expenditure have been again classified into the two important components of ‘productive’ and ‘social’
infrastructure. Total infrastructure investment is the sum of these two categories.
Productive infrastructure consists of the categories of Agriculture, Forestry and Fishery,
Industry, Energy and Handicrafts, Irrigation, Research and Technology, Special Works (plus those of
Athens/Thessaloniki), Transportation (plus those for Railways), Water/Sewage Works, and
Prefectural Works/Programmes. Social infrastructure is comprised by Education, Health and Welfare,
Housing, Public Administration, and Tourism sub-categories.
The dominant approach for the estimation of the public capital stocks in the infrastructure
literature is the perpetual inventory accounting method. This approach uses the flow of infrastructure
investment for the estimation of stocks, assuming that a percentage of the existing stock has been
depreciated. The formula for the estimation of the capital stocks to be used here is:
Gt= (1 -δ )Gt_, + It I
where Gt is the end of year public capital stock in year t, δ is the geometric rate of depreciation, and It
is real investment in public capital during years t. (For more details on this fairly standard method of
capital stock construction see for example Holtz-Eakin 1993b).
It is apparent that three pieces of information are necessary for the estimation of public capital
stocks. The real investment in infrastructure (It) can be approximated by the monetary sums paid by
the PIPR. Regarding the geometric rate of depreciation (δ ), it was assumed that the rate remained
constant for all of the examined period, and that it equals 0.05. This is similar to the depreciation rate
used by Bajo-Rubio and Sosvilla-Rivero for the estimation of the Spanish infrastructure stocks (1993,
Appendix 1). This figure is arbitrary, but it has chosen assuming that Greek infrastructure capital is
not much different from the Spanish, and that in addition the depreciation rate for fixed assets
(buildings) given by the NSSG (see Xenaki 1997) is not dissimilar. Finally, as there were no existing
stock estimates at regional level, it was decided to build on the first year of the available data17.
The output data comprise the regional Gross Domestic Product (GDP). These data were
provided by the National Statistical Service of Greece (NSSG) and are supplied with a breakdown
into the various sub-categories which comprise the regional GDP. These are Agriculture-Farming-
Forestry-Fishery (henceforth Agriculture), Mines, Electricity-Lighting-Water Companies (henceforth
Electricity), Constructions, Transportation-Telecommunications (henceforth Transportation),
Commerce, Banking-Insurance-Land Estate sectors (henceforth Banking), Housing, Public
Administration, Health-Education (henceforth Health), Services (which includes various categories of
services), and Manufacturing. As the impact of public capital on the latter has been analysed in detail
17 This is a common practice in the estimation of capital stocks. See for instance Corrales and Taguas (1991).
13