1
The Impact of MNE Investment on
Employment in Ireland
Brian O’Malley
University of Amsterdam Amsterdam School of Economics Faculty of Economics and Business
Student Number: 11737573
Email: brian.omalley@student.uva.nl Date: 28/06/2018
Number of Words: 12946
Course: Master Thesis Economics
Track: International Economics and Globalisation Supervisor: Dr Dirk Veestraeten
2
Statement of Originality
This document is written by Brian O’Malley who declares to take full responsibility for the contents of this document.
I declare that the text and the work presented in this document are original and that no sources other than those mentioned in the text and its references have been used in creating it.
The Faculty of Economics and Business is responsible solely for the supervision of completion of the work, not for the contents.
3
Contents
1. Introduction ... 4
2. Recent Developments in Ireland ... 7
3. Literature Review on Foreign Direct Investment in Ireland ... 17
4. Regression of Employment in Ireland ... 24
5. Conclusions ... 40
References ... 45
4
1. Introduction
Ireland is a small, open economy that has punched above its weight in terms of attracting Foreign Direct Investment (FDI) since the 1980s. While Ireland moved into the post-World War II period slowly, several endogenous and exogenous factors aligned, and Ireland began an extremely rapid period of convergence to economic levels experienced by first world countries from the end of the 1980s. It is now seen as a prominent player in international markets, ranking 17th worldwide for its ease of doing business and 4th for trading across
borders (Doingbusiness.org, 2018). Ireland’s top 5 multinational investors in 2016 were the global conglomerates Microsoft, IBM, BASF Group, Google and Tesco (The Irish Exporters Association, 2016) while other world-renowned companies such as Apple, Pfizer and HP already have a significant production presence in the country. According to the Top 150 Born in Ireland 2016 report, the leading industry for multinational enterprises (MNEs) was the technology sector, which held six of the top twenty spots while the healthcare and pharmaceuticals sectors held five.
The methodology followed in this thesis consists of an empirical analysis based on previous work developed by the existing literature. Recent studies have investigated the factors that led MNEs to invest in Ireland and how the economy was rocked by the GFC. The intended contribution of this paper is therefore to extend the existing literature by explaining the stability of MNE investment to employment in Ireland. My research question is as follows: “Is employment created by MNEs in Ireland stable?”
I will carry out a regression for the time period 1983-2016. It will see the employment rate as a function of several variables. These variables will include the standard determinants of employment in an economy such as the interest rate, Gross National Product (GNP), the population growth rate and lagged employment rate for example. A dummy variable will be included to describe whether the economy is in a period of recession or not. The foreign direct investment represents the inward FDI stock as a percentage of GDP. This will show its relative importance to the Irish economy as it will not be affected by general inflation with its effect always shown in terms of the most general indicator of the economic performance of the country. A measure of competitiveness will be included to contextualise the position and competitive advantage of Ireland in international markets at each period. It is important
5 to look at this study in terms of employment as these MNEs do not contribute largely to Irish GNP due to profit repatriation nor do they rely on the Irish credit market to any real extent, features that I will elaborate on later.
There will also be dummy variables included in my regression to describe the “The Changing Sectoral Pattern of FDI Stocks” as discovered by Barry (2007). As my regression begins in 1983, the first of the four stages – where Ireland began to move away from protectionism at the end of the 1950s – will not be relevant to my work. Instead, the first period that I will observe is from the point of Ireland’s accession to the European Communities in 1973 until the late 1980s where the country shifted into higher-tech sectors. The next phase is that of the 1990s which saw a global high-tech boom and the acceleration of the offshoring of services. Finally, the remainder of my time period is described by Barry to have been dominated by the substantial offshoring of research and development (R&D) functions by MNEs.
I will look at the interaction effects between FDI and recessionary periods, and how this affects employment. I will then be able to comment on how stable the employment created by FDI is. This is a particularly pertinent question today when the potential jobs being created in Ireland by firms relocating from the UK due to Brexit are investigated but also the challenges of the looming global trade war and US tax reforms that are looking to bring US-born firms back to redomicile at home. The findings of my research will also me to comment on the worthiness of Ireland’s aggressive policy pursuing MNE investment and whether, for instance, Ireland’s low corporate tax can be justified, even against the extreme pressure it puts the country under from a number of EU leaders (The Irish Times, 2017). It will allow me to deliver advice to the Irish government and, in fact, many other small economies with regards to dealing with these opportunities and threats optimally.
The remainder of this thesis is structured as follows. Chapter 2 will provide a brief overview of the Irish economy over the past few decades, providing some basic information such as on its size and employment levels. Chapter 3 presents an overview of the existing literature on the recent history of the Irish economy and the development of FDI flows with the aim to provide the background for the rest of the paper. It also provides a critical assessment of previous work on MNE investment in Ireland and indicates the intended contribution of this paper. Finally, it examines the features that make Ireland an attractive location for FDI.
6 Chapter 4 is the empirical chapter of the thesis, describing the variables chosen and
combining them into a single regression in order to determine the importance of MNE investment for employment in Ireland. It goes on to looks at the results of the regression, assessing the impact of MNE investment on employment in Ireland and how stable this employment is. Finally, chapter 5 examines the future opportunities and threats that may affect the flows of MNE investment in Ireland through major political and macroeconomic changes worldwide. It also presents the main conclusions drawn for the analysis along with advice for policymakers.
7
2. Recent Developments in Ireland
The Irish economy has experienced substantial changes since the early 1980s – enduring two severe recessions but also experiencing the unprecedented growth of the ‘Celtic Tiger’ period. At the beginning of the period of my regression, 1983, Ireland sat behind EU and the Organisation for Economic Co-operation and Development (OECD) levels of wealth but now finds itself as one of the richest countries, per capita, in the world. The economic and so, hence, the social landscape of the country has altered hugely, and this chapter is aimed at explaining some of these changes.
The evolution of the population of Ireland is the first port-of-call when it comes to Ireland’s recent developments. There were roughly 3.5 million people in Ireland in 1983 but this number has risen dramatically to its current level, approaching 4.75 million – an increase of over 35%.
Source: The World Bank
3 .5 4 4 .5 5 1983 1990 1995 2000 2005 2010 2016 Year
8 Some of this increase can be attributed to immigration levels in Ireland of recent years. Since the mid-1840s, when a brutal famine hit the country, Ireland has seen huge swathes of people emigrating in search of a living. The USA, UK and Australia were among the most common destinations. However, as Ireland emerged from the recessionary years of the 1980s, this emigratory trend finally began to reverse. The ‘Celtic Tiger’ period brought hope to the country and attracted new migrants to the country as well as enticing former
emigrants to return home.
The 2004 enlargement of the European Union also played a significant role in attracting new immigrants into Ireland. 10 countries joined the union which opened the option of visa-free migration for the new members. Immigration spiked over the following 5 years, as seen in Figure 2, only coming to a halt with the onset of the Global Financial Crisis. In fact, in 2007, Ireland experienced a net migration level of over 100,000 migrants to the country. This spike in immigration explains a large portion of the steep increase in the population of Ireland in the mid-2000s as seen in the previous graph. This figure only begins in 1987, the first year the data became available.
Source: The World Bank
-5 0 0 5 0 1 0 0 N e t M ig ra ti o n ( '0 0 0 s ) 1987 1990 1995 2000 2005 2010 2016 Year
9 Another key development over the past 35 years has been the huge increase in the level of female participation in the labour force. Female participation levels were stagnant during the 1980s, sitting just below 35%, but the upturn in the economy saw an influx of females deciding to join the workforce. By the turn of the millennium, this number had increased by over 12 percentage points to 47%. The peak rate of female participation was reached in 2007, just before the onset of the recession, with 54.5% of females partaking in the labour force. While it had taken Ireland longer than most OECD and European Union countries to catch this trend, once the process began, it quickly caught up to these averages. Male participation rates fell slightly during the years studied, but this fall was outweighed by the increased contribution of females to the labour market. The total participation rate
witnessed a remarkable trough-to-peak rise of almost 11.5 percentage points.
Source: The World Bank
4 0 6 0 8 0 P a rt ic ip a ti o n R a te ( % ) 1983 1990 1995 2000 2005 2010 2016 Year
Female Participation Male Participation Total Participation
10 The unemployment rate in Ireland since 1983 has followed the cycles of the economy but the chart below shows the true success of the ‘Celtic Tiger’ period in reducing
unemployment in the run-up to the millennium. From a 1987 peak above 18%,
unemployment steadily improved over the 1990s to a level of just 4.3% in 2000. In fact, unemployment levels dropped even lower the proceeding year, falling as low as 3.7%. These years were described as an economy in ‘full employment’ as the unemployment rate
remained below 5% until the onset of the recession in 2008. Again, the procyclicality of unemployment shone through as many businesses were forced to shut down while huge swathes of employees lost their jobs. Unemployment spiked, reaching nearly 15% in 2011 and 2012 as the economy adjusted. However, as Ireland has emerged from its financial crisis and austerity, unemployment has tracked downwards again and fell back below 7.9% in 2016.
Source: The World Bank
Conversely, employment rates fell below 44% in the mid-1980s and did not begin to see regular growth until 1993. As the ‘Celtic Tiger’ period came into full swing, employment
0 5 1 0 1 5 2 0 U n e m p lo y m e n t R a te ( % ) 1983 1990 1995 2000 2005 2010 2016 Year
11 rates continued to grow and reached a peak of 61% in 2007, coinciding with the peak of the boom. The rise in participation rates, as described above, in the period after the recession of the 1980s can be used to explain some of the overall increase in the employment rate. The GFC saw employment rates plummet back as far as 51% in 2011 and 2012. While rates have improved since the darkest days of the GFC, they are still a long way from the heights of the last decade.
Source: The World Bank
Looking at the progress of Ireland’s GNI per capita provides an opportunity to compare the performance of the Irish economy with some of its international counterparts. In 1983, Ireland sat significantly below the European Union and OECD averages, with an annual GNI per capita of US$5,705 versus US$6,857 and US$8,952 respectively. As the recessionary days of the 1980s came to a close, Ireland roughly matched the increases in GNI per capita of the two groups before surpassing the European average in 1997. It took until 2001 to better the OECD average but from this point, Ireland’s GNI per capita diverged substantially from its counterparts, climbing all the way to $53,000 in 2007 and 2008. This compares to averages
4 3 .7 7 6 1 .0 8 5 0 5 5 E m p lo y m e n t R a te ( % ) 1983 1990 1995 2000 2005 2010 2016 Year
12 of just $37,000 and $38,000 for the OECD and EU countries respectively. While it can be seen from the graph below that the recession of the late 2000s hit Ireland harder than the European Union and OECD averages, Ireland also recovered to a far stronger position, reaching back above $53,000 in 2016. This level far exceeds the two comparisons used here – 30% above the OECD average and 39% higher than the EU average.
Source: The World Bank
The following graphs focus on the FDI to GDP (FDI:GDP) ratio in Ireland to represent the flows of FDI received by Ireland. Firstly, the 15 years beginning in 1983 are shown, a time when the level of FDI inflow was relatively low as Ireland battled through the recession of the 1980s. It was not until 1990 and 1991 that Ireland’s inflows escaped these lows,
climbing from a level of FDI of 0.2% to 2.7% of GDP. By 1997, Ireland’s FDI had reached 3.3% of its overall GDP. 0 2 0 0 0 0 4 0 0 0 0 6 0 0 0 0 G N I p e r c a p it a ( $ ) 1983 1990 1995 2000 2005 2010 2016 Year
Ireland GNI per cap OECD GNI per cap EU GNI per cap
13 Source: The World Bank
1998 saw the level nearly quadruple and set in motion 18 years of volatility in the level of the FDI:GDP ratio but with a far higher average rate than the previous years in the study. A negative level in 2004 came as a result of foreign multinationals operating in Ireland divesting more than they invested in the country, but this quickly rebounded in 2005 with FDI inflows reaching 22% of GDP. There was a record inflow of FDI in 2015 as FDI climbed above 70% of GDP, more than double any previous year on record. The Irish Times (2017) reported this increase “to be due to financial and corporate restructuring rather than new investments”, citing “tax-inversion deals, which allow companies to shift profits to the lowest tax jurisdiction in a merger.” The size of the level of FDI:GDP ratio in Ireland gives a good understanding of the importance of MNE investment in Ireland and the role it plays in the Irish economy.
0 2 4 3 1 F D I: G D P 1983 1990 1997 Year
Fig 7: FDI:GDP 1983-1997
14 Source: The World Bank
There were a number of factors that combined to make Ireland the attractive destination it is today for FDI and that have played major roles in the development of the FDI:GDP ratio detailed above. Some of these factors are inherent to the country while others were a result of explicit government policy aimed at enticing MNE investment to Ireland. Below are lists of both factors: 0 2 0 4 0 6 0 8 0 F D I: G D P 1997 2000 2005 2010 2016 Year
Fig 8: FDI:GDP 1998-2016
15
Table 1: Factors Attracting FDI Under Government's Control
Corporate tax rate Access to EU market Good industrial relations
Education
Western Europe standards of governance
Increasing contact with international agencies since late 1950s – IMF and World Bank Extension of Industrial Development Agency functions - State Incentives
Trade liberalisation
Table 2: Factors Attracting FDI Outside Government's Control
Cultural connections with the USA English-speaking
Atlantic location
Demographics/Population growth Cultural Factors (work ethic, strike culture)
Ireland began moving towards convergence with Western European averages at the
beginning of the 1970s when the government set in place initiatives to support and promote international trade in Ireland. Ireland’s accession to the European Union (EU) in 1973, its comparatively high level of education, native English-speaking and low corporate tax level of 12.5% have all played vital roles. Growth accelerated and the period from 1994 to 2008 became known as the ‘Celtic Tiger’ era as Ireland was one of the top-performing economies in the world. These conditions combined to attract MNE investment into the country and my thesis will investigate the employment created by this investment.
16 The ‘Celtic Tiger’ period came to an abrupt end in 2008 as the Irish economy began to crumble under the Global Financial Crisis. The housing market that had been booming pre-crisis ran into trouble while banks that were hugely exposed to mortgage debt followed suit. One of the causes of the recession in Ireland was the failure of the Central Bank of Ireland to regulate the market adeptly. The failure to regulate quickly became evident as six Irish banks required re-capitalising. Of these six, only Bank of Ireland was not entirely
nationalised. The European Commission, the European Central Bank and the International Monetary Fund, collectively known as the Troika, were called in to aid the Irish government. Austerity was introduced by the Irish government while the National Pension Reserve Fund was also deployed in the bank recapitalisation effort. Ireland’s credit rating took a severe beating, falling from the highest possible pre-crisis level all the way down to junk status.
The decisions of Irish governments to promote FDI in Ireland obviously came with an opportunity cost. I will investigate the stability of the employment created by MNEs in Ireland to determine whether the aggressive policy of Irish governments to create an FDI-friendly economy was worthwhile, at the cost of, for example, higher corporate tax
revenues that could be collected if Ireland were to move closer to the EU average corporate tax rate. I will consider both boom and recessionary periods to determine the stability of employment in Ireland.
17
3. Literature Review on Foreign Direct Investment in Ireland
The Irish economy has experienced remarkable growth since the 1980s but particularly from the late 1990s to the onset of the recession in 2008. This growth saw the country converge to economic levels associated with a first world country and was the product of the
alignment of many factors and conditions. Much of the literature has focussed on the journey and how the economy has arrived at its present point. The review of the relevant literature below is divided into six sections: the roots of FDI in Ireland, the GFC, why Ireland is an attractive location for FDI, the distinction between gross national product vs gross domestic product, variables affecting employment and competitiveness.
A) The Roots of Foreign Direct Investment in Ireland
One of the leading papers on the subject of multinational investment in Ireland is
Offshoring, Inward Investment and Export Performance in Ireland by Barry & Bergin (2012).
It provides a timeline of foreign direct investment in Ireland since the 1950s and highlights the economy’s move into higher value services and production over the years. Whelan (2013) explains this timeline in further detail; examining the demographics, the
macroeconomic policy and the housing boom of the period which are identified as key elements in understanding the context of the recession.
Barry & Bergin (2012) outline the areas that are most FDI-intensive, namely ICT,
pharmaceuticals and medical devices in the manufacturing sector along with software and IT and international financial services in the internationally traded services segments. They also detail the role that the increase in globalisation has played in the growth of the Irish economy along with contextualising just how strong the Irish FDI position is. They describe some of the key factors behind Ireland’s success in attracting export-platform FDI with geographic proximity to the US and EU membership identified as significant determinants. Agglomeration and demonstration effects are also found to contribute to Ireland’s
attractiveness as an export-platform location – Barry & Bergin (2012) use the results of Barry & Bradley (1997) to demonstrate the influence of the presence of key market players in attracting newcomers to set up in Ireland.
18 Furthermore, the IMF Selected Issues paper on Ireland (2017) and The Economist (2014) highlight the importance of the Industrial Development Authority1 in the convergence of the
Irish economy to Western European levels and investigates the sources of the FDI inflows received by Ireland. US investors dominate FDI stock while the EU also plays a prominent role in FDI. The source of these inflows is important as we see the threats to US firms investing abroad by the US tax reforms and the looming global trade war.
B) Global Financial Crisis
On the other hand, it is impossible to examine the last 40 years of the Irish economy
without addressing the crash that Ireland experienced upon the arrival of the GFC. The crash in the Irish economy of the late 2000s is extensively reported in the literature (IMF, 2017; Lane 2014). Whelan (2013) explains how the housing market unwind contributed to the crisis in the banking system. The role of the Central Bank of Ireland is questioned in its failure to protect the economy through surveillance and regulation while the failure of Anglo Irish Bank and Irish Nationwide is another issue brought to the table.
While the banking crisis caused significant damage to the financial system and the economy, FDI is less vulnerable to these strains as MNCs in Ireland “predominately rely on external sources … to fund activities, such that the distressed state of the domestic banking system has not directly damaged funding mechanism for this sector” (Lane 2014, p. 18). Another hypothesis, suggested by Godart, Görg & Hanley (2011) in the abstract of their paper “is that foreign multinationals are less linked into the Irish economy” and therefore they are “more likely to leave once the economy is hit by a negative shock”. They go on to describe reasons why foreign firms are not more likely than Irish firms to leave during a crisis because they have operations outside Ireland though. They also usually have exports from Ireland, so their performance is not solely reliant on the domestic economy. The authors suggest that there may be a greater stability of FDI due to the sunk costs and longer-term focus that comes with setting up in a country, unlike the other types of international capital flows such as portfolio investment. However, they go on to find evidence counter to this hypothesis.
1 The Industrial Development Agency is the state agency, founded in 1949, tasked with attracting and
19 Firstly, building on previous research about foreign-owned companies’ linkages to upstream firms – while acknowledging their higher export rate – Godart, Görg & Hanley (2011) also address the sunk costs involved in setting up in a country. They investigate service and manufacturing firms and found no difference between the probability of an Irish-owned and a foreign-owned company leaving Ireland during the crisis. I will explore the behaviour of Irish employment during the GFC to establish the validity of these claims and to conclude whether the jobs created by MNE investment are stable or not.
C) Why Ireland is an attractive location for FDI
A list of the reasons that combined to make Ireland an attractive location for FDI is included in chapter 2 and these reasons have received extensive coverage in literature. I will provide further detail on three of these factors below: demographics, education and the corporate tax regime. Demographics are included because, as seen in chapter 2, Ireland has
undergone huge population changes as well as participation rate adjustments since 1983. Education has also played a huge role in attracting FDI to Ireland with highly-skilled industries such as the technology sector and healthcare and pharmaceuticals dominating MNE investment. These industries necessitate high levels of education to fulfil staffing requirements and the evolution of education in Ireland is telling. Finally, Ireland’s corporate tax regime is the elephant in the room when it comes to FDI and is a very contentious policy internationally.
i. Demographics
The Irish economy experienced rapid growth through the late 1980s and early 1990s in a period that became known as the beginning of the ‘Celtic Tiger’ years. This expansion can be broken into three major parts – a higher population, a higher fraction of the population at work and increased output per worker. These features aligned leaving Ireland “By the early 1990s…an enormous capacity to grow far faster than it had been doing” according to Whelan (2009, p. 3). Ireland underwent a baby boom through the 1970s that peaked in 1980, contributing to a larger workforce. Female participation rates lagged many European counterparts since the end of the second world war but finally picked up as Ireland emerged from the recession of the 1980s. While the Irish economy experienced the beginning of the
20 ‘Celtic Tiger’ period, job creation only slightly outsized the swelling labour supply meaning also that wage growth never reached an excessive level and it was not until 2000 that ‘full employment’ was reached. Honohan & Walsh (2002) also reported productivity growth of just under 3% per annum over the 1990s which all combined to increase the attractiveness of Ireland as a destination for MNE investment.
ii. Education
An improvement in the level of education in Ireland was a key component in attracting FDI into the country. Barry (2007) identifies an OECD report issued in 1965, “Investment in Education”, that was “informed throughout by the perspective that education was a means by which society invested in its own future”, as a key turning point for education in Ireland. Having produced a scathing description of a system that saw “over half of Irish children left school at or before the age of thirteen”, major changes were introduced. Free secondary education and free access to special transport networks for all second-level schools were introduced in 1967. The third level system was also overhauled with the introduction of the technical-education system. The idea of sub-degrees, which were shorter than standard university degrees, was launched and were practically orientated around the needs of local industry. By 1978, Ireland had the second-highest proportion of third-level students taking sub-degrees around the world (Barry, 2007). The IDA and the Manpower Consultative Committee worked carefully together to ensure that graduates possessed the requisite skills for the FDI that Ireland was attracting. An example of how prepared the Irish workforce became through the reform in the education system can be seen in the “Institute for Management Development, World Competitiveness Yearbook for 2005”. In that report, Ireland was ranked 2nd of 60 OECD and medium-income developing countries in response to the statement “the educational system meets the needs of a competitive economy”.
iii. Corporate Tax Regime
A tax on profits of manufacturing industry of 10% was put in place at the end of the 1970s (Barry, 2007). This rate was extended to computer software in 1984 and again to qualifying activities in the newly-founded Irish Financial Services Centre in Dublin in 1987. Other services remained under a corporate tax rate of 32%. Under pressure from the EU, the Irish government decided to harmonise their corporate tax rate for all sectors in 1998. This new
21 rate was implemented from 2003 at 12.5% and represents one of the lowest corporate tax regimes in Western Europe. In fact, it is the effective corporate tax rate that matters most, which Buijink, Janssen & Schols (1999) find Ireland’s rate to be around one half of the EU average.
D) The distinction between GNP vs GDP
The open nature of the Irish economy and the strong presence of foreign companies means that common international measures such as GDP, that may, for instance, accurately describe the state of the US economy may not be as reliable for Ireland. The strong presence of foreign firms in Ireland means that a large amount of the profits earned in Ireland are repatriated and thus does not benefit the country as much as GDP would
suggest. To combat this, Barry (2000) used GNP instead of GDP as GNP is the money value of all the goods and services produced by the residents of a country. A Department of Finance official publication (2014) highlights the differences between GDP and GNP/Gross National Income (GNI) in Ireland. With Ireland, “net factor flows are invariably large and negative” because “the returns on inward investment are much greater than the returns on Irish investment in the rest of the world” (p. 14). Indeed, this is supported by the IMF (2017) when they state that the gap between Ireland’s GDP and GNP is “among the highest in the world”. Lane (2017) adds to this discussion with a more technical analysis of the gap for the Irish economy. I will therefore use GNP in place of GDP as one of my explanatory variables.
E) Variables Affecting Employment
Tunah (2010) considered the macroeconomic variables which cause unemployment for Turkey. It was found that real GDP, the consumer price index and the lagged unemployment rate all had a significant impact on the current unemployment rate. Aurangzeb (2013) also examined the determinants of unemployment rates, but this time for Asian countries: India, China and Pakistan. Inflation, GDP, the exchange rate and population growth were all found to be significant variables in the case of these three countries. These are important findings as they provide the basic variables for the regression I will run on the level of employment in Ireland.
22 Another study that is closer to Ireland was carried out by Stanila, Andreica & Cristescu (2014) on employment rates in EU countries. They create two clusters of European
countries using Hierarchical cluster analysis on which panel data models are estimated. GDP and trade openness – a measure of imports and exports as a percentage of GDP – are each only found to be statistically significant for one of the clusters, but both provide a positive effect on the employment rate. Average gross earnings are found to be positive and significant for the two clusters. Remittances were also investigated but this would be to accommodate poor, Eastern European countries that rely on emigrant workers to bolster their economy by sending home earnings. Therefore, as Ireland does not rely heavily on remittances as some of its poorer European counterparts do, it will not be of interest to my study.
Scarpetta (1996) also highlights the importance of unemployment benefits and employment protection regulation on the employment rate finding both to be significant determinants of the employment rate. I will therefore include measures of employment protection
regulation in my regression. I will use the OECD Employment Protection Database’s estimator of strictness of employment regulation as a proxy for these variables.
F) Competitiveness
Another important variable in Ireland’s international trade performance and its
attractiveness to FDI is competitiveness. A stronger level of competitiveness means that costs are falling for exporting firms which, in turn, helps to increase profits. The IMF 2017 Article IV Consultation (2017) rates Ireland’s global competitiveness ranking quite well because of a “sound legal and regulatory system, a strong workforce and flexible labor [sic] market, and low corporate income taxes” (p. 5). Rashid & Akram (2017) explore the UK market, using the real exchange rate as a measure of competitiveness. They find a statistically significant effect of the real exchange rate on the level of employment in the country.
Ireland, however, is a slightly more complicated case than the UK due to the fact that it became a member of the Eurozone in 1999. I will label this period with a dummy variable to represent this change in policy. I will also include a variable for the real exchange rate that
23 combines the real exchange rate of Ireland with its two major trading powers outside of the EU – the USA and the UK – for the whole period of the regression.
24
4. Regression of Employment in Ireland
In order to analyse the impact of MNE investment on employment in Ireland, I will carry out a regression analysis study. The dependent variable of the regression will be the rate of employment. The variables I have selected come from a wide array of origins – from previous studies of employment and unemployment to literature on the drivers of the Irish economy – and aim to explain the development of the employment rate in Ireland since 1983. To begin, I have included thirteen independent variables, one interaction term and one lag of the dependent variable which will be analysed for statistical, intuitive and economic significance.
A) Unit Root Analysis
I began my regression analysis by studying each of the variables graphically to check for potential trends or structural breaks. Next, three different tests were used – the Augmented Dickey-Fuller, the Phillips-Perron and the Dickey-Fuller GLS – to test for stationarity and unit roots in each of the variables. It was found that the employment rate, GNI, female
participation and population growth all contained unit roots. Table 3 below is a summary of the results of the Augmented Dickey-Fuller and the Phillips-Perron tests which both have null hypotheses of a unit root being present. In each case, none of the critical values are breached so it is not possible to reject the null hypothesis and therefore it can be said that a unit root is present in each of the three variables.
25
Table 3: Unit Root Analysis 1
Variable Test Statistic 1% Critical Value 5% Critical Value 10% Critical Value Augmented Dickey-Fuller Test Employment Rate -1.090 -4.306 -3.568 -3.221 GNI 0.291 Female Participation -0.426 Population Growth -1.700 Phillips-Perron Test Employment Rate -5.688 -23.524 -18.508 -15.984 GNI -2.747 Female Participation -3.633 Population Growth -8.177
In order to combat the issue of a unit root, first-differencing is applied to the three series. Table 4 below provides a summary of the results of unit root testing on the first-differenced series. At the 10% significance level, the null hypothesis of a unit root can be rejected for the first-differenced variables of GNI, female participation and population growth.
According to the Augmented Dickey-Fuller test, the presence of a unit root in both female participation and population growth can also be rejected at the 5% level of significance while the Phillips-Perron test allows the same conclusion to be drawn at the 5% level for GNI.
The issue of a unit root in the first-differenced version of the employment rate is not as straightforward, however. The three tests that have been used suggest the presence of a unit root even after first-differencing. Double-differencing presents issues, though, from both an intuitive and an economic angle. Along with the fact that unit root tests do have some problems during the crisis and that the rejections are borderline, I have decided to work with the first-differenced series for the dependent variable, the employment rate, in my regression. This means that the dependent variable can now be described as the growth of the employment rate.
26
Table 4: Unit Root Analysis 2
Test Statistic 1% Critical Value 5% Critical Value 10% Critical Value Augmented Dickey-Fuller Test Employment Rate -2.583 -4.316 -3.572 -3.223 GNI -3.525 Female Participation -3.673 Population Growth -3.695 Phillips-Perron Test Employment Rate -12.232 -23.396 -18.432 -15.936 GNI -20.029 Female Participation -17.925 Population Growth -18.386
27
B) The Variables
Of the fifteen independent variables included in the analysis, six are dummy variables. Firstly, there are variables that show when Ireland became a member of the European Monetary Union (EMU) as part of the Maastricht Treaty of 1992, and when Ireland adopted the euro as the national currency. The treaty itself entered into force at the end of 1993, with 1994 being its first full year in existence. Therefore, the dummy variable has been assigned a value of zero from the beginning of the regression until 1993 while 1994 through to the end of the regression has a value of one. The euro only officially became the single legal tender of its members in January 2002, but it was introduced initially as a real currency with a single monetary policy from the start of 1999. Additionally, there are two dummy variables included that are related to the changing sectoral patterns in the Irish economy discovered by Barry (2007). Barry identifies four stages that are already detailed in the literature review. The first period is not included as it precedes the regression period used while the second and third period are labelled with dummies (‘htsectors’ and ‘htboom’). The last period, from 2000 onwards, is not included to avoid the problem of the dummy
variables trap due to the presence of the dummy variables ‘htsectors’ and ‘htboom’.
Additionally, it is particularly important to identify periods of economic recessions when investigating the stability of employment in Ireland. It is these periods where the stability can truly be tested versus the overall level of employment in the economy. Changes in FDI during crisis periods will be examined to determine whether they are
employment-disturbing factors and whether they affect the change in the employment rate.
A dummy variable (‘crisisdummy’) is used to identify the crisis periods in the Irish economy. The period from 2008 until 2012 is highlighted to represent the GFC. The GFC first hit Ireland in 2008 and required a bailout from, amongst others, the Troika as previously detailed. This bailout formally ended in early 2013 and, therefore, 2012 is the last year labelled as a crisis year. The recession of the 1980s is slightly more difficult to define in range. In my
regression, the period from 1983 until 1986 has been chosen, bookended by negative annual GDP growth. Another reason to select this period was that the employment rate fell considerably over this period too, from 46.9% down to 43.8%. 1986 saw an end to this downward trend, having plateaued since 1985, and the employment rate then begins its rise, converse to the fall in unemployment reported in chapter 2.
28 The lagged value of the dependent variable – the growth of employment rate –
(‘lemployment’) is important to include as growth or contraction of the rate of employment last year may influence this year’s rate of change. For example, in the case of a boom or a recession, trends can take shape meaning that last year’s change in the employment rate may affect this year’s change.
The next variable included is inflation, measured as the Consumer Price Index (CPI) on an annual basis. It is a basic percentage. GNI is included to represent the strength of the economy over the regression period. The reason for choosing GNI over GDP has been explained previously in chapter 3 and it is included after first-differencing as the change in GNI, with its units measured in millions of US Dollars.
The Employment Protection Indicator (EPI) is a measure created by the OECD to explain the level of protection afforded to employees in a country and how easily they can be
dismissed, or new employees can be hired. It is measured on a scale from zero to five, from loosest to strictest. Ireland has fared somewhere in the middle of the list of OECD countries and other major countries in terms of EPI over the period of the regression. Figure 9 gives a picture of the Irish level of EPI from 1983-2016.
Source: OECD 1 .2 5 1 .3 1 .3 5 1 .4 1 .4 5 E P I 1983 1990 1995 2000 2005 2010 2016 Year
Fig 9: EPI
29 The theoretical and practical appeal of including such a variable is clear. However, it seems to take the appearance of a dummy variable over the time period in question as it is virtually constant over time with a single drop in the crisis period. It is found to have a unit root according to the tests used above but first-differencing would just leave a spike at either end of the recession. This could lead to collinearity with the crisis dummy variable. The fact that it sees so little change also indicates that it most likely does not have a large effect on the change in the employment rate over the period and therefore I have omitted it from the regression.
The interest rate (‘irate’) included is the 3-month fixed interbank market rate, also known as the Euribor rate. It is the rate at which euro interbank term deposits are offered by one prime bank to another, within the euro area. The interest rate has been generally
decreasing over the regression period but is found not to have a unit root according to the Augmented Dickey-Fuller and the Phillips-Perron tests. This general downward trend shows how international capital markets grew and credit became far more accessible to
governments and private agents. The trend reversed, however, as the rate charged during the GFC spiked from the levels of the mid-2000s due to rising fears over the ability of some European countries to repay debt and to the worsening business climate. It did not take long for the rates to fall back down to all-time lows though as the ECB attempted to stimulate the economy in the EMU area by providing cheap credit. 2015 saw a record low ECB marginal interest rate of 0.3% as the ECB flirted with the zero lower bound.
Foreign Direct Investment (‘fdi’) is recorded as net inflows of FDI as a percentage of GDP. This variable has already been reported on in chapter 2 and dwindled below 1% for the 1980s. It exploded from the onset of the 1990s, reaching a rate of over 25% by the turn of the millennium. As mentioned in chapter 2, 2015 saw a huge spike in the level of FDI:GDP, reaching above a level of 70%, more than double any previous or later levels. While a valid explanation is provided for this phenomenon in chapter 2, the fact that this is a significant outlier from the remainder of the data could cause issues from a data analysis point of view. Therefore, a dummy variable ‘fdispike’ has been included in the regression to control for this outlier. This dummy has a value of zero for each of the years in the regression except for 2015 which is assigned a value of one.
30 The rate of change in the female participation rate is included, having originated as the female participation rate (‘fparticipation’). This is the percentage of the female population, aged 15+, who are taking part in the labour market. Chapter 2 shows the huge
developments that took place with this rate over the course of the regression period and therefore it is a potentially important variable in the explanation of the employment rate. Population growth (‘popgrowth’) is included as the year-on-year percentage change in the whole population. The real exchange rate (‘realxrate’) is the real effective exchange rate – “a nominal effective exchange rate index adjusted for relative movements in national price or cost indicators of the home country, selected countries, and the euro area” (World Bank, 2018) – expressed on the base year of 2010. It is found to be significant in several the different unemployment studies detailed in section F of Chapter 3 so should be included in this regression to begin. It serves to provide an estimate of the international
competitiveness of the Irish economy over the years.
Finally, the interaction term (‘c.crisisdummy#c.fdi’) is included to answer the research question – whether employment created by FDI is stable or not. It is an interaction between FDI investment and the crisis periods. This variable will allow me to determine whether FDI-created employment is, in fact, more stable than the overall employment rate. The baseline recession variable is also included to begin with as it is likely that employment rates differ between periods of recession and the remainder of the period of the regression. The same can be said for the FDI variable which is included due to the likelihood of employment rates being higher when FDI is higher and vice versa. Not controlling for these could lead to omitted variable bias.
∆𝐸𝑚𝑝𝑙𝑜𝑦𝑚𝑒𝑛𝑡 𝑅𝑎𝑡𝑒𝑡 = 𝛽0+ 𝛽1. ∆𝑃𝑜𝑝𝑢𝑙𝑎𝑡𝑖𝑜𝑛 𝐺𝑟𝑜𝑤𝑡ℎ𝑡+ 𝛽2. 𝐶𝑃𝐼 𝑡+ 𝛽3. ∆𝐺𝑁𝐼𝑡+ 𝛽4. 𝐶𝑟𝑖𝑠𝑖𝑠 𝐷𝑢𝑚𝑚𝑦𝑡 + 𝛽5. 𝐸𝑢𝑟𝑜𝑡+ 𝛽6. 𝐸𝑀𝑈𝑡+ 𝛽7. 𝐹𝐷𝐼𝑡+ 𝛽8. 𝐹𝐷𝐼 𝑆𝑝𝑖𝑘𝑒𝑡+
𝛽9. ∆𝐹𝑒𝑚𝑎𝑙𝑒 𝑃𝑎𝑟𝑡𝑖𝑐𝑖𝑝𝑎𝑡𝑖𝑜𝑛𝑡+ 𝛽10. ∆𝐸𝑚𝑝𝑙𝑜𝑦𝑚𝑒𝑛𝑡 𝑅𝑎𝑡𝑒𝑡−1+
𝛽11. 𝐻𝑖𝑔ℎ𝑒𝑟 𝑇𝑒𝑐ℎ 𝑆𝑒𝑐𝑡𝑜𝑟𝑠𝑡+ 𝛽12. 𝐻𝑖𝑔ℎ 𝑇𝑒𝑐ℎ 𝐵𝑜𝑜𝑚𝑡+ 𝛽13. 𝑅𝑒𝑎𝑙 𝐸𝑥𝑐ℎ𝑎𝑛𝑔𝑒 𝑅𝑎𝑡𝑒𝑡+ 𝛽14. 𝐼𝑛𝑡𝑒𝑟𝑒𝑠𝑡 𝑅𝑎𝑡𝑒𝑡+ 𝛽15. 𝐶𝑟𝑖𝑠𝑖𝑠 𝐷𝑢𝑚𝑚𝑦𝑡. 𝐹𝐷𝐼 𝑡+ 𝜀𝑡
31
C) The Regression
The results of the regression (table 5) include all the variables detailed above and serves to explain the employment rate in Ireland. This will then provide the opportunity to investigate the stability of employment created by MNEs in Ireland. The high R-squared figure for the regression, at a level of 0.890, indicates that the data are close to the fitted regression line that the model explains a large amount of the variability of the response data around its mean. The relatively low level of the root mean-squared error, 0.669, also validates the quality of the fit. It can therefore be said that this concise model offers a good fit for the data.
Table 5: Regression Describing the Employment Rate in Ireland
Source: Own calculations based on CSO, World Bank & OECD data
_cons 2.094947 2.82596 0.74 0.469 -3.89582 8.085713 c.crisisdummy#c.fdi -.0555349 .058331 -0.95 0.355 -.179191 .0681213 realxrate -.0201735 .0189315 -1.07 0.302 -.0603064 .0199595 htboom 1.61303 .5848483 2.76 0.014 .3732075 2.852853 htsectors 1.555677 .6412781 2.43 0.027 .196228 2.915125 D1. .1051681 .1471215 0.71 0.485 -.2067155 .4170516 lemployment D1. .544254 .2819557 1.93 0.071 -.0534653 1.141973 fparticipation fdispike -2.045369 1.573365 -1.30 0.212 -5.380753 1.290015 fdi -.0024884 .0158862 -0.16 0.877 -.0361656 .0311888 emu -.3228689 .7581921 -0.43 0.676 -1.930164 1.284426 euro -.0143077 .5861223 -0.02 0.981 -1.256831 1.228216 crisisdummy .2027762 .5628489 0.36 0.723 -.9904102 1.395963 irate -.2283978 .1083571 -2.11 0.051 -.4581046 .0013089 D1. .0000649 .0000401 1.62 0.125 -.0000201 .0001499 gni cpi .1313302 .10987 1.20 0.249 -.1015838 .3642443 D1. .1061239 .4867105 0.22 0.830 -.9256564 1.137904 popgrowth D.emprate Coef. Std. Err. t P>|t| [95% Conf. Interval] Robust Root MSE = .66946 R-squared = 0.8904 Prob > F = . F(13, 16) = . Linear regression Number of obs = 32
32 Examining the regression above, only the two dummy variables representing the periods of sectoral patterns described by Barry (2007) are found to be statistically significant at the 5% level. Extending this to the 10% level of significance, the interest rate and the change in female participation also become statistically significant with p-values below 0.100.
Looking at the coefficients of the regression, it can be seen that the interest rate has a negative effect on the employment rate. This means that as credit becomes more
expensive, the growth of the employment rate falls. For each increase of 1% in the interest rate, the change in the employment rate falls by 0.228%. This makes sense from an intuitive point of view as businesses become more conservative when the cost of credit is high. With higher costs in one area, firms would look to cut costs elsewhere in order to maintain their competitiveness and one avenue for this is by laying off employees. This would, in turn, lead to a fall in the employment rate. The higher borrowing costs may also put a stop to planned investments that otherwise could have created further employment.
An increase in the female participation rate, on the other hand, has a positive effect on the dependent variable with a one unit increase in the change in the female participation rate leading to a 0.544% rise in the level of the growth of the employment rate. This means that, approximately, for every 1% rise in the number of females partaking in the workforce, the overall growth of the employment rate will rise by 0.544%. This makes sense from an
intuitive point of view as the more females partaking in the workforce, the more that can be employed and the major changes in this variable over the regression period, as outlined in chapter 2, mean that this relatively large coefficient value is of no surprise.
Both periods of sectoral patterns described by Barry (2007) have made positive
contributions to the growth of the employment rate. He described the first period where the country was shifting into higher-tech sectors, running from the beginning of the
regression until the end of the 1980s, as a period which had a positive effect on the growth of the employment rate, adding 1.556%. The next period, which saw a global high-tech boom and the acceleration of the offshoring of services, had a similarly positive effect on the dependent variable, with an extra 1.613% worth of growth added to the baseline. The successful transition of the Irish economy from the recession of the 1980s into the ‘Celtic Tiger’ period involved moving into higher-tech, higher-value sectors which led to substantial job creation in the country. The ‘Celtic Tiger’ period saw a further spike in employment as
33 the high-tech sector into which the Irish economy had just transitioned experienced a
boom. Ireland also benefited as a recipient of services from the global offshoring trend.
However, while not statistically significant, there are further variables which play an
important role in an economic sense and hence can be considered of economic significance in describing the dependent variable. A number of variables in the regression also share correlations which may dampen their modelled effect on the growth of the employment rate and maybe even their statistical significance too.
The crisis dummy is one of these variables as the employment rate differs between periods of economic recession and ‘normal’ and boom times. Removing the crisis dummy would wrongly attribute the fall in the employment rate that takes place around recessions to another variable or lead to a biased error term. The coefficient for the crisis dummy is quite surprising though, at a value of 0.203, which suggests that the growth of the employment rate increases during recessionary periods. This is obviously counterintuitive as the opposite would actually be expected. However, looking at the p-value of 0.723, it can be seen that the coefficient is highly insignificant. The 95% confidence interval again stretches well below zero which means that 0 cannot be precluded. There are also a number of other variables with which the crisis dummy could share correlations such as GNI – which would be expected to fall during times of crisis – and interest rates – which tend to spike in these periods. These correlations may explain the small positive coefficient for the crisis dummy variable.
FDI is another variable that is found to be statistically insignificant with a p-value of 0.877, far greater than the 5% or 10% level of significance. The FDI spike’s coefficient of -2.045 is also not significant from a statistical viewpoint, with a p-value of 0.212 exceeding the 5% and 10% significance levels. It suggests that the change in the employment rate was a large, extra negative effect as a result of the spike in the FDI:GDP ratio during the year 2015. The value of the FDI coefficient is also quite close to zero while the 95% confidence interval straddles zero which may be a surprising finding. It does have economic significance,
however, as the more FDI invested in Ireland in a year, the higher the expected employment rate. For this reason, I have run some correlation testing between the FDI variables and the time-dummy variables that are already included to investigate whether these dummies may capture some of the effects of the FDI level.
34
Table 6: Correlations with FDI
Variable Correlation with FDI
Member of the euro 0.636
Member of the EMU 0.519
Barry’s higher-technology sectors -0.406
Barry’s high-tech boom -0.306
Barry’s R&D offshoring 0.607
Source: Own calculations based on CSO, World Bank & OECD data
Pearson’s correlation coefficient for the correlation between both FDI and membership of the euro, and FDI and EMU membership are positive and fairly strong, as is the case with Barry’s final period of research and development offshoring. Intuitively, this makes sense as each of these three dummies have a starting value of zero and end the regression period with a value of one, each one kicking into play between 1994 and 2000. As was previously analysed in chapter 2, FDI also followed a growth trend from the beginning of the period until the turn of the century before levelling out over the next decade and finally growing again from 2011. Therefore, using both the statistical and intuitive logic, it is evident that some of the effect of FDI on the growth of the employment rate may be contained in these three dummies.
The first two periods of Barry’s sectoral patterns have negative correlations with the FDI variable but neither have particularly large absolute values at 0.406 and 0.306 which means that their impact can be ignored. The other three dummy variables listed above have much larger Pearson correlation coefficients with the FDI variable though. Both the euro and EMU dummies have small negative coefficients, at -0.014 and -0.323 respectively, which, coupled with their positive correlations with FDI, would suggest that FDI has a very small extra negative effect on the growth of the employment rate than the FDI coefficient alone
suggests. However, the coefficient for Barry’s final period dummy has a much more positive implication for the effect of FDI on the growth of the employment rate. There is no labelled coefficient for this period so as to avoid the dummy variables trap but this means that the effect of the final variable is included in the intercept. With a positive correlation coefficient
35 of 0.607 suggesting that a substantial degree of the effect of the FDI variable is contained in Barry’s final period dummy, the intercept coefficient of 2.095 indicates that FDI in fact has quite a large positive effect on the growth of the employment rate. While the 95%
confidence interval is quite wide and does reach below zero, the presence of the coefficient above zero and the upper bound having a far greater absolute value than the lower bound implies that a large amount of the positive effect of FDI on the dependent variable is contained in the intercept coefficient.
There is also an argument that many types of FDI do not bring much employment. This argument is founded on the idea that high-tech FDI can involve large monetary investments but may not be labour-intensive due to the technological processes involved. This type of MNE investment, which dominates the FDI landscape in Ireland, creates far less
employment than FDI related to manufacturing that would require a large human capital outlay. With technological giants such as Google, Apple and Hewlett Packard making up a huge amount of the FDI invested in Ireland, there is plausible weight to the argument that FDI in Ireland does not actually create much employment and this may impact the
unexpected coefficient for FDI. There is also the case of tax-inversion deals where
companies reincorporate abroad by having an Irish-owned company purchase its currents operations. This allows it to reduce their tax burden through the relocation of their tax domicile which means that FDI increases without leading to much of an effect on employment.
The real exchange rate also has a negative effect on the dependent variable as an increase in one unit of the real exchange rate leads to the change in the employment rate dropping by 0.202%. This makes sense from an intuitive point of view as an increase in the real
exchange rate means a reduction in the level of Ireland’s competitiveness and therefore the country’s exports are hit, leading to a lower level of employment.
The growth rate of GNI has a very small impact on the growth of the employment rate according to the regression, with a coefficient of 0.000. While its p-value exceeds the 10% level, it still is relatively low at just 0.125. Again, it looks like some of the effects of GNI on the employment rate are absorbed by other variables as was suggested in the case of the crisis dummy earlier. A positive effect of GNI on the employment rate would be expected as
36 greater levels of GNI indicate times of economic prosperity which, in turn, suggest higher levels of employment.
The inflation rate is another variable that is found to be statistically insignificant. With a coefficient of 0.131, it implies that a higher level of inflation leads to an increase in the amount of people employed in the country. This is believable due to the fact that the inflation rate was quite low during the regressionary period, particularly after the first three years where it only breaks above the 5% level on one occasion. With such a confined range for inflation, it can then be argued that periods of very low or negative inflation can be considered recessionary periods while periods of higher inflation suggest economic prosperity. This analysis then suggests that CPI is another indicator of the economic
performance of Ireland and therefore it will share features with other explanatory variables such as the growth rate of GNI and the crisis dummy variable. The high levels of inflation in the first three years of the period also will mean that there will be a degree of correlation between it and the first of Barry’s periods of sectoral patterns which has a value of one for the length of the 1980s.
The change in the population growth variable has a p-value of 0.830 that means that it is quite a distance from any level of significance and its very small coefficient of just 0.106 implies that it has no real effect on the growth of the employment rate. It makes sense from an intuitive point of view that its influence on the dependent variable is negligible as an increase in the population growth rate does not necessarily mean that more people become available to work. Often, the increase in the population growth comes from a baby boom, as was explained in chapter 3, which means that it would take approximately two decades before this population growth has an impact on the employment situation.
Membership of the EMU and the euro are also quite far from any level of statistical significance with p-values of 0.676 and 0.981 respectively. Both dummies have small negative coefficients that imply membership of both the EMU and the euro led to a fall in the employment rate but neither have a large enough effect to analyse. These dummies also share correlations with a few of the other explanatory variables, most especially each other, as they come into play as the ‘Celtic Tiger’ period is beginning which therefore dilutes their effect on the dependent variable.
37 Next, the lagged value of the growth of the employment rate is found not to be statistically significant, with a p-value of 0.485 according to the regression above. However, it does have an economic and intuitive relevance in the regression as it represents trends in the changing of employment rates over the years. While it may have been expected to be more
significant and even slightly larger, its coefficient value of 0.105 means that a growth of 1% in last year’s employment rate means that an extra 0.105% is added to this year’s growth. This positive effect though would be expected as periods of economic growth and economic decline tend to last for a few years in a row and therefore a positive year last year would indicate a better chance of a positive year this year. As with many of the other variables, some degree of collinearity would exist here as this explanation clearly describes trends in the economic landscape that are already a feature of other independent variables such as the crisis dummy.
The constant, β0, is again not statistically significant at the 5% or 10% confidence levels in
this equation with a large p-value of 0.469. Its coefficient of 2.095 is quite a distance from zero which suggests that the baseline is positive meaning that if all other variables have a value of zero, the growth of the employment rate would be 2.095%. This indicates the positive, but not significant, trend in the growth of the employment rate since the beginning of the regression that was touched upon in chapter 2 when the unemployment rate was explored.
Finally, in order to answer the research question, the coefficient of the most importance is that of the interaction term. The interaction term is required to determine the impact of MNE investment on the stability of employment in Ireland. The interaction term is a
combination of the FDI and crisis dummy variables and can be interpreted as the additional effect of FDI on employment during a crisis. The coefficient of this term is -0.056 with another insignificant p-value of 0.355. Adding this coefficient to the crisis dummy
coefficient, the total effect of the crisis on employment can be determined and a very small positive overall effect is observed at just under 0.200. This suggests that crisis periods did not take away employment but from a simple observation of the data it is evident that this is not the case. Therefore, I look to the insignificance of both the crisis dummy and the interaction term to explain this contradiction along with the collinearity between these two variables and a number of the other explanatory variables contained in the regression.
38 These factors clearly combine to explain the negative effect of crisis periods on the
employment rate that is not visible from the direct coefficients here.
Looking at the interaction term from an isolated point of view, its insignificant and minute negative coefficient means that FDI has a negligible approximate addition to employment during a crisis. This suggests that movements in FDI are not employment-disturbing factors during crises. FDI does not appear to have any particular effect on the stability of
employment during recessionary periods and instead it looks like MNE-created employment behaves in the same manner as the overall employment in the economy.
As can be seen from the recent developments in Ireland described in chapter 2 and informed by the literature review in chapter 3, there are many factors that attract MNE investment to Ireland. Nevertheless, it is now clear that this type of investment has no real effect on the stability of employment during recessionary periods. While these factors will entice companies to Ireland and, in doing so, create employment in the country, the new jobs created will not be any more secure than the rest of the employment in the economy during recessionary periods.
My results concur with the empirical findings of Godart, Görg & Hanley (2011), discussed in section B of chapter 3 when they report that there was no difference found between the probability of an Irish-owned and a foreign-owned company leaving Ireland during the GFC. While they do acknowledge the reliance Ireland has on foreign multinationals, their analysis finds that MNEs and comparable Irish firms do not act any differently in terms of exit
behaviour. They also admit that foreign multinationals do not add any additional instability to the economy. These findings would be consistent with the indifference I have found between the stability of MNE-created employment and the overall employment in the economy.
However, the findings of my thesis do run in contradiction to some of the literature mentioned in section B of chapter 3 too. Both Lane (2014) and Godart, Görg & Hanley (2011), in their opening background section, suggest that MNE investment is more stable than regular investment in the economy. Their arguments are based on the fact that MNEs predominantly rely on external, international sources for funding and based on the common use of Ireland as a base for manufacturing export goods and services which both mean that
39 the performance of MNEs is not solely reliant on the domestic economy. They also propose that the sunk costs and the longer-term focus that comes with an MNE setting up in a country would imply a greater stability to the employment they create. However, my empirical research has found no difference in this level of stability.
40
5. Conclusions
In modelling the determinants of the growth of the employment rate in Ireland, several explanatory variables were included with only a few found to be statistically significant. At the 10% level, the interest rate, the change in female participation in the labour force and both of Barry’s periods of sectoral patterns were found to be statistically significant. A number of other variables were found to have economic and intuitive significance in the explanation of the dependent variable and a regression of fifteen variables was constructed.
In addressing the research question – “Is employment created by MNEs in Ireland stable?” – it appears that employment created by MNEs in Ireland is no more stable than the total employment in the country. Periods of economic recession were examined, and it was found that changes in FDI during crisis periods are not employment-disturbing factors nor do they affect the change in the employment rate.
However, it is indisputable how important MNE investment is for employment in Ireland. Below is a table detailing the size of the labour force and the total number of people
employed in Ireland over the past three years. The importance of FDI for employment in the country can be seen here with the amount of jobs in the country created by and dependent on MNE investment. With the help of the IDA, the amount of jobs created from FDI sources over the past three years has amounted to over 23,000 (IDA, 2018). With the total number employed in Ireland rising by just over 145,000 in this same period, nearly 16% of the overall growth has stemmed from FDI sources. This underlines the role FDI investment has had in the recovery from the recent GFC and demonstrates just how important it is to Ireland.
41
Table 7: Labour Force Statistics
Category 2017 2016 2015
Labour Force (‘000) 2,374.8 2,331.1 2,292.2
Total Employed (‘000) 2,230.8 2,163.5 2,085.4
Total Employed in FDI (‘000) 210.4 199.9 187.1
Total Employed in FDI (% of Labour Force) 8.9 8.6 8.2
Total Employed in FDI (% of Total Employment) 9.4 9.2 9.0 Source: Central Statistics Office Statbank & IDA Ireland
A number of opportunities and threats exist today in relation to FDI flows worldwide that are outside of Irish control. It is clear that these will play a major role in the performance of the Irish economy over the coming months and years due to the intrinsic relationship between the Irish economy, international trade and FDI. As was illustrated in chapter 2, FDI makes up a very significant portion of Irish GDP and it is evident how important it is to the Irish economy. To underline the complexities of the situation, I will briefly examine one of these opportunities – that of Brexit – but while also acknowledging the potential threat it poses to Irish FDI flows at the same time. On the other side, I will describe how the looming global trade war – one that seems to be worsening by the day as I write this chapter – and American tax reform may have dire effects for the Irish economy.
The issue of Brexit, from an economic point of view, very much has two sides to the
argument. On one hand, if Britain were to exit both the EU and the European Single Market (ESM), Ireland would be stripped of one of its largest trading partners which could
potentially involve Irish FDI taking a severe hit. While the Irish government has worked hard over the past number of years to loosen its reliance on the UK for trade purposes,
expanding trade to new EU members in particular and further afield, the UK still receives 13.8% of Ireland’s exports (CSO, 2016), and remains Ireland’s second largest export partner (BBC, 2017). Imported goods from the UK also make up over 25% of total Irish imports. The lack of a free trade agreement would make Irish exports more expensive to UK customers and would likely mean they would look elsewhere to source a large portion of these goods and services. This would make Ireland a far less attractive destination for MNE investment
