Abstract
This introductory essay aims to provide a comprehensive overview of a collection of 17 articles, previously published in the Urban Studies journal, now consolidated as a ‘virtual special issue’. The articles contribute to numerous strands of what has, over the decades, become an extremely voluminous literature concerning the interplay between population change and productivity within cities. It is important to acknowledge that the volume of this literature makes it effectively impossible to summarise effectively in a single essay. This article instead presents a contextual overview, a glimpse into the various thematic strands arising from the literature, and an introduction to each of the selected virtual special issue articles themselves. It briefly highlights the unique contributions of these articles to the overarching objective of the Urban Studies journal, which is to advance our understanding of the urban condition. Yet, it is also worth noting that the literature does not neatly segregate into distinct strands. Indeed, several contributing authors within the virtual special issue rightly emphasise the significant interdependencies between population change, migration patterns, specialisation, firm location, urban and natural amenities, agglomeration economies, and productivity. This essay attempts to progress logically from one strand to another, with a specific focus on labour and housing markets, as well as transportation costs or systems, as mediators of these complex relationships. Finally, the essay concludes by bringing attention to several strands of future inquiry that might prove fruitful and would be of interest to the readership of this journal.
Introduction
In this virtual special issue (VSI henceforth) a compilation of articles previously published in the Urban Studies journal is brought together and placed within the broader context of existing literature. This endeavour showcases the distinctive contributions made by the Urban Studies journal to advancing knowledge in general and serves as a concise overview of various strands of literature, offering an introduction to the selected articles within the VSI. As we delve into the line-up of articles, careful consideration is given to identifying remaining knowledge gaps. Naturally, this draws attention to the potential for additional future contributions that can enhance our comprehension of the nexus of urban issues highlighted by the VSI contributions.
As implied by the title of the VSI, the central themes explored within the collection centre on agglomeration, population change (including migration), innovation, and productivity. The thematic threads are examined through the lens of comprehending urban housing markets. The articles in the line-up serve a dual purpose: they either contribute to one of the aforementioned strands of literature or shed light on how housing markets mediate the influences of these forces. The VSI is driven by the idea that the role of housing systems in facilitating or impeding agglomeration, population change, migration patterns, business start-ups, innovation, and productivity gains has, thus far, received inadequate attention within the academic discourse. Through its broad, interdisciplinary appeal to scholars in the field of urban studies, this journal has already contributed substantially towards unravelling the complex connections within and between numerous strands of literature.
In the following sections the article summarises the literature on: amenities and migration; amenities and the ‘attractiveness’ of urban places; urban accessibility and productivity; and the operation of urban housing and labour markets. It will be obvious that the concept of ‘amenity’ plays a central role in the narrative, but less obvious that a consistent and analytically rigorous approach has been adopted in the conceptualisation and/or analysis of amenities. Indeed, we note that ‘amenity’ is distinct from ‘amenities’, and that the latter has been broken down to further sub-categories including urban, natural, and cultural aspects. An observation repeated later in the article is that whether or not ‘amenities’, as they have been defined and measured in previous studies, actually give rise to ‘amenity’ depends on the characteristics of individuals, places and geographical contexts concerned. Furthermore, the related question as to whether amenity influences economic and demographic decisions and urban system outcomes has been approached with a diversity of methodological and empirical tools in previous studies.
As we progress through the subsequent sections it may become apparent that a surprisingly large, wide, and diverse set of physical, cultural, sociological, economic, and natural phenomena and processes have variously been categorised as amenities in previously published contributions. Some of these are more contestable than others. Natural aspects such as geographical features conducive to good views or accessibility to walks in nature (natural amenities), proximity to urban parks, green space, recreation opportunities (urban amenities) or to museums (cultural amenities) seem less contentious. Proximity to ‘creative’ and/or ‘like-minded’ people, recent graduates or young people seem more contentious, at least when examined through an urban or housing economics lens. For example, on the question of high housing prices these have been variously described both as a disamenity and as a proxy for multiple positive amenity effects in the literature. Which definition is more appropriate presumably depends on the disciplinary lens adopted in addition to the specific research question that motivates the researcher(s). Perhaps our difficulty in answering such questions stems from a more general failure of the literature to rigorously articulate what is meant by ‘amenity’ and ‘amenities’, to define what the tests should be, and the assumed directions of causality.
At the risk or hope of providing provocation, and thus helping to trigger further debate, a suggested analytical framework begins by defining what amenity is, and what amenities might include, and excludes other constructs and processes from consideration. However, before introducing a suggested framework we begin with a general introduction to the literature at the point at which it seems most logical – from a housing and urban economics viewpoint: the interplay between ‘amenities’ and population change. We note that the literature on the latter focuses mainly on migration (of people and their households rather than firms, industries, or other non-residential economic activities). In the next section we also begin to examine the diversity of ways in which scholars have attempted to define ‘amenities’ in the course of urban and regional research.
Amenities, migration, and population change
Rapid urbanisation and its profound impact on the world’s population have been subjects of great interest for researchers over many years and across various branches of study. Over this time, a vast body of literature has explored different facets of complex urbanisation processes. One prominent line of inquiry, which dates back at least four decades, focuses on the role of amenities in shaping migration patterns and explaining why some urban places have become more popular and populous over time. Prominent studies such as Graves and Linneman (1979) have argued that changing preferences, and hence demand, for location-specific urban amenities are linked to rising per capita incomes and lead to the rising importance of migration to higher amenity destinations over time (see also Clark et al., 2003; Clark and Cosgrove, 1991; Graves, 2014; Poston et al., 2009). These studies shed light on the link between amenity-driven migration and economic factors, highlighting the importance of understanding this dynamic process.
Examining migration flows in the USA, Partridge and Rickman (2006) found that labour market shocks only accounted for a portion of the temporal variation that they observed. This suggests that other factors influence migration decisions beyond purely economic considerations. Furthermore, Artz et al. (2021) conducted a study focussing on small towns in Iowa, USA. They revealed that concentrations of educated individuals, higher incomes, and local competition significantly encourage new start-ups. These findings underscore the multifaceted nature of migration patterns and the diverse factors that shape them. Yet, they also flag an important blurring of the boundaries between studies of residential and non-residential migration. Cities change over time through the migration of people and their households, but also through the locational choices of established and newly forming firms and industries. But clearly there are likely to be numerous feedback loops between these processes both spatially and over time. We return to this point later in the article. For the time being, we focus on population change, and specifically residential migration.
Yet the relationship between amenities and population change, through migration, is not straightforward. Different studies have presented varying perspectives on the relative importance of amenities compared to other factors. For example, Greenwood and Hunt (1989) emphasised the role of jobs and wages as more influential factors in labour migration within US cities (see also Darchen and Tremblay, 2010; Scott, 2010). Glaeser et al. (2001) found that urban amenities play a stronger role than natural amenities in attracting highly skilled individuals. Hsieh and Liu (1983), on the other hand, argued that inter-regional migration is more heavily influenced by amenities in the short-run, and the pursuit of better social life in the long-run. These contrasting viewpoints highlight the complexity of the phenomenon and the need for a nuanced understanding of the interplay between amenities and migration and its influence on population change.
Another dimension of this discussion revolves around identifying specific demographic groups motivated to migrate to urban areas with perceived high levels of (urban and/or natural) amenities. 1 Some studies have shown that student migration decisions to US cities are influenced by natural amenities (Dotzel, 2017), and Glaeser and Gottlieb (2006) proposed that income and education levels may shape individuals’ willingness to pay for proximity to certain urban amenities, such as museums. The potential circularity of this argument arises from the assertions that certain population cohorts are more likely to relocate in response to the availability of amenities, and that the provision (thus existence) of amenities is more likely given the presence and concentration of those cohorts.
Moreover, lifecycle influences and tax settings have been identified as additional factors that intersect with amenity-driven migration decisions. Studies such as Clark and Hunter (1992), Jensen and Deller (2007), and Conway and Houtenville (2003) have explored how age-specific considerations and tax policies affect the importance of amenities in migration choices. According to Nelson and Ehrenfeucht (2020), highly educated individuals, despite their general high levels of mobility, experience notable periods of immobility over their lifetimes. Their migration decisions are shaped by a complex interplay of economic, social, and amenity factors. Notably, individuals with higher skill levels exhibit a greater propensity to migrate to cities and regions offering urban amenities. The significance of these amenities may vary depending on the distance of the move (Gu and Shen, 2021; Liao and Wang, 2019; Yu et al., 2019; Zheng, 2016; Zheng et al., 2020).
The complicated interplay of urban, natural amenities and economics factors is further emphasised by Chi and Marcouiller (2013) who focussed on in-migration patterns to remote rural areas in Wisconsin, USA. Their findings revealed that natural amenities played a role in attracting individuals to rural areas but did not significantly influence migration to urban areas. Moreover, they contend that in-migration to rural areas is more closely tied to regional economic growth trends rather than amenities. In addition, Niedomysl and Hansen (2010), who conducted a study utilising survey data from Sweden, highlighted that employment opportunities outweighed leisure, recreational, and cultural factors (amenities) in influencing migration decisions. Importantly, this effect was particularly pronounced among highly educated workers (thus implying a feedback loop from concentrations of economic opportunity to potential productivity gains arising from the migration of more productive workers). These findings are consistent with other research, such as Chen and Rosenthal (2008) and Ulrich-Schad (2015), further emphasising the dominant role of employment opportunities in shaping migration choices, particularly for individuals with higher levels of education.
Geographical context also matters, and particularly the largely unanswered question about the applicability of North America-dominated findings to other parts of the world. While studies conducted in North America have formed a substantial part of the literature, there is a growing body of literature that focuses on other regions. Rodríguez-Pose and Ketterer (2012) expanded the analysis from the US to European cities and highlighted the relatively minor role of urban amenities in European migration research. They pointed out that Europe’s historical and climatic factors, as well as cultural and administrative barriers, may result in a different distribution and significance of amenities compared to North America.2,3 In a similar vein, Cheshire and Magrini (2006) argued that language, cultural, and institutional differences among European countries challenge the assumption of perfect mobility and equalisation of returns between cities. Their findings suggested a two-step migration process involving the selection of a national destination followed by the choice of a specific city.
This introductory summary of one ‘strand’ of the literature (as we have conceptualised it) concerns the interplay between amenities and migration-induced population change. Scholars have emphasised the role of urban and natural amenities in attracting individuals and fostering economic growth. However, unravelling the relative significance of amenities compared to other determinants, discerning the impact across different demographic cohorts, and unpicking the spatial variations inherent in this interplay requires further exploration. By delving into these nuanced aspects, we can gain a deeper understanding of the multifaceted dynamics at play in urbanisation processes. In addition, to this point we have made no real effort to scientifically or methodically define what is meant (and what is not meant) by the terms ‘urban amenities’ or ‘natural amenities’. We also return to this issue later in the article.
Related VSI studies
We now turn to exploring some contributions made by articles previously published in the Urban Studies journal. As highlighted previously, some scholars have attempted to stratify the literature to create two principal groups: those that emphasise urban and natural amenities and their role in attracting creative people and those that emphasise the importance of agglomeration. In a survey of the literature, Niedomysl and Hansen (2010) draw particular attention to the work of Jacobs (1969) and Florida (2002, 2005) and the notion that cities produce amenities that, in turn, attract talented people. They attribute the agglomeration arguments of Glaeser (1994, 2003) as emphasising a narrower aspect of creativity (skills/university education), which in turn brings economic growth. One such study in this tradition represents an important contribution to the literature and is chosen as the first paper lined up in this VSI.
Paper 1, by Glaeser and Gottlieb (2006), argues that American cities in the 1970s were driven by migration towards better climates and urban sprawl driven principally by the rise of the automobile. Indeed, they go further by noting that car ownership facilitates sprawl and emphasise that ‘cars require low densities and decentralisation’ (Glaeser and Gottlieb, 2006: 1279). As transportation costs decreased significantly during the 20th century, producer-orientated cities that clustered around natural resources were replaced by consumer cities driven by preferences for amenities. They also argue that a significant decline in crime rates increased the willingness to pay for living in cities, leading to a reduction in real incomes (income after controlling for housing costs). In their framework, population mobility ensures equilibrium in which places with higher incomes possess disamenities (such as higher housing prices or longer commute times) to offset them (see also Adamson et al., 2004; Glaeser and Mare, 2001; Winters and Li, 2017). Others have found that wage premia persist even after accounting for higher living costs in more attractive destinations (Garretsen and Marlet, 2017; Kemeny and Storper, 2012).
Paper 2, by Hamidi and Zandiatashbar (2019), rehearses the transformation of US cities towards knowledge-intensive businesses reliant on highly educated and creative workers, thought to be particularly attracted by cities’ diversity and high amenities. The novelty in this contribution is the argument that urban form plays a mediating role in the relationships between economic restructuring and migration. Both the city and the neighbourhood scales are seen to be important, but in different ways. The authors focus on urban compactness – an urban form that they argue as conducive to walkability, land use diversity, accessibility, promoting social capital, and reducing poverty and exclusion. They cite walkability as being particularly attractive to millennials, who they describe (following Florida, 2014) as the ‘driving force behind knowledge production’ (Hamidi and Zandiatashbar, 2019: 1583). Their empirical work finds that firm clustering is the most important variable in determining innovation capacity at the local level. Walkability and transport connectivity are also important. At the regional level, compactness becomes more important, and the conclusion is that urban sprawl suppresses innovation.
Paper 3, by Osland and Thorsen (2013), examines the joint effects of urban amenities and labour market accessibility on the spatial structure of housing prices in a Norwegian case study. Extending earlier work on US cities (Giuliano and Small, 1991; McMillen, 2004), this study focuses on a case study in which housing values diminish with distance from a single recognised urban centre (Stavanger) but may also be influenced by proximity to one of two identified sub-centres. The argument is that some households’ location decisions may be affected by the availability of employment opportunities in a sub-centre, reflecting the reality that households often comprise more than one economically active member. The hedonic analysis shows that the presence of a sub-centre within a 20 minute drive time impacts housing prices, building to an 8% discount from the more urban sub-centre and a 28% discount from the rural sub-centre. Overall, their conclusion is that urban sub-centres do not seem to be particularly important in explaining variations in housing prices within the urban area but that they matter more in rural housing markets.
In paper 4, Storper’s (2010) analysis critically examines the use of ad hoc combinations of variables in previous studies aimed at explaining variations in urban growth rates. Through an extensive literature review, Storper highlights the inherent interdependence and simultaneity among residential and firm locations, as well as specialisation. Building upon the literature on comparative growth economics, he argues for a more comprehensive framework that encompasses specialisation, human capital, and institutions to shed light on the mechanisms driving city growth. Notably, Storper emphasises the need for historical context in understanding urban economics, new economic geography, and the ‘new urbanisation economies’ approach to comprehending urban growth. His article underscores the importance of specialisation and discourages viewing it as the sole determinant of differential city growth and instead recognises it as an independent process. The discussion once again underscores the interplay of endogeneities and prompts relevant inquiries about the causal sequence. For instance, Storper observes that skilled and creative individuals are drawn to amenities, yet their migration often precedes the emergence and accumulation of such amenities. Additionally, he strongly advocates for considering institutions as a crucial explanatory factor for differential growth, dissecting this concept into three fundamental elements: the business climate, the shaping of labour force participation and effort levels, and the fostering of problem-solving abilities. These factors are believed to impact the identification and exploitation of new economic opportunities substantially.
The first four studies in this VSI therefore offer some useful initial insights to the intricate relationships between amenity factors, migration decisions, and population dynamics (see Table 1 for a concise summary). The next section aims to begin unpacking what is meant by ‘amenity’ and ‘amenities’ in a variety of ways set out, assumed or conceptualised in our review of the literature. We follow a similar format to that followed in this section, beginning with an overview of the literature before moving on to explain the key contributions made by the articles selected for the VSI. We conclude the section by suggesting a disciplined approach to conceptualising and defining urban and natural amenities, and for setting up research questions relevant to the domain of urban studies.
Summary of the first four VSI articles.
One of the themes that emerges from the first four VSI articles is that what constitutes an amenity may depend on geographical context. This breaks down further to the regional level (particularly in terms of applicability of findings outside North America) but can also break down even further to the urban/rural dimension. The question of what might constitute an amenity is then a nested one, because there are national contexts in which most or all cities possess a particular factor that could be considered an amenity in a different national context in which that factor is present in a subset of cities. The importance of defining the direction of causality is a second important theme to emerge from this section. As we have seen, some population cohorts may be particularly attracted to places or cities with an abundance of features from which those people derive utility. Yet, the creation and/or concentration of those features considered to be amenities may also depend on the presence and density of those population cohorts. This seems to suggest an important dimension to our proposed conceptual framework which we return to in our conclusions.
Amenities and the attractiveness of urban places
In this section we extend the discussion on defining amenity and amenities and introduce a second set of Urban Studies articles that make contributions to the debate. We begin by acknowledging that cities possess and have accessibility to both positive and negative amenities that have a range of effects on both the utility of households and on firms’ productivity (Graves, 2014). Differences in the type and quantity of amenities across cities and their regions can influence wages and housing costs. The literature contains many studies that have attempted to define a range of urban and natural amenities and/or test their influence on a range of urban outcomes. We also noted from our survey of the literature that there exists a wide range of factors that apparently or arguably contribute to cities’ attractiveness, although certain groups and themes are more commonly observed. This includes cultural and leisure amenities such as restaurants, theatres, cinemas, architectural and historical features, streetscapes, and monuments (Audretsch et al., 2021b; Backman and Nilsson, 2018; Buch et al., 2014; Falck et al., 2011; Garretsen and Marlet, 2017). Natural and climatic amenities include open space, blue space, climate, and accessibility to striking natural habitats (Backman and Nilsson, 2018). However, it is important to acknowledge that several studies caution against the arbitrary selection of amenity variables, overreliance on available data, or neglecting the influence of historical factors or other path dependency issues (Audretsch et al., 2021a; Storper and Scott, 2009).
As mentioned earlier, the importance of urban and natural amenities as predictors of migration within North America is widely recognised, but there is much weaker recognition of the significance of amenities to a range of urban and regional outcomes elsewhere (see Cheshire and Magrini, 2006, with reference to the European context). Nevertheless, studies have shown that regional amenities continue to be vital determinants of area attractiveness (to people and their households) even after controlling for economic factors (Bonasia and Napolitano, 2012; Germani et al., 2021; Nilsson, 2014; Rodríguez-Pose and Ketterer, 2012). The importance of amenities may vary based on the distance of the move and the motivations for doing so (Biagi et al., 2011). It also seems likely that what matters to people in the context of economic migration to cities is highly context-dependent, and this is revealed not just in the variables that show up as statistically significant in econometric modelling, but in the choice and definition of urban and natural amenity variables to begin with. In addition, disamenity such as crime or pollution may be important determinants of outward migration (de Sousa, 2014; Dustmann and Okatenko, 2014). Berger et al. (2008) found that higher levels of quality of life implied by urban and natural amenity variables were a significant determinant of inter-city migration flows, and housing prices, for Russian cities in the year 2000.
The conceptual linkages between aspects of diversity in urban places and economic productivity are explored in numerous contributions that primarily build on the work of Florida (2002, 2005). In a useful summary, Grant and Kronstal (2010) refer to melting pot, talent, Bohemian, and diversity indices as attempts to measure the attractiveness of urban places to the ‘creative classes’ of potential migrants. Yet, their empirical work found that migration levels were higher, and motivations much more complex than could be explained with reference only to these measures. Studies in this vein inevitably raise more questions than they can answer, including whether certain amenities can be considered as such for all population cohorts, or whether they are only amenities in the eyes of a subset of the population. Another question concerns whether concentrations of certain population cohorts can really be considered as an amenity in its own right, or whether this is too much of a stretch of the concept of amenity. We return to this issue in our discussion and conclusions.
Some studies have found that measures of diversity, cultural norms, and place-based culture are important predictors of differences between cities in terms of new firm start-ups, innovation, and/or entrepreneurship (Audretsch et al., 2021b; Audretsch and Belitski, 2013; Cheng and Li, 2012; Hopp and Stephan, 2012; Huggins and Thompson, 2014; Lee, 2015; Lee et al., 2004; Penco et al., 2020). Huggins and Thompson (2014) and Penco et al. (2020) are particularly good examples of approaches to building measures and indexes of culture and talent. Audretsch et al. (2021b) also found that the ethnic diversity of the labour force is a more important predictor of new firm start-ups per capita than the diversity of firms.
Lehmann et al. (2017) studied Germany’s 69 largest urban areas (all of which have a population over 100,000) to find that new firm start-ups are influenced by both proximity to concentrations of creativity and the availability of smart finance (venture capital transactions). Perhaps the most important question raised by studies of linkages between amenities and innovation or entrepreneurship concerns the order of causality. In other words, does the presence of concentrations of amenity really affect those outcomes, or is there an intervening process which concerns the relationship between amenities and the migration of people likely to engage in innovation and/or entrepreneurship?
A number of studies of migration between Chinese cities have found that environmental factors play a role in the settlement decisions of migrants (Cui et al., 2019; Gao and Sam, 2019; Li et al., 2017, 2020a) and regional levels of innovation (Zhang and Chung, 2020). Some studies have found that adverse environmental conditions, such as air pollution, contribute to the outflow of migrants from some Chinese cities (Chen et al., 2017). Others have found that employment, income, and career opportunities greatly outweighed urban amenity factors in migration decisions in Chinese cities in the early 2000s (see Liu and Shen, 2014a). The balance between economic factors and urban or natural amenities may also differ between economic sectors. For example, Song et al. (2016) examine a set of natural and cultural amenities, including climate, sea access, economic openness, policy environment, and infrastructure which they build into a composite total amenities variable. They found that total amenities play a more important role in Chinese regional attractiveness for IT service sector workers than those employed in IT manufacturing.
Other studies have attempted to analyse the importance of urban and natural amenities in relation to the underlying values possessed by different groups of individuals. One such example is the fifth article in this VSI –Hirschle and Kleiner (2014)– who examined migration to European regions, focussing on a number of aspects of Schwartz’s (2006) Human Value Scale. They examined 14 items belonging to four cultural factors (hedonism, individualism, social harmony, and traditionalism), with data collected between 2002 and 2010 within the European Social Values Survey. They found that although structural factors – particularly unemployment – were the strongest determinants of net migration, social harmony followed by hedonism were also important, statistically significant factors. However, they found that while hedonism had the expected direction of effect (proxying for places with a more vibrant culture), social harmony had an unexpected negative effect. They do not draw firm conclusions about the reasons for this but offer a number of explanations, including that places offering anonymity (indifference to being different) can be seen as having a particular form of amenity by some prospective migrants. This may also suggest that what counts as an amenity may be dependent on the context of the individual and their values in addition to other socio-economic characteristics, as well as the geographical context itself.
VSI paper 6 is by Echeverri-Carroll and Ayala (2011). This study builds on extensive literature on city scale, agglomeration, and productivity by considering the separate impacts of city scale and of knowledge spillovers on wages. They note that the US cities with the greatest density of human capital, as proxied by patents per capita, are not the largest ones but tend to have a population below 1.5 million. Following Wheaton and Lewis (2002), they reflect that knowledge spillovers occur when workers in the same industry group learn from each other and that city-scale triggers an incentive for workers to become more specialised. Both effects can be described as localisation economies, and both have the potential to lead to a gain in productivity. An important aspect of Echeverri-Carroll and Ayala’s (2011) contribution to the literature is their recognition of potential endogeneities between city scale, human capital, and innovation. Numerous studies show that increasing a city’s average education level, or concentration of employment in skill-intensive industries, raises wages overall. Yet, as the authors point out, there is also a possibility that more highly educated workers are motivated to relocate to cities with a higher population of skilled workers. This begs the recurring question of whether concentrations of certain population cohorts can be considered as an amenity, or whether such concentrations are merely a proxy for some other process(es) or phenomena, such as the existence of higher wages, wider choice of employment or more rapid promotion possibilities.
Paper 7 in the VSI series, by Qian (2013), unpacks the impacts of diversity and tolerance, arguing that Florida’s (2002, 2005) creativity index – based on technology, talent, and tolerance – confounds very different influences on innovation and entrepreneurship (see also Qian et al., 2013, who argue that human capital, new knowledge, and entrepreneurship are endogenous to each other, and largely explained by five exogenous factors: agglomeration, industry specialisation, quality of life, presence of universities, and social diversity). The central argument is that while innovation and entrepreneurship are driven by human capital and/or the concentration of a creative class, diversity and tolerance exert indirect effects by shaping the attraction and retention of talent. However, it is essential to differentiate between tolerance and diversity. Tolerance refers to an individual’s standard of accepting deviations from that standard, whereas diversity is a broader concept encompassing the inclusion of various groups. By examining 276 metropolitan statistical areas (MSAs) in the United States, Qian et al. (2013) demonstrate a strong correlation between tolerance and diversity. The study reveals that the most diverse MSAs tend to rank lower in terms of tolerance. Nevertheless, when both tolerance and diversity are included in an entrepreneurship model, only one of these factors proves statistically significant.
Furthermore, Qian et al. (2013) find a statistically significant negative relationship between the number of university faculties per capita and entrepreneurship, which they attribute to a displacement effect. Aamoucke’s (2016) study on German cities uncovers a strong association between the presence of higher education institutions and the formation of technology sector start-ups.
To conclude this section, our review of the literature suggests a nuanced portrayal of the nexus between amenities and city attractiveness, underscoring the intricate, multi-layered dependencies that influence this relationship (see Table 2 for a summary of the VSI papers discussed in this section). Amenities, both urban and natural, emerge as critical determinants of urban allure, acting as levers not only affecting household utilities but also shaping firm productivity and decision-making. This relationship, however, is not static but contingent upon a complex interplay of factors, including economic conditions, environmental considerations, cultural norms, and each city’s unique historical and spatial context. Caution is urged in regard to the selection and interpretation of amenity variables, suggesting a need for a more nuanced and contextual understanding. The discussion underscores the pivotal role that diversity, tolerance, and knowledge spillovers play in amplifying or attenuating the impact of amenities, thereby affecting the overall attractiveness of cities. This section thereby also contributes to the ongoing discourse in urban economics (and the Urban Studies journal itself) by highlighting the interconnected, complex nature of the amenity–city attractiveness relationship. Figure 2 below provides an illustration of how city attractiveness and amenities are interlinked and how they interact with households and firms. In the subsequent subsection, we delve deeper into the intricate relationship between urban accessibility as an amenity and its impact on firms, entrepreneurship, and innovation.
Summary of the first four VSI articles.
Urban accessibility as an amenity
There are various strands of literature that consider the influences of urban/natural amenities directly on urban economic outcomes, without explicitly considering the central mediating roles of residential households. The purpose of this section is to examine the relationship between urban accessibility and economic productivity. As an intervening step, we first consider the mechanisms through which amenities (as a broader concept) directly and indirectly influence productivity by influencing entrepreneurship.
Many studies of the connections between urban/natural amenities, agglomeration, and productivity have adopted a viewpoint centred on the individual worker, student, retiree, or household. Urban and/or natural amenities may influence productivity indirectly by altering innovation and entrepreneurship as intervening factors (Glaeser et al., 2010; Low and Isserman, 2015; Zhang et al., 2020). This strand of the literature is relatively sparse by comparison to other strands of agglomeration economics studies, and there have been previous calls for further research (see Glaeser et al., 2010, e.g.). However, the literature concerning the interface between amenities, firm behaviour, and productivity is not entirely absent. Some studies have shown that specific industry sectors are more sensitive to urban disamenities, including high crime rates (Rosenthal and Ross, 2010) or positive amenities, including the presence of younger people, an abundance of graduates, and low housing prices (Dorfman et al., 2011). Rickman and Wang (2017) examined patterns in US regional population growth in the US and concluded that, on average, natural amenities have a more powerful attractor effect on firms rather than households and that this leads to impacts on productivity through agglomeration economies. However, they also find that in larger metropolitan areas, households are more strongly attracted to natural amenities than firms.
There may be a case for further investigation to delve into the connections between urban and/or natural amenities, agglomeration, innovation, and their subsequent impacts on entrepreneurship, firm behaviour, and productivity. Yet, in our discussion of the panoply of amenities affecting entrepreneurship and productivity one element asserts itself with considerable weight: urban accessibility (encompassing elements such as transportation costs, employment accessibility, and proximity to the city centre) in influencing innovation, entrepreneurship, firm behaviour, and ultimately, productivity.
A body of literature highlights the significance of transportation costs and urban accessibility. Notably, Schmidt and Courant (2006) found that natural amenities (‘nice places’) outwith the metropolitan statistical areas (MSAs) but within accessible distance generate wage compensating differentials. Bayer et al. (2009) show that studies of willingness to pay for better air quality may under-estimate such effects when the costs of moving are high. Comparing a discrete approach to standard hedonic analysis, they find that willingness to pay is much higher for the former, based on an analysis of US MSAs.
Given its interdisciplinary focus on urban geography, planning and development, housing markets, and transportation, the Urban Studies journal has been at the forefront of contributing to this line of research. The next sub-section suggests this by highlighting the role of urban accessibility as a critical amenity in shaping innovation and entrepreneurship. It does this by providing a summary of four papers chosen for inclusion in this VSI.
VSI papers on urban accessibility, productivity, and entrepreneurship
The following articles examine how factors related to urban accessibility (including transportation costs, public transit improvements, employment density, and the concentration of metropolitan employment) impact productivity, firm location, and job matching within urban areas.
Paper 8, by Melo et al. (2017), tests the relationship between urban agglomeration and productivity for the 50 largest metropolitan areas in the United States. The main innovation in their analysis is the use of both employment density and employment accessibility, a measure of employment density which takes into account transportation accessibility proxied by automobile drive-time. They find an elasticity of between 0.7% and 1%, but that very similar results are obtained for employment density and employment accessibility. This, they argue, is suggestive of the benefits of agglomeration decay with a broadly linear functional form with distance from the city centre. However, although they find that the benefits of agglomeration could extend as far as 60 minutes’ drive-time from city centres, they find that the majority of the benefit accrues to the first 20 minutes. This supports the view that the local or urban neighbourhood level is of particular importance. The focus of their study on the largest metropolitan areas means that the importance of city scale is not extensively tested, and may also help to explain their finding that there are no obvious threshold effects.
In Paper 9, Chatman and Noland (2014) postulate that investment in public transit improvements may lead to higher productivity arising from the larger labour pools and better job matching that occur through higher city scale and greater employment density. They are careful to observe several open questions and are cautious in the interpretation of their empirical results. For example, they note that improved public transit systems could lead either to larger and higher density cities or, in the presence of cheaper land and the possibility of urban sprawl, to decentralisation of population and/or firms. They also point out that any increase in employment density that occurs in response to investment in public transit does not automatically signal that firms are seeking out the advantages that occur through improved spatial proximity.
The empirical work in this article examines three measures of agglomeration rather than a single variable. Specifically, it examines city centre employment density, the concentration of metropolitan employment within city centres, and metropolitan population relative to labour force size. The dataset includes 364 metropolitan areas in the United States. Although the study includes a range of models and approaches, the researchers report that the most reliable measure of agglomeration relates to the effect of city centre employment density on wages. They find that investment in public transport increases wages, but the effect is stronger for small metropolitan areas and for metropolitan areas with better pre-existing public transit systems.
In paper 10, Osman et al. (2019) argue that the presumption of a negative relationship between traffic congestion and economic performance has been lightly tested empirically. In a study of new firm formation in the San Francisco Bay area, the article examines five broad industry groups (advertising, biotechnology, computer systems design, IT manufacturing, securities) on the basis that these are seen as the main driving force behind economic growth. Grocery stores are included as a control group. On testing different distance decay functions, the authors find that new firms are more likely to start up near other firms in the same industry for five of the six industry groups. However, this localisation effect only holds true for two sectors – computer systems design and security. On their analysis of driving speed, used as a proxy for congestion, their analysis controls for several intermediating factors, including median incomes, ethnicity, and zonal population, and finds that new firms are generally attracted to rather than repelled from localities that suffer from congestion. They explain this by concluding that the urbanisation and localisation gains to productivity arising from starting up in proximity to own and other industry sector firms must outweigh the diseconomies arising from traffic congestion.
Paper 11, by Houston (2005), examines the application of the skills mismatch hypothesis to the UK context, carefully noting the importance of understanding commuting patterns and the definition of both housing and labour markets. In particular, one of the assumptions underpinning skills mismatch – that workers are inherently mobile across the entirety of metropolitan labour markets – is brought into question. The conceptual model set out in this paper describes a logic in which workers seek employment in job-rich neighbourhoods within the metropolitan labour market and may relocate close by in order to maximise total wages after paying transportation costs. Although this process leads to higher housing costs within concentrations of employment, Houston (2005) argues that housing and labour markets do not fully clear, resulting in higher equilibrium unemployment rates in relatively inaccessible urban neighbourhoods. Moreover, the analysis demonstrates that the incentive for workers to move closer to employment naturally increases with transportation costs with the result that house prices in employment-rich neighbourhoods are driven up. When transportation costs are low, workers are more likely to commute from a wider set of metropolitan neighbourhoods, thus reducing the compensatory role of differential housing prices. Table 3 summarises the key findings of these VSI papers.
Summary of VSI articles 8 – 11.
Cities and their housing and labour markets
Houston (2005) provides a neat introduction to this section because housing markets do not simply comprise physical structures but are deeply intertwined with socio-economic factors concerning labour markets, individual preferences, and population dynamics. The housing market acts as an intermediary and significant influencer in the dynamics of urban markets, affecting and being affected by migration patterns, city scale, economic opportunities, commuting patterns, residential choices, and labour market decisions. Through its contributions to the literature, the Urban Studies journal has played a leading role in recognising the intermediating roles played by housing markets and systems and how they shape relationships between migration, population change, city scale, and productivity.
Indeed, in the first article in the VSI, Glaeser and Gottlieb (2006) note that population decline in cities may lag economic decline due principally to the physical longevity of the housing stock and the continuation of its utility in the face of declining economic opportunity. Other contributions to this VSI have emphasised the role played by housing affordability and tenure in mediating migratory flows. Housing markets also help to shape commuting patterns, locational preferences, and choices. Studies of residential choice trade-offs show that housing affordability, neighbourhood quality, and commuting time are dominant factors (see e.g. Frenkel et al., 2013).
Paper 12, by Jones and Leishman (2006), is an empirical study of a system of local housing markets in the West of Scotland. The article reviews a collection of articles that examine the regional house price ripple effect in the UK. Although that strand of literature is beyond the scope of this VSI, the article itself provides a contribution to understanding how house price shocks may be transmitted between local housing markets in the context of a metropolitan region. Specifically, the authors argue that sets of local housing markets that are strongly connected through residential migratory flows are also more likely to share housing price dynamics. The analysis in this article uses tests for lead–lag relationships and cointegration analysis to confirm this supposition. This study also argues that residential migratory flows are more significant at sub-regional spatial scales (between local housing market systems) than at regional scale, at least in the UK. The implications of this for planning and housing supply are significant in the context that some local housing systems were found to be closed (i.e. with house price dynamics apparently independent of the dominant metropolitan market), while others were found to work inter-dependently. This point is also made in the literature on spatial housing submarkets (see Galster, 1996; Keskin and Watkins, 2017; Leishman et al., 2013, in this journal). The underlying argument is that plentiful new-build supply in one spatial housing submarket tends to reinforce price differentials within the metropolitan market if there are other spatial submarkets with higher price points and low levels of new-build supply.
Paper 13 in the VSI is by Hincks and Wong (2010), who point out that the interplay between residential and employment decisions in the UK is weakening over time and becoming more complex than it was in the past. They contribute a range of theoretical insights about interactions between housing and labour markets, commuting, migration, and self-containment before moving on to make a series of practical observations which now appear to be accelerating in the aftermath of a pandemic. For example, they cite trends away from suburban living and city centre employment, the rising importance of dual career households, the increasing incidence of urban living and non-urban employment, and trends in working and commuting across different suburbs. They predicted that the rise of ICT and flexible working practices would continue to break down the spatial connection of housing and labour markets.
Hincks and Wong (2010) provide a detailed and practical case study, based on Northwest England, demonstrating how migration and commuting data can be used to construction Housing Market Areas (HMAs). Drawing on a previous framework set out by Brown and Scott (2012), they identify four guiding principles which include that HMAs should have a close association with travel-to-work areas (TTWAs), reflect supply and demand of housing as revealed by migration flows, and exhibit a high degree of self-containment. They also recommend a continuous process of consultation with real estate agents, which emphasises the potentially dynamic nature of both housing and labour market boundaries.
Paper 14, by Thorsnes et al. (2015), examines the direct effects of natural amenities on housing prices and the indirect effects via housing market renewal and restructuring in the context of Dunedin, New Zealand. Using cross-sectional data, their hedonic estimates find a strong negative premium (to an approximate distance of 400 m) of proximity to state housing developments and a strong positive premium (to an approximate distance of 1 km) of beach accessibility. These premia dissipate quickly after these threshold distances. Testing for the effect of views on housing prices, they find that the variable is insignificant but speculate that the negative externality of exposure to strong and/or cold winds may offset the desirability of views. They find a strong premium (3.3% per additional hour) associated with exposure to the sun, a factor they describe as being of importance given the cold, damp climate. Furthermore, this study tests the notion that state housing developments deliberately sited in areas of high natural amenities acted to displace the siting of higher income housing as the city’s suburbs developed and populated, and commuting costs declined. They find that distance from the CBD is not significant in their models of market housing but becomes so when distance from state housing development is also included in the model. The results suggest a polycentric effect in which distance from the CBD shows up as a price discount and distance from concentrations of state housing as a price premium.
Paper 15, by Lux and Sunega (2012), contributes to the literature on regional labour market adjustment, focussing on the role played by housing systems in facilitating or impeding labour migration. A case study of the Czech Republic examines whether regional differences in housing costs and in homeownership rates lead to impacts on labour migration and unemployment rates. This study stands out in terms of its innovative methodology for dealing with the endogeneity between tenure choice and the decision to migrate for economic reasons. The authors reason that individuals with a higher propensity to choose to migrate will logically have a higher propensity for choosing private renting in preference to home ownership. To deal with this, they use housing attitudes survey data to model individuals’ propensity to migrate if faced with a long-term unemployment situation at their current residence.
The empirical analysis finds that regional variation in housing affordability and migration only matters in the case of regional migration flows to Prague and only for those educated to university level. The authors interpret this to mean that higher housing costs are only an impediment to workers with high human capital and a potential desire to migrate towards the high-technology centre of the nation’s economy. Interestingly, they find that housing prices are a stronger impediment to migration than private rents, which may also suggest that such workers are motivated by prospects of entering home ownership, in addition to the income potential of technology jobs.
Royuela and Vargas (2009) is paper 16 in the VSI. It is one of several applied housing market studies that have empirically explored the close relationships between housing and labour markets. Based on a Spanish case study (Catalonia), the authors develop algorithms based on commuting patterns and migration patterns, ultimately concluding that commuting patterns produce more consistent spatial housing market areas (HMAs). Hincks and Wong (2010) emphasised that while businesses make location decisions based on profit potential, household location decisions partly reflect the quality-of-life factors, with households locating where the needs of some or all of its individuals are met. Where there is tension between these needs and employment choices, this usually results in longer commute times. It is worth reflecting that the use of migration or commuting data to reveal HMA boundaries may have had stronger justification before, or certainly during, the SARS-CoV2 pandemic. Certainly, the underlying argument is that individuals and their households will tend to live relatively close to their place(s) of work and study. Royuela and Vargas’s (2009) innovation is to demonstrate that commuting flows, rather than migration patterns, can also be used to reveal the boundaries of HMAs. They argue that the degree of self-containment appropriate to defining those boundaries can be set in such a way as to maximise the homogeneity of housing prices within each HMA after controlling for distance.
Chakrabarti and Zhang’s (2015) paper is the 17th in the VSI and is motivated by what they describe as a commonly perceived policy concern that unaffordable housing markets reduce local employment growth by rendering regions unattractive to workers and firms. They develop a theoretical model in which workers maximise their utility functions through consumption of housing, a composite good, and urban amenities. Firms’ production costs depend on the wage rate and urban rent. The predictions of the model are that housing costs will be higher and wages lower in cities with higher levels of amenities. Firms’ production costs remain constant between high and low-amenity places because higher urban rents are offset by lower wage costs in high-amenity cities. Central to their argument is the idea that the elasticity of land supply may be relatively high early in the development of cities but reduces as cities expand. In high-amenity cities, previous waves of population growth have already shifted land rents to high levels and reduced the elasticity of land supply. With housing affordability defined as the ratio of housing costs to wages, affordability is poor in high amenity cities. The knock-on impact of high cost and inelastic land supply is a lower level of employment growth. Table 4 provides a summary of the key findings of the final group of VSI papers.
Summary of VSI articles 12 through 17.
Discussion and concluding comments
The articles chosen for inclusion in this VSI represent only a fraction of the contributions that the Urban Studies journal has made to understanding the interplay between urban amenities, migration, business growth, innovation, productivity, or urban housing markets. Yet, the hope is that the chosen selection will help illuminate these connections and demonstrate further research areas that may interest the readership of the journal. While some strands of the literature are now very well developed, others are less so and have the potential to be fruitful lines of further inquiry.
One of the more challenging issues that researchers must grapple with concerns the conceptualisation of social, demographic, and economic processes and the directions of causality being presumed therein. Most of the studies reviewed in this article have attempted to do so explicitly, but in some the presumptions are implicit. Neither do the various ‘strands’ of literature necessarily reconcile with each other conceptually. The presence and/or concentration of amenities in regions, towns and cities has been theorised as a causal factor in the movement of both people (and their households) and firms. In the case of the latter, amenities have also been linked to the creation of new firms or the rate of technological advance of firms and their industries. Furthermore, the presence, concentration or growth of population cohorts conducive to new economic activity have been likened to an amenity in its own right, with resulting feedback effects both to the movement of ‘creative people’ and to the propensities for firms to start up and to innovate. The potential circularity of these arguments, or at least the existence of various feedback loops, makes it important for researchers to be clear about what factors might count as an urban or natural amenity, to whom and in what geographical context, and to which social, economic, or demographic process(es).
In standard (or ‘neoclassical’) economic thinking, people and their households make decisions based on their utility maximisation objectives, but firms and industries are driven by business objectives such as profit or sales revenue maximisation, or the desire to build market share. We therefore argue that the cornerstone of a conceptual approach to the analysis of amenities in urban studies should begin with the assumption that urban and natural amenities affect the choices made by people rather than firms. Movements of firms, industries, firm start-ups, and innovation are further down the chain of causality. Such processes may be affected by the presence or concentration of skilled and/or ‘creative’ people which have partly been brought about by amenities, but the amenities do not have a direct causal link to those processes.
Questions concerning how housing markets, systems and outcomes affect processes of urban change have also been lightly researched in the context of the voluminous literatures referred to in this article. Many studies of amenities and population change assume an equilibrium between income and employment opportunities on one hand, and both housing and transport costs on the other. An abundance of urban and/or natural amenities in one city or region compared to another is argued to lead to migration of those population cohorts with a strong appetite for them, and whose presence and concentration lead to higher levels of economic activity. The feedback effect inflates incomes and housing prices further. The equilibrium is struck when the repellent effect of high housing costs dissuades further migration to those places. Yet there are several strands of housing economics literature (not reviewed in this article) that argue against such frictionless adjustment. Studies of regional economics and ripple effects suggest that housing costs are slow to adjust downwards – a phenomenon referred to as a one-way ratchet or tendency for housing prices to be ‘sticky’. In addition, the literature on housing submarkets is primarily motivated by the idea that housing markets are slow, and sometimes unable, to adjust to a new equilibrium based on changes to amenities, housing prices or migration. A review of these literatures is beyond the scope of this article, but it is sufficient to note that they exist and that they argue against the idea that urban housing markets adjust fully, quickly, or at all in response to urban economic conditions. Indeed, this is reflected explicitly in one of the articles reviewed here and attributed in part to the longevity of the housing stock. Housing markets and the ways in which they function are almost certainly one of the most important intervening factors between interactions of urban/natural amenities and population change, so are instrumental by extension in influencing the feedback loops to change in economic activity discussed earlier.
Defining how urban and natural amenities should be defined appropriately is also important to the conceptual framework. It is clear from our review of the literature that a great many different things could be considered as amenities, but the definition depends on the national, regional, geographical context, the geographical scale, and even the definition of the population cohorts that are of interest to the study. It seems obvious that something can only be considered an amenity if it is not present in all locations. Thus, in a country that has well planned cities, walkability or access to public green space would not be considered an amenity in the sense that is used in many of the studies reviewed in this article. Such features only become urban amenities in the event that some cities lack them. This is also a reminder to consider the assumptions being made by the researcher in terms of the mechanics of the migration process itself. Are individuals assumed to choose national, regional then urban destination when deciding to migrate? If so, what could be important in terms of amenities may be quite different than had we assumed that migrants are already in a specific country and region and are making a final choice about urban destination.
The cost or possibility of producing or replicating amenities may also be important. If the creation of replicable amenities, such as restaurants, cinemas, and night-clubs, simply requires a concentration of people who think these amenities are important then the effects they exert on urban outcomes may be less important than the existence of costly or impossible to replicate features such as access to picturesque countryside or an abundance of historic buildings. This reflection also borrows from the housing submarkets literature in which it is generally recognised that some physical aspects of housing neighbourhoods (such as an abundance of stone-built properties with large gardens) are more difficult to replicate than others (proximity to an urban park or a well-ranked public school).
Our reflection on the potential applicability of the housing economics, and housing submarkets, literature to understanding the amenities/migration/agglomeration nexus is a fitting place to draw these reflections to a close. The Urban Studies journal welcomes submissions that make further contributions to all the research questions and issues touched upon in the VSI articles, but studies that place the housing system at the forefront of these debates are particularly welcomed.
Footnotes
Acknowledgements
The authors are grateful for comments on earlier drafts of this article from Jon Bannister, Evert Meijers and Tony O’Sullivan. The views expressed in the paper together with responsibility for any errors remain with the authors.
Declaration of conflicting interests
The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
Funding
The author(s) received no financial support for the research, authorship, and/or publication of this article.
