his real wage to be stable.

5

This basic idea leaves much room for questioning. First the basic assumptions can be questioned. Do all human beings want to maximise their utility? This is already a bit tricky, since one could argue that many people do things that are not beneficial for their well-being. The rational expectations theorist would argue that either it is actually good for the person or that the person simply does not know what is good for him. Another thing is that utility is maximized in terms of money. So the highest wages attainable gives the highest utility. It seems fair to say that at least some people can have other values than money to maximise their utility.

6

(11)

is to say that the future is already sealed in the present, which is very dangerous to do, since it excludes possibilities other than the expectations and this might lead to the wrong basis for decisions. The current crisis is the most prevalent example as the expectations were that house prices would rise endlessly and in the end they did not.

Apart from that, Shaw himself already confesses that the equilibrium will not be reached (Shaw 1984: 108). An important reason for this is the money illusion that is, according to him, a fact of life. (Shaw 1984: 113). This is to say that people can value money with a discrepancy to its real value, i.e. misjudge the real wages. This says that the rational expectation hypothesis is a utopian benchmark and I think it should not have a place in economics other than the role of benchmark. The role as benchmark is one of the important features of the NAIRU theory that will be discussed in the following section.

2.2 The NAIRU story

In his book The rise of unemployment in Europe, a Keynesian approach, Stockhammer does not consider the New Classicals as the real competitor for the Keynesian theory regarding explaining unemployment. The real challenger is the NAIRU-model, but because of the many similarities with the New Classical theory, I will treat the NAIRU-model as a NAIRU-model within the New Classical theory. NAIRU means: Non-accelerating inflation rate of unemployment.

The implication of the model is that when unemployment is higher than the NAIRU inflation will increase, if the unemployment is lower, inflation will fall. At the NAIRU, the inflation is stable. NAIRU is the point of unemployment, the market will move towards, given the market rigidities. (Dimand 2008).7

This model is not market clearing. This is distinct from the New Classical model. In the short run the model determines inflation and in the long run determines the actual unemployment and output level. Product markets adjust to the natural output level. In short the model has six important features:

7

(12)

1. Unemployment has an effect on real wages.

2. Distribution (of income) does not effect the level of demand. 3. The real balance effect has an expansionary effect.

4. In the long run the system in anchored in a long-run equilibrium rate of unemployment. 5. The long term equilibrium rate of unemployment is independent of demand factors. 6. Higher wage flexibility and lower bargaining power of workers reduce unemployment. (Stockhammer 2004: 27).

The explaining feature of this theory lies mainly in point 6. Unemployment comes to be because labour unions and workers have become more rigid in negotiating their wages, and social welfare such as unemployment benefits. In the Keynesian framework this is voluntary unemployment (see section 1.1). The first point has been discussed in section 1.2. The NAIRU theory agrees with the New Classicals and disagrees with Keynes. The same goes for point 2.

The first NAIRU-model was developed by Phelps (Phelps 1968). In this model, there are two basic assumption. The first is that firms offer higher wages to prevent the high costs for searching new employees (Phelps 1968: 683). The second assumption is that Phelps related this to the Philips curve change of inflation to the change of unemployment. Within this relation only a high inflation may lead to decrease in unemployment. (Phelps 1968: 696). Friedman gave the model its name giving concept of the natural rate of unemployment:

“At any moment in time there is some level of unemployment which has the property that it is consistent with equilibrium in the structure of real wages…The ‘natural state of unemployment’…is the level that would be ground out by the Walrasian system of general equilibrium equations, provided that there is embedded in them the actual structural characteristics of the labour and commodity markets, including market imperfections, stochastic variability in demands and supplies, the costs of gathering information about job vacancies and labour availability, the costs of mobility and so on.”

(Friedman 1968: 8).

The main difference between the New Classical model and the NAIRU model is that in the New Classical model labour markets will clear unless there are impediments such as minimum wages. The NAIRU model is more a bargaining model, assuming intrinsic conflict within the market and therefore there will be no market clearing (Stockhammer 2004: 59).

(13)

unemployment to the inflation, so the unemployment is determined by mark-ups and wage-push factors. Phelps thought only unanticipated inflation was related to unemployment. (Dimand 2008). This explains why the NAIRU can be in between the Keynesian and the New Classicals theories. Product markets will passively adjust to this new equilibrium. (Stockhammer 2004: 61).

2.3 Market rigidities

New Classical theory assumes markets clear if they are left alone. We have seen some assumptions above: all market agents are rational, all market agents wish to maximize their utility and unemployment and real wages are strongly correlated. How then, is it possible that there is something as unemployment? First of all, there is frictional unemployment, such as season unemployment or unemployment in between jobs, miscalculation of the demand by managers. Second, there can be rigidities in the market that prevent the market to work as it should, i.e. the correlation between unemployment and wages is hurt.

As Siebert puts it:

Any labour market is surrounded by an array of institutional arrangements that form a complex web of incentives and disincentives on both sides of the market. For example, demand for labor is determined not only by the conventional market elements like output prices and the productivity of labor, but also by specific regulations relating to work time, layoffs, or other matters, and by taxes that raise the cost of paying workers. The supply of labor is partly determined by the reservation wage of potential workers, which in turn is shaped by institutions like the minimum wage, the level and duration of unemployment, welfare and social security payments. (Siebert 1997: 39).

(14)

2.4 Subconclusion

The New Classical theory assumes that people are able to predict the future. This would imply that employees are able to maintain their real wages and the trade-off between inflation and unemployment will not come to be. New Classical therefore reject the Philips curve. The underlying theory for this assumption is the rational expectations hypothesis, put forward by Shaw. He claims that people act rationally and predict the future based on the knowledge of the past.

This would also mean, given that markets are not regulated, the labour market will clear. Policy to reduce unemployment therefore, should be set to reduce market rigidities. The perfection of markets also implies that financial markets are stable and not the cause of unemployment.

(15)

III Empirical part

3.1 Summary theoretical part

In the theoretical part two theories have been identified that can explain unemployment. These are the Keynesian theory and the New Classical theory, more specifically the NAIRU theory. There are two main differences between these theories. The first is a disagreement about the possible trade-off between unemployment and inflation, i.e. about whether the Phillips curve is valid. The second is that the theories identify different causes for unemployment.

The first difference is explained by the difference of real and nominal wages. Keynes has claimed that employees negotiate nominal wages, so that when inflation rises, labour gets relatively cheaper and unemployment will decline. The extreme form of the New Classical theory, with the perfect market theory of Shaw, claims that this lasts only for a very short period and that employees will increase their nominal wages as not to have their real wages declined. The result will be that the share the firms have to pay to labour will be the same as in the old situation and therefore the unemployment will be the same. All that there is left is the actual inflation. New Classicals therefore would, when being concerned with unemployment, not advice money creation, while Keynes would.

Keynes and the NAIRU-theory both claim that inflation has an effect on unemployment in the short term. For the long term however, the NAIRU theory claims, as advocated by Friedman, (Friedman 1968) there is a fixed point of unemployment to which the labour market will aim. The NAIRU-point is determined, and this is the second difference, by market rigidities. The theory claims that if there would be no limitations to the market, there would be a minimum of unemployment. If market rigidities, such as minimum wages, employee-protection or government spending on labour market policy are added, unemployment will rise.

(16)

consumption will also go up, but not by as much as wages have gone up. This causes a negative difference between the revenue of the aggregate firm and the wages the aggregate firm has to pay. If this is not to result in an increase unemployment, investment in the aggregate firm has to increase and this is Keynes point for labour market policy. In short, there are two discussions at hand. The first is about an assumption of the theories: is there a trade-off between unemployment and inflation? The second is about the identification of the drivers of unemployment: are these market rigidities as the New Classicals claim? Or is a lack in effective demand the root cause for unemployment, to follow Keynes? It should be noted here that it is possible that both causes may play a role. It can be that both labour gets relatively more expensive on account of market rigidities, and that there is a lack of revenue to pay the wages as the effective demand declines.

3.2 Methodology

This leaves a few things to be tested. This is first the assumption of the theories. This is basically uncertainty. Uncertainty itself can of course not be tested, but the specific case about which the theories deal: uncertainty about the inflation in the future and the trade-off between inflation and unemployment that may possibly follow. This trade-trade-off is represented in the Phillips curve and this can be tested. Second, the causes of unemployment are to be tested, causes that either confirm the Keynesian or the NAIRU theory. This leads to the following hypotheses:

1a If the New Classicals are right, there is no relation between unemployment and

inflation

1b If Keynes is right, there is a negative relation between unemployment and inflation 2a If the New Classicals are right, there is a positive relation between unemployment and

market rigidities

2b If Keynes is right, there is a negative relation between unemployment and effective

(17)

It must be noted that although the theories are competing, they are also covering different areas. The NAIRU theory seeks the causes for unemployment in the supply and Keynes sought those in the demand.

For the hypotheses 1a and 1b a model for co-integration is required. The basic model for this is:

1. ut = β0 + β1

∆INFL

t + εt

ut is unemployment in the year t, INFLt is inflation and

εt

possible white noise. After this

it is important to test whether the data fits this model, so if the variables are non-stationary but the differences between the variables are non-stationary. This will be done with the Dickey-Fuller test. I will test if the addition of the variable exports minus imports to control for shocks in the world economy, such as the oil price shock in the eighties. In model 1c I will use the change in gross capital formation as a control variable. This variable has according to Stockhammer the biggest explainatory power and will come back in the models 3 and 4. This leads to an extention of model 1:

1.b ut = β0 + β1∆INFLt + β2 Ex-Im + εt

1.c ut = β0 + β1∆INFLt + β2 ∆GCF + εt

The test for hypotheses 2a and 2b will have the form Stockhammer has used, but some variables will change. I will first test the hypotheses apart. And when both models have explanatory power, I will test the models together:

2. ut = β0 + β1GSt + β2JPt + β3UD + β4ut-1 + εt

3. ut = β0 +

β1∆GCF + β2LC + β3ut-1 + εt

(18)

There variables represent unemployment (ut), government spending on labour market

policy (GS), a rank of job protection (JP), union density (UD), a lag for unemployment (ut-1), the change in capital accumulation (∆GCF) and total wages (LC), inflation (INFL), exports minus imports (Ex-Im), the change in capital accumulation (∆GCF) and total wages (LC). In the next section these variables are discussed.

The expected signs are, for the variables in the last model:

β1 < 0, β2 > 0, β3 > 0, β4 >

0, β5 < 0, β7 < 0, β8 > 0.

3.3 The variables

The dependent variable unemployment is the unemployment rate of the workforce per country, as measured by the OECD. Inflation is the inflation of the consumer prices. The government spending on labour market policy is the total amount of spending such as: training, employer incentives and unemployment benefits. It is expressed in a percentage of GDP. The variable job protection is the mean of three ranks, that rank from 1 to 6: protection of permanent workers against (individual dismissal), regulations on temporary forms of unemployment and specific requirements for collective dismissal. The higher the value the higher the protection. Union density show the ratio of workers that are member of a labour union.

These last three variables are in line with the NAIRU story, are rigidities in the market, so that the labour market does not function perfectly. This can be caused by governments spending on the labour market and job protection. Due to these inputs, it is possible that firms are not flexible enough to adjust the amount of workers they hire. The union density is usually a good measure for, among other things, collective agreements. In short, it gets harder for firms to pay lower wages per employee, and this might cause a rise in unemployment.

(19)

Stockhammer furthermore uses a variable for the accumulation of capital by firms and claims that this is related to unemployment. It seems to make sense: if firms have less money to pay for wages, unemployment will rise. For this variable I will use the Gross capital formation. It makes sense to exclude the investment in financial assets, as the GCF does, if ones wants to have a look at the relation of investment to unemployment. The GCF is given in absolute values, but I will use the change per year in percentages, because the absolute values are hard to compare with the rates of unemployment.

The source for all the variables is the OECD, except for the inflation, which was found in the database of the World Bank.

3.4 Data

Data of 17 countries are used. These are all OECD countries and I have decided to leave 13 countries out of the original set of 30 countries. Luxembourg did not have data about unemployment, and the other countries either had a low count of observations, or else had observations that were utterly strange. New Zealand for example had an unemployment measured of 0. The countries of the sample of 17 countries can be found in Appendix A. The data of the 17 countries that are left over, can be summarized as follows:

Variable Observations Mean St. Deviation Minimum Maximum

U 676 5.25 3.26 0.2 16.9 INFL 779 6.06 9.59 -13.85 131.83 Ex-Im 876 -0.38 6.52 -68.11 22.16 GS 355 2.19 1.45 0.1 7.19 JP 324 1.83 0.90 0.21 3.49 UD 744 41.16 20.33 7.8 83.9

GCF 658 1.07 0.01 0.67 1.47 LC 593 4.95 10.28 -10.22 114.93

Table 1: data description

(20)

observations, 31 are below 1 and only 7 below 0.5, and this is acceptable. The inflation has a mean of 6, but a very high maximum, but only 3 observations are above 100 and 7 above 50. For the variable Ex-Im it seemed attractive to adjust the variable to its change per year, but the data is more extreme in that way, so it does not help. The data as it is now shows a negative mean, meaning that all the 17 countries together have exported less value than they have imported considered over the whole time period.

Governments spent on average 2.19% of GDP on labour market policy, with a minimum of 0.1% and a maximum of 7.19%. About the job protection variable it is known that it ranks from 0 to 6. The mean is 1.83, the minimum only 0.21 and the maximum 3.49. In union density there are big differences, the minimum is 20.33, while the maximum is 83.9. The difference in the gross capital formation circles around 1, with a mean of 1.07 and a minimum of 0.67, a maximum of 1.47. The positive minimum means the gross capital formation is ever increasing. The change in labour compensation has a negative minimum, so is not always increasing. For now it suffices to see that the mean of the change in labour compensation is smaller than the change in inflation, which would imply that the aggregate real wages over time have decreased. The maximum is quite extreme, the labour compensation has more than doubled, but only 7 observations of the 593 are above 50. The independent variables have been checked for correlation, and done so per model:

Correlation U INFL Ex-Im ∆GCF

U 1

INFL -0.12 1

Ex-Im 0.03 -0.01 1

∆GCF -0.09 0.04 0.01 1

(21)

Correlation GS JP UD Ut-1

GS 1

JP 0.41 1

UD 0.71 0.32 1

Ut-1 0.54 -0.03 0.24 1

Table 3: correlation model 2

Correlation

GCF LC

GCF 1

LC 0.02 1

Table 4: correlation model 3

Correlation INFL GS JP UD Ex-Im Ut-1

GCF LC

INFL 1 GS -0.10 1 JP -0.11 0.40 1 UD -0.02 0.71 0.30 1 Ex-Im -0.12 0.26 0.41 0.22 1 Ut-1 -0.17 0.53 -0.09 0.18 0.12 1

GCF -0.14 -0.10 -0.10 -0.41 -0.03 0.22 1 LC 0.28 -0.18 0.11 -0.06 -0.03 -0.26 -0.14 1

Table 5: correlation model 4

(22)

3.5 Tests

The first tests consider hypothesis 1a. For hypothesis 1a a possible relation between unemployment and inflation is tested. I will use model 1. The initial regression uses pooled data of all countries and the following tests are per country, and thus focusing on the relation of inflation and unemployment per country. The models 1, 1b and 1c are used as well. The results are presented below and the Stata-output can be found in Appendix B.

Country INFL INFL (Ex-Im) INFL (∆GCF)

All -0.05** -0.05** -0.05** (0.004) (0.004) (0.001) Australia -0.11 -0.15 -0.11 (0.276) (0.170) (0.277) Austria -0.39** -0.39** -0.32** (0.000) (0.000) (0.000) Belgium 0.19 0.09 0.17 (0.268) (0.418) (0.329) Canada 0.21 0.87 -0.06 (0.838) (0.440) (0.505) Denmark -0.003 0.18 -0.005 (0.980) (0.427) (0.969) Finland -0.52** -0.01 -0.59** (0.000) (0.995) (0.000) France -2.54** -2.55** -2.46** (0.000) (0.000) (0.000) Germany -0.55** -0.48** -0.53** (0.017) (0.032) (0.022) Ireland 0.08 -0.02 0.05 (0.695) (0.732) (0.807) Japan -0.19** -0.09** -0.15** (0.000) (0.025) (0.000) Korea -0.19 0.01 -0.23* (0.182) (0.905) (0.088) Mexico 0.004 -0.001 0.004 (0.442) (0.411) (0.461) Norway -0.27** -0.32** -0.27** (0.000) (0.000) (0.000) Sweden -0.39** -0.22 -0.44** (0.000) (0.766) (0.000) Switzerland -0.45** -0.27** -0.23* (0.001) (0.013) (0.084) United Kingdom -0.12 -0.22** -0.13 (0.106) (0.002) (0.104) Unites States 0.15** 0.13* 0.15* (0.041) (0.091) (0.058)

(23)

The results of the tests with the pooled data show a significant and negative relation. This confirms hypothesis 1b that the Philips curve is indeed valid and there is a trade-off between unemployment and inflation. The uncontrolled results per country show 9 significant results, of which 8 have the sign that support hypothesis 1b. When the model is controlled for the exports minus the imports, all 7 significant results show the expected sign. In the model controlled for the change in gross capital formation shows 8 significant results with the expected sign.

This would lead to conclude that hypothesis 1b can be accepted and that there is a trade-off between unemployment and inflation, and that this relation is negative.

One problem might be that the regressions are spurious, because both the time-series for unemployment and inflation are non-stationary. If this is the case can be tested with the Dickey-Fuller test, which I will apply for the 10 significant results.

Country DF U DF INFL Austria -2.958 -2.938 Finland -2.938 -3.000 France -3.000 -2.938 Germany -2.938 -3.000 Japan -2.927 -2.938 Norway -2.966 -2.938 Sweden -2.938 -2.938 Switzerland -2.975 -2.938 United Kingdom -2.966 -2.938 United States -2.927 -2.938

Table 7: Dickey-Fuller tests for stationary time-series

The critical value for the model I used and for a 5% confidence interval is -3.37 and if the tested values are higher than this critical value, the time-series are stationary. Based on the above, it can be said that all the significant values found in the regression have stationary time-series and are therefore not spurious.

(24)

lines of inflation and unemployment do not establish a clear image of whether the development is the same. Therefore I will test the relation between unemployment and inflation, in the models 1, 1b and 1c, for different periods of time. The results are below:

Time period INFL INFL (Ex-Im) INFL (∆GCF)

1955-1969 -0.42** -0.23** -0.08 (0.000) (0.012) (0.352) 1970-1979 0.10* 0.07 0.06 (0.086) (0.193) (0.351) 1980-1989 -0.02 -0.02 -0.03 (0.486) (0.508) (0.249) 1990-1999 -0.15** -0.15** -0.15** (0.002) (0.002) (0.002) 2000-2009 -0.33** -0.30** -0.34** (0.003) (0.006) (0.002)

Table 8: results for hypothesis 1a, per time period, * = significant at a 10% level, **=significant at a 5% confidence level.

It must be noted that model 1c in the first period has only very little observations. Apart from that a very clear distinction can be made. In the periods from 1955 to 1969 and from 1990 to 2009, there is a significant negative relation between unemployment and inflation. In the period between 1970 and 1989 there is no significant relation between unemployment and inflation.

This result might be a first step to a daring hypothesis: the validity of economic theories is

based on time-periods and circumstances. This can be applied to Say’s law, but maybe the hypothesis can become more general and account for all economic theories.

(25)

Variable 1955-2009 1980-1990 1990-1999 2000-2009 JP -0.10 -0.17 -0.03 (0.126) (0.141) (0.686) GS 1.87** 0.22** 0.46** 0.25** (0.000) (0.044) (0.000) (0.001) UD -0.01** -0.01 -0.01* -0.01** (0.007) (0.315) (0.064) (0.032) Ut-1 0.84 0.92** 0.82** 0.86** (0.000) (0.000) (0.000) (0.000) JP -0.95** -1.53** 0.10 (0.000) (0.000) (0.585) GS 1.87** 2.64** 2.13** 1.40** (0.000) (0.000) (0.000) (0.000) UD -0.04** -0.13** -0.05** -0.03** (0.000) (0.000) (0.000) (0.000)

Table 9: results regressions for hypothesis 2a, * = significant at a 10% level, **=significant at a 5%

confidence level.

In the table above the most clear result is that of government spending on labour market policy. It is significantly and positively related to unemployment. This means that if expenditures on labour market policy rises, unemployment rises. This is in line with the NAIRU theory. Union density shows persistently a negative sign, which is not expected. Job protection is only significant in the model without the lagged unemployment and gives the unexpected negative sign.

One problem is that the data for the time period 1955-1980 is lacking for government spending. The job protection rank starts only in 1990.

(26)

Country ∆GCF LC Ut-1 Obs. All -5.87** 0.004 0.95** 501 (0.000) (0.341) (0.000) Australia -1.84 -0.04 0.81** 31 (0.341) (0.350) (0.000) Austria 0.48 0.001 0.91** 29 (0.760) (0.968) (0.000) Belgium -6.54** 0.03 0.90** 25 (0.008) (0.676) (0.000) Canada -7.10** 0.05* 0.91** 35 (0.000) (0.052) (0.000) Denmark -6.95** 0.87** 0.98** 39 (0.001) (0.045) (0.000) Finland -7.60** -0.03 0.89** 39 (0.000) (0.321) (0.000) France -7.21** -0.07 0.96** 13 (0.031) (0.443) (0.000) Germany -8.03** -0.10** 0.80** 39 (0.000) (0.011) (0.000) Ireland -5.88** 0.43 0.96** 22 (0.000) (0.341) (0.000) Japan -2.96** -0.001 0.90** 37 (0.000) (0.866) (0.000) Korea -6.38** -0.06 0.68** 18 (0.001) (0.179) (0.002) Mexico -4.65** -0.004 0.92** 20 (0.028) (0.590) (0.000) Norway -1.89** -0.05** 0.84** 37 (0.025) (0.007) (0.000) Sweden -5.77** -0.04* 0.90** 39 (0.000) (0.061) (0.000) UK -4.92** 0.01 0.96** 36 (0.004) (0.569) (0.000) US -7.67** 0.08** 0.94** 37 (0.000) (0.001) (0.000)

Table 10: regression results hypothesis 2b, * = significant at a 10% level, ** = significant at a 5% level

(27)

Since both models 2 and 3 have a significant variable, and thus neither theories can based on this claim the total explanatory power, it makes sense to test the variables in the total model: model 4. Below are the results for tests with model 4, with the countries pooled, for the same time periods as above and both with and without the lagged dependent as independent variable: Variable 1955-2009 1980-1989 1990-1999 2000-2009 INFL 0.21 0.05 0.02 -0.02 (0.287) (0.251) (0.417) (0.664) GS 0.19** 0.14 0.29** 0.13* (0.002) (0.154) (0.001) (0.097) JP -0.04 -0.82 0.027 (0.550) (0.425) (0.710) UD -0.004 -0.003 -0.01** -0.002 (0.134) (0.728) (0.029) (0.389) Ex-Im -0.001 0.03 -0.01 -0.001 (0.854) (0.305) (0.591) (0.743) ∆GCF -6.91** -4.95** -8.88** -3.71** (0.000) (0.000) (0.000) (0.000) LC -0.02* -0.07** -0.03 -0.02 (0.084) (0.006) (0.125) (0.110) Ut-1 0.92** 0.91** 0.93** 0.88** (0.000) (0.000) (0.000) (0.000) INFL -0.25 0.25 -0.01 -0.16 (0.668) (0.203) (0.933) (0.103) GS 1.89** 2.19** 2.15** 1.36** (0.000) (0.000) (0.000) (0.000) JP -1.14** -1.78** -0.84 (0.000) (0.000) (0.661) UD -0.05** -0.18** -0.06** -0.04** (0.000) (0.000) (0.000) (0.000) Ex-Im 0.03** 0.244** 0.07 0.001 (0.043) (0.040) (0.135) (0.900) ∆GCF 1.75 4.13 1.16 1.75 (0.275) (0.323) (0.610) (0.385) LC -0.03 -0.16 -0.02 -0.06 (0.395) (0.180) (0.702) (0.131)

Table 11: regression results model 4, * = significant at a 10% level, ** = significant at a 5% level

(28)

the model without the lagged variable, and in the model with the variable it is significant over the whole time period and in the nineties.

(29)

IV Conclusions and recommendations 4.1 Conclusions

The goal of this thesis is to give an advice for policy regarding unemployment. For that purpose two theories were scrutinized: Keynesian theory and the New Classical theory. It became clear that there are two important differences between these theories. They have a different assumption they start off with and they indentify different causes of unemployment.

The assumption Keynes held was that there is a trade-off between unemployment and inflation, because future inflation can not be predicted and therefore real wages can not be negotiated. The New Classicals claim based on the rational expectations hypothesis that the future can be predicted. The Keynesian theory can be expressed in the Philips curve. I have tested the relation between unemployment and inflation, both with and without control variables. The results show a significant and negative relation, which confirms the Keynesian theory and hypothesis 1b.

When the relation is tested in different time periods, the relation between unemployment and inflation is not significant between 1970 and 1989. It is the time period in which Keynes had lost its appeal and then probably for right reasons. From 1990 on to now, the Keynesian theory shows significant results again, but its appeal has not yet returned. I have tested the causes of unemployment first per theory. The New Classicals claim that market rigidities, so everything that interferes with the market, cause unemployment. Government spending on labour market policy proves to increase unemployment significantly. Keynes believed involuntary unemployment can be caused by a falling behind of effective demand. Investment is therefore expected to be negatively related to unemployment. The change in gross capital formation showed this negative significant relation. This means that hypotheses 2a and 2b are both confirmed.

The conclusion is that in some periods in time, and as well as over the whole time, there is a trade-off between unemployment and inflation. And that an increase in government expenditure on labour market policy increases unemployment, while an increase in gross capital formation decreases unemployment.

(30)

Furthermore, based on this research, it hardly makes sense to invest in labour market policy, since this increases unemployment. The right policy would be to increase the gross capital formation. This can be done by lowering the interest rate and to make investments as a government. But perhaps the strongest advice is to take Keynes serious again.

4.2 Recommendations

That government spending on labour market policy would increase unemployment is of course an unwelcome message. Governments wish to be able to decrease unemployment. It can be recommended to look at these expenditures more closely. It can be that some of the expenditures actually decrease unemployment. Maybe it a time-lag can be included in the model to determine if the expenditures have an effect after several years.

Another point is that of causality. It is possible that because of high unemployment rates, governments spent a lot on labour market policy. The same goes for the change in gross capital formation, maybe investments decrease after unemployment rises. This leaves room for more research.

(31)

Literature

• Akaike, H. 1969. Fitting autoregression for prediction Annals of the institute of statistical mathematics 21: 243-7.

• Review 21: 107-12.

• Borio, C. 2008. The financial turmoil of 2007-?: a preliminary assessment and

some policy considerations. BIS Working paper No. 251. Bank for International

Settlements.

• Carter Hill, R., Griffiths, William E. and Lim, Guay C. 2008. Principles of

econometrics.

• Congdon, T., 2007. Keynes, the Keynesians and monetarism.

• Claesens, S., Kose, M. A., and Terrones M. E., 2008 What happens during

recessions, crunches and busts? IMF Working Paper No. 08/274. International

Monetary Fund.

• Diamond, D. W. and Rajan, G.R. 2009. The credit crisis: conjectures about

causes and remedies. Unpublished.

• Dimand, R.W., Edmund Phelps and the modern macro-economics Review of political economy. Vol. 20. (2008). 23-39.

• Engle, R.F. and C.W.J. Granger. 1987. Cointegration and error term correction:

Representation, estimation and testing Econometrica 55: 215-76.

• Friedman, M., 1968 The role of monetary policy American Economic Review Vol. 58 1-17.

• Glen, P. J., The efficient capital market hypothesis, chaos theory, and the

insider filing requirements of the securities exchange Act of 1934: the predictive power of form 4 filing’ Fordham Journal of corporate and financial law.

Vol. XI 85-160.

• Islam, F., et al. 2003. The empirics of U.S. Philips Curve: a revisit American Business Review 21: 107-12.

(32)

• Krugman, P., How did economists get it so wrong’ New York Times September 6, 2009.

• Minsky, H. P. 1992. The financial instability hypothesis. Jerome Levy Economics Institute of Bard College Working Paper No. 74.

• Phelps, E., 1968 Money-wage dynamics and labor-market dynamics Journal of political economy, Vol. 78 678-711.

• Shaw, 1984 Rational expectations.

• Sheehan, B., 2009 Understanding Keynes’ General theory.

• Shiller, R.J. 2003 From efficient markets theory to behavioural finance Journal of economic perspectives 83-104.

• Siebert, H., 1997 Labor market rigidities: at the root of European

unemployment Journal of economic perspectives Vol. 11 37-54.

• Skidelsky, R., 2009 The return of the master.

• Stockhammer, E., 2004 The rise of unemployment in Europe, a Keynesian

approach.

• Turner, D., et al. 2001. Estimating the structural rate of unemployment for the

OECD countries OECD Economic Studies No. 33: 171-216.

Referenties

GERELATEERDE DOCUMENTEN

This study will explore two different categories of political apologies: the apology for a personal scandal for which the politician is personally responsible versus the apology for a

Om deze vraag te beantwoorden heeft een onafhankelijk instructeur de kinderen drie keer beoordeeld op het uitvoeren van de basiselementen die leiden tot de Onderzoekend leren

This paper offers a set of multidimensional data models and analysis techniques that can be used to detect the most prevalent known fraud types and should prove useful in

consistency in brand advertisements and customer-based brand equity, brand attitude and brand image and whether this relationship is moderated by brand familiarity, a 3 (level

Resultaten lieten zien dat dimensies van temperament of van opvoedgedrag afzonderlijk niet van invloed waren op de emotieregulatie van peuters maar dat juist de combinatie van

After exploring the experiences and perceptions of employees, the empirical data yield the results which assist in the development of the wellness facilitation coaching

Editorial 3 The article ‘A survey on topological properties, network models and analytical measures in detecting influential nodes in online social networks’ represents

Special attention is paid to the stern tube lubricant, as stern tube seals are shown to be unreliable when operating with some of the new Environmentally Acceptable