LogarithmiC ModeliNg

Logarithmic Modeling

In the empirical analyses below we also applied the variable HUMRES which stands for expenditures in education, training and R&D called human resources in the categorization of the EC. According to the theoretical framework outlined in the previous chapter, technological change depends to a large extent on local/regional factors of innovation Logarithmic Modeling. Thus the unit of empirical investigation applying equation (3.1) should be some sub-national geographical entity. Since the lowest level of spatial aggregation of the type of data we need for analysis is the county the selected unit of analysis is Hungarian counties. The spatial unit is denoted by i while t stands for time in equation (3.1).

(3.1)TFPGRi,t = a0 + a1KNATt + a2RDi,t + a3KIMPi,t + a4INFRAi,t + a5EDUi,t + ei,t,

where

TFPGR is the annual rate of growth of Total Factor Productivity at the county level,

KNAT is domestically available technological knowledge accessible with no geographical restrictions (A in equation (2.5)),

RD stands for private and public regional R&D (H in equation (2.5)),

KIMP is imported technologies,

INFRA is investment in physical infrastructure,

EDU is investment in human capital (education and training),

e is the stochastic error term.

To implement equation (3.1) in an empirical analysis we relied on different data sources. KNAT is measured by the number of patents available in Hungary obtained from the Hungarian Patent Office. In empirical estimations we measured RD alternatively either by R&D employment or by R&D expenditures aggregated from data at private, public and university research institutes. The Hungarian Central Statistical Office provides these data. The measure of KIMP is the share of foreign direct investments in total private investments. To measure foreign direct investments we used data on the number of firms in different size groups and percentage of firms in manufacturing. Data come from regional and county statistical Yearbooks published by the Hungarian Central Statistical Office. Investments in infrastructure measure INFRA. Data on infrastructure investments include investments in transportation, telecommunication, health care and utilities. Data sources are regional statistical yearbooks. HUMCAP is measured by all (private and public) expenditures on education and training. Data sources are Hungarian National Accounts by the Central Statistical Office. All the variables measured in monetary terms are in 1995 Hungarian Million Forints (Baden, 2002).

To empirically generate a variable measuring the growth in TFP we followed the solution developed in the growth accounting literature. TFP levels for each county are calculated from a constant return to scale Cobb-Douglas ...