Abstract
Government involvement in facility construction is typically justified on the basis of ex-ante predictions of economic impact resulting from events hosted at the new or upgraded facility. This paper examines the impact of facility construction on construction sector employment and real GDP across 15 New Zealand cities between 1997 and 2009. Results from static and dynamic models indicate that certain types of facilities had short-term (during construction) positive impacts on construction sector employment growth, although only stadium projects generated positive post-construction employment impacts. There is also little in the way of empirical evidence to suggest that new or upgraded facilities had any significant impact on local area real GDP either during or post-construction.
Acknowledgements
I would like to acknowledge the support of colleagues in the School of Economics and Finance at Massey University, and sincerely thank two anonymous referees and Dorian Owen at the University of Otago for their insights and constructive criticism in reviewing this manuscript. The usual disclaimer applies.
Disclosure statement
No potential conflict of interest was reported by the author.
Notes
1. Carlino and Coulson's (Citation2004) findings were later re-examined and the results found to be sensitive to changes in the model and the data used (Coates, Humphreys, & Zimbalist, Citation2006).
2. Fully referenced versions of these tables are available upon request from the author.
3. The location quotient is measured using the following formula: .
If the location quotient equals 1, this means that TLA i has the same concentration of employment in the construction sector as the nation. This can be interpreted to mean that the employment in the construction sector in the TLA meets the needs for local demand. If the location quotient is less than 1, this can be interpreted to mean that employment in the construction sector for the TLA is insufficient to meet local demand. Likewise, if the location quotient is greater than 1, this suggests that local employment is more than sufficient to meet demand.
4. Parameter estimates for the TLA fixed effects, period fixed effects and TLA-specific time trends are not reproduced in the tables, but are available upon request from the author.
5. My sincere thanks to one of the referees for this suggestion.
6. Several alternative estimation procedures exist for ‘long’ (large T) panels – including the LSDV and corrected LSDV (LSVDC) approaches (Kiviet, Citation1995). A case could be made for applying GMM methods to the dynamic models, but with ‘long’ panels this can lead to other econometric problems (Judson & Owen, Citation1999), so coefficients were estimated using the LSDV and LSDVC methods. Similar estimates were obtained for both approaches, although they are reported only for LSDV – the complete results are available upon request from the author.
7. The significant negative coefficient on the lagged GR_EMP_CONST variable likely reflects the seasonal nature of employment – an increase in growth in one quarter is likely to lead to a decrease in growth in the next (and vice versa).