A quick look at the relationship of output and employment by nationality in Bahrain

NOTE: Edits have been made towards this article on the 30th of November 2021 to reflect corrected data due to data-collection mistakes previous to publication. The new results do not alter the outcomes of the article’s content significantly. The original published dataset and article are available by request.


Introduction


What is the relationship between output and employment in Bahrain? This article takes a quick look at the theory and actual results of what output and employment looks like by nationality in Bahrain. Does an increase in production result in higher employment rates amongst expatriates and Bahrainis? If so, why? Why do we see a relationship between output and employment? These are the items that we will address.

This article is broken down into three sections. The first section is the theory of employment and output, the second section is data and results, and the third is the implications and recommendations.


What is the relationship between employment and output?


How are employment and output related to one another? In simple macroeconomic terms, for a firm to produce output (products and services), they need two inputs. Those inputs are capital (machinery, equipment, buildings, and other items) and labor (employing people). Therefore, increased production should be associated with increased capital and/or labor when an economy produces goods and services.

Thus if there is a relationship between an increase in output and labor, we can measure this relationship by looking at the changes in output and the changes in employment. This relationship is known as Okun’s Law, wherein 1962, Arthur Okun theorized that if output increases above (or below) its natural rate, this means that so too will unemployment (in our case, employment) rates will decrease or increase (as unemployment and output are inversely related) below or above their natural rates (Owyang, Vermann, and Sekhposyan, 2013). It is essential to ask, why can’t we simply have no unemployment (or full employment for those who are seeking jobs) and simply “max out” output? This is because potential output measures the economy’s productive capacity when the labor market is at its natural unemployment rate. This rate is believed to always be above zero, which will vary from theory to theory (example: people may move from job to job is often referred to as Frictional unemployment). Therefore the potential output will always be lower than the maximum output in a sustainable manner.

From Okun’s paper, he finds that if real GNP (not GDP, but similar) is held at zero, the unemployment rate will grow at 0.3 percentage points on average in the US. This relationship has held more or less the same across time, with some variations in specific periods. Overall what do we expect when we test Bahrain for Okun's Law?

First, we do not have unemployment rate data; thus, we can’t look at how Okun’s law is viewed in a typical fashion. Second, using a figure such as the employment ratio is problematic as for expatriates, this rate is more or less stable and only has slight variations across time. Therefore in this article, we will simply measure the relationship between a percentage change in output and a percentage change in total Bahraini employment and Expatriate employment (this is not as good as a measurement of unemployment rates, labor force participation rates, increases in hours worked per employee, or labor productivity, however, for the given limitation in data and unique structure of Bahrain’s labor market will allow us to see such a relationship). Our results indicate that Bahraini employment responds less towards changes in output, while expatriate employment rates are more responsive.


Data we will use, treatments that will be done to our data, and model used to estimate.


The data we will use is real GDP per quarter from IGA, and total employment by nationality being Table A from LMRA. For GDP data, we will first do what’s called “seasonal adjustments” towards the Quarter on Quarter Real GDP, then we will use the total yearly real GDP and calculate the percentage change from quarter to quarter. For Total Bahraini and Expatriate employment, we will also seasonally adjust the data from quarter to quarter. The seasonal adjustments are made using X13-ARIMA. We use data from Q4 2009 to Q4 2020. The data below reflects changes in percentage terms and not in absolute unit changes.


WHAT DOES THE DATA TELL US?


First, let’s take a look at the relationship between output and total employment.

We see a relationship between changes in output and changes in total employment from the scatter plot. If we are to use the slope of the fitted line, the graph indicates that a 1% increase in output is related to a 0.35% increase in total employment. However, it is essential to note that using the slope here is not the most appropriate way to measure the relationship. There are more sophisticated ways to test the actual relationship. We will continue to use the slope for comparative purposes to show a rough relationship between different employment types. Furthermore, if we were to look at our R^2 figure (a measurement which allows us to judge the “goodness of fit for our linear trend line”, we see that the “goodness of fit” is rather low.

Next, let’s look at the relationship between output and Bahraini employment.

Here, we see a relationship that is weaker than that of total overall employment. When using the slope again, we see that the relationship here is that for a 1% increase in GDP, we would see a 0.25% increase in total Bahraini employment. Furthermore, we see that the R^2, which measures how well our trend line fits the data, is also relatively low as well.

Last, let’s take a look at the relationship between real GDP and total expatriate employment.

Here, we see that expatriate employment is more sensitive towards changes in output. Using the slope, a 1% change in total GDP indicates that expatriate employment increases by 0.46%. Compared to Bahraini employment, it appears that output is generally caught more in expatriate employment than Bahraini employment.

Once again, it is important to iterate that such uses of scatter graphs and slopes only informally look at the relationship between changes in GDP and employment and thus should not be taken as actual results of how employment levels move with output. More robust methods are available but are too advanced for an article and are generally reserved for research papers to explain in detail and with more precise estimates.


What could the graphs above indicate?


What we see with these results is somewhat what we expected. First, we see total employment does increase with total output. However, we see some indications of how Bahraini employment and expatriate employment moves with changes in output. We see that percentage changes of Bahraini employment move slower with percentage changes in output. As for the case of expatriates, we see that the data indicates that employment moves more with percentage increases in output. Thus, we can see that output is generally caught more in expatriate employment than Bahraini employment.

Why is it that we see indications that Bahraini employment changes slower with changes in output? Potentially, regulatory reasons may be a reason why Bahraini employment grows slower than expatriates. First, it is important to note that Bahrainis have minimum wages across all forms of skill levels, which increase with skill level (egovernment, n.d). In comparison, no minimum wage exists for expatriate workers. Furthermore, employers pay a higher GOSI contribution for Bahraini employees than expatriate employees (12% vs. 3%) (PWC, 2021). A final example is that as per article 110 if employers are forced to dismiss employees due to reorganization, downscaling of business, or partial closure, local Bahraini employees must retain if it is possible to terminate a foreign employee who has the same competence and experience as the local employee instead (Abbas & Mohamed, 2021). Overall, the above potential regulatory burdens may make it difficult to justify hiring a Bahraini with similar skill levels when expatriate labor may be more inexpensive.

Another explanation is that a significant portion of the Bahraini workforce is employed in the public sector. As of Q4 of 2020, we see that approximately 32% of Bahraini’s are employed in the public sector. Generally speaking, governments may prioritize employment above all other objectives; therefore, if we were to look at the relationship between Government services output and public sector employment changes, the relationship between the two may be weak. Compared to the private sector, objectives are different where the private sector typically prioritizes profit maximization. Thus, testing specifically output less government services (a rough proxy for private-sector output) and private sector employment of Bahraini may yield a stronger relationship.

It is important to note that the graphs above indicate that the goodness of fit is low, meaning that there are potentially more factors that may affect employment growth in the Kingdom for expatriates and nationals. Therefore while this quick look gives us some intuitive information on how employment growth moves with GDP growth, more sophisticated and in-detailed analyses are needed.


What could be appropriate policies?


From the data, there are potential appropriate policies we can ascribe. The first of which is in regards to reducing hiring costs surrounding Bahraini. With current minimum wages, a possible recommendation is to mitigate any increases in minimum wages across the different skill categories. This way, this will ensure that rises in base wages do not increase relative to alternative opportunities, being expatriate employment of similar skill level or capital deployment. Furthermore, another policy recommendation would also be to reduce costs associated with social insurance. This can be done by decreasing the cost of social insurance contributions paid by employers for Bahraini employees. 

The last set of potential policy recommendations is to look at increasing costs associated with hiring expatriates. This could include possible hiring caps and increasing employers’ social insurance contributions for expatriate employees. The combination of decreasing potential costs associated with Bahraini employees and increasing costs associated with expatriate employees may be a sensible way of ensuring that employers’ overall costs do not substantially increase. Employers may be more incentivized to hire locally rather than expatriate labor. 

By changing employers’ incentives on what nationality base while minimizing costs associated with hiring overall labor, these potential policy recommendations may shift Bahraini employment contributions towards the labor force in future periods. The next steps towards more specific policy recommendations are to test the relationship between output and employment in more depth and sophistication.

 
2012-02-09manual-typewriter-0209stock.jpg

ABOUT THE AUTHOR