Do jobs follow people or people follow jobs? A meta-analysis of Carlino–Mills studies

ABSTRACT

Do jobs follow people or people follow jobs? A meta-analysis of Carlino–Mills studies. Spatial Economic Analysis. This study examines the classic question as to whether ‘jobs follow people’ or ‘people follow jobs’ by performing a meta-analysis of 321 results from 64 Carlino–Mills studies. It is found that the results are highly divergent, but that more results point towards ‘jobs following people’ than towards ‘people following jobs’. When it comes to the reasons for the variation in results, we find that the results are mostly shaped by the geographical location, spatial resolution, and population and employment characteristics present in the data, as well as by the model’s specification, its functional form and the spatial weight matrix specification.

摘要

工作随人、抑或人随工作？卡利诺—米尔斯（Carlino–Mills）研究的元分析。Spatial Economic Analysis. 本研究透过对六十四个卡利诺—米尔斯研究的三百二十一项结果进行元分析，检视 “工作随人” 抑或是 “人随工作” 这个古典的问题。本研究发现结果高度歧异，但有较多的结果指向 “工作随人”，较少指向 “人随工作”。而在有关研究结果变异的原因方面，我们发现这些结果大多由数据所呈现的地理区位、空间分辨率以及人口与就业特徵所形塑，并取决于该模型的特殊性、功能形式和空间加权矩阵特殊性。

RÉSUMÉ

Est-ce que les emplois suivent les gens ou est-ce que les gens suivent les emplois? Une méta-analyse des études de Carlino–Mills. Spatial Economic Analysis. À partir d’une méta-analyse de 321 résultats provenant de 64 études de Carlino–Mills, cette présente étude cherche à examiner la question fondamentale, à savoir si, oui ou non, ‘les emplois suivent les gens’ ou ‘les gens suivent les emplois’. Il s’avère que les résultats obtenus divergent sensiblement. Toujours est-il que davantage des résultats laissent indiquer que ce sont ‘les emplois qui suivent les gens’ plutôt que ‘les gens qui suivent les emplois’. Alors, qu’est-ce qui explique la variation dans les résultats? Il s’avère que les résultats sont influencés dans la plupart par l’emplacement géographique, par la résolution spatiale, et par les caractéristiques de la population et de l’emploi qui figurent dans les données, ainsi que par la spécification du modèle, par sa forme fonctionnelle et par la spécification de la matrice de poids déterminée à l’échelle spatiale.

RESUMEN

¿Siguen los trabajos a las personas o las personas a los trabajos? Un metaanálisis de los estudios de Carlino-Mills. Spatial Economic Analysis. Mediante un metaanálisis de 321 resultados de 64 estudios de Carlino-Mills, en este estudio examinamos la cuestión clásica de si ‘los trabajos siguen a las personas’ o ‘las personas siguen a los trabajos’. Observamos que los resultados son marcadamente divergentes, pero que más resultados demuestran que los ‘trabajos siguen a las personas’ y no que ‘las personas siguen a los trabajos’. En lo que respecta a los motivos para la variación en los resultados, vemos que los resultados están definidos sobre todo por la ubicación geográfica, la resolución espacial y las características de población y empleo presentes en los datos, así como por la especificación del modelo, su forma funcional y la especificación de la matriz del peso espacial.

KEYWORDS:

• Population–employment interaction
• jobs–people causality
• simultaneous equations
• adjustment model
• Carlino–Mills model
• meta-analysis

关键词:

• 人口—就业互动
• 工作—人的因果关系
• 联立方程式
• 调适模型
• 卡利诺—米尔斯模型
• 元分析

MOTS CLÉS:

• interaction population–emploi
• causalité emplois–gens
• équations simultanées
• modèle d’ajustement
• modèle de Carlino–Mills
• méta-analyse

PALABRAS CLAVES:

• interacción entre población y empleo
• causalidad entre trabajos y personas
• ecuaciones simultáneas
• modelo de ajuste
• modelo Carlino-Mills
• metaanálisis

JEL:

• C3
• O18
• R11
• R14
• R23

DISCLOSURE STATEMENT

No potential conflict of interest was reported by the authors.

ORCID

Gerke J. Hoogstra http://orcid.org/0000-0002-0983-1017

Jouke van Dijk http://orcid.org/0000-0002-9522-5987

Raymond J. G. M. Florax http://orcid.org/0000-0002-2808-5953

Notes

1 To illustrate the importance attached to this publication, it was identified by Isserman (2004) as the most cited regional science article of 1987.

2 The theoretical foundations of the model are extensively described in many other studies and are not further discussed here. For more details, see Carruthers and Mulligan (2007) and Mulligan et al. (1999).

3 For simplification, the framework does not include multiple equations for subgroups of jobs or people, or additional equations for other possible dependent variables such as income or wages.

4 Here, ‘unique’ is taken to mean that a study must show a different value for at least one variable included in our meta-regression analysis.

5 In our study, weighting by dataset makes more sense than weighting by study, the approach usually employed in meta-analytical studies. In our sample, several CM studies had utilized the same data, and so the results of these studies are quite likely to be similar. Note, however, that results are not necessarily similar within individual studies because the same subset of data are not always used across the various model estimations. Later in our meta-regression analysis, we address this overlap of data when we control for possible interdependency among the study results.

6 Cut-off values of 2.9 and 607.9 square miles were used to distinguish between these different area observations, which were determined by calculating the mean and standard deviation of the average land area size (in natural log square miles) of the spatial observations from 321 datasets. Note that spatial areas such as US census tracts and US counties may significantly differ in size. For example, only US census tracts in highly urbanized areas tend to be classified as small-area observations and only US counties in rural areas fall under large area observations. For most of the CM studies, external internet data sources were used to obtain the necessary information on area size.

7 Initially, we also considered a range of other categories of location-specific variables, including natural amenities and recreational facilities, demographical characteristics (e.g., age and ethnic composition), labour market characteristics (e.g., unemployment and skill levels), and location characteristics (e.g., central locations and distance to the urban core). These categories turned out to be non-significant predictors and were therefore excluded from the final model.

8 See also note 5 above.