1. REGIONAL DEVELOPMENT WEEK 2

2. Recap Last week we have mentioned the evolution path of regional economic models. Traditional trade theory New trade theory New Economic Geography Models This week we will talk about some important models that are quite important in regional economic theory and policy in detail.

3. Traditional Tools for Measuring and Evaluating Regional Economic Performance Regional economic development policy is basically about the allocation or reallocation of resources to enhance the economic performance of industries. Planners and policy makers need to be able to measure and evaluate that performance.

4. Thus it is necessary to: Measure the degree to which economic activity and employment in a region is related to serving local demand as against serving demand external to the region (i.e. exports). Assess a region’s overall performance relative to that of other regions. Assess which industry sectors are performing better in the region. Assess a given sector’s efficiency relative to other industry sectors’ performance in the region.

5. Measures of Concentration There are several measures used in empirical studies to investigate geographical and industrial concentration within and across countries/regions. We will only consider some basic measures.

6. Herfindahl Index H= Where; s ij c denotes share of employment in industry i in region j in total employment of industry i:

7. The Herfindahl index is a measure of industrial concentration. Its main advantage is the computational simplicity. On the other hand Herfindahl index does not take the areas of the region into account, it assumes they all have same sizes and it is also sensitive to the number of firms in each industry

8. The Dissimilarity Index for Regional Specialization DSR j = Where; s ij s denotes share of employment in industry i in region j in total employment of region j and s i denotes share of total employment in industry i in total employment and calculated as follows:

9. The Dissimilarity Index for Industrial Concentration DCR i = Where; s j denotes share of total employment in region j in total employment and calculated as follows;

10. Krugman Specilization Index KSI= where k and l are two different regions. The Krugman specialization index, compares two regions and identifies how specialized or despecialized these regions are.

11. Gini Coefficient for Regional Specialization GINI j s = where, R i = λ i indicates the position of the industry i in the ranking of R i in descending order.

12. Gini Coefficient for Industrial Concentration GINI i c = where, C j = λ j indicates the position of the region in the ranking of C j in descending order

13. Economic Base Theory Economic base theory (EBT) is an easily understood traditional body of thought in the field of regional economic development. EBT views an economic system as composed of two parts: one, called non-basic, is viewed as producing for local consumption; the other, called basic, is viewed as producing goods and services primarily for external consumption, (that is for export from the region).

14. Economic development theorists believe that the critical cornerstone of a regional economic system is its basic economic activities. Thus by expanding (export) base activities, the local regional economy not only expands employment and earnings in the region directly, but also expands employment and earnings indirectly. As a consequence, the primary focus in sectoral targeting analysis is on basic economic sectors and activities.

15. A complimentary approach is entitled import substitution, whereby goods and services are imported to support basic and, in some cases, non-basic production. With this approach, industry sectors that are insufficiently developed to support local basic activity are targeted for investment and development. By expanding these sectors, the relative importance of basic sectors often can be increased.

16. Measuring the Economic Base of a Region A variety of techniques have been developed to separate economic systems into their basic and their non-basic parts. The simplest method is to sort industry sectors into those that are primarily basic and those that are primarily non-basic. Subsequent efforts have relied on location quotients, which are measures estimating the importance of industry sectors to the local economy relative to their importance in a larger reference economy

17. Two types of location quotients: One is measured the minimum requirements approach, where that locality which has the least employment (earnings or some other indicator of scale) in a sector becomes the base against which the same sector in all other regions is compared. Alternatively, location quotients can be computed in terms of some reference area, (e.g. the nation), whereby the contribution to the basic part of the economy is measured as the part that is greater than the proportional amount found in the reference area.

18. Calculating Location Quotients Location quotients are well known measures of the relative importance of sectors compared to their importance in a larger frame of reference as described above. Location quotients (LQ) are computed as follows: LQ ir = (E ir /E r )/(E iN /E N ) Measures of scale other than employment can be used; for example earnings and gross regional product GRP

19. LQ>1, means a higher concentration in the region than in the country and LQ>1.25 considered as an initial indicator of regional specialization.

20. An Example: Industrial Targeting in Northern Virginia A case study application of industrial targeting analysis is a study of the Northern Virginia region in the United States. The time period for this study was 1988–1993. One objective of the study was to identify and evaluate the performance of the primary technology intensive industry sectors.

21. An Example: Southeast Anatolia Region (1980-2000) High point cluster 1980Driver industriesLQShare in employment (%) Food, beverages and tobacco Slaughtering, preparing and preserving meat 6.556.13 Dairy products2.080.98 Vegetable and animal oils and fats2.403.79 Grain mill products1.591.96 Prepared animal feeds2.400.96 Distilling, rectifying and blending spirits 16.785.98 Tobacco1.379.06 Textile Spinning, weaving and finishing textiles 1.6329.39 Carpets and rugs4.074.5 ChemicalsPetroleum refineries33.3728.31 Plastic products not classified elsewhere 2.062.97

22. High point cluster 1980Driver industriesLQShare in employment (%) Textile Spinning, weaving and finishing textiles 3.7755.08 Made-up textile goods except wearing apparel 1.585.16 Carpets and rugs11.898.60

23. Mediterranean Region High-point cluster 1980Driver IndustriesLQShare in employment (%) Textile Spinning, weaving and finishing textiles 2.6347.49 Basic metal industriesIron and steel basic industries3.6724.79

24. High-point cluster 2000Driver industriesLQShare in employment (%) Textile Spinning, weaving and finishing textiles 2.0830.37 Manufacture of wearing apparel except leather and fur 1.249.95 ChemicalsPetroleum refineries1.650.73 Manufacture of plastic products not classified elsewhere 1.493.95 Non-metallic mineral products Manufacture of glass and glass product 3.063.70 Manufacture of cement, lime and plaster 1.391.31 Basic metal industriesIron and steel basic industries4.1517.57

25. Shift-Share Analysis A simple descriptive, quick and relatively inexpensive technique for analyzing regional growth and decline over time is shift-share analysis. This technique enables the assessment of a region’s overall performance relative to other regions.

26. The Traditional Shift-Share Model The traditional shift-share model measures regional growth or decline by decomposing it into three components: National share (NS): that is, that part of change attributable to overall national trends Industrial mix (IM): that is, that part of change attributable to the industrial composition or mix of the region Regional shift (RS): that is, that part of change attributable to regional advantage or competitiveness.

27. Early shift-share models outlined in Perloff et al. (1960) focused on total regional employment and had only two components: Total shift (TS), expressed as:

28. Differential Shift (DS), expressed as: where: ei and Ei respectively are regional and national employment in industry i; e and E respectively are regional and national total employment in all industries; and t-1 is the initial period and t the end period (e.g. inter-censal dates) of the analysis.

29. Dunn (1960) introduced to the model differential rates of growth in individual industries, to give what is known as the ‘proportionality effect,’ which is equivalent to the industry composition or mix (IM) effect referred to above. Ashby (1967) introduced a three-component model of regional change, incorporating national share (NS), industry mix (IM), and regional shift (RS)

31. This classical shift-share model—which has been used extensively by economists, geographers, regional scientists and planners in regional analysis— thus emphasizes not only the role of regional change for a region- specific industry, but also the regional shift or competitive component as a measure of the relative performance of the region for a specific industry.

32. A position shift is interpreted as being associated with the comparative or competitive advantage of the region for that industry, or vice versa. The partition of regional change into the three components—NS, IM and RS— was intended to enable researchers to study the sources of change separately.

33. An Example: Regional and Axial Shifts in the United States Spatial Economy In the early 1980s the United States, the Northeast and Midwest regions were loosing out to growth in the Southeast, South, Southwest and West. This was evidenced by population movements as well as shifts in industrial location and employment. The Northeast always was a net out-migration region. What was of significance was the sudden change in pace and destination of population movement beginning in the 1970s. While the West and Southwest showed gains in population, the dramatic growth appeared in the South. At least some observers interpreted this as a direct transfer from the (old) North to the (new) South

34. Shift-share analyses were computed for all States and for seven primary transportation industries in order to compare the rates of growth among the different states and in particular to compare the regional competitiveness of the states in these industries.

35. Total Transportation, Communications and Public Utilities Local and Interurban Passenger Transit Trucking and Warehousing Water Transportation Air Transportation Pipelines Transportation Services

36. The general pattern among the results was that in all transportation sectors job loss was occurring due to industry mix and further loss in the Northeast and Midwest due to a loss of State competitive share (see Table 3.4). Transportation services were the highest growth part of the transport sector (see Table 3.5). The general trend shows a loss in competitive share from North to South and West even though, in absolute terms, employment remained highest for most transport sectors in the North, because of the heavy industrial concentration there.

37. Two criteria were used to identify high performer States: the competitive share had to at least equal the national growth component; the absolute growth in employment must be at least 2 percent of absolute growth for the industry nationally.

38. Critiques and Extensions of Traditional Shift-Share Analysis Despite its continued widespread use, shift- share analysis has been heavily criticized for having temporal, spatial, industrial aggregation, theoretical content and predictive capability deficiencies

39. Shortcomings Dawson (1982) lists six shortcomings of the traditional shift- share model: Changes in the industry mix in the national economy are not taken into account. This is a weighting problem as changes from the beginning to the end of the period of time over which change is being measured may have quite different weights or opposite signs for the industrial mix and the competitive effects. Results are sensitive to the degree of industrial and regional disaggregation. The differential industry component is unstable over time, and the degree of instability varies among industries.

40. Growth resulting from inter-industry linkage and secondary multi-sector effects are not explicitly isolated but are included in the competitive component (RS), whereas they should be included in the industry mix component (IM). The differential component (RS) may be influenced by relatively spurious causes, including the incorrect classification of firms, product heterogeneity within firms, and transfers of production between separate sites of individual firms. The technique provides no information on the capacity of a region to retain growing industries or on how to attract them in the first place (Richardson 1978).

41. Modifications and Extensions Researchers have incorporated the shift-share model into other statistical forecasting methods, including: (a) Analysis of Variance (ANOVA) models (Berzeg and Koran 1984, 1978). (b) A multiplicative model of shift-share (Theil and Gosh 1980; Kurre and Weller 1989). (c) Univariate autoregressive integrated moving average (ARIMA) time series models. (d) A linear model of shift-share analysis (Knudsen and Barff 1991).

42. Another trend in shift-share analysis is the use of econometric models developed by Emmerson et al. (1975) and by Berzeg and Koran (1978). These are early forms of the information-theoretic approach developed by Theil and Gosh (1980).

44. Next Week Total Factor Productivity Approach (3.4.3)