The second predictor IAi – or infrastructure and international accessibility – is an own composite index and requires some explanations. Literature deploys different indicators for the state of infrastructure (cf. chapter 4.1.5). Often telecommunication variables (such as fixed telephone lines per capita) are found to have the strongest and most significant impact on FDI-location decisions (cf. e.g. Bellak et. al 2007). This might be – to some extent – a reasonable point, especially for cross-country studies, but the causality seems unclear in my opinion.

Telecommunication and Telephone Lines

To make the long story short, fixed telephone lines per capita of course correlate strong and significant with both GDP per capita in Romania and the number of foreign firms on county level on Romania.[1] On the other hand, it could also be argued that higher economic activity and higher GDP per capita have a positive influence on the number of fixed telephone lines. A telephone line is a prerequisite for most business activities and hence, agglomeration of firms will increase the frequency of fixed telephone lines per capita. But that does not necessarily mean that the sheer availability of telephone lines attracted the firms.Fixed lines might be available at other places but are simply less solicited due to an otherwise less attractive business environment. Furthermore, a higher per capita GDP will make the access to telecommunication services more affordable for households.

For sure, the frequency of subscribers to such services is also a matter of availability and to be considered infrastructure. But a fixed telephone line in Romania is relatively equally available in Bihor, Arad or Sibiu. Thus, the differing frequencies of subscribers in these counties are more likely to be explained by differing income and urbanization levels (i.e. via the size of average households, not necessarily via availability only) and different degrees of economic activities. Due to the difficulties in detangling the causalities in the telecommunication sector the variable will not deployed here. The paper will focus rather on more unambiguous transport and traffic related variables as it is unlikely that FDI-inflows to Romania caused the existence of European Roads or airports several years before.[2]

Roads and Road Density

Often used as a proxy for infrastructure is overall road density, others consider only certain kinds of roads. Accordingly, a first starting point could be the road density of the counties in Romania. However, eyeballing the data for Romania shows that this approach would be misleading. A simple regression for the 24 counties from the NewsIn 2008 monitoring yields only a weak effect and does not even fit the data appropriately (R² = 0.0143). The regression improves somewhat after the share of modernized roads is deployed as predictor variable for infrastructure. Nonetheless, the relation leaves the biggest part of the variance unexplained (R² = .3413) and overall road densities do not differ dramatically, anyhow. Does this mean that infrastructure does not matter either?

Most probably not. Ianoș (2006) linked FDI-performance in Romania to the course of main traffic roads, and NewsIn (2008) identified shortages in infrastructure likewise explicitly as major explanatory factors for local economic performance and FDI-perfomance. As seen in the case of Sibiu (cf. footnote 67 on page 71) “infrastructure” from an economic point of view can be considered (relatively) well even if network densities (29.4 % for Sibiu in 2005, INS 2007) are relatively low. Sibiu profits from its position as intersection of main traffic roads (typically the European Roads are considered as such) and the existence of an international airport. Likewise does Cluj, further supported by the current building of the motorway “Autostrada Transilvania”. So do Constanța (both further backed by an airport and two important maritime ports) and the Capital Bucharest (two airports). Thus, all have a relatively strong infrastructure.

Image 4.1: Simple Regression – GDP per Capita and Modern Roads, by County 2005

Infrastructure as captured by Road density in Romania

Source: Data from INS 2007; own graphic, own calculations

Currently, Romania features only a short highway but several projects are under way. The best working roads are the “national roads” [“drumuri naționale”, maybe comparable to German country roads (“Landstraße”)]. In general, roads in Romania are in a bad state (cf. SAR 2007) what is also reflected in the share of modernized roads (cf. Figure 4.1).[3] There are two classes of national roads in Romania (clasa A and clasa B). Some national roads of both classes are allocated to “international” (this is to say heavy) traffic – but not all.[4] Even the allocated roads are actually not capable for this purpose (cf. Guvernul României 2005: 81f.) but alternatives are missing and as a result they suffer from a heavy over-utilization which in turn is not helpful for maintaining their anyway critical state. To sum up, what really differs significantly across the counties in Romania is not the statistical tracked overall density of roads but the endowment with infrastructure in the sense of international transportation links.

Airports, Ports and other Traffic Links

International traffic links are to be found on the map on ARIS’ website (Agenția Româna pentru Investiții Străine, 2008). ARIS is a legal Romanian agency, which seeks to be the “Government leading body for providing consultative services to foreign investors” (ARIS, 2008). Provided, among other information, is an interactive infrastructure map containing roads, airports and ports, planned and existing highways plus railway connections. This map is the main source for the infrastructure index IAi. Most, but not all European Roads are considered on this map; especially only few class B European Roads were considered. While this pre-selection on behalf ARIS may actually be an useful indicator of which roads are “advertisable”, I gathered road data nevertheless from Microsoft Map Point 2006, European edition. To guard against misunderstandings: beneficiaries of this measure are solely the less well endowed counties.

This focus on European Roads can be backed by some examples. First, a look on the official traffic news, issued by the Compania Națională de Autostrăzi și Drumuri Naționale din România (CNADNR) for today (2008-04-11) announces three interruptions of traffic on national roads (all due to reasons like “Alunecare de teren din cauza precipitațiilor”)[5] and impeded traffic at further 16 locations on national roads for similar reasons (just in two cases due to modernization and construction works).  Second, another illustration of the state of the Romanian road network might be expressed by the rule of thumb that you can account for one hour of driving to make some 50 or 60 km (average for all kind of roads).

The IBRD (2006: 15) estimates that road transportation costs in Romania are 30 % – 60 % higher than they should be. Furthermore, Nokia might serve as example. Nokia’s new plant at the place of Jucu is already connected to the infrastructural relatively well endowed local municipality Cluj-Napoca nearby via a national road and features, furthermore, a local railway station on double electrified rails.[6] Nonetheless, these existing connections of the industrial park Tetrarom III did not seem to fit the needs of Nokia and its suppliers as their demand for another highway, mainly parallel to already existing national roads, towards the airport and next motorways (currently under construction) might indicate (cf. e.g. Cireașa 2008).

Infrastructure Index

Overall, the state of the national roads is not optimal. On the other hand, they are Romania’s best working roads and some of them – termed “European Roads” – make part of the Pan-European Transport network and are open to heavy traffic. The state of the other roads is worse and hence, firms seem to find it important to be near as possible to European or at least national roads.

Hence, I suggest an index for “infrastructure”, consisting of five single indicators:

  1. Number of airports and ports (both, maritime and fluvial, according to ARIS)
  2. Intersection of European Roads (according to Microsoft Map Point 2006, EE)
  3. Number of European Roads (according to Microsoft Map Point 2006, EE)
  4. Number of motorways (existent or planned, according to ARIS)
  5. Important border position (Hungarian border or Black Sea)

While well working ports might be cheap means of transports, airports are additionally a time saving mean, mainly for human traffic, thus indispensible for MNEs and their personnel flows. Hence, for each airport and port a county is scored with one point.[7] As Brașov’s airport is currently still under construction it was scored with 0.5 points.[8] The county Ilfov does neither possess an airport nor another kind of port but as it actually contains Bucharest (two airports) and its surface represents just a small fraction of other counties’ surface it was scored with one point.

The position as intersection of European Roads was likewise rewarded with one point. No half points were given but if the intersection was next to the border of another county each of them gained one point. The intersection indicator captures the fact that some counties are traversed in all four directions by European Roads while others are just crossed or strived. Furthermore, an intersection of European Roads evidently increases the transportation opportunities.

Likewise, each European Road crossing a county was scored with one point with the exception for roads which just crossed a small part of the county without passing a main city. These counties were scored with 0.5 points (Neamț). Again, each motorway (existing or planned for the near future) crossing a county was scored with another one point. No half points were given. The planned routes for the highways to be build have been known for a long time now and many are currently worked at, though the effectivity of implementation is not always thrilling (cf. e.g. Turnock 2006).[9]

The position next to the Hungarian border or the Black Sea was scored with another point and applied only to the western counties (Arad, Bihor, Satu Mare and Timiș) and the two counties at the Black Sea (Constanța and Galați). This indicator accounts for the availability of high quality transportation facilities (at least for transit purposes) beyond the Hungarian border (cf. e.g. Kohagen 2007 a or NewsIn 2008) or at the Black Sea (cf. e.g. Kohagen 2007 b or NewsIn 2008).


The railway system was not considered in the infrastructure index in order to keep things simple. According to the ARIS map the beneficiaries would have been again the already better endowed counties anyhow and furthermore just some 28 % of all goods are transported on railways (cf. Guvernul României 2005: 84) despite their greater (potential) effectivity. Human traffic for MNEs is unlikely to use the Romanian railway system.

All values enter the infrastructure index without being weighted as every kind of weighing would be the same questionable as their non-weighing.

Academic Research paper and Study of the Economy of Romania and Romanian Business

Usually deployed control variables for FDI are wages or unemployment but they can be skipped in the Romanian case as economic activities – FDI in particular – within Romania tend to behave contra-intuitive and mostly avoided regions with higher unemployment and lower wages so far (cf. chapter 3.5).


[1] However, with regard to the three chosen outcome variables, the impact loses dramatically size and significance once the other predictors are deployed. These in turn remain both strong and significant. Cf. Appendix 6  for a model including fixed telephone lines per capita.

[2] Large-scale FDI-inflows were sometimes accompanied by solicitations for infrastructural enhancements. However, this phenomenon is rather a new one as huge FDI-inflows arrived relatively late in Romania. Furthermore, these enhancements will not be considered in the predictor variable but just these facilities, which were already available or planned before.

[3] In turn, the share of modernized roads improves considerable if only national roads are considered and reaches typically levels of over 90 % per county (cf. INS 2007).

[4] Of course, other kinds of roads are also open to “international” traffic but they are more restrictive concerning transport masses and dimensions. I would like to thank the Romanian Ministry of Transports for clarifications.

[5] Slide of soil due to rainfalls [SH].

[6] Cf. e.g. http://www.cjcluj.ro/harta5.html for the national road and http://www.cjcluj.ro/harta3.html for the railway station. Both maps are exposed on the County councils web site.

[7] Again, there exist some ports or stations which are not considered on ARIS’ map. The reason is probably the simple fact that some 85 % of all inland ports are totally obsolete (cf. Guvernul României 2005: 87). Thus, I preferred to follow the ARIS (2008) map in this case.

[8] As in the case of the motorways to be build, this measure can be justified with rational expectations. The notion of rational expectations is based on the argument that agents do not react only to the experiences gained in the past but also with reference to currently available data about the future (for more information on rational expectations cf. e.g. Burda & Wyplosz 2002: 101f.). Thus, investors could be attracted by a location due to their knowledge about the current building of a new motorway or airport. The proceedings of Nokia and Ford are neat examples for this mechanism.

[9] This for, one motorway to be built (Autostrada Est-Vest) was not considered as it was for a long time unknown whether it will be built and which will be the exact route (cf. newsiasi.ro 2007). Accordingly, the map on ARIS 2008 does not mention the highway.