I would like to thank two anonymous reviewers and the editors of the journal for useful comments and suggestions. All remaining mistakes are my own.
The views expressed in the article are my own, and do not necessarily reflect those of PPMI
- 1 I consider CEECs to be 11 of the member states that have joined the European Union (EU) since 2004: (...)
1This article poses the question of how incentive investment policies, particularly in the form of state aid, helped drive openness to foreign direct investment (FDI) inflows among the Central and Eastern European Countries (CEECs)1. FDI is a key tenet of the economic growth policies of these countries, and has contributed to the region’s integration into European and global markets. Between the fall of the Soviet Union and the 2008 economic crisis, foreign investment flows in the CEECs grew at a pace faster than both developing and developed countries (Figure 1), while its share of global stock rose from 0.6% to 3.5% (Szent-Iványi, 2017, p. 2). This has led several authors in the comparative capitalism literature to talk about dependent (Nolke & Vliegenthart, 2009), Porterian (Drahokoupil, 2009, i.e., based on high-quality industrial investment), or FDI-based and -led market economies (Bohle & Regan, 2021; Myant & Drahokoupil, 2011).
Figure 1: FDI openness in the CEECs compared to other world regions (1990-2018)
Source: Author's elaboration based on UNCTAD data. (Developed countries in Europe, America, Asia and Oceania; developing countries in Africa, America, Asia and Oceania)
2The research question this article seeks to answer is borne out of two seemingly contradictory findings in the literature. On the one hand, qualitative case studies on the CEECs’ investment policies often emphasise the importance of subsidy packages in attracting foreign investors (e.g. Medve-Bálint, 2014; Tőrös et al., 2017); on the other hand, quantitative analyses offer a mixed record in support for government policies as a tool of investment promotion and attraction (e.g. Morisset, 2003; Oman, 2000).
3This study takes advantage of the high-quality data provided by the European Commission to distinguish the effect of different types of aid on FDI openness, and the policy trade-offs that they entail. While state aid is an important element of a government’s toolkit to achieve a broader economic agenda (see Schito, 2021a), this policy is hardly monolithic, and can be employed for contradictory goals, differently benefitting domestic producers and foreign investors. Here, I offer a novel analysis of the interplay between two typical tools of the CEECs’ economic growth policies – state aid and FDI – and of the circumstances under which state aid becomes part and parcel of their investment policies.
4The CEECs constitute an ideal regional focus due to their peculiar political economies. As Bandelj (2008, 2009) notes, this is a region where economic activities of this kind are recent and have come into existence de novo as the very pillars of a broader market-creating process. Following the massive sell-offs due to privatisations, foreign ownership in the CEECs is far higher than in the rest of the EU member states (Eurostat, 2016). For its part, state aid is an encompassing policy tool that can be employed as an effective investment incentive on multiple fronts: it covers not only cash grants, but also financial and fiscal incentives such as tax reliefs, guarantees, soft loans and equity subsidies, thus giving policy-makers leeway on how to best entice foreign investors (Buigues & Sekkat, 2011).
5In the next sections, I first review the literature on FDI and investment incentive policy, from which I formulate the hypotheses to test. I then present the data and the methodology, and carry out the statistical analyses. The findings show that the policy trade-offs of the different state aid tools play out on two levels. First, while regional aid may indeed be associated with more FDI openness, other types of state aid measures are less likely to be seen as attractive to foreign investors. Second, changes in the overall composition of aid expenditures, as required by EU competition policy have radically altered the state aid-FDI openness nexus, suggesting a declining importance of investment measures for the CEECs’ economic growth policies. Hence, the article uncovers important contradictions both between the differing goals of state aid and investment measures for the governmental economic agenda, and between the different levels of governance at which such policies are regulated and implemented, which are discussed in the conclusion.
- 2 Negative effects include potential over-dependence on multinational corporations, which reduces the (...)
6FDI is investment made by a company in a foreign (host) country and in which the company assumes control of at least 10% of the host firm (UNCTAD, 2009, p. 35). FDI leads to a long-term relationship between the foreign company and the host country, reducing the speculative purposes of the investment and fostering a potentially long-term relationship between the two. Through FDI, the firm gains a new growing market, usually associated with lower labour costs and much economic potential to untap, while the host country reaps the benefits of the inflows of capital, higher R&D&I (research and development and innovation) output and positive spill-overs, leading to higher competitiveness and economic development2.
7Among the CEECs, inward FDI policy started picking up pace in the immediate post-socialist decade of transition mainly thanks to mass privatisation programmes. Although such programmes were pivotal to creating a market economy, several governments were still reluctant to open up their gates, preferring instead strategies of national capital accumulation (Bandelj, 2008; Drahokoupil, 2009; Meyer, 1995). Starting from the late 1990s and early 2000s, as this programme had come to an end, an EU-driven process re-oriented CEECs-bound FDI by favouring greenfield investment to establish industrial plants, which became the main engine of growth, particularly among the Visegrad-4 (Drahokoupil, 2009, p. 47; Meyer & Jensen, 2005). This process was possible thanks to the backing of the EU, who wanted to eliminate as many barriers as possible within the Single Market, thus inviting a large swathe of Western European investors into the CEECs (Medve-Bálint, 2014, 2015). The final period, marked by the 2008 economic crisis, is instead characterised by a decline of importance in FDI as the main driver of economic growth (Galgóczi et al., 2015; Szent-Iványi, 2017), which is being replaced by EU structural funds as a source of foreign financing (Bohle, 2018). Ever since, as Figure 2 shows, in none of the three main regions in which the new member states are situated – the Baltics, the Visegrad-4 and Southern Eastern Europe (SEE) – has the level of FDI inflows (as a share of %GDP) come back to the heights of the mid-2000s.
Figure 2: FDI openness in the CEEC regions (2000-18)
Source: Author's elaboration based on data from national banks
8In order to attract FDI as part of their economic growth policy, host countries’ governments may deploy several incentives. These can be financial, such as outright grants and loans; fiscal, which mostly concern the lowering of corporate tax rates or tax deferrals/holidays; or they can take the form of subsidised infrastructure and services or exemption from certain national environmental or labour regulations (UNCTAD, 2003, p. 123). Among the CEECs, all have, at one point or another, employed investment incentives (see Anton, 2017; Cass, 2007; Éltető & Antalóczy, 2017; C. Jensen, 2006; Tőrös et al., 2017). For instance, Czechia amended its 1998 investment laws multiple times between 2012 and 2021, as did Bulgaria; Romania recently amended its scheme of investment support for promoting regional development; in 2018 and 2019, Slovakia and Poland also introduced laws to promote investment3.
9In the EU, such incentives fall for the majority under the banner of state aid. State aid is regulated by the Directorate General for Competition (DG COMP) as part of a long-standing European tradition that puts competition policy at the heart of the functioning of the Single Market (Schito, 2021b). However, not all state aid has the same effect on competition. The political science literature often distinguishes between ‘good’ and ‘bad’ aid (Blauberger, 2009). The former, often also called ‘horizontal aid’, does not discriminate between the sectors in which undertakings operate and aims to achieve common policy objectives that are in line with the agenda of the European Commission (e.g. the Lisbon Agenda, the Europe 2020 Strategy, and most recently NextGenerationEU)4, such as smart, sustainable and inclusive growth, or climate and digital transition. This type of aid is aimed at generating positive spill-overs in terms of R&D&I, environmental protection, or employment, for instance, and is therefore more likely to be approved. The latter, known as ‘sectoral aid’ is thought to be most distortive of competition because it is often aimed at rescuing firms in difficulty or non-profitable sectors (e.g. shipbuilding, steel, coal), and is therefore ‘bad’ due to the limited possibility of positive spill-overs. During the transition period, sectoral aid in the CEECs has been described as ‘unplanned’, ‘reactive’, and ‘with strong crisis management features’ (Hölscher et al., 2014). A third type of aid, called ‘regional aid’, is usually regrouped under the horizontal banner, but it is somewhat more restrictive in terms of geography, since it only applies to the less economically developed regions of a member state (Schito, 2021b).
10Today DG COMP identifies 14 main objectives. These are: regional development, environmental protection including energy savings, R&D&I, training, sectoral development, SMEs (small and medium enterprises) and risk capital, culture, employment, social support to individual consumers, heritage conservation, R&R (rescue and restructuring), compensation of damages caused by natural disasters, promotion of export and internationalisation, and closure aid. Of these, regional development goals can be understood as being regional aid; R&R operations and sectoral development are closest to our understanding of sectoral aid, whereas the remaining objectives can be categorised as horizontal aid (Schito, 2021b).
- 5 See Commission Guidelines on Regional State aid, Communication 2021/C 153/01, para 178-188 on maxim (...)
- 6 Own calculations using data from DG COMP.
- 7 Figures A1-A4 in the Supplementary Material show the trend by country for the three different types (...)
11An important peculiarity of regional aid is that there exist exceptions to state aid law that allow for higher ceilings in state aid allocations when the measure is aimed at promoting the development of an economically backward area5. Since the level of GDP per capita is much lower in the CEECs than their Western counterparts, state aid is often justified as contributing to the country’s regional development (Medve-Bálint, 2014). Indeed, regional aid reaches peaks of over 70% of total aid allocations among the CEECs (71.1% in the Baltic countries, 73.1% in the SEE countries, and 70.7% among the Visegrad countries) and averages 27.1% of the total in the 2000-18 period (21%, 23.5%, and 34.4% respectively), compared to 17.7% in the EU-156. Éltető and Antalóczy (2017) further suggest that EU membership may have forced the CEECs to increase their regional aid allocations to cope with the more competitive economic environment. The numbers seem to support this claim: since joining the EU, average regional aid expenditures rose to 22.2%, 24.4%, and 37.3% of the total allocations for Baltic, SEE, and Visegrad countries respectively7.
- 8 The subsidy package was given in the form of regional aid in accordance with Art. 107.3(a) TFEU, se (...)
- 9 Respectively, case SA.28405 and case SA.22246.
12This put the CEECs in a strong position to attract FDI by using regional development policies. As an example, when Jaguar-Land Rover (JLR) was venue-shopping for a new automotive plant, it eventually chose Slovakia over Canadian, Polish, and Austrian sites thanks to the numerous advantages this country offered: tariff-free access to the EU, relatively low-wage workers, strong supply base, low-cost logistics, and upgraded infrastructure (Jacobs, 2017). Most of all, the Deputy Prime Minister of Poland, Slovakia’s most direct rival for the investment, conceded that Slovakia had offered an amount of state subsidies that his country was not willing to match (Reuters, 2015)8. Likewise, Ślusarczyk and Kot (2012) show that between 2004 and 2010, regional state aid allocations in Poland increased over five-fold, going almost hand-in-hand with increases in FDI. For instance, regional aid played an important role in attracting the French IT multinational Atos and the Japanese Orion Electric to Poland. In both cases, the rationale behind the aid was the development of the Kujawsko-Pomorskie region, which at the time suffered important socio-economic handicaps and high unemployment rates9.
13Further, the European Commission legitimised investment competition policies by approving most of the incentive schemes and by setting regional state aid ceilings to be compatible with EU regulation (Medve-Bálint, 2015, p. 75). Thus, if shifting state aid allocations towards regional development is a strategy conducive to a more aggressive inward FDI policy, a first hypothesis is:
H1a: Higher percentages of regional aid over total allocations lead to more openness to FDI in the CEECs.
14Yet, the use of government policy to foster FDI has found contradictory evidence in the economic policy literature (Morisset, 2003; Oman, 2000). Among the CEECs, Carstensen and Toubal (2004) find that lower levels of corporate tax rate led to higher openness to FDI. Beyer (2002) and Bevan et al. (2004), instead, note that the development of FDI was detached from government incentives and competition policy. Rather, as Bandelj (2009, 2010) shows, it was the establishment of investment promotion agencies (IPAs) that contributed the most in accelerating investment during this uncertain period. Cass (2007) further suggests that the different roles that IPAs take on impinges on their effectiveness in generating investment.
15Fewer studies have focused on the period since the 2000s. These are mostly based on case studies of a limited number of countries in the region and in which the different types of available incentives are described and analysed (e.g. Anton, 2017; Éltető & Antalóczy, 2017; Tőrös et al., 2017). In several cases, regional aid is mentioned explicitly, but little evidence exists with regards to the effectiveness of regional aid schemes (Drahokoupil & Galgóczi, 2015, p. 34; Tőrös et al., 2017, p. 214). Although these findings are mostly limited to the Visegrad countries, given the commonalities in FDI and state aid trends across the different regions in Central and Eastern Europe as expounded above, it is reasonable to assume that they may apply beyond the Visegrad-4. Thus, an opposite hypothesis on the use of state aid as part of the country’s investment policy whereby such measures fail to crowd-in FDI is:
H1b: Shifts towards regional aid have no significant effect on the CEECs’ openness to FDI.
16The other two types of aid – horizontal and sectoral – are less likely to be employed to attract foreign investment. Horizontal aid is often used to achieve governmental policy objectives such as improving green technologies or decreasing unemployment (Schito, 2021a). Sectoral aid, for its part, is often aimed at redressing the consequences of technological developments in favour of unproductive firms (Blauberger, 2007). This type of aid, therefore, can work as a safety net for the losers of globalisation, thus taking on a protectionist hue (Hwang & Lee, 2014). For instance, prior to joining the EU, Poland and the Czech Republic undertook important restructuring measures concerning shipyards and the banking sector, respectively, in order to avoid their bankruptcy in anticipation of the competitive pressures that EU membership would entail (Blauberger, 2007). A more recent example involves a €96 million loan to a shipyard firm in Croatia in 2019, with the aim of safeguarding 1,800 jobs after the firm posted continuous financial losses as a result of the 2008 economic crisis10.
- 11 Case SA.34462.
- 12 Case SA.38454.
- 13 More exactly, increases in horizontal (sectoral) aid crowd out the sum of regional and sectoral (ho (...)
17In other cases, aid earmarked for sectoral development may be employed regardless of the forces of globalisation, but simply as a way to facilitate the development of certain economic activities or of certain economic areas, as per Article 107.3(c) TFEU. In March 2012, the Latvian authorities notified the Commission that they wanted to undertake measures aimed at promoting cultural heritage conservation, with an estimated €1.5 billion project11. More recently, the Commission used the same legal basis to allow aid granted by the Hungarian authorities for the construction of two nuclear power plants12. Thus, though for different reasons, horizontal and sectoral aid should correlate with FDI negatively, since higher percentages of these two types of aid necessarily crowd-out the third type, regional aid13. Hence:
H2: Higher percentages of horizontal or sectoral aid over total allocations lead to less openness to FDI in the CEECs.
18These hypotheses will be tested by means of regression analysis. This section first describes the variables that will be included in the statistical model and how they are operationalised, starting with the dependent (FDI openness) and independent (state aid types) variables, before introducing several political and economic controls. Table 1 summarises these data, including a short description and a data source for each variable. The last part of the section presents the estimation procedure that will be employed for analysis.
- 14 Some scholars calculate the relative importance of FDI to a host country in terms of population, si (...)
19The dependent variable is FDI inflows over gross domestic product (GDP) for each year. This variable measures a country’s openness to, or reliance on, foreign capital, but not the amount of FDI inflows (Li, 2009)14. FDI/GDP ranges from a minimum average of 1.8% in Slovenia to a maximum of 7.5% in Bulgaria for the period 2000-18, with a peak of 27% in Bulgaria in 2007 and a disinvestment of 11% in Hungary in 2015.
20The main independent variable is the type of state aid (regional, horizontal, sectoral) measured as the percentage of total aid for each year. The three types are exclusive, meaning that increases in one type necessarily lead to decreases in the other two. Since not all state aid is equally likely to be reserved to foreign investors, this operationalisation allows to test the relative importance of regional aid (which should be more likely to attract foreign investment) compared to horizontal and sectoral aid (which is less likely to be employed for such purposes). In almost all the countries under consideration horizontal aid is the most common type, although some exceptions exist. In Croatia, the shipbuilding industry is still an important recipient of R&R aid, and in Latvia high levels of sectoral aid were influenced by the 2012 scheme to promote culture. In the Visegrad countries, on the contrary, since the main goal has been to attract investment in complex manufacturing, regional aid has often matched expenditures for horizontal objectives.
21I control for several political and economic factors, aimed at providing alternative explanations to variation in FDI openness. First, N. M. Jensen (2006) notes that the market size, trade openness and level of economic growth of the host country should be positively associated with FDI openness, since these factors determine the economic potential that multinational corporations can untap. High debt levels should also make a country more open to FDI because of its need to alleviate the fiscal pressure. This was exactly what happened in Hungary in the early 1990s when mass sell-offs to foreign investors were undertaken as a response to large levels of debt (Stark & Bruszt, 1998, p. 54). Two other important economy-related features are also included. One is the level of corporate taxation, which, in some CEECs (e.g. Hungary, Lithuania, Czechia) is among the lowest in Europe. Despite the mixed evidence, low levels of corporate taxations are in some cases found to be important incentives for a host country’s inward FDI policy. The second is the relative unit labour cost (RULC) compared to the eurozone countries. Low RULCs constitute an important comparative advantage for these countries: decreases in RULCs improve the country’s competitive position relative to their trading partners in the euro area, which may push investors from these countries to choose CEECs for their greenfield investment over other locations (Carstensen & Toubal, 2004).
- 15 EU membership has been shown to foster FDI openness thanks to both the ideational (i.e., market com (...)
22A parallel strand in the FDI literature further suggests that political stability, the protection of property rights and lack of corruption make host countries more attractive to multinational corporations (N. M. Jensen, 2003, 2006; Li & Resnick, 2003; Wei, 2000). This is because, since FDI involves important fixed costs at the beginning and spans several years, foreign investors may want to ensure that their investment is not compromised by sudden changes in the political environment, or corrupt bureaucrats and politicians. Part of the literature on FDI further suggests that the ideological leaning of the government matters. According to Pinto (2013), ‘the left loves FDI’ and vice-versa because of the distributional consequences of FDI, which is more likely to foster labour demand, thus catering to the constituencies of labour-based parties. Although appealing in theory, this is less likely to apply to the CEECs due to their communist past, which meant that the main reformers were centre-right proponents such as Estonia’s Mart Laar, Hungary’s József Antall and Slovakia’s Mikuláš Dzurinda, just to name a few examples. A final institutional feature that may be relevant is the electoral system. For instance, Garland and Biglaiser (2009) find that in Latin America, another region with several developing countries, FDI is more likely to favour candidate-centred electoral systems, since it is here that narrow interests can more easily champion incoming FDI to respond to local demands. As a result, more proportional electoral systems are less likely to push the host country to open up to FDI15. Table 1 below provides descriptive statistics, together with a description of and sources for each variable.
Table 1: Descriptive statistics and sources
Variable
|
N
|
Mean
|
St. Dev.
|
Min
|
Max
|
Description
|
Source
|
FDI openness
|
202
|
4.29
|
4.24
|
-11.42
|
27.9
|
Net FDI inflows as %GDP (in euros)
|
Several national bank websites
|
Regional aid
|
181
|
27.52
|
18.72
|
0
|
73.10
|
Regional aid as %total aid
|
Own calculations from DG COMP (2021)
|
Horizontal aid
|
181
|
46.14
|
26.81
|
1.48
|
95.77
|
Horizontal aid as %total aid
|
Own calculations from DG COMP (2021)
|
Sectoral aid
|
181
|
26.34
|
28.19
|
0
|
96.97
|
Sectoral aid as %total aid
|
Own calculations from DG COMP (2021)
|
Trade openness
|
197
|
114.31
|
32.34
|
54.02
|
183.99
|
Total trade as %GDP
|
Armingeon, Engler and Leemann (2020)
|
Real GDP growth
|
208
|
3.23
|
4.06
|
-14.81
|
11.89
|
Growth of real GDP, %change from previous year
|
Armingeon et al. (2020)
|
log(GDP)
|
208
|
10.77
|
0.98
|
8.6
|
13.06
|
Logged GDP in million euros
|
Eurostat (2021a)
|
Debt/GDP
|
209
|
41.96
|
24.01
|
6.65
|
102.77
|
Gross general government debt as %GDP
|
Armingeon et al. (2020)
|
Left government
|
208
|
32.74
|
32.71
|
0
|
100
|
Cabinet posts held by left-wing parties as %total cabinet posts
|
Armingeon et al. (2020)
|
Disproportionality
|
208
|
6.51
|
3.18
|
1.2
|
17.37
|
Index of electoral disproportionality according to
Gallagher (1991)
|
Armingeon et al. (2020)
|
Corporate tax rate
|
209
|
20.21
|
5.56
|
9
|
40
|
Corporate tax rate
|
Tax Foundation (2021)
|
RULC
|
209
|
2.69
|
7.39
|
-14.5
|
63.1
|
Relative Unit Labour Cost of the CEECs relative to their eurozone parties. Change in % from previous year
|
Eurostat (2021b)
|
Property rights
|
209
|
1.22
|
0.49
|
-0.07
|
2.19
|
Degree to which the state owns and controls capital (scale 0-4)
|
Coppedge et al. (2021) and
Pemstein et al. (2021)
|
Political Stability
|
187
|
0.68
|
0.31
|
-0.38
|
1.3
|
Perception of the likelihood that the government in power will be destabilised (scale -2.5 to 2.5)
|
Coppedge et al. (2021) and
Pemstein et al. (2021)
|
Corruption
|
198
|
48.23
|
7.63
|
24.78
|
60.54
|
Bayesian index of perceived corruption as abuse of public power for private gain (scale 0-100)
|
Standaert (2015)
|
Notes: all right-hand side variables are lagged by one year; 2000-18 sample
Source: The Author
- 16 Much of the econometrics literature suggested that ECMs could only be employed when cointegration i (...)
23To estimate the effects of these covariates on the sample countries’ openness to FDI, I employ a time-series-cross-sectional (TSCS) regression on a panel dataset ranging from 2000 to 2018 for 11 CEECs. Since the hypotheses suggest that inward FDI policy will respond to changes in the composition of state aid allocations, the most appropriate statistical model of choice is the so-called Error Correction Model (ECM, see De Boef & Keele, 2008). The ECM assumes that the dependent variable is in an equilibrium relationship with the independent variables, but that when these variables change, it produces an ‘error’ in the dependent variable, which is corrected over time either as a new equilibrium is reached or as the shock fades (Ehrlich, 2007, p. 592). Hence, short-run changes in Υt are a function of both short-run changes in Χt and how much Χt and Υt were out of equilibrium in the previous time period (Beck & Katz, 2011). The ECM allows to directly estimate the rate at which changes in order to return to equilibrium after a change in Χt (De Boef & Keele, 2008, p. 189)16. The functional general form of the ECM is:
24where the dependent variable is the rate of change, as evidenced by the Δ; the λ coefficients are first differences in the right-hand side variables; and the β coefficients represent lagged variables β0 is the coefficient for the lagged dependent variable (LDV) and is known as the error correction coefficient, representing the rate at which equilibrium is restored (Keele et al., 2016), whereas βkΧ is a vector of different control variables. While β captures the effect of a once-off change in a variable dispersed across future time periods, λ measures the immediate effects of such changes (Rickard, 2012, p. 1177). The total long-run effects (also called long-run multiplier, LRM) are represented by the ratio between the lagged independent variable of interest and the negative of the lagged dependent variable (De Boef & Keele, 2008). While the literature on the politics of FDI often assumes that there exists a temporal gap between government policy in terms of incentives and foreign investors’ responses, the ECM allows to explicitly test the dynamics of such a lag to estimate how quickly shifts in state aid policy are followed by changes in FDI openness.
25For each type of aid (regional, horizontal, and sectoral), I estimate two models with different fixed effects: one is the standard country-fixed effects and the other is region-fixed effects (the three regions are the SEE, the Baltics and the Visegrad-4), since the literature suggests that these regions have opened up to FDI in very different ways (Bohle & Greskovits, 2012). Diagnostics tests also confirm that there is a significant difference between pooled and country-grouped estimates. Further tests suggest that neither serial nor cross-sectional correlations are present, but also that the data-generating process may suffer from some heteroscedasticity, so I estimate all equations using panel-corrected standard errors (PCSE, see Beck & Katz, 1995). The general information matrix (GIM) test devised by King and Roberts (2015) also shows little to no model misspecification for all three main independent variables using the baseline models, thus suggesting that the use of robust standard errors will not severely bias the results. Finally, the variance inflation factor (VIF) shows very low levels of multicollinearity (VIF < 3 for all three independent variables). The next section presents the results of the empirical tests.
26Table 2 presents the main results for regional, horizontal and sectoral aid using the ECM. The table shows that, in line with H1a, shifts towards more regional aid are most likely to foster shifts in FDI openness. One standard deviation of a once-off change in regional aid (i.e., the differenced variable) is responsible for an immediate effect on FDI openness of 18.6% in Model 1 and 20.6% in Model 2. This same once-off change in one standard deviation in regional aid dispersed across future time periods (i.e. the lagged variable) amounts to a change in FDI openness of 32.7% in Model 1 and 28.3% in Model 2. The total impact of changes in aid allocations on FDI openness, both immediate and distributed over future time periods, is given by the LRM. The LRM is significant only for regional aid. It suggests that the total effect of a percentage rise in regional aid over total allocations generates an increase in FDI openness by 0.104 percentage points in Model 1 and 0.074 in Model 2.
27This seemingly small effect in the long run may be due to the immediateness with which the shock adjusts, which is captured by the median and the mean lag lengths. The median represents the first lag at which at least half of the adjustment occurs, whereas the mean calculates how long it takes to adjust back to equilibrium – the smaller the median and the mean, the faster the adjustment (De Boef & Keele, 2008). In Model 1 the median is 2 and the mean is 0.79, with 47.1% of the adjustment taking place within the first year, 17.6% in the second year, and 5.8% in the third year, totalling around 70% of the adjustment within three years. In Model 2, the median is 1 and the mean is 0.33, with 72.9% of the adjustment taking place in the first year, 13.7% in the second and 2.4% in the third year, accounting for 90% of the total adjustment. Indeed, the error correction rate, as given by the coefficient for the LDV, sits closer to -1 than 0, which points to a steady, but still fairly quick, return to long-run equilibrium in both cases, even more so for Model 2 (Keele et al., 2016). Hence, inward FDI policy seems to respond rather quickly to shifts in state aid policy, suggesting that the bulk of the regional aid effects might be immediate and that this type of aid is often purposely employed as an investment incentive. Finally, while both horizontal and sectoral aid present negative signs as expected, neither is significant at an acceptable level of confidence, suggesting that in these cases the hypothesised relationships may be spurious.
28Three implications are noteworthy. First, what much of the literature on FDI has noted regarding the importance of regional aid as investment incentives among the Visegrad-4 may indeed be expanded to all CEECs. None of the countries in the sample constitute outliers in terms of FDI openness, and there are no influential observations pulling the results in favour of one of the three regions. However, this association should not be read as strictly causal: the regression results say nothing about the effectiveness of government intervention. Second, and as expected, neither horizontal nor sectoral aid seem to have been employed as an investment incentive. The nature of these types of aid is not particularly favourable to foreign investors and is likely employed to achieve different objectives, often favouring domestic producers. Finally, the non-significant statistical association between these two types of aid and FDI openness means that H2 (the negative effect of horizontal and sectoral aid on FDI openness) is not supported. Hence, while these measures do not foster FDI openness, neither are they necessarily protectionist or to the detriment of foreign investors.
Table 2: ECM on the effect of aid allocations on FDI openness (2000-18)
|
|
|
(1)
|
(2)
|
(3)
|
(4)
|
(5)
|
(6)
|
|
FDI openness(t-1)
|
-0.673***
|
-0.817***
|
-0.617***
|
-0.805***
|
-0.623***
|
-0.786***
|
|
(0.071)
|
(0.071)
|
(0.071)
|
(0.073)
|
(0.070)
|
(0.073)
|
|
|
|
|
|
|
|
Regional aid(t-1)
|
0.070***
|
0.061***
|
|
|
|
|
|
(0.024)
|
(0.023)
|
|
|
|
|
|
|
|
|
|
|
|
Horizontal aid(t-1)
|
|
|
-0.006
|
-0.027
|
|
|
|
|
|
(0.015)
|
(0.018)
|
|
|
|
|
|
|
|
|
|
Sectoral aid(t-1)
|
|
|
|
|
-0.026
|
-0.009
|
|
|
|
|
|
(0.017)
|
(0.020)
|
|
|
|
|
|
|
|
Trade openness(t-1)
|
-0.001
|
0.010
|
0.011
|
0.016
|
-0.003
|
0.033
|
|
(0.015)
|
(0.030)
|
(0.016)
|
(0.032)
|
(0.017)
|
(0.030)
|
|
|
|
|
|
|
|
Log(GDP)(t-1)
|
-1.439
|
-4.028**
|
-0.851
|
-2.622
|
-1.426
|
-4.892**
|
|
(0.900)
|
(1.976)
|
(0.934)
|
(2.385)
|
(0.961)
|
(2.388)
|
|
|
|
|
|
|
|
Real GDP growth(t-1)
|
0.485***
|
0.347*
|
0.456**
|
0.352*
|
0.489**
|
0.305
|
|
(0.184)
|
(0.181)
|
(0.192)
|
(0.188)
|
(0.190)
|
(0.189)
|
|
|
|
|
|
|
|
Debt/GDP(t-1)
|
0.017
|
0.079***
|
-0.0001
|
0.069**
|
0.010
|
0.063**
|
|
(0.019)
|
(0.027)
|
(0.019)
|
(0.028)
|
(0.019)
|
(0.028)
|
|
|
|
|
|
|
|
Left government(t-1)
|
0.012
|
0.024*
|
0.006
|
0.015
|
0.008
|
0.013
|
|
(0.010)
|
(0.013)
|
(0.011)
|
(0.013)
|
(0.010)
|
(0.013)
|
|
|
|
|
|
|
|
Disproportionality(t-1)
|
-0.456***
|
-0.680***
|
-0.375***
|
-0.672***
|
-0.408***
|
-0.673***
|
|
(0.127)
|
(0.161)
|
(0.130)
|
(0.168)
|
(0.130)
|
(0.169)
|
|
|
|
|
|
|
|
Corporate tax rate(t-1)
|
-0.099
|
-0.161
|
-0.100
|
-0.143
|
-0.110
|
-0.135
|
|
(0.101)
|
(0.140)
|
(0.104)
|
(0.143)
|
(0.104)
|
(0.145)
|
|
|
|
|
|
|
|
RULC(t-1)
|
0.119
|
0.228**
|
0.037
|
0.185**
|
0.062
|
0.181*
|
|
(0.093)
|
(0.091)
|
(0.093)
|
(0.093)
|
(0.093)
|
(0.095)
|
|
|
|
|
|
|
|
Property rights(t-1)
|
0.697
|
-0.867
|
0.172
|
-1.350
|
0.393
|
-1.096
|
|
(0.804)
|
(1.260)
|
(0.816)
|
(1.294)
|
(0.820)
|
(1.324)
|
|
|
|
|
|
|
|
Political stability(t-1)
|
-4.412***
|
1.164
|
-2.960*
|
2.112
|
-3.650**
|
2.176
|
|
(1.523)
|
(2.024)
|
(1.502)
|
(2.004)
|
(1.564)
|
(2.074)
|
|
|
|
|
|
|
|
Corruption(t-1)
|
-0.111**
|
0.038
|
-0.045
|
0.107
|
-0.043
|
0.067
|
|
(0.055)
|
(0.115)
|
(0.054)
|
(0.117)
|
(0.051)
|
(0.123)
|
|
|
|
|
|
|
|
ΔRegional aid
|
0.049**
|
0.054***
|
|
|
|
|
|
(0.022)
|
(0.020)
|
|
|
|
|
|
|
|
|
|
|
|
ΔHorizontal aid
|
|
|
-0.003
|
-0.022
|
|
|
|
|
|
(0.016)
|
(0.016)
|
|
|
|
|
|
|
|
|
|
ΔSectoral aid
|
|
|
|
|
-0.023
|
-0.010
|
|
|
|
|
|
(0.019)
|
(0.019)
|
ΔTrade openness
|
0.104**
|
0.125***
|
0.134***
|
0.138***
|
0.126***
|
0.157***
|
|
(0.046)
|
(0.046)
|
(0.047)
|
(0.047)
|
(0.047)
|
(0.047)
|
|
|
|
|
|
|
|
Log(ΔGDP)
|
-0.183
|
0.078
|
-0.145
|
0.051
|
-0.179
|
0.065
|
|
(0.497)
|
(0.469)
|
(0.519)
|
(0.485)
|
(0.512)
|
(0.490)
|
|
|
|
|
|
|
|
ΔReal GDP growth
|
0.359**
|
0.267*
|
0.323**
|
0.261*
|
0.353**
|
0.226
|
|
(0.153)
|
(0.150)
|
(0.160)
|
(0.155)
|
(0.158)
|
(0.156)
|
|
|
|
|
|
|
|
ΔDebt/GDP
|
-0.025
|
0.003
|
-0.029
|
-0.005
|
-0.015
|
0.011
|
|
(0.080)
|
(0.080)
|
(0.083)
|
(0.083)
|
(0.082)
|
(0.084)
|
|
|
|
|
|
|
|
ΔLeft government
|
0.007
|
0.013
|
0.010
|
0.012
|
0.010
|
0.011
|
|
(0.015)
|
(0.014)
|
(0.015)
|
(0.015)
|
(0.015)
|
(0.015)
|
|
|
|
|
|
|
|
ΔDisproportionality
|
0.006
|
-0.118
|
-0.002
|
-0.131
|
-0.040
|
-0.150
|
|
(0.143)
|
(0.145)
|
(0.150)
|
(0.150)
|
(0.147)
|
(0.148)
|
|
|
|
|
|
|
|
ΔCorporate tax rate
|
-0.429**
|
-0.439***
|
-0.493***
|
-0.489***
|
-0.495***
|
-0.475***
|
|
(0.168)
|
(0.159)
|
(0.174)
|
(0.163)
|
(0.172)
|
(0.166)
|
|
|
|
|
|
|
|
ΔRULC
|
0.024
|
0.087
|
-0.016
|
0.067
|
-0.006
|
0.068
|
|
(0.075)
|
(0.073)
|
(0.078)
|
(0.076)
|
(0.078)
|
(0.078)
|
|
|
|
|
|
|
|
ΔProperty rights
|
-3.121
|
-5.245*
|
-1.535
|
-4.568
|
-1.529
|
-3.164
|
|
(3.093)
|
(2.935)
|
(3.181)
|
(3.055)
|
(3.131)
|
(3.012)
|
|
|
|
|
|
|
|
ΔPolitical stability
|
-3.318
|
0.429
|
-2.083
|
1.298
|
-2.710
|
1.143
|
|
(2.093)
|
(2.184)
|
(2.121)
|
(2.190)
|
(2.129)
|
(2.243)
|
|
|
|
|
|
|
|
ΔCorruption
|
-0.126
|
0.420
|
0.072
|
0.597
|
0.071
|
0.626*
|
|
(0.375)
|
(0.365)
|
(0.386)
|
(0.369)
|
(0.380)
|
(0.373)
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
LRM
|
0.104***
|
0.075***
|
0.001
|
0.033
|
0.041
|
0.011
|
|
(0.034)
|
(0.028)
|
(0.025)
|
(0.022)
|
(0.027)
|
(0.025)
|
|
|
|
|
|
|
|
|
Region-FE
|
✓
|
|
✓
|
|
✓
|
|
Country-FE
|
|
✓
|
|
✓
|
|
✓
|
Observations
|
132
|
132
|
132
|
132
|
132
|
132
|
R2
|
0.580
|
0.681
|
0.543
|
0.661
|
0.554
|
0.652
|
Adjusted R2
|
0.471
|
0.565
|
0.425
|
0.537
|
0.438
|
0.526
|
F-Statistic
|
5.314***
|
8.193***
|
4.580***
|
7.482***
|
4.782***
|
7.207***
|
|
Note: OLS with Panel-corrected standard errors in parentheses; * p<0.1; ** p<0.05; *** p<0.01.
Source: The Author
29Among the control variables, six results stand out. First, corporate tax rate bears a negative sign as expected, but is only significant in its differenced form, suggesting that its effect is limited to immediate impacts that are not dispersed across future periods, which once again is suggestive of a quick change in inward FDI policy. Second, growing markets with trade potential seem to generate a positive feedback loop with inward FDI policy, as evidenced by the positive and significant coefficients for trade openness (in the immediate) and GDP growth (both immediate and across future time periods).
30Third, contrary to the theoretical argument advanced by Pinto (2013), left-wing governments in the CEECs do not seem to love FDI more than centre or right-wing ones, on average. Despite presenting positive signs in all models, the coefficient is never significant at the 5% level. As suggested earlier, this may be a consequence of the main liberalisation proponents being centre-right figures. Fourth, when country-fixed effects are included, higher level of debt always show a positive and significant relationship with FDI openness. Hence, FDI can be an important source of debt refinancing for these countries, which usually have low fiscal and budgetary capacity, due to their small size. As was mentioned earlier, this was the case for Hungary in the 1990s, but the phenomenon seems to have been relevant also in similar, more recent moments, such as the 2008 economic crisis.
31Fifth, unlike findings from studies focused on the transition period (Bandelj, 2009; Bevan et al., 2004), the institutional environment does not seem to have an important effect on the decision of foreign investors to turn to these countries. Indices of corruption, political stability and property rights are found to be either insignificant or are contrary to expectations, and the trend is generally erratic. While this may on the one hand suggest that this is due to reduced differences in the political-institutional environment, perhaps a more compelling explanation is that the effect of such institutional factors is moderated by EU membership, which can work as an important guarantor of legitimacy and stability for foreign investors. Finally, the analysis shows that the findings by Garland and Biglaiser (2009) seem to travel well across the globe. As in Latin America, more majoritarian and candidate-centred electoral systems are associated with more FDI openness, though the stickiness of electoral institutions should make us wary of interpreting these results without further analyses – indeed the immediate effect of electoral institutions is statistically insignificant.
- 17 Communication from the Commission to the European Parliament, the Council, the European Economic an (...)
32The next step is then to understand how shifts in state aid and inward FDI policies have changed the relationship between the two. Between 2000 and 2018, two key events have contributed to altering this relationship. The first one is the effect of the 2008 economic crisis, which completely changed the trajectory of FDI inflows among the CEECs, as also attested in Figure 2. This exogenous shock has been analysed at length in the literature (Bohle, 2018; Galgóczi et al., 2015; Szent-Iványi, 2017), and will not be repeated here. The second event has received less interest from scholars, but it is equally important in analysing the state aid-FDI nexus. In 2012 the Commission approved the so-called State Aid Modernisation (SAM) initiative17, which re-organised state aid control on the very basis of the economic crisis. The Commission claimed there was a need for a better use of scarce resources, and promoted those aid schemes with more important positive spill-overs and long-term sustainability, such as environmental protection and R&D&I. As a consequence, the bulk of the increase in total aid spending in the CEECs has been driven by spending for environmental protection and energy savings due to the inclusion of specific renewable energy schemes. This has resulted in a 25 percentage points increase in horizontal aid expenditures over total aid allocations (Schito, 2021b). Hence, since the SAM, as horizontal aid crowds-out the other two types, and as other sources of financing other than FDI are sought out (Bohle, 2018), the relationship between regional aid and FDI openness is expected to falter.
- 18 The LDV takes values of -1.06, -0.99, and -1.00 for regional, horizontal, and sectoral aid respecti (...)
- 19 Further t-tests support the use of the restricted ECM compared to its general form. The GIM also sh (...)
33Table 3 shows the results for the post-SAM period, between 2013 and 2018. The general ECM form employed above is no longer valid for this sample. In all cases, the error correction coefficient approximates or even equals -1, which suggests that Υt adjusts immediately and completely following shocks in Χt (Keele et al., 2016)18. In such cases, the Υt process is not dynamic and no long-run equilibrium exists. Thus, I employ a restricted form of the ECM which only includes the differenced variables, i.e., those whose effect on the dependent variable is immediate (De Boef & Keele, 2008 succinctly call this form ‘Differences model’)19.
Table 3: The effect of post-SAM aid allocations on FDI openness (2013-18)
|
|
|
(1)
|
(2)
|
(3)
|
(4)
|
(5)
|
(6)
|
|
ΔRegional aid
|
0.051
|
0.052
|
|
|
|
|
|
(0.037)
|
(0.043)
|
|
|
|
|
|
|
|
|
|
|
|
ΔHorizontal aid
|
|
|
-0.020
|
-0.016
|
|
|
|
|
|
(0.032)
|
(0.039)
|
|
|
|
|
|
|
|
|
|
ΔSectoral aid
|
|
|
|
|
-0.034
|
-0.042
|
|
|
|
|
|
(0.046)
|
(0.050)
|
|
|
|
|
|
|
|
ΔTrade openness
|
0.159
|
0.208
|
0.164
|
0.250
|
0.241
|
0.363
|
|
(0.164)
|
(0.218)
|
(0.176)
|
(0.244)
|
(0.172)
|
(0.219)
|
|
|
|
|
|
|
|
Log(ΔGDP)
|
0.456
|
-0.193
|
0.410
|
-0.297
|
0.478
|
-0.379
|
|
(0.708)
|
(1.660)
|
(0.724)
|
(1.703)
|
(0.725)
|
(1.680)
|
|
|
|
|
|
|
|
ΔReal GDP growth
|
-0.217
|
-0.273
|
-0.175
|
-0.214
|
-0.195
|
-0.224
|
|
(0.404)
|
(0.470)
|
(0.412)
|
(0.479)
|
(0.412)
|
(0.473)
|
|
|
|
|
|
|
|
ΔDebt/GDP
|
0.139
|
0.055
|
0.121
|
0.055
|
0.175
|
0.120
|
|
(0.138)
|
(0.181)
|
(0.145)
|
(0.193)
|
(0.148)
|
(0.189)
|
|
|
|
|
|
|
|
ΔLeft government
|
-0.015
|
-0.019
|
-0.014
|
-0.018
|
-0.016
|
-0.022
|
|
(0.026)
|
(0.029)
|
(0.027)
|
(0.030)
|
(0.026)
|
(0.029)
|
|
|
|
|
|
|
|
ΔDisproportionality
|
0.270
|
0.359
|
0.339
|
0.475
|
0.365
|
0.523*
|
|
(0.233)
|
(0.290)
|
(0.232)
|
(0.289)
|
(0.225)
|
(0.260)
|
|
|
|
|
|
|
|
ΔCorporate tax rate
|
-0.869**
|
-1.214**
|
-0.878**
|
-1.268**
|
-1.003**
|
-1.415***
|
|
(0.378)
|
(0.448)
|
(0.396)
|
(0.477)
|
(0.390)
|
(0.454)
|
|
|
|
|
|
|
|
ΔRULC
|
0.243
|
0.223
|
0.256
|
0.249
|
0.235
|
0.236
|
|
(0.196)
|
(0.215)
|
(0.200)
|
(0.219)
|
(0.201)
|
(0.218)
|
|
|
|
|
|
|
|
ΔProperty rights
|
4.469
|
2.088
|
4.230
|
1.178
|
2.967
|
-0.314
|
|
(6.161)
|
(7.137)
|
(6.354)
|
(7.428)
|
(6.287)
|
(7.157)
|
|
|
|
|
|
|
|
ΔPolitical stability
|
-5.495
|
-6.323
|
-6.206
|
-7.711
|
-8.136
|
-10.019
|
|
(5.529)
|
(6.180)
|
(5.783)
|
(6.635)
|
(5.548)
|
(6.045)
|
|
|
|
|
|
|
|
ΔCorruption
|
-0.727
|
-0.025
|
-0.629
|
-0.160
|
-0.794
|
-0.455
|
|
(0.858)
|
(1.155)
|
(0.880)
|
(1.213)
|
(0.888)
|
(1.164)
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
Region-FE
|
✓
|
|
✓
|
|
✓
|
|
Country-FE
|
|
✓
|
|
✓
|
|
✓
|
Observations
|
49
|
49
|
49
|
49
|
49
|
49
|
R2
|
0.321
|
0.409
|
0.292
|
0.379
|
0.294
|
0.391
|
Adjusted R2
|
0.042
|
-0.091
|
-0.000
|
-0.147
|
0.003
|
-0.124
|
F Statistic
|
1.149
|
1.498
|
1.000
|
1.322
|
1.012
|
1.393
|
|
Note: OLS with Panel-corrected standard errors in parentheses; * p<0.1; ** p<0.05; *** p<0.01.
Source : The Author
- 20 According to De Boef & Keele (2008), the ECM is equivalent in form to the Autoregressive Distribute (...)
34The results, however, show that, following the changes in priorities in the use of state aid as established by the SAM, the relationship between changes in state aid use and FDI openness is no longer robust. The coefficients for the three types of aid all retain the expected sign, but in no case are they significant. The only consistently robust result concerns the use of corporate tax rate policies: as in Table 2, lower tax rates have an immediate effect on FDI openness, though their effect across future time periods is unknown. Nevertheless, the small number of observations in the sample and the lack of consistently significant effects suggest caution in the interpretation of these results20.
35In sum, these analyses advance current scholarship in several ways. First, they show the importance of having a more fine-grained look at the contradictory goals of government incentives (particularly state aid) and of the policy trade-offs they entail. Second, they highlight the usefulness of modelling the temporal effects of government policy, which offer information about the speed of adjustment of a host country’s inward FDI policy following changes in other government policies. And third, they emphasise the frailty of the state aid-FDI nexus that for a long time represented a key tenet of the economic growth policies of the CEECs.
36The goal of this article was to provide a quantitative assessment of the policy trade-offs of different state aid tools on the CEECs’ inward FDI policies. Because CEECs are capital-scarce economies, it is imperative for them to attract foreign capital (Tőrös et al., 2017). This paper showed how, across different models and specifications, there is a positive, strong association of regional aid allocations with FDI openness among the CEECs, suggesting that shifts in state aid policy towards more regional allocations tend to be conducive to more FDI openness. The hypothesised negative effect of horizontal and sectoral aid on FDI openness has instead found less support.
37These results help make sense of the seemingly paradoxical findings regarding the nexus between government policy in terms of investment incentives and FDI openness. Past studies did not analyse the association between regional aid and FDI openness in a systematic manner (Ślusarczyk & Kot, 2012); found uncertain effects for FDI incentives (Beyer et al., 2002); or fell short of providing evidence about the impact of regional aid expenditures on FDI openness that could apply to all the CEECs (Drahokoupil, 2009; Medve-Bálint, 2015; Tőrös et al., 2017).
38Hence, by disaggregating state aid according to the likely usefulness of the measure to the host country’s inward FDI policy, a different picture emerges whereby a theoretical and empirical bridge is created between the qualitative studies stressing the importance of regional investment policies in the CEECs on the one hand, and the mixed findings from the quantitative analyses of government policies for investment promotion on the other.
39Two key implications follow. First, incentives constitute only part of a country’s economic growth policy, and they can be deployed with contradictory goals. Thus, when examining the effects of government policy on FDI attraction, particularly in the field of state aid, scholars should be explicit about the goal for which the policy is designed and whether trade-offs exist between the different tools employed. Second, this study underscored the frailty of the association between state aid measures and FDI openness following state aid control policy reform. Recently, governments in several CEECs have adopted an increasing number of investment acts to attract FDI. However, as the present analysis shows, these policies may be incompatible both with the recent shift towards horizontal aid, which foreign investors may see as less attractive, and with the growing centrality of EU structural funds in the economic growth of several CEECs. Hence, if host countries’ governments are to use investment incentives effectively, they need to reconcile not only the contradictions in their adopted strategies for economic growth policy, but also those between the different levels of governance at which these policies are regulated.
40Finally, the article presents some limitations worth underlining. First, one should take this study for what it is: an analysis of the association of trade-offs in state aid measures with a country's openness to or reliance on FDI, and not of the level of FDI inflows into a host country. As Li (2009) forcefully argued, the two are not interchangeable. The article does not infer that a higher share of regional aid expenditures over total aid will lead to more FDI in the host country, but merely that said country will be more open to – or reliant on – FDI, as compared to a situation in which more aid is allocated for different purposes (e.g. environmental protection or R&R). This is also reflected in the post-SAM period, where the relationship became weaker as changes in state aid control policy switched the priorities on the use of state aid.
41Second, the research design employed here cannot establish causality. The article does not uncover the causal mechanism that links the employment of the aid measure to the effectiveness of its intended purpose (either as investment incentive or as protectionist shield). These two limitations notwithstanding, the implication remains clear: shifts in the composition of aid allocations in favour of regional aid are suggestive of a more aggressive inward FDI policy for the host country.
42The third limitation is the geographical horizon to which such findings can apply, for two reasons. First, the EU is unique in the world in its regime of state aid control. Second, even within the EU, the CEECs represent an exception. Aside from the Benelux countries and Ireland, no other EU-15 member states has as FDI-driven an economy as the CEECs (Schito, 2021b). Very few countries in the world can be referred to as FDI-dependent, and even then, the methods and tools of FDI attraction differ wildly. For instance, in his in-depth case studies of the liberalisation of FDI in Argentina and South Korea, Pinto (2013) finds that in the former, FDI liberalisation was due to the changing pattern of labour influence and support, particularly towards the Peronist regime; in the latter case, instead, FDI policy was driven by the preferences of capital owners, with labour only playing a minor role. Yet, the fact that findings may not travel well geographically due to the specificities of the EU competition regime and the CEECs does not mean that they cannot travel at all. This study represents an opportunity to develop the research agenda in this direction and increase the external validity of the findings.
43Lastly, the range of economic activities for which the CEECs have attracted FDI has considerably expanded in the last decade, ranging from complex manufacturing to financial services, to information technology, to accommodation (Hunya, 2015). This article is unable to capture such a dynamic diversity. Scaling the analysis of the FDI-state aid nexus at the sectoral level can help us better understand when and how government policy aimed at the development of specific economic sectors becomes part and parcel of the host country’s inward FDI policy, and whether the government’s goals are contradictory or entail policy trade-offs. Although data at the sectoral level exist (though for a smaller number of years), a new and more specific theory needs to be developed. This article can serve as a starting point in this direction.