Contemporary structure of software employment in metropolitan Tel Aviv
The diminishing impact of distance on the spatial organization of employment
Nurit Alfasi, Shaul Krakover
Department of Geography and Environmental DevelopmentBen-Gurion University of the Negev
Software firms by size, 2006
425 out of 552 software firms in Israel are located at the Tel Aviv metropolitan area
Employees Firm numbers Percentages
1-10 175 41.2
11-26 98 23.1
26-50 57 13.4
51-100 29 6.8
101-250 19 4.5
251-500 10 2.4
501-1000 5 1.2
1001-2000 4 0.9
> 2000 3 0.7
Unknown 25 5.9 Source: Dan & Bradstreet report, 2006
Objective
Examination of the distance factor as a spatial organization power Utilizing five nodes:
Old CBD
New CBD
The location of the three largest firms employing more than 2000 workers each.
Old CBD
New CBD
Ness IT2450 employees
Matrix2150 employees
Team2077 employees
Methods employed to examine relationships with distance:
1. Major rings and sectors using LQ.
2. Aggregation by 5 km.-wide rings.
3. Single node – log-linear regression
4. Piecemeal linear regression
5. Piecemeal log-linear regression
6. Multi-focal regression
Software firms, 2006Location quotients of employees in software firms
Location in metropolitan area
Employees in software firms
LQ Employees, excluding the three largest firms
LQ, excluding largest firms
Core (Tel Aviv Yaffo) 6,290 1.4 3,840 1.1
Inner ring 10,584 1.4 8,434 1.4
Northern Section 4,556 3.3 2,406 2.2
Eastern Section 5,963 1.9 5,963 2.4
Southern Section 65 0.0 65 0.0
Middle ring 11,071 1.1 8,994 1.2
Northern Section 1,811 0.8 1,811 1.1
Eastern Section 8,349 3.4 6,272 3.3
Southern Section 911 0.2 911 0.2
Outer ring 1,593 0.2 1,593 0.3
Northern Section 1,222 0.4 1,222 0.5
Eastern Section 366 0.2 366 0.3
Southern Section 5 0.0 5 0.0
The software industry is concentrated at the north-eastern parts of the metropolitan area, mainly in the core area and in the inner and middle rings.
Software firms, 2006
Location quotient < 1
1 < Location quotient < 3
3 < Location quotient
Distance Bands Old CBD New CBD Team Matrix Ness-IT
0-5 4,013 17,208 17,215 5,416 16,850
5-10 15,245 6,319 8,653 17,548 6,769
10-15 6,489 3,785 2,122 2,833 3,951
15-20 2,685 1,095 125 3,049 1,485
20-25 92 1,085 1,380 252 413
25-30 976 38 35 405 67
30-35 30 5 5 32 -
35-40 5 - - - -
Beta -0.702 -0.837 -0.827 -0.660 -0.886
Adjusted R2 0.41 0.64 0.62 0.32 0.73
Prob. 0.052* 0.019 0.022 0.106* 0.019
Aggregation by distance bands
* Not significant
Aggregation by distance bandsNumber of employees by distance bands to the nodes
Distance decay curves of total employment originated from New CBD and Ness-IT produce statistically significant linear relationship with distance.
0
5,000
10,000
15,000
20,000
0-5 5-10 10-15 15-20 20-25 25-30 30-35 35-40
Distance bands
Em
plo
ye
es
Old CBD
New CBD
Team
Matrix
Ness-IT
N.S. – Not significant *** Significant at the 0.01 level
Distance to Intercept B Significance Adjusted R2
Old CBD 2.687 0.065 N.S. 0.000
New CBD 1.786 -0.023 N.S. 0.000
Team 3.177 -0.206 *** 0.055
Matrix 3.269 -0.221 *** 0.022
Ness-IT 3.418 -0.339 *** 0.049
Single Node analysisLog employees with distance
Three segments of firms: 1) larger than 750, 2) larger than 100 to 750, and 3) 1 to 100. Each segment fitted independently
Piecemeal linear regressionDistance to
Intercept Slope- b significance Adjusted R2 N Size of firms
Old CBD 1983 -36.081 N.S. 0.000 8 Large343 -5.288 N.S. 0.019 33 Medium
19 0.049 N.S. 0.000 357 Small
New CBD 1860 -32.573 N.S. 0.000 8 Large341 -6.429 * 0.054 33 Medium
20 -0.011 N.S. 0.000 357 Small
Team 1867 -48.427 N.S. 0.004 8 Large332 -6.807 * 0.081 33 Medium
22 -0.258 N.S. 0.002 357 Small
Matrix 2023 -46.587 N.S. 0.011 8 Large374 -8.811 ** 0.136 33 Medium
22 -0.254 N.S. 0.003 357 Small
Ness-IT 2041 -79.999 N.S. 0.198 8 Large340 -7.845 ** 0.111 33 Medium
22 -0.270 N.S. 0.003 357 SmallOnly for medium size firms there is a slight decreasing slope indicating a decrease in the number of employees with distance.
N.S. – Not significant * Significant at the 0.10 level ** Significant at the 0.05 level.
Piecemeal analysisLog employees with distance
The old and new CBD does not show signs of organizational power. All three large firms exercise a small but significant amount of organizational power on other large and medium firms.
N.S. – Not significant * Significant at the 0.10 level ** Significant at the 0.05 level.
Distance to Intercept Slope- b Significance Adjusted R2 N
Large and medium firms – more than 100 employeesOld CBD 5.344 0.048 ** 0.099 41
New CBD 6.224 -0.041 * 0.062 41
Team 6.169 -0.044 ** 0.089 41
Matrix 6.324 -0.045 ** 0.085 41
Ness-IT 6.200 -0.048 ** 0.100 41
Small firms – 1-100 employeesOld CBD 2.532 -0.006 N.S. 0.000 357
New CBD 2.560 -0.011 N.S. 0.016 357
Team 2.648 -0.021 ** 0.010 357
Matrix 2.687 -0.021 ** 0.011 357
Ness-IT 2.655 -0.023 ** 0.011 357
Multi-focal analysisLog employees with distance
Distance to Intercept B Sig. B Sig. B Sig. Adjusted R2
Two nodesTeam + Matrix
3.576 -0.199 *** -0.202 .004 0.072
Team + Ness-IT
3.466 -0.147 *** -0.222 .007 0.070
Matrix + Ness-IT
3.652 -0.149 ** -0.299 .00008 0.057
Three NodesTeam + Matrix + Ness-IT
3.727 -0.155 *** -0.166 .021 -0.172 .040 0.080
N.S. – Not significant ** Significant at the 0.05 level *** Significant at the 0.01 level
1. Proximity to specific nodes appears to have diminutive organizational power on employment in software.In particular, the Old and New CBDs do not display organizational power on the location of the software firms when analyzed by exact location.
However, significant though weak regression results were produced for the location of the firms with respect to their distance to the three largest firms. Team, located in the suburb of Petah Tiqwa has the most powerful organizational effect as reflected by the highest R squares.
2. Medium and large software firms (100+ employees) show some organization patterns around the three largest firms.Whereas, analysis did not procure any distance profiles for small firms. Thus, while large and medium firms are somewhat sensitive to the impact of distance, small firms are less organized.
Conclusions
Conclusions
4. Clearly there is a need to employ other techniques to reveal the structure of agglomeration in metropolitan areas.
3. Do we face new locational decision realities within urban spaces? Israel’s software firms tend to concentrate in the TA metropolitan area in the north-eastern sectors of the inner ring. However, the impact of distance within this area seem to be marginal and there is no central node that organizes the spatial distribution.This corresponds with Frenkel’s, Shefer’s and Roper’s (2003) findings about the tendency of innovative hi-tech businesses in Israel to locate in metropolitan areas. With diminishing distance relationships does the pattern correlates with Porter’s (1990, 1998, 2000) notion of clusters?