Empirical Evidence On Okuns Law Economics Essay

The study of Okun’s law is still an important theory. The relationship between the increase in unemployment and the decrease in a country’s GDP that is the gross domestic product refers to as the Okun’s law. Actually, Okun assumed that a 1% rise in the growth rate above the trend rate of growth would lead only to 0.3% in the reduction of unemployment. On the other hand, a 1% increase in unemployment will represent approximately more than 3% loss in GDP growth. This theory is named after the economist, Arthur Okun who in 1962 was the first person who has made a dept observation about the relationship between economic growth and unemployment. During the period 1962, Okun presented two empirical relationships linking the rate of unemployment and real output. Both of them were simple equations which were used as the rules of thumb since that time and thus since that time the presented equations have become associated with his name, Okun.

The Okun’s first relationship, that is, the difference version take into consideration how changes in the unemployment rate from one quarter to the next moved with quarterly growth in real output. The equation is: Change in the unemployment rate = a + b*(Real output growth). This connection captures the correlation between output growth and movements in unemployment for that period. That is, how output growth differs simultaneously with changes in the unemployment rate contemporaneously. The variable b is called the Okun’s coefficient. The proportion “-a/b” indicates how the rate of output growth is consistent with the stable unemployment rate, that is, how rapidly the economy would normally need to develop to preserve a given level of unemployment.

Despite the fact that Okun’s first relationship relied on macroeconomic figures, his second relationship, that is, the gap version relate the level of unemployment to the gap between potential output and actual output. This equation takes into consideration how the economy would produce under full employment conditions. Hence Okun’s second relationship took this form: Unemployment rate = c + d *(Gap between potential output and actual output).

The parameter c can be inferred as the unemployment rate linked with full employment and the coefficient d would be positive to be traditional to the connection above.

Hence just to keep the unemployment rate constant, the rate of GDP growth must be equal to its potential growth. To reduce unemployment, therefore, the rate of GDP growth must be above the growth rate of potential output (Tatom, 1978).

The relationship stated above, that is, between economic growth and unemployment is one among the most famous in macroeconomics theory and has been held for several countries and regions mainly, in developed countries (Lee, 2000; Christopoulos, 2004; Daniels and Ejara, 2009). However, there are very few attempts of testing the theory of Okun’s law in developing countries and mostly in small island countries. Therefore, another motivation of this study is to test the validity of Okun’s law for a developing country like Mauritius. That is, if Okun’s law is valid for Mauritius, this would then imply whether or not unemployment can be reduced by boosting growth.

The problem of unemployment is quickly being assumed as an unsafe proportion because in many countries there has been a great boost in unemployment rates in recent decades. Clearly are many reasons for the increase in unemployment as mention partly below:-

The first reason is Technology. The increase in technology has contributed greatly to the changes in the demand for labour. Machines have substituted jobs in some occupations.

Globalisation is also one of the reasons, since 1950; because of trade liberalisation there has been a huge increase in the movement of capital across the world. That is, a reduction in barriers to international trade has made possible for countries to make use of their comparative advantages and natural resources. For instance, developing economies are relatively having unskilled labour and so they are capable to produce labour intensive goods, such as textiles, cheaper than the developed countries. Thus, the developed world takes advantage of the cheaper labour intensive products in exchange for their exports of goods and services produced by skilled labour and capital.

Therefore developed economies had to decrease their home production of products such as textiles leading to a shift in demand for labour from low-skilled and manual jobs towards higher-skilled occupations and vice-versa.

Moreover unemployment have increased to a large extent around the world after the sharp rise in oil price during the 1970s and also because of the collapse of the Bretton Woods system of fixed exchange rates. But however unlike countries, unemployment in many European countries has never returned to the low levels as seen during the Golden Age after the Second World War. The explanation for unemployment is that industrialized economies became more unstable and are more regularly subject to shocks such as rises in oil price or exchange rate fluctuations.

In other words, the problem has both macroeconomic and structural dimensions. Hence, because of these factors, many economies are not able to create sufficient jobs to absorb all the additional job seekers. Many studies have analyzed the effects of GDP growth on unemployment. As Carre and Drouot (2004) stated, ‘economists have been interested in studying the relationship between growth and unemployment at least since the beginning of the industrial revolution.’

3.2 Empirical evidence on Okun’s Law

Sögner and Stiassny (2000) investigated Okun’s law for 15 Organization for Economic Cooperation and Development (OECD) countries and verify for it’s the structural stability, for the period from the year 1960 to 1999 and for Germany it was 1960 – 1989. They used data on employment and the labor force they check whether structural instability is caused either from the demand side or the supply side.

From their results they conclude that the reaction of the unemployment to changes in GDP are different, this depends on which country they are considering. By applying their methods ( where continuous changes are considered ), they obtained changes in Okun’s law for the Switzerland, Germany, Denmark, Finland, France, Great Britain, Japan, the Netherlands, Norway and Sweden. For Austria, Belgium, Canada, Italy and the US they inferred a stable Okun relationship.

Moreover, by considering labor demand and labor supply as functions of GDP growth they were are able to deal out with changes in Okun’s law to demand and supply side changes. In an early pace the OLS estimates are gave (uses – Cochrane-Orcutt method) for 15 OECD countries. And next the GDP development connection was investigated by way of Bayesian methods and Kalman Filtering. Furthermore the authors additionally use the Markovchain Monte Carlo (MCMC).By using their econometric analysis they found that, for the majority of countries the changes in the Okun coefficients are mostly due to a raised in labor demand (reaction of employment) on GDP variations. According to their results, only for France, Switzerland and Germany the changes in Okun’s law were enforced by labor force effects. In Italy all the relationships are stable over time.

They derived a negative correlation for 15 OECD countries between their two measures and the estimated employment coefficients. Hence this implies that countries with a highly protected labor market actually reveal a low reaction of employment to GDP variations this is mainly due to labor hoarding that is, not to lay off unneeded workers during the period of recession to ensure that skilled and experienced workers will be available after the recession.

However in this paper they have selected a group of counties that is Switzerland, Germany, Denmark, Finland, France, Great Britain, Japan, the Netherlands, Norway, Sweden Austria, Belgium, Canada, Italy and the USA, but in my case it will be only one country that is Mauritius. Most of the countries selected are found to be developed countries whereas Mauritius is a developing country. Moreover employment and labour force have been use as their main variable.

There has been wide literature on the issue of growth and unemployment. MOOSA (2008) studied the validity of Okun’s law in four Arab countries, the objective of his paper was to estimate Okun’s coefficient, and explore the validity of Okun’s law, for four Arab countries: Algeria, Egypt, Morocco and Tunisia for the covering period 1990 to 2005. In his article, he uses two models for measuring Okun’s coefficient: the gap model and a modified version of the growth rates model.

The Gap model related the level of joblessness to the gap amid possible output and actual output. In potential output, Okun tend to recognize how far the economy should produce “under conditions of maximum employment”. Whereas, modified version of growth rate model is an adjusted form that includes a variable rate of period preferences. Hence, this is disparate from the standard growth rates model in the sense that the dependent variable is the log level of joblessness rather than its percentage change.

Nowadays, economists are focusing on the possibility of asymmetry in the relationship between output-unemployment as represented by Okun’s law. In the paper of MOOSA (2008), the meaning of asymmetry is that the reaction of unemployment to output growth is different when the economy is growing from compared to when the economy is contracting. Two models are used for measuring Okun’s coefficient in the study of MOOSA (2008): the gap model and a modified version of the growth rates model.

The first model use in this journal, that is, the gap model uses dynamic and static of cyclical unemployment on cyclical output. The cyclical variables of output and unemployment are calculated by applying the Hodrick-Prescott (1997) filter (HP filter) to the observed time series and then are converted into trends and cycles. The Hodrick-Prescott (HP) method was primarily developed to smooth period sequence, That is, to get a smooth (long-term) component

Contrary to the gap model, another form of variation in model specification is the use of the first-difference model. In the gap model output and unemployment are measured in terms of the cyclical components (for example, Lee, 2000), but in the first difference model, the output and unemployment variables are expressed in first differences, that is, growth rates. One of the problems with the selected gap model is the choice of the decomposition or de-trending method, thus different estimates of the trends and cycles are produce. The author have de-trended its variables so as to get rid of long-term trends and hence to emphasize on short-term changes.

The second model used by MOOSA is a structural time series version of the growth rates model. The main idea behind equation of structural time series version of the growth rates model is that unemployment can be explained in terms of its variables and the growth rate of output. One of the assumptions in this model is that, the trend and cycle are assumed to be uncorrelated while the irregular component is assumed to be white noise. The model is estimated in a time-varying parametric (TVP) framework to capture any variation in the coefficient thus the Okun’s coefficient has a time subscript.

For the four countries, he found that output growth does not translate into employment gains, that is, the Okun’s coefficient he found was to be statistically insignificant. This means that, the results presented in his study suggest that unemployment and output are unrelated in the four countries studied irrespective of the model used to estimate Okun’s coefficient.

There are three possible reasons that can explain for the finding that Okun’s law is not valid for the four countries examined in his paper. The very first reason is that unemployment in these countries is rather structural instead of being cyclical. Structural unemployment is job loss caused not because of lack of demand, but rather by changes in demand patterns and obsolescence of technology thus requiring retraining of workers and huge investment for new capital equipment. Time frame of study may shed light whether these were actually the case for the 4 countries

The inflexibility of the labour markets in these countries is second explanation for the Okun’s law to be insignificance. This is particularly because the labour market is controlled by the government as the main source of demand for labour hence the rate of causing unemployment is not so high.The structures of these economies are the final reason, which is dominated by the government and perhaps one sector (for example, the oil sector in Algeria). That is, growth in this sector will not reduce unemployment if the leading sector is not labour-intensive. As the developed economies are more diversified than the developing economies. Okun’s coefficient tends to be higher for the former than the latter.

However no matter what the reason is for the Okun’s coefficient to be not valid, it can be concluded that the lack of growth does not explain the unemployment problem in the four countries examined in this paper as the government controls the labour market.

But however the author studied for four countries which are likely to be developing economies but they are rich in oil and natural gas reserves, phosphates and agricultural produce, and receives large amount of foreign currencies from the tourist industry unlike Mauritius. And moreover in these economies government control the labour market which is not the case of Mauritius.

Haririan et al, (2002), studied on the long-run relationship between GDP growth and unemployment for the period 1975 to 2005 covering the following selected Middle East and North Africa (MENA) countries: Turkey, Egypt, Israel, and Jordan. The literature has usually shown a negative liaison between GDP growth and unemployment. The MENA region covers a range of economies with varied economic structures and labour markets. It can be said that there are some economic differences between oil-producing and non-oil-producing countries in that region. However, today, one of the most important economic and social problems of the MENA region as a whole is the high rate of unemployment. Hence for this reason, some economists believe that there is a strong inverse relationship between the GDP growth and unemployment in the region.

They found that the relationship between GDP growth and unemployment is very multifaceted and there are not very clear. According to their results, the impact of GDP growth on unemployment is not very strong. Their empirical evidences for Turkey, Israel, Jordan and Egypt support a negative relationship between GDP growth and unemployment. According to their paper, one of the main reasons for the high unemployment rates in the MENA region is the high population growth of these countries. Although the last few decades the countries have witnessed decline in the population growth, still the population growth rate remains very high. Thus, high population growth rate is one of the reasons for the persistent high unemployment rates in the countries according to this paper. The four countries analyzed had an average population growth rate of 1.93 percent in 2006, compared to the OECD average of 0.6 percent. They also believe that another reason for the high unemployment rates in the MENA region is the economic performance of these countries. .

In the analysis on the long-run relationship between GDP growth and unemployment, the most common approach in recent economic literature, that is, Hodrick-Presscott Filter have been used by the authors. The benefit of this model is that it does not fit a linear trend, but it fits an optimal, smoothed and most likely non-linear trend, thus this gives the researcher the opportunity to clear the effects of long-run persistent shocks on trends. This filter is better when the researcher have to de-trended time series data before the estimation.

The authors had preferred to carry out a cross-country comparison while analyzing the impact of GDP growth on unemployment for country groups. This method is widely used in the literature

Example: Harris and Silverstone, 2001 and Zellner, 1962. Using this technique, they can search whether there any significant cross country differences for the sample. Moreover they have used the Augmented Dickey Fuller test (ADF test), An increased Dickey-Fuller test is an examination for a unit root in a period sequence sample. It is specially used for larger and extra complex set of period sequence models and in this case the null hypothesis is non-stationarity.

Akaike Information Criteria and Swartz Criteria were used while deciding the number of lags.

According to their results, the series strongly are stationary in the 1st difference. In relation to the results presented in their tables for the four countries, GDP growth affects unemployment negatively. Akaike Data Criteria is a measure of the comparative goodness of fit of a statistical model and the Schwarz Criterion is a criterion for selecting amid proper econometric models. According to the authors, this relationship is stronger in Egypt and Turkey than in Israel and Jordan. Despite that the relation between GDP and the unemployment is weakest, but is still statistically significant. The results may also indicate that Turkey and Egypt are doing well in creating jobs while growing. In contrast, the results might also signify Israel’s growth mainly depends on productivity.

The relationship between GDP growth and unemployment found in the MENA countries is not as simple, according to the paper it is very complex. Due to institutional conditions in the labour market and demographic factors the relationship between these two variables may not exist. In fact for the MENA countries, the high population growth rate is one of the major reasons for high unemployment in these countries. Similarly for the Arab countries, whatever model which have been used for estimating Okun’s coefficient, the results shown in the study for the four Arab countries suggest that output and unemployment are not related. According to the paper these are the reasons for which can be suggested for the invalidity of Okun’s law:

People living in the countries do not have the skills to do the available jobs,

The labor markets have become inflexible due to the dominant role played by the government in the Arab countries.

SODIPE and OLUWATOBI (2012) investigated on the economy of Nigeria. The objective of their article was to study the existence of an Okun-type relationship for the Nigerian economy for the period 1970-2009. According to the journal, a time series data was employ on unemployment and GDP for the period 1970 to 2009. Furthermore they estimated the relationship of unemployment rate on the GDP variables to observe the relationship of unemployment rate with the growth rate of GDP in Nigeria.

The estimation was made using the gap method, where unemployment is related to the deviations of output from potential GDP. Moreover unemployment and national income was included in the model. On the other hand, the second option is the use of Okun’s first-difference method, thus this will test the relative sensitivity of output to unemployment changes. To identify the Okun’s law coefficient for the Nigerian economy, they had employed the co-integration technique. After having tested for the order of integration of the series, a co-integration representation was carried out. The difference version take into account how the unemployment rate vary from one quarter to the next moved with quarterly growth in real output. The co-integration technique was developed by Engle and Granger in 1987, it is used to account for non-stationarity and it offer more powerful tools when the data sets are of limited length.

The Augmented Dickey Fuller and Phillip Perron tests were used to test for the order of integration. In addition, his order of the co-integration was determined. They found that unemployment rate and real GDP growth rate are not co-integrated. According to their study, the Augmented Dickey Fuller (ADF) and the Phillip Perron (PP) unit root test were useful in testing for the stationarity of the time series. Phillips-Perron (PP) unit root test can be applied at first difference of each series while analyzing time series data.

Both variables were tested at first difference because the result was showing that none of the variables is stationary. The results presented in the table for the first difference was positive that is both ware stationary. Hence, their co-integration test was conducted at first difference.

Additionally the Johansen Co-integration test was used in order to find out the existence of a long run relationship between unemployment and gross domestic output and the type of relationship that exists between them. The Johansen test is applied when there is more than one co-integrating relationship, not like the Engle-Granger method.

However the authors concluded that a long run relationship exists between unemployment and

GDP and the relationship exist between them are negative in Nigeria. Moreover they have used co-integration techniques which have been criticized as being confusing, unreliable, and dangerous.

Lal et al, (2010), investigated on Okun’s law and test if it is valid for some Asian countries like China, Pakistan, India, Srilanka and Bangladesh. For this purpose they have used a time series data for the period 1980 to 2006. The basic aim for this paper is to determine the correlation between the unemployment gap and the output gap for the chosen countries. According to them, they have selected these Asian countries because their characteristic like labor characteristics, natural resources, geographical and climate conditions are more or less the same.

Another motivation for their study is to check whether the relationship between the measures of unemployment gap and output gap is statistically important in long-run and the short run. Therefore for this paper the unit root test and Engle Granger (1987) cointegration approach have been used to verify the long run relationship and the stationary of variables. Between the two standard model specifications of Okun’s law, the Gap model has been selected for advance analysis of the okun’s law. They have taken the data like Data of unemployment and output, Gross Domestic Product and GDP deflator from the from the World Bank dataset (WDI).

They have used the predictable technique of ordinary least square (OLS) as the majority of the macroeconomic variables are non-stationary series, hence by using the OLS this provides opportunity of co- movement between variables or spurious regression. In this paper they have remove the non stationary variables by differencing the time series parameters. Error correction modeling gives short run dynamics which thus tries to find out the causal liaison in the short run.

The authors have used the Fully Modified Ordinary Least Square (FMOLS), once the order of cointegration has been decided, for the long-run elasiticities and to get asymptotic efficiency that is, normal distribution. Ordinary least squares, that is, OLS, is a technique for estimating the unknown variables in a linear regression model. This process diminishes the sum of squared vertical distances between the observed responses in the dataset and that which are predicted by the linear approximation.

According to this paper, the results found by the authors do not support the proposition of the Okun’s law in the developing countries because of the problem of asymmetry. Moreover due to many fluctuations in rates of inflation in Asian region, the natural rate of unemployment in developing countries particularly Pakistan and India cannot be estimated.

3.2 Conclusion

It can be concluded that most of the paper examine uses the time series data. Moreover many authors have make use of group of counties while analyzing the relationship between GDP growth and unemployment. Most of the results have found to be either negative relationship or no relationship between GDP growth rate and unemployment rate especially in developing countries.