In light of stagnating regional prosperity and widening inter-regional disparities in many economies, this paper advocates for a fresh research agenda aimed at deepening our comprehension of regional economic development. This is achieved by exploring various conceptual perspectives, their practical applications, and suggestions for future research directions. Traditional viewpoints perceive development as a consequence of, and reliant upon, local economic structures. Higher regional performance is correlated with specific mixes of regional industries and human capital. This paper contends that for a deeper understanding of the mechanisms propelling regional economic development, adopting a relational approach is beneficial. Such an approach pays heed to the networks among economic actors across diverse spatial scales, from local to global. These networks generate knowledge, access to technologies, resources, and markets, thereby catalyzing income growth. To aid regional policy agendas, it is imperative to move beyond identifying regularities that shape development and delve into distinct regional pathways through systematic comparative analyses of local contextual and institutional conditions.
Introduction
The pursuit of economic development in stagnating or declining city regions has long been a focal point for economic geographers and regional economists. Recent years have seen a resurgence of interest in this topic due to the escalating disparities in income and employment opportunities within national urban systems. This paper delves into the complexities of regional economic development, aiming to shed light on different models, their limitations, and avenues for future exploration.
Industrial Structure and Regional Performance
It is widely acknowledged that variations in industrial composition are linked to diverse levels of regional economic development. However, there is less consensus on the ideal mix of industries and localized resources that promote economic growth. Previous studies have explored this. Whether higher levels of regional industrial specialization or a more diverse industrial structure lead to elevated regional prosperity. Influential research by Glaeser et al. (1992) examined economic growth in U.S. cities, testing hypotheses related to Jacobs, Marshall-Arrow-Romer, and Porter’s externalities. Their analysis indicated that city industries with more local competition and higher industrial diversity tended to exhibit faster growth. Another strand of research, pioneered by Frenken et al. (2007), focuses on “related” and “unrelated variety” in local industrial structures. This approach posits that a large number of related industries, termed “related variety,” stimulates economic growth by creating new markets and employment opportunities. On the other hand, “unrelated variety,” which measures the diversity of employment at an aggregate level, does not spur growth. Empirical studies using models similar to Glaeser et al. (1992) have found positive associations between high local-related variety and regional employment growth, while specialization effects drive productivity growth.
Human Capital and Regional Development
Apart from industrial structure, the role of human capital in regional development cannot be overstated. Regions with a high proportion of college-educated workers tend to exhibit stronger economic performance. However, the mechanisms through which human capital influences development remain underexplored. Research by Giannone (2017) and Moretti (2004) highlights the importance of skilled labor in driving regional prosperity.
A Relational Perspective
This paper advocates for a relational perspective in understanding regional economic development. Such an approach focuses on the economic linkages and connections within and across territories, from local to global scales. These networks facilitate the exchange of ideas, technologies, and access to markets, thereby fueling regional growth. While structural conditions play a crucial role, interactions among economic actors often drive development.
Contextual and Institutional Influences
To comprehend why regional development paths diverge despite similar structural characteristics, it is essential to consider the role of contextual and institutional factors. Comparative analyses of these influences, coupled with individual actions, offer insights into the dynamics of economic growth. By examining specific geographical cases, researchers can uncover the unique triggers for growth or decline within regions. The discussed models raise several questions regarding the modeling of regional economic development about industrial structure. One of the key issues is the emphasis on regional industrial diversity as an optimal structure for promoting economic growth, a relationship that remains less than clear. Kemeny & Storper (2015) argue that previous research’s focus on relative specialization (the share of employees in a specific industry) rather than absolute specialization (the total number of employees in an industry) obscures the fact that the benefits of specialization stem from the overall economies of scale within a city-regions industries, not just their relative sizes. They also note significant wage level differences even at a granular level (e.g., the 6-digit SIC industry level) across regions. This suggests that boosting employment in a high-income industry in one region may not necessarily have uniformly positive effects across all regions. Therefore, it becomes crucial to consider a broader range of potential influences on regional economic development and understand why certain industries might thrive in some cities more than others. In section 5, we will explore how specific institutional and contextual conditions can influence the relationship between industrial structure and economic development. This includes the idea of opening up enclosed regions to incorporate development stimuli from external sources. Another unresolved issue is that while approaches centered on industrial structure are effective in connecting regional development with past industry configurations, they often fall short of explaining the specific changes that lead to observed shifts in development. For example, the studies mentioned earlier link growth in regional income levels to structural indicators in previous periods but do not delve into the changes in these indicators. To illustrate, Table 3 presents some analyses from the related variety literature for U.S. Metropolitan Statistical Areas (MSAs), where related and unrelated variety are calculated based on NAICS data. The first two columns of Table 3 associate related and unrelated variety levels in the base year 2010 with changes in regional development indicators between 2010 and 2017 (with employment density as a control). The regional development indicators include per-capita wage and salary income growth rates and employment growth rates. Interestingly, this model closely echoes key findings of previous studies in the United States. Notably, the table shows that related variety in the base year is significantly and positively linked with employment growth but not with income (productivity) growth. Moreover, it finds a positive relationship between unrelated variety and income (productivity) growth. While previous studies suggest that policies should promote related variety, the implications are less clear. The subsequent columns in Table 3 introduce a different model formulation, focusing on the changes in independent variables about regional development. This aims to understand the changes associated with shifts in development, rather than assuming development solely follows from a past status. Interestingly, the reformulated model shows that the change in related variety is not significant, while the change in unrelated variety is negatively associated with income growth. This suggests that increases in unrelated variety may lead to decreases in per-capita income. While the table implies a potential link between employment (income) growth rates and cities with high levels of related (unrelated) variety, it doesn’t confirm a causal relationship. It could be, for instance, that employment and related-variety levels are correlated but that the mix of related industries in cities has remained relatively stable. In such a scenario, related variety may not explain changes in employment over time, which might be prompted by different processes. While we agree that an urban area’s industrial structure shapes future growth opportunities, the existing industrial structure may not fully reveal how shifts in economic development occur and what economic actions drive them. This calls for a research agenda to deepen our understanding of underlying processes and to include economic relationships across various spatial scales that can ignite development.
Human Capital and Regional Performance
In parallel to industrial structure, scholars widely acknowledge that differences in human capital strongly correlate with variations in economic development across regions. Human capital can be conceptualized in various ways, with one common measure being the share of workers with a college degree. Studies consistently find a positive relationship between a region’s overall human capital and its level of economic development. For instance, Moretti (2004) discovered that an increase in the share of college graduates in a city results in higher nominal wage incomes for workers across all education levels in that city. This suggests that when non-college-educated and college-educated workers’ labor is not interchangeable, an increase in college graduates’ numbers boosts the productivity of non-college-educated workers. Moreover, non-college-educated individuals can learn from their college-educated counterparts, leading to enhanced productivity and income levels (Moretti, 2004). Co-location of college-educated workers can facilitate better exchanges of knowledge in both quantity and quality, thereby driving up productivity and income levels (Davis & Dingel, 2019).
Figure 1 depicts the relationship between a city’s income and its human capital levels, defined here as the share of college-educated workers, across 479 Core-Based Statistical Areas (CBSAs) in the United States. CBSAs combine Metropolitan Statistical Areas (MSAs) with an urban core of at least 50,000 people and micropolitan statistical areas with 10,000 to 50,000 people. Panel A) of Figure 1 illustrates changes in the natural logarithm of per-capita wage and salary incomes between 2010 and 2017 relative to the share of college-educated workers in 2010. This design mirrors that of Glaeser et al. (1992) and Frenken et al. (2007) in Tables 1 and 2. The sizes of the circles represent the total CBSA employment in both plots. Panel B) showcases changes in the share of college-educated workers and per-capita income between 2010 and 2017. Similar to columns 3 and 4 of Table 3, Panel B) goes further by relating economic development to changes in the independent variable, not just its past state as in Panel A). The positive slopes of the fitted regression lines in both panels underscore two points: a) cities with higher shares of college-educated workers in 2010 tended to experience larger improvements in per-capita income between 2010 and 2017, and b) cities with greater changes in their share of college-educated workers generally saw more significant shifts in income levels. However, there are evident outliers in both plots, indicating that factors beyond human capital also significantly influence economic development differences across city regions.
Consistent with Figure 1, prior research demonstrates a robust positive relationship between regional human capital levels and economic development. Additionally, there’s evidence that the benefits of co-location among college-educated workers within a city region may be increasing over time. Cities with high shares of college-educated workers in 1980 have continued to attract more of these workers, leading to a concentration of talent in certain urban areas. This trend has been linked to the widening income gap among U.S. cities (Giannone, 2017) and to enhance overall quality of life in already prosperous cities due to the concentration of college-educated workers (Diamond, 2016). It is widely accepted that the clustering of college-educated workers is strongly tied to regional economic development, particularly in already successful urban areas. However, while a skilled labor force is undoubtedly vital for thriving regions, the precise mechanisms linking this workforce to economic production and growth stimulation remain less clear. Many existing models often overlook the mechanisms and agents behind development, warranting further investigation. An important unresolved question in the realm of human capital approaches is whether the growth of college-educated workers (or a broader shift in human capital) in a city precedes or follows economic development. Storper & Scott (2009), for instance, propose two views: one where jobs move to areas with higher human capital, and another where people move to regions with better job opportunities. This distinction is crucial for understanding whether the influx of highly skilled workers causes economic development or whether their migration reflects deeper processes of job creation and opportunity, making these cities attractive places to move to. Alternatively, underlying processes and migration patterns might mutually reinforce each other. Similar to the discussions on industrial structure, understanding the true drivers of development and external impulses is essential for a comprehensive understanding. Generally, it is agreed upon that the growth of college-educated workers, as well as other forms of urban human capital such as the “creative class” (Florida, 2002; 2017) or diverse birthplace backgrounds (Ottaviano & Peri, 2006; Buchholz, 2021), is associated with positive trends in regional economic development (Glaeser & Gottlieb, 2009). Panel B) in Figure 1 supports this argument. However, this relationship is not automatic. The rise in the share of these workers in a city sets off co-evolutionary processes where new interactions and learning capabilities lead to innovation, economic growth, and increased migration, thereby boosting urban human capital further. Like industrial structure, human capital plays a critical role in regional economic development by shaping the quality of interactions that can foster new technologies and jobs, but this potential is realized only when these interactions materialize. Understanding causality in regional economic development thus necessitates identifying and examining the mechanisms behind economic growth (Sayer, 1992). Conventional models often fall short in this aspect, as they primarily associate regional development with local structures while overlooking the potential benefits of dynamic linkages within and across regions, as well as their contributions to human capital formation. In the subsequent sections of this paper, we will adopt a relational perspective to extend and complement existing models, highlighting the potential of such linkages as indicators of economic actions and interactions that drive regional development.
A Relational Perspective of Regional Economic Development
Studies on regional economic development frequently adopt an explicit evolutionary perspective, aiming to explore the “processes by which the economic landscape […] changes from within over time” (Boschma & Martin, 2010a: 6 f.). Typically, this involves identifying localized structural conditions that influence regional production and innovation (Boschma & Martin, 2010b). As discussed earlier, conventional model designs often focus on the impact of regional structural conditions at a single point in time on future development. While it’s evident that broad regional structures shape future development possibilities, the limited attention to the actual mechanisms triggering this development, and how these unfold through linkages at various spatial scales, renders such explanations incomplete. To address this gap, we employ a relational perspective that investigates how actors, including firms acting collectively, develop networks through which growth impulses are transmitted and collective action is mobilized to stimulate economic development (Bathelt & Glückler, 2011; Glückler, 2013). Specifically, we use a region’s connectivity at different spatial scales as an indicator of such economic activities (Bathelt & Buchholz, 2019; 2021). Although comprehensive data on direct and extensive economic interactions across all regions within a country is not available, we can identify specific network structures spanning from local to global scales. In our analysis of the U.S. urban system, we utilize firms’ subsidiary networks within a city region as an indicator of growth triggers transmitted to or created within that region. These subsidiary networks can be viewed as long-term infrastructures facilitating the exchange of goods, knowledge, and human capital, both locally and with other regions domestically and internationally. While not the sole measure of connectivity (Glückler, 2013; Cano-Kollmann et al., 2016), these networks are crucial for understanding how resources and market opportunities integrate into regional production and innovation networks. Global subsidiary linkages, in particular, offer access to new technologies and markets, but they also entail risks due to fundamental institutional differences that need to be navigated (Cantwell et al., 2010; Glückler & Bathelt, 2017). Regions without firms capable of managing these global risks may instead establish domestic linkages within a more homogenous national institutional context, where trust is easier to build (Trippl et al., 2009). While the potential gains from extended domestic market access can be substantial, the opportunities to access fundamentally new knowledge and market segments may be limited (Li & Bathelt, 2018). Local-level networks are also crucial for triggering collective action and catalyzing regional development potential. These networks have long been central to studies on industry agglomeration and clusters and are vital for integrating external knowledge and market opportunities into local production and innovation networks. An excellent database for capturing firms’ subsidiary networks at various spatial scales in the United States is the LexisNexis Corporate Affiliations database. This database provides information on ownership hierarchies for the largest public and private firms from 1993 to 2017. For each firm, we identified the locations of headquarters and majority-owned subsidiaries within city regions in the United States (CBSAs) and in other countries. This enabled us to construct comprehensive subsidiary networks and compute a series of network indicators at the local, national, and international levels, reflecting the channels through which economic growth impulses are transmitted and created. In our model, we used the following variables: 1) Scale, representing the number of firms in the LexisNexis database within a CBSA to capture potential agglomeration and scale effects promoting knowledge creation and productivity increases; 2) Local Connectivity, indicating the subsidiary linkages within a CBSA; 3) National Connectivity, measuring how many other U.S. CBSAs a particular CBSA is connected to through domestic firm linkages; and 4) International Connectivity, capturing the linkages to foreign-country capitals, global cities, and other cities a particular CBSA has. National and International Connectivity can also be subdivided into inward and outward components, depending on whether the subsidiary linkages are directed into or out of a city, respectively. These variables were computed annually from 1993 to 2017, with all connectivity measures normalized by the number of firms in a CBSA to represent connectivity levels relative to a city’s overall size.
Impact of Connectivity Measures on Regional Income Levels
Table 4 presents panel regressions, both with and without the distinction between Inward and Outward Connectivity, to estimate the effects of connectivity measures on the logarithm of per-capita wage and salary income in a Core Based Statistical Area (CBSA). The model incorporates CBSA and year-fixed effects, providing insights into how changes in the independent variables (all in logged form) influence changes in the dependent variable over time. This methodology mirrors the models in panel B) of Figure 1 and columns 3 and 4 of Table 3. Our approach diverges from many conventional quantitative models of regional development in three key aspects: firstly, we emphasize changes in independent variables as catalysts for development; secondly, we measure the tangible linkages between actors that drive the creation of new technologies and labor demand; and thirdly, we include the impacts of extra-regional linkages on development. The positive signs of all coefficients in Table 4, accompanied by generally low p-values (except Local Connectivity), underscore two critical points. Firstly, it is vital to consider the actual connections between actors, such as firms, when modeling development. Economic progress is fundamentally catalyzed by the actions and interactions of workers, firms, and organizations. Secondly, delving beyond the intra-regional level in the study of regional development is essential, as pivotal triggers often stem from extra-regional connectivity and related economic impulses, both domestically and globally.
Table 4 highlights that urban-regional income levels not only rise in conjunction with intra-regional processes, such as agglomeration effects but also benefit from growing domestic and international connectivity in subsidiary networks—especially with outward linkages exhibiting generally higher significance levels compared to inward ones (Bathelt & Buchholz, 2019). These models are robust, showcasing the efficacy of a relational approach to regional economic development. Such an approach accounts for the actual connections driving the transmission and creation of knowledge, as well as enhancements in productivity (Bathelt & Buchholz, 2021). The precise nature of these relationships at any given time is pivotal for growth and innovation within a city, driving changes in development. While a relational perspective aligns with the industrial structure and human capital approaches discussed earlier, they can also be formulated and tested in relational terms. For instance, Frenken et al. (2007) propose that related variety serves as a proxy for inter-industry knowledge spillovers. Incorporating such linkages into model designs enables the examination of how cities with a higher number of college-educated workers foster improved knowledge exchange environments, ultimately enhancing workers’ productivity and income levels (Moretti, 2004; Davis & Dingel, 2019). In the subsequent section, we argue that despite capturing various linkages between firms (and their workers) within and outside their local contexts, significant forces driving disparities in development across regions remain unexplained. This discrepancy often hinges on local contexts and institutional settings, which wield considerable influence in generating distinct development dynamics.
Geographical Contexts, Institutional Settings, and Regional Development
Despite the noteworthy findings in Table 4, a notable proportion of changes in per-capita income across cities remains elusive. The low R2 values in our models with city and time-fixed effects—though not explicitly included in the R2—underscore this gap, a common challenge in longitudinal models of regional economic development (refer to Table 3 and Figure 1). As previously emphasized, the three approaches discussed herein are complementary, with past structural conditions shaping the potential for social and economic relations to yield economic benefits. Yet, even with meticulous measurements of diverse relationship types among various actors, mediated by historical structural conditions, substantial developmental shifts remain unaccounted for. This suggests the presence of other elusive or intricate influences contributing to changes in regional economic development, leading to deviations from predicted paths or fostering new developmental dynamics. The regularities identified through large-scale macro analyses, such as the regression models applied in this paper, primarily illuminate average relationships across numerous cities and timeframes. Consequently, highly aggregated quantitative variables at the city level may fall short of explaining specific deviations in development dynamics for each unique regional case. However, it is surprising that few studies venture beyond model regularities to investigate geographical cases at the grassroots level. Why do certain regions conform to or deviate from model regularities, and what underlies these discrepancies?
There exists no universal answer to the causes of substantial regional economic development residuals. Nonetheless, it is widely believed that difficult-to-measure, context-specific institutional factors wield considerable influence (Rodríguez-Pose, 2013). For instance, Storper et al. (2015) find that commonly cited factors—such as immigration, cost-of-living, urban human capital, public policies, or chance—fall short in explaining the varying levels of development between San Francisco and Los Angeles city regions. Instead, their study suggests that these differences stem from a blend of inter-organizational and interpersonal networks, alongside institutional settings. Specific contextual conditions may elucidate why certain cities and regions underperform or overperform concerning their structural characteristics.
To illustrate this point, Figure 2 presents variations in income changes for U.S. CBSAs concerning the predictions of the three distinct models discussed herein: panel A) references the industrial structure model (Table 3, column 3), panel B) corresponds to the human capital model (Figure 1, panel B), and panel C) reflects the relational model (Table 4, column 1). Only CBSAs from each dataset that experienced average changes in their independent variables for the respective development models are plotted (i.e., changes within 0.5 standard deviations from the mean change of all independent variables). Income changes are standardized. Ideally, if each model provided a perfect explanation, income changes would closely align with model predictions. However, substantial variations, as depicted in the box-and-whisker plots in Figure 2—highlighting the 25th, 50th, and 75th percentiles of the data alongside extreme outliers—indicate otherwise. A significant deviation from the predicted income change of ‘0’ suggests that the underlying model struggles to explain regional development effectively for a particular city region. The figure indicates that while each of the three regional development models makes good predictions for income development in many CBSAs, predictions for other CBSAs often diverge from actual income changes. It is precisely in these cities where a closer analysis of contextual and institutional conditions becomes crucial to decipher their regional development dynamics, indicative of diverse development mechanisms.
For Example
Odessa, Texas, emerges as an outlier in the relational model (panel C), exhibiting considerably higher income increases than predicted by the model. This CBSA, deeply entrenched in the fossil fuel industry, witnessed substantial income growth post-financial crisis from 2010 to 2017 despite experiencing average changes in connectivity levels. However, Odessa navigated a preceding economic downturn by diversifying into other energy sectors and logistics (Federal Reserve Bank of Dallas, 2018). The city’s focused reliance on the fossil fuel industry created a unique development context that defied the linear logic of the models employed.
Conversely, Flint, Michigan, depicted in panel C, portrays a contrasting scenario, with income development lagging behind predicted changes. On the surface, Flint epitomizes the devastating effects of deindustrialization processes on the U.S. automobile industry. However, based on the connectivity variables in model 1 of Table 4, one would have anticipated significantly higher income changes than those observed. Flint’s economic decline corresponded with a loss in economic diversity (reduction in unrelated variety), the persistence of the former industrial structure (marginal increase in related variety), and network linkages that lost their capacity to spur development. Research on deindustrializing cities within the U.S. Rust Belt underscores how the structure of civic networks shaped distinct capacities for city-regions to mobilize actors and resources, branching into new economic activities (Safford, 2004). Moreover, several studies highlight racism’s role in Flint’s economic decline and that of other deindustrializing U.S. cities, often with sizable Black populations, linked to austerity measures, private sector divestment, constraints on local government autonomy, and punitive policies targeting Black individuals (Pulido, 2016; Hackworth, 2020).
Conclusion
In our quest to deepen the understanding and explanation of regional economic development, we commenced this paper by underlining the necessity to move beyond models solely focused on the impact of broad structural attributes. Instead, we advocated for a perspective that delves into how cities and regions are intricately woven into networks of relationships spanning multiple spatial scales. Undoubtedly, structural facets such as the industrial composition or the endowment of human capital in a regional economy play pivotal roles, shaping the potential for city-regions to cultivate new capabilities and economic prowess. However, we critiqued existing approaches to regional economic development for their tendency to overly emphasize localized structural features, often at the expense of the dynamic relationships and interactions that foster the transmission and generation of fresh knowledge and market access—ultimately spurring income growth and developmental progress. Embracing a relational viewpoint necessitates the consideration of both intra-city-region connections and extra-local linkages, both domestically and globally. We contend that without these conduits for developmental impulses, it becomes less probable that a city boasting high levels of human capital or an advantageous industrial structure will undergo significant developmental strides. As elucidated in this paper, commonplace industrial structure variables align with city-regions experiencing regional economic expansion, yet shifts in these variables may transpire gradually, yielding heterogeneous effects. While advocating for a relational stance on regional economic development marks a progressive stride, we also posit that exhaustive quantitative models can only unveil average relationships between independent and dependent variables. Deviations persist, hinting at distinct developmental dynamics and logics. The models we’ve discussed represent abstract generalizations, devoid of specific contextual and institutional conditions crucial for development. Hence, we advocate for a future research trajectory that places greater emphasis on systematic comparative examinations of cities and regions (Storper et al., 2015; Pike, 2021). This approach integrates quantitative and qualitative analyses into narrative frameworks illustrating divergent economic development trajectories. This endeavor will enable a deeper comprehension of the mechanisms guiding established pathways and the genesis of new impulses for economic development. A fusion of expansive statistical analyses and comparative case studies holds significance from a regional policy perspective. It equips policymakers with insights into cross-regional regularities alongside nuanced understandings of unique local contexts, thereby fostering economic development. This proposition assumes particular importance for small, lagging city regions, which have garnered considerable scholarly and public interest yet have received comparatively less attention in the regional economic development literature than their larger, above-average development counterparts.
Leave feedback about this