06Appendix

Linear (OLS) results

This report uses four OLS models to assess the role of different factors on the amount of new economy and clustered new economy activity across TTWAs. The dependent variable in Model 1 and Model 2 is the number of new economy firms per 10,000 working-age residents in a TTWA. The dependent variable in Model 3 and Model 4 is the number of clustered new economy firms per 10,000 working-age residents.

Logistic regression results

This report uses logistic regression models to assess the degree to which different factors alter the odds that an LSOA has a hotspot. In all models, the dependent variable is therefore a binary outcome reflecting whether or not an LSOA has a hotspot. The coefficients reported in the table are odds ratios.

In Models 2 to 4, the impact of the urban and Greater South East dummies (included as controls) on hotspot formation appear negative. This is because the other independent variables, such as jobs density or access to skilled workers by trains, contain the bulk of the advantages offered by cities and Greater South East locations. Without them, as can be seen in Model 1, the advantages of urban and Greater South East locations are substantial.

Some of the analysis of the impact of universities, and the results displayed in Figure 12 and Figure 14, originate in marginal effects (calculated in STATA) derived from these models.