4.2.2 Romanian Institutions

After the used variables have been introduced we will deal now with the impact of the quality of Romania’s Institutions Gi. The assessment was separated from the other test for two reasons. First, the public  administration test considers only 24 counties while the test of infrastructure and human capital includes all counties of Romania in order to increase precision. The  second reason is illustrated in Figure 4.2.  A simple regression in the form

Yi= α + βGi + e

using the NewsIn score for the 24 monitored local Romanian institutions as predictor and GDP per capita 2005 as outcome turns out to be a thankless exercise unless a more Popperian attitude is adopted. The relation between institutions and GDP per capita is weak (β = .058)  and of a low significance (.788). The explained variance is neglectable (R² = .003).  The constant α (RON 12,540.48) is even nearly as high as average GDP per capita  on county level itself (RON 13,271.81).[1] The situation does not improve substantially if the other outcome variables are  deployed.

Image 4.2: Linear  Regression – GDP per Capita in Romanian Counties and Quality of Institutions

Source: Data from INS  2008 and NewsIn 2008; own graphic, own calculations

A further look on the distribution does not suggest a  non-linear or any other kind of trend; therefore further tests concerning Romania’s institutions will be skipped.  We rather see most monitored authorities achieving good rankings but per capita GDP in Romania varies dramatically even among same ranked counties. Additionally, the worst result for public administration (City Hall of Buftea, county Ilfov obtained  zero points) is accompanied with the second highest per capita GDP, thus, far  above the average GDP per capita. This works also the other way round. The poorest county of the sample (Teleorman – RON 8,011.35 per capita GDP in 2005)  excelled with a score of 17 points concerning the quality of its institution (City Hall of Alexandria).

Quality of Institutions on the Local and Regional Level in Romania

There might be several reasons for this result. First  of all it seems striking that local Romanian institutions  did not perform that bad (mean  = 14.54 out of 20 possible points, standard  deviation  = 5.28). Actually, most of them did rather well and only six out of 24 authorities were scored below the high average. One  might suspect sponsorship effects as the local institutions could be interested  in pleasing the agency NewsIn. Then again, this notion fails to explain why  some of the authorities were not interested in making a good impression while  others were. Another objection would be that the administration score is from  2007 but GDP per capita data is from 2005.

Indeed, it might be possible that  certain incidents (e.g. EU-accession) changed the behavior of the institutions in recent time. Such a change can, of course, not be reflected by GDP data from  2005. On the other hand, the given county order with regard to GDP per capita from  2005 is not new and probably  unchanged, thus, GDP data for 2005 nonetheless may serve as a proxy.[2] One could also argue that the sample was too small and the results are, hence, randomly. Furthermore, it is obvious that the indicator is a neat but weak one as  it covers only few dimensions of several possible dimensions concerning  quality. of institutions. Therefore, the result has to be taken with some caution  and is by no means a blueprint for Romania’s authorities. But it nonetheless  supports the frequent notion that things in Romania often work better than  expected. While the range and variation of achieved scores confirm the  expectation that public administration in Romania can be and (too) often is deficient,  the large share of well scored cases suggests that more often things get easily  done.

This is by the way in line with my personal experience gathered while  working at this paper. I contacted Romanian authorities thrice and got at least  answered or easily helped out. Ironically, I cannot say the same about  Transparency International (Romanian Office), which did not even bothered to  answer my mail.

Footnotes

[1] This  average value refers only to the current sample. The slight differences between  the figures in the table and the equation in the Excel generated chart are due  to rounding errors.

[1] Furthermore,  the ranking order of the counties can be expected to be unchanged as CNP  projections on NUTS-III level do not indicate major changes.