Printer Friendly

A spatial analysis of small- and medium-sized information technology firms in Canada and the importance of local connections to institutions of higher education.

As part of community/regional development policy, governments in Canada attempt to create conditions that stimulate the formation of high-technology clusters and, in so doing, often encourage firm--university/college liaisons. Information technology (IT) is an important segment of the high-technology sector, and small- and medium-sized Canadian IT firms are disproportionately attracted to large metropolitan areas (with Toronto, Ottawa-Hull, Vancouver, Calgary, Montreal, Kitchener and Edmonton being the most noteworthy). A nearest neighbour analysis suggests that small- and medium-sized IT firms are clustered within the metropolitan setting, and a nearest neighbour hierarchical spatial clustering technique demonstrates that intra-urban IT agglomerations can be objectively identified. The linkages between small- and medium-sized IT firms and higher education institutions are, on average, not strongly entrenched within Canada's IT culture, although many of these firms still connect with universities or colleges through co-operative programs and other means of employee recruitment and via general networking with faculty members. Thus, governments may be able to support IT cluster formation by encouraging firm-university/college connections that centre on student participation.

Dans le cadre de la politique en matiere de developpement communautaire et regional, les instances gouvernementales canadiennes veulent crier des conditions favorables d la croissance de grappes de haute technologie et, ce faisant, qui peuvent soutenir le developpement de relations entre les entreprises et les universites et colleges. Les technologies en information (TI) sont un domaine important du secteur de la haute technologie. De facon disproportionnee, les petites et moyennes entreprises en TI au Canada se retrouvent dans les grandes metropoles (entre autres Toronto, Ottawa-Gatineau, Vancouver, Calgary, Montreal, Kitchener et Edmonton). Une analyse de proximite permet de demontrer que les petites et moyennes entreprises en TI se developpent en grappe d I'interieur des limites des metropoles. De plus, il est possible de constater, d I'aide d'une technique de regroupement spatial hierarchise fondee sur la proximite, que les regroupements d'entreprises en TI peuvent etre identifiees de maniere objective au niveau intra urbain. Les liens entre les petites et moyennes entreprises en TI et les institutions d'enseignement superieur ne sont pas bien etablis, en general, dans le contexte canadien des TI. Toutefois, plusieurs de ces entreprises s'unissent avec des universites et colleges par I'intermediaire de programmes cooperatifs, par le recrutement d'employe et par le resautage avec des membres d'une faculte. Les gouvernements peuvent donc appuyer I'etablissement de grappes en TI en faisant la promotion de relations entre les entreprises et les universites et colleges qui mettent I'accent sur la participation des etudiants.


Recent research on so-called 'learning regions' represents continued efforts to understand the workings of the innovation process and the conditions that encourage the clustering of high-technology activity in certain areas (Wolfe 2002). It is often assumed that firms involved in similar or complementary activity benefit from being spatially proximate to one another through transaction cost savings and via the achievement of external economies of scale advantages. As Wolfe and Gertler (2003) summarize, the agglomeration of similar economic activity potentially allows each firm within the cluster greater access to suppliers, skilled labour and a common knowledge base. The transmission of 'common knowledge' is central to the idea of a learning region. Knowledge can be diffused through formal and informal interpersonal or employee-related contacts amongst the firms (which may result from direct face-to-face interaction or more indirectly from the research ambience generated by, for example, a nearby leading firm), or knowledge may be dispersed through relationships with local universities or other institutions of higher learning and/or research (Wolfe and Gertler 2003).

The formation of high-technology clusters and learning regions has strong local/regional development overtones, and certain government organizations aim to stimulate the transmission of knowledge amongst firms and research nodes (such as universities). In Canada, government organizations such as the National Research Council (NRC) provide resources to research and development (R&D) firms in various high-technology sectors, such as aerospace, biotechnology and information and communication technologies. The main approach of the NRC is to set up technology centres in communities across Canada that can dispense technological advice and support to innovative small- and medium-sized enterprises (SMEs) and also to encourage university/college-industry projects. Its goal is to create 'technology clusters' that incite the formation of spin-off companies and start-up firms so that eventually a critical mass of skilled people, expertise, capital and entrepreneurial drive can contribute to a higher quality of life for local communities (NRC Canada 2003). This growth-pole/growth-centre strategy also has foundation in Myrdal's (1957) 'circular and cumulative causation' ideas in which once the technology cluster is established, growth will ideally become self-reinforcing.

Information technology (IT) firms are an important component of the high-technology business segment in Canada. In a report prepared for Statistics Canada, Beckstead et al. (2003) found that Canadian employment in information and communication technology increased over 70 percent from 1990 to 2000 (with most of this growth occurring in large urban areas). In addition, this sector was a major contributor to new business creation. In 2000, 74 percent of people employed in IT-related activities worked for a firm established after 1990. The 1990s represent the boom years for the IT industry collectively, and since 2000, the sector experienced a considerable slow-down with adverse effects on employment and investment portfolios. However, as Lahey (2003) reports, International Data Corporation (Canada) predicts that an 'inevitable upgrade cycle' will occur (between 2003 and 2007) as IT hardware and software products and services will have to be replaced and that much of this renewed activity will be handled by Canadian small- and medium-sized IT firms.

Given the importance of the IT sector, continued research is needed and a greater understanding of the spatial choices of, especially, small- and medium-sized IT firms would be desirable. Thus far, limited attention has been given to appreciating the geographic tendencies of IT firms in Canada, particularly at the intra-urban scale and for the specific case of small- and medium-sized firms. The results of this study provide progress within this regard.

Empirical research has also been remiss in specifically defining the spatial boundaries of high-technology clusters. As Wolfe and Gertler (2003) point out, researchers will typically stress the role of geographic proximity, but often the spatial boundaries of high-technology agglomerations are either not firmly delineated or too widely assessed, at perhaps the provincial or national level, when precise local analysis may be more appropriate. This is a significant concern because meaningful analysis is contingent on properly appraising the spatial extent of these agglomerations. For instance, IT companies are ordinarily found throughout the metropolitan environment, and in the absence of a systematic approach, a 'cluster of IT firms' could be discerned in very arbitrary terms. Consequently, identifying, through statistical substantiation, where IT clusters are found within Canadian metropolitan areas is another objective of this study.

In addition, while researchers continue to debate the importance of university influence in sustaining technology clusters, some fundamental concerns remain. Langford (2002) concludes that the measurement of interaction between firms and universities, and more generally the role that universities play in the innovation process, tends to be indirect and that carefully designed surveys would provide more poignant information. Some authors have addressed this concern (such as Romijn and Albaladejo 2002; Chamberlin and de la Mothe 2003), and this analysis continues in this vein by integrating the results of a survey that details the extent of higher education institution (HEI) linkages with small- and medium-sized IT firms in Canada.

In general, there is continuing need for empirical analyses to evaluate the location tendencies of firms as technological advances 'shrink the world'. With access to sophisticated communication infrastructure, firms can perform many transactions at a distance, and as Dicken (2000) explains, there is an emerging literature that views 'place' as largely irrelevant as firms become 'hyper-mobile'. While this likely overstates the situation, it may be precarious to assume that all firms (technology-based or otherwise) require cluster membership for success, and for those that do, it is not immediately obvious where these clusters are situated. For the case of modestly sized IT firms in Canada, this study's systematic approach appends the literature by suggesting that location decisions continue to produce distinct geographic patterns. Moreover, clearer cognisance of where intra-urban IT clusters are located vis-a-vis institutes of higher education and the nature of IT firm-university/college linkages can provide insight for government organizations that attempt to encourage these liaisons in the spirit of community/regional development. Prior to detailing the results of this study, a discussion on the concepts relating to agglomeration economies, high-technology clusters and higher education liaisons is offered and, following that, a brief review of the 'high technology SME' literature is given.

Agglomeration Economies, High-Technology Clusters and Firm-Higher Education Institution Liaisons

High-profile technology agglomerations in Silicon Valley and Route 128 have had very dramatic impacts in Santa Clara County and in Boston (Saxenian 1994). No doubt these and other 'success stories' influenced the growing body of research that seeks to better understand the benefits gleaned by individual firms that cluster in space and the extent to which these advantages of spatial proximity (or agglomeration economies) incite growth and prosperity within the surrounding region (Beeson 1987; Petrakos 1992; McCann 1995; Mun and Hutchinson 1995; Selting etal. 1995; Fujita and Thisse 1996; Krugman 1997; Zheng 1998; Higano and Shibusawa 1999; Abdel-Rahman 2000; Belleflamme et al. 2000; Gordon and McCann 2000; Kloosterman and Lambregts 2001).

Commonly, agglomeration economies are defined as decreasing average costs that accrue to individual firms located in areas that have a high concentration of activities (Hanink 1997). The economies (or savings) achieved by each individual firm are a result of lower costs of spatial interaction and are dependent on the enlarged output of all firms in the agglomeration. These external economies of scale have normally been categorized in two ways: localization economies (or the benefits that extend to similar firms doing similar things at a single location) and urbanization economies (or the benefits that accrue to all firms in urban locations). Localization economies are savings that are external to the firm but internal to the agglomeration of similar industry, and urbanization economies are external to both the firm and industry but internal to the city (Moomaw 1988).

In many ways, the concepts relating to agglomeration economies parallel the notion of 'clusters'. Porter (2000, 254) describes a cluster as 'a geographically proximate group of interconnected companies and associated institutions in a particular field, linked by commonalities and complementarities'. Prevezer (1997) provides a consolidation of the literature and recognises that high-technology innovative clusters can encourage additional firm membership through specialized labour, specialized inputs, knowledge spillovers and market/user accessibility.

* specialized labour is typical of clusters, and this technical or scientific expertise can be attractive to small, innovative companies or to managers of large companies;

* specialized inputs and the disproportionate presence of producer service firms that possess specialized equipment, research tools or specific technologies may be an attraction to new innovative firms in a particular sector;

* knowledge spillovers are more likely to occur in clusters and may include highly technical information that is exchanged via working relationships between firms (or through liaisons with research institutions such as universities), or knowledge may be transferred less formally through the overall innovative ambience of the cluster;

* market/user accessibility is enhanced if firms are proximate to the users of their product or service. Not only are there cost savings in terms of distance, but firms also have the ability to readily react to changes in market demand.

Specialized labour, specialized inputs and industry-specific knowledge spillovers are localization economies that would benefit a distinct cluster of companies (such as those involved in biotechnology or IT). Tacit industry-specific knowledge spillovers may be particularly important to highly innovative firms and these 'face-to-face contacts' are very sensitive to distance:
   Tacit knowledge is the relatively unformulated, interiorized,
   even unsuspected know-how that becomes
   formulated in conversations, in response to questions
   and in processes of speculating, hypothesizing
   or brainstorming. Disclosure of tacit knowledge
   demands conditions of spatial proximity among
   interlocutors of consequence to the problem or
   project at hand (Cooke 2002a, 180).

Market/user accessibility and general knowledge transfers are most likely achieved in larger urban centres and, depending on the circumstances, could be termed either localization or urbanization economies. These and other urbanization economies, such as well-developed transportation and communication services, a diverse labour market and varied financial and municipal services, continue to encourage the clustering of economic activity in urban areas despite the presence of various urban diseconomies of scale (such as traffic congestion, crime and high land values) (Wheeler et al. 1998; Feldman 2000).

Knowledge transfers and spillovers that can occur from firm-university or firm-college relationships (or, collectively, HEI linkages) are often assumed to be commonplace in areas of high-technology firm concentration, and it is presumed that these relationships promote the cluster's agglomeration economies and attractiveness to other firms. Huggins and Cooke (1997) categorize the economic impact of universities on regions in terms of expenditure (by the faculty, students, staff and visitors) and knowledge. The authors state that because access to knowledge has become an important component in urban and regional development policies, the university is recognized as a key node within the knowledge network and thereby crucial to community well-being. In a similar vein, Meyer and Hecht (1996) state that universities are in fact community growth poles and demonstrate that areas in Canada that have universities exhibit a higher level of economic well-being relative to other areas.

Firm-HEI relationships can take many forms and can range from relatively informal connections, including faculty exchanges (for lectures or consulting) and student internships, to very formal and committed alliances that are formed to develop new products or technologies (Mead et al. 1999). Some connections are purely financial in nature (such as with firms that provide endowments to universities or colleges for such things as graduate student funding or lab upgrades), or associations may be more interactive and 'hands on' in which both the HEI and industry personnel address immediate industrial issues (Santoro and Chakrabarti 2002). Most HEIs will pursue at least some industrial connections, but the approach undertaken may only be ad hoc in nature. Yet, many associations are part of a carefully planned community/regional development strategy, such as with research/science parks, where the university or college often plays the role of an 'anchor' and is a catalyst to high-technology company spin-offs (Berry 1998). While there has been research on the pitfalls of greater firm-HEI interaction, such as less time for students (Stephan 2001; Poyago-Theotoky et al. 2002) and academics accepting restrictions on the disclosure of their research (Cohen et al. 1998), the predominating belief is that the advantages of these relationships are potentially multitudinous. As Lee (2000) outlines, there are several reasons why academics are attracted to industrial partnerships and these include the ability to access additional funds for academic research, to supplement research and teaching with practical insight and to create student internships and job placement opportunities. Firms also have much to gain from these collaborations as academics can help solve specific technical or design problems and develop new products and processes. In addition, firms benefit from having access to university or college resources (such as labs, libraries and workshops), and closer ties with facility may enhance recruitment aspirations (Lee 2000). For a more comprehensive literature review of HEI-industry knowledge transfers and forecasts on the nature of future collaborations, consult Agrawal (2001) and Langford (2002).

The Specific Case of Technology-Based SMEs

Empirical study commonly shows that innovative SMEs have a strong tendency to locate within clusters. Several areas of the world have been investigated, including Canada (Britton 1996, 2002; Niosi and Bas 2001; Niosi 2002), China (Belderbos and Carree 2002), Finland (Schienstock and Tulkki 2001), France (Lemarie et al. 2001), Germany (Gertler 1996; Zeller 2001), India (Kharbanda 2002), Japan (Yamawaki 2002), Sweden (Carlsson 2002), the United Kingdom (Huggins and Cooke 1997; Cooke 2001, 2002b; Romijn and Albaladejo 2002) and the United States (Audretsch and Feldman 1996; Prevezer 1997; Varga 2000; Audretsch 2001; Carlsson 2002; Cooke 2002b; Owen-Smith et al. 2002). The investigation of cluster formations continues to evolve, and in Canada, the methodology employed has moved towards more qualitative, place-specific case studies. Notable additions to the literature, which at least in part examine the role of SMEs, include wireless telecommunications and global positioning systems in Calgary (Langford et al. 2003), information and communication technology in New Brunswick (Davis and Schaefer 2003), IT agglomeration in Cape Breton (Johnstone and Haddow 2003), technology clusters in Ottawa (Chamberlin and de la Mothe 2003), and the multimedia industry in Montreal (Tremblay et al. 2002) and Toronto (Mills and Brail 2002). Evidence suggests that technology-based SMEs benefit from urbanization and localization economies, but the nature of these advantages can vary with industry type (Prevezer 1997; Carlsson 2002) and with the stage in an industry's life cycle (Audretsch and Feldman 1996).

According to the literature, there are a number of criteria that encourage the formation of high-technology clusters that are made up of, primarily, small- and medium-sized firms. Specifically, these companies are attracted to areas that feature a strong research ambience, enhanced not only by a critical mass of similar firms but also by affiliations with universities and/or government programs and the presence of complementary firms that offer specialized services (Prevezer 1997; Audretsch 2001; Cooke 2001; Mills and Brail 2002; Tremblay et al. 2002; Yamawaki 2002; Johnstone and Haddow 2003). Other factors that can stimulate the growth of SME clusters include the presence of a large leading firm in the area (Cooke 2001; Yamawaki 2002), the availability of a talented labour force (Mills and Brail 2002; Davis and Schaefer 2003; Langford et al. 2003) and close proximity to venture capital operations and other financial institutions (Audretsch 2001). While many technology-based firms will serve national and international markets, researchers are virtually united on emphasizing the benefits of a strong local market in maintaining a cluster as well. It is also typically stated that as these clusters prosper, the marked inflow of new firms and the expansion of the local talent pool will reinforce community/regional well-being.

Where analysis has been far less conclusive relates to the necessity for technology-based SMEs to be in close proximity with university/college resources. Romijn and Albaladejo (2002) found that while the majority of their sample of firms (thirty-three small software development and electronics manufacturing companies) were spin-offs from a university or a former public laboratory, the need to be near a leading university was not a strong factor in location decisions. Audretsch (2001) came to similar conclusions regarding the spatial pull of the local university, and Audretsch and Feldman (1996) and Lemarie et al. (2001) found that SME-university local linkages are only of great consequence when the firm, and/or the overall industry, is young. However, Huggins and Cooke (1997), Prevezer (1997), Cooke (2001) and Owen-Smith et al. (2002) determined that the local university contributed 'knowledge impacts' that helped to propel regional industrial clusters. For example, Huggins and Cooke (1997) explain that academic expertise at Cardiff University has been shared with both large foreign multinationals and small local firms. They argue that geographic proximity is significant as the local academics at Cardiff bring a unique indigenous perspective that firms could not obtain from abroad.

The following study appends the SME-oriented literature by considering the spatial properties of small- and medium-sized IT firms in Canada, statistically delineating the spatial extent of IT clusters within Canadian cities and analyzing the magnitude of IT firm connections with local universities and colleges.

Small- and Medium-Sized IT Firms in Canada

The primary data set and its characteristics

A sizeable sample of 2,074 IT firms was compiled from CATAlliance's (2003) Techno-Connect database. This unique source provides an Internet-based forum for IT companies to advertise their products and services and thereby encourages connection with potential customers. Along with each company's location, most records include information on product and IT specialization, the number of employees (less than fifty, fifty to ninety-nine, hundred to 500 or greater than 500), the year of establishment and direct contact (via a telephone number and/or a link to the company's web site). CATAlliance also lists some American and European companies in their database, but these cases were not utilized in this investigation. Therefore, the data set compiled for this analysis is made up of all IT companies that were classified as operating in Canada and that had registered with CATAlliance before 21 February 2003.

There is no appreciable bias in this sample, and the 2,074 companies utilized in this study provide a good representation of the universe of IT firms in Canada. There is no membership fee, and as long as Internet access is available, any IT company has the ability to advertise with CATAlliance. Companies are encouraged to indicate their area of specialization and the following eight IT activities are listed on the web site: software developer, software publisher, IT consulting, IT distributor, value-added reseller (VAR), multimedia content development, original equipment manufacturer and Internet service provider. Yet, as this is a user-maintained site, it is possible that some non-IT companies are included within the database and this could overstate IT intensity in any given location. If this has occurred, it is likely that these cases would represent a negligible portion of the sample used in this study. The web site is explicitly set up for companies and individuals seeking IT products and services for their business needs, and firms would have little to gain by advertising on this site if their operations do not involve IT-related activity.

On the basis of this sample, it is clear that IT companies in Canada tend to be modestly sized. Of the 2,074 firms, 92.9 percent employed less than fifty people and a further 5.2 percent had fifty to hundred employees. All firms in this study had less than 500 employees. Admittedly, there is some discrepancy over what exactly denotes an SME: Romijn and Albaladejo (2002) define SMEs as having less than 250 employees, Cohen et al. (2002) used a 500-employee benchmark, and Yamawaki (2002) used both a revenue and a 300-employee distinction. Thus, if one subscribes to Cohen et al.'s more liberal definition, then all of the firms in this analysis could be enumerated as SMEs. At any rate, the vast majority of the cases in this data set would fall under most SME employment thresholds, and certainly the 1,927 companies that listed less than fifty employees (or roughly 93 percent of the sample) can assuredly be classified as small firms.

Table 1 summarizes the primary functions of the Canadian IT companies used in this study, and these activities closely correspond with the definition provided by the North American Industry Classification System. As Davis and Schaefer (2003, 123) explain, the information (and communications) sector is composed of:
   equipment and component manufacturing, goods-related
   services (primarily wholesalers, distributors
   and leasers) and intangible services (software publishers,
   cable program distributors, telecommunications,
   carriers and resellers, information and data-processing
   services, equipment repair and computer
   systems design). Information technology training is
   also conventionally included in the information and
   communications sector.

In terms of CATAlliance's definitions, software development was the most common undertaking. However, consulting activities, adding value to existing products (VARs), developing multimedia content, IT-related hardware manufacturing and software publishing activities were also listed with reasonable frequency by IT firms. Of the 884 companies that provided detailed information on their activities, 44.1 percent specialized in one specific IT business type, 26.1 percent were involved in two primary activities, and 29.8 percent were quite diversified business entities involved in at least three of the IT functions listed in Table 1.

For most cases, the database also includes the year of firm establishment. Some of the companies were created before 1970 (4.2 percent), and an additional 10.3 percent were formed between 1970 and 1979. However, the majority of the companies were established more recently, either during the 1980s (44.5 percent) or after 1990 (41.0 percent). Since the inception of CATAlliance's Techno-Connect web site in 1999, it is inevitable that some of the companies listed in the database have since gone out of business. However, as each of these locations were at some point deemed attractive enough to start a business, it is reasoned that all cases should be included in this analysis even if some locations were ultimately not profitable.

The geographic properties of small- and medium-sized IT firms in Canada

As will be demonstrated, small- and medium-sized IT companies in Canada are clustered at two scales: these firms are disproportionately present in a few census metropolitan areas (CMAs) in Canada, and notable clusters of IT companies can be identified within these key CMAs. Moreover, these clusters are often in the vicinity of an HEI. Other Canada-wide studies have confirmed that large metropolitan areas are associated with greater amounts of high-technology activity in general (Britton 1996; Industry Canada 2003) or with biotechnology-related (Niosi and Bas 2001) or information and communications technology and science-based activity (Beckstead et al. 2003) specifically. This study is unique in its emphasis on small- and medium-sized Canadian IT firms, in systematically identifying IT clusters (via statistical substantiation) within CMAs, and in administering a survey to understand the nature of the linkages between modestly sized IT firms and institutions of higher education.

As summarized in Table 2, these SMEs are strongly attracted to urban economies, and more specifically, metropolitan centres are the most frequent targets. Of the top fifteen urban centres, only the consolidated area of Fredericton is not classified as a CMA. Yet, this polarization of IT activity becomes even more striking with movement up the urban hierarchy. Toronto, Ottawa-Hull, Vancouver, Calgary, Montreal, Kitchener and Edmonton collectively housed 76.9 percent of modestly sized IT firms in Canada, and six of seven of these centres comprise the top six in terms of population (Kitchener was ranked tenth in 2001) (Statistics Canada 2003). The dominance of Toronto, particularly, and also Ottawa-Hull cannot be overstated as roughly a third of the IT firms in this sample were stationed in the Toronto CMA, and almost half of the nation's total were located in either Toronto or Ottawa-Hull. Correspondingly, Ontario led all provinces by housing 59.7 percent of Canada's small- and medium-sized IT firms.

While the strong attraction of IT activity to Canada's largest urban centres is clear, it is perhaps more instructive to understand the spatial tendencies of these firms within the metropolitan environment. To this end, both a general nearest neighbour analysis and a nearest neighbour hierarchical clustering procedure were performed for the top seven IT centres in Canada (Toronto, Ottawa-Hull, Vancouver, Calgary, Montreal, Kitchener and Edmonton). For both approaches, points of latitude and longitude were used to denote the location of each firm, and the analyses were carried out using the Crimestat statistical package (Levine 2002). Statistics Canada's (2000) postal code conversion file was used to match each firm's postal code with the centroid of the relevant enumeration/dissemination area to create the latitude and longitude co-ordinates.

The computation and significance evaluation of a nearest neighbour index allows a researcher to determine whether an observed pattern of points deviates from a theoretical (random) distribution sufficiently enough to be considered significantly clustered or dispersed in space (Ebdon 1987). The mean nearest neighbour distance is the average distance between all points and their nearest neighbours. The mean random distance (or the theoretical distance for a random distribution) is equal to 1/(2 [square root of (p)]), where p is the number of points divided by the area. With Crimestat, the actual area of the study region can be entered (and a proportional rectangle is calculated), or the software will derive an area based on the rectangle defined by the minimum and maximum X and Y (longitude and latitude) points. (1) The nearest neighbour index, then, is the ratio of the mean nearest neighbour distance to the mean random distance. This index can range from zero (perfectly clustered) to 2.15 (perfectly dispersed), and a completely random pattern is indicated by a value of one. The mean nearest neighbour distance less the mean random distance divided by the standard error yields the test statistic which can be compared with the appropriate critical value of a standard normal deviate to test for significance (Ebdon 1987).

Table 3 provides the results for the nearest neighbour analysis for IT firms located within seven CMAs in Canada. Because the objective was to ascertain whether IT firms are significantly clustered within these urban environments, one-tailed tests were appropriate and all of the nearest neighbour evaluations yielded highly significant results (as indicated by the p-values). A clustered pattern of IT activity in all seven CMAs is confirmed, as the nearest neighbour indices are consistently well below the value of one (or perfect randomness) and in all cases the test statistic is negative.

Notice that a 'significantly clustered' result was achieved either when using the actual CMA area or when area was calculated by maximum and minimum points. Logically, the nearest neighbour index is smaller, and indicative of a greater degree of IT concentration, in situations when actual CMA area was used to generate the mean random distance. In some cases, the difference is quite large (see particularly Calgary and Edmonton), and this suggests that IT activity is rare, if not completely absent, at the margins of these CMAs. By contrast, with Toronto, particularly, but also with Montreal, Kitchener and Vancouver, the nearest neighbour indices are reasonably similar, and this would indicate that IT firms can be found throughout these CMAs. As will be illustrated, the intensity of IT activity within all of these CMAs is greatest towards the core, which corresponds with the result of an overall clustered pattern.

In efforts to objectively identify where the key clusters of IT firm activity are located within the metropolitan environment, Crimestat's nearest neighbour hierarchical spatial clustering technique was employed. This clustering routine group points together on the basis of spatial proximity, and initially, first-order clusters are distinguished. Each cluster must have a minimum number of predetermined cases (the default is ten points). In addition, each cluster is composed of cases that are closer together than the user-defined threshold distance, which is 'the confidence interval around a random expected distance for a pair of points' (Levine 2002, 33). Essentially, the threshold distance is synonymous with the probability of selecting any two points by chance. The default probability is 0.1, and this means that fewer than 10 percent of the pairs could be expected to be this close by chance. Thus, any pair of nearest neighbours that are closer than the random/ chance distance will be allocated to a cluster, but this cluster will not be retained by the routine unless there are at least ten cases in total that meet the threshold distance criteria. Hence, the likelihood of forming a first-order cluster of ten points is not high; all of these points must be closer in space than would be the case by chance. Or, put another way, if there is a 10 percent chance of two points being closer than what would randomly be the case, then the probability of retaining this cluster steadily declines with three, four or ten points (Levine 2002).

Depending on the nature of the data set and on how the user defines the threshold distance and the cluster membership minimum, a hierarchy of clusters may result. In the next phase, the first-order clusters are reduced to points (cluster centres) and appraised for second-order clustering, and the process continues until either a single cluster is achieved or the user-defined threshold stops the process. Each cluster can be graphically illustrated as standard deviational ellipses. In general, the first-order clusters are designed to highlight local 'hot-spots' of activity, and the second--and higher-order clusters distinguish important regional areas (Levine 2002).

Using the aforementioned defaults to determine threshold distance and cluster membership, the results of the nearest neighbour hierarchical spatial clustering analysis for IT firms are shown in Figures 1-7 and summarized in Table 4. In each figure, the first-order, and in some cases the second-order, clusters are represented by the standard deviational ellipses. These ellipses are two standard deviations from the cluster's centre of minimum distance and will on average capture more than 99 percent of the cases (Levine 2002). Also displayed are the locations of any university or college campus that had at least 3,000 full time students. To determine campus locations and student enrolment, the Association of Universities and Colleges (2003) and the Association of Canadian Community Colleges (2003) were consulted and some main campuses were telephoned directly.

Given that Toronto led all other CMAs in attracting IT companies, it is not surprising that the Toronto metropolitan area also had the most first-order clusters (see Figure 1). The most dominant agglomerations of IT companies are situated in the southern part of the CMA and include areas of the downtown core, York and Forest Hill. These four clusters contain 150 (or 21.3 percent) of the total 704 IT companies sampled in the Toronto CMA, and a strong presence of HEIs in the vicinity is also noteworthy. Other important IT agglomerations can be found to the northeast of the Toronto core, from North York and Willowdale to Richmond Hill. In addition, the second-order cluster distinguished this entire band of first-order clusters as an important regional agglomeration of IT activity. To the west, two IT accumulations were detected in the area of Lester B. Pearson International Airport. While suburbs such as Markham, Mississauga and Scarborough are station to many IT companies, none of these locations attained enough critical mass of spatially proximate IT cases to be highlighted by the cluster routine.


Like the Toronto situation, Ottawa's downtown core (the northernmost first-order cluster on Figure 2) is the single most important area for IT agglomeration in the CMA, but other zones of significance can be identified. While the downtown cluster contains sixty-three of the CMA's 291 cases (or 21.6 percent), suburban locations such as Bell's Corners (to the east of the downtown core) Kanata (furthest east), Blossom Park (to the southwest) and a locale between Nepean and the MacDonald--Cartier International Airport (the southernmost first-order cluster) were also isolated as key IT first-order accumulations. This entire area, in addition to a small adjacent section of Hull just on the northern side of the Ottawa River, comprised the regional cluster. Interestingly, all of the HEIs were enclosed within this second-order cluster indicating that regional, rather than local, proximity to these institutions may be important to IT companies in the Ottawa-Hull CMA.


While IT firms are present in good numbers in North Vancouver and Burnaby, only Vancouver's core and the Richmond suburb to the south were isolated as significant first-order clusters by the nearest neighbour hierarchical spatial clustering routine (see Figure 3). The city's downtown core holds forty-four IT companies (or 25 percent of the CMA total). Although there does not appear to be a strong spatial connection with the HEIs, the two clusters are quite close to the Vancouver International Airport (located just west of the Richmond cluster).


As with Vancouver, Montreal's IT pattern features two noteworthy clusters: one in the downtown core and the other in a suburban location (at St. Laurent), which is in the vicinity of an international airport (see Figure 4). The downtown agglomeration contains an impressive 29.2 percent of the CMA's IT firm membership, but unlike Vancouver, this core cluster is also home to no less than eight of the CMA's fourteen major HEIs.


Both CMAs in Alberta display a very polarized pattern of IT firm settlement (see Figures 5 and 6). Only one cluster was formed in Calgary, but it is clearly a dominant agglomeration with 37.1 percent of the CMA's sampled IT firms located in this central location. Edmonton's downtown also has a commanding influence on IT firm settlement (with 31.1 percent of the metropolitan total), but a second 'attached' cluster can be found just to the south. All of Edmonton's major HEIs are found in the central IT cluster, whereas Calgary's larger campuses, while still quite concentrated, are not spatially proximate to the core IT cluster.


The CMA of Kitchener contains three noteworthy urban communities (Kitchener, Waterloo and Cambridge), but Waterloo's higher emphasis on high-technology activity is mimicked by the results of the nearest neighbour hierarchical spatial clustering routine which allocated both first-order clusters to this section of the CMA (see Figure 7). While the two clusters are spatially continuous and collectively very important, the southern agglomeration (essentially Waterloo's core) contains a greater number of IT companies and is home to two major HEIs.


Table 4 summarizes some of the information that can be derived from the maps. Unquestionably, there is a strong propensity for IT companies to cluster in a few identifiable locations within these metropolitan environments. Vancouver housed the lowest proportion of total IT companies within its clusters, but even so, roughly one-third of the CMA's IT companies are located in only two first-order clusters. Montreal, Kitchener and Edmonton also had two first-order clusters but with even higher proportions of IT companies stationed within these agglomerations. IT activity in Calgary also exhibits considerable concentration; the CMA's sole cluster has generated considerable influence on location choice. While Toronto and Ottawa--Hull had comparatively more clusters, polarization of IT activity is still very apparent as over half of all IT companies sampled in these CMAs were attracted to these clusters.

In general, the distance between these first-order clusters and an HEI is not great. In fact, the mean distance between an IT cluster's centre and the nearest HEI within these CMAs was not more than 5.14 km, and in many cases, this distance was considerably less (see Table 4). In fact, there are several examples of where the distance between an IT cluster's centre and a university or college is less than 1 km. One might infer that the close spatial proximity of IT clusters to institutions of higher education would signify an abundance of strong working relationships between IT firms and the education sector. However, the following section's results indicate that this is not necessarily the case.

Higher education and IT firm connections: survey results

A phone survey was conducted from 4 June to 17 June 2003, in which firms were randomly selected from the 2,074 IT companies' data set. Phone numbers were retrieved from CATAlliance's Techno-Connect database. In all, 880 calls were made to reach the goal of 500 respondents. As it turned out, 380 calls were made to companies that were unwilling to participate in the survey or to phone numbers that were not in service. While the key question related to the nature of firm--university/ college connections, each IT firm surveyed was also queried on head office and R&D locations. There is considerable support in the literature for the notion that R&D and external connections are greatest at either a firm's head office or at the location of R&D-oriented subsidiaries (for a summary of the literature, see Malecki 1997). Indeed, one can hypothesize that small--and medium-sized IT firms that are locally owned and perform local (in-house) R&D might have a greater propensity to pursue linkages with institutions of higher education. Therefore, the specific questions asked were as follows:

* Where is your company's head office located?

* Where does most of your company's R&D occur?

* In doing business, have you had any contact with local universities or colleges? In what capacity?

The vast majority (86.6 percent) of the IT companies surveyed were in fact head offices. A few of the companies revealed that they were headquartered elsewhere in Canada (5 percent) and 8.4 percent were subsidiaries of a foreign company. Interestingly, the location of R&D activity provided similar proportions: 85.6 percent of the IT firms questioned performed only local (in-house) R&D, 6 percent were involved in both local and non-local R&D and 8.4 percent did their R&D elsewhere (either somewhere else in Canada or in a foreign locale). By far, the most 'typical' company contacted via the phone survey was a locally owned IT firm that performed its R&D in-house. Overall, 414 of the 500 firms surveyed (or 82.8 percent) had these characteristics.

In total, there were 558 IT firm-higher education contacts amassed from the 500 phone questionnaires. As summarized in Table 5, 237 (or 42.5 percent) of the IT firms stated that they had no involvement with local universities or colleges. The most frequently stated liaison was with co-operative programs. Yet, it is noteworthy that while 214 firms stated that co-operative linkages were pursued, seventy-four of these firms (or roughly a third) had to discontinue their involvement. The most common reason for abandoning co-operative activities was due to company downsizing and the reduced ability to pay their portion of a student's wage. General networking with universities or colleges, such as attending presentations or meeting informally with university/ college faculty, was indicated as important by several of the IT firms. Also, some firms specified that they kept in contact with institutions of higher education to recruit graduates. There were a few cases in which the university or college was a client of an IT firm. Typically, this involved the purchase of specialized software packages or employing the IT firm to upgrade information or communication services. Close collaborations with university or college faculty on very specific projects (such as in the area of product development) occurred but were comparatively quite rare. Other infrequent connections were the use of campus resources (typically for library research), and a few of the IT firms employed personnel who taught a course at the local university or college.

Despite the fact that the majority of IT firms were locally owned and did most of their R&D in-house, affiliations with universities and colleges were, in aggregate, quite modest with co-operative placements being the only contact that can be termed exemplary. This trend is similar for the subset of firms that had non-local head offices and did most of their R&D elsewhere, but slight variations are apparent (see column 4 of Table 5). Interestingly, these firms were more likely to have contact with local universities and colleges and, even though the proportions are not large, were also more likely to recruit recent graduates, to use campus facilities and to have personnel teach a course. As result, one cannot conclude that IT firms that are locally owned and perform most of their R&D in-house have a greater propensity to develop local higher education liaisons.

A number of the IT firms interviewed (183) were members of first-order clusters and thereby located within key IT agglomerations within Toronto, Ottawa-Hull, Vancouver, Montreal, Edmonton, Calgary or Kitchener. It is striking how similar the contact-type proportions for these 'first-order cluster' firms are in comparison with the entire sample of 500 firms (as shown in column 5 of Table 5). Therefore, while it is clear that IT firms have an inherent tendency to concentrate within intra-urban clusters and that universities or colleges are often close by, membership in these clusters does not seem to influence how, or if, these firms interact with HEIs.


Advances in communication technologies have allowed increasingly more transactions to take place over large distances (Mills and Brail 2002), but many firms still exhibit strong clustering tendencies in terms of location choice. In assessing how firms are related to territory, Dicken and Malmberg (2001) argue that competitive firms are not spread evenly through space but are connected in localized clusters and Porter (2000) claims that location and cluster participation remain critical to a firm's success. For the specific case of small- and medium-sized Canadian IT firms, this study suggests (via statistical substantiation) that these firms are concentrated in space. Although as Wolfe and Gertler (2003) point out, the fact that firms are geographically clustered does not necessarily explain why, when or under what circumstances spatial proximity matters. Hence, the research presented in this paper invites subsequent investigation.

In terms of the location patterns of small- and medium-sized Canadian IT firms, clustering appears to have occurred at two levels. On one hand, a large metropolitan area bias is displayed. Toronto and Ottawa--Hull were the two most outstanding examples, but Vancouver, Calgary, Montreal, Kitchener and Edmonton have also captured a disproportionate share of Canadian IT activity. On the other hand, clustering within these seven metropolitan areas was also revealed (by means of a nearest neighbour analysis). Moreover, a nearest neighbour hierarchical spatial clustering procedure systematically identified the positions of these intra-urban IT clusters. The results derived from this technique insinuate that the downtown core, for the aforementioned seven metropolitan areas, continues to be a vital node for IT activity. In addition, for all of these CMAs (except Calgary), significant clusters were found in some 'older' suburban locations that are situated reasonably close to the more dominant downtown IT agglomerations. Conversely, the technique illustrates that there is a comparative absence of statistically significant IT accumulations in outer suburban locales and in more peripheral areas of these CMAs.

The spatial arrangement of these small- and medium-sized IT firms, particularly in favour of more central locations within the Canadian metropolitan environment, seems to imply that many of these companies are attracted to areas of dense economic activity. Moreover, most of these clusters were near (or included) institutes of higher education and some were in the vicinity of an airport. It could be hypothesized that these clustering tendencies are suggestive of other urbanization and localization advantages. Future research could more precisely determine the intrinsic spatial characteristics of these IT intra-urban clusters and more completely describe the nature of any agglomeration economies advantage (which may include proximity to markets, skilled labour, venture capital firms and so on).

Certainly other factors, beyond external economies of scale benefits, could be contributing to the spatial pattern of IT activity within larger CMAs as well. For instance, urban zoning regulations and the availability of office space limit location choice. The high number of downtown IT clusters identified by the nearest neighbour hierarchical spatial clustering analysis may be reflecting not only agglomeration advantages but also office availability in these core areas. Yet, the relative absence of significant IT agglomerations in more peripheral locations does not follow the recent trend of suburban office development in larger Canadian cities (Gad and Matthew 2000) so correspondence does not appear to be exact. An interesting corollary study, then, could be to assess agglomeration advantage as constrained by office space supply.

As stated, major universities and colleges are typically located within, or close to, these IT clusters but the overall significance of small-and medium-sized IT firm-higher education liaisons in Canada remains ambiguous. Based on a survey of 500, it was established that 42.5 percent of small--and medium-sized IT firms have had no working contact with a university or college. Of those firms that did work with an institution of higher education, participation in co-operative programs was by far the most common linkage, but general networking and graduate recruitment were also listed as important by several IT companies. On average, the location of the firm's head office and/or principle area of R&D had very little impact on the nature of firm-education linkages. Also, firms that were members of intra-urban IT clusters did not approach university/college relationships with any appreciable distinctiveness.

As Cohen et al. (2002) point out, those who have studied university/college-firm connections are usually led to the policy questions: how much government support should these relationships receive and should bridging mechanisms be encouraged? Governments in Canada continue to develop policies that encourage high-technology agglomeration and facilitate greater integration between high-technology industry and HEIs. The results from this research suggest that when small--and medium-sized IT firms have pursued working relationships with universities or colleges, it has usually been to secure in-house 'people' resources, more so than to develop highly structured faculty liaisons in the area of product development, systems design or other R&D-related projects. Indeed, other studies have demonstrated that growth in small technology firms in Canada may be impaired by a lack of skilled IT workers (Hilson 2003); hence, it is not surprising that co-operative relationships, graduate recruitment and general networking with faculty (which may also involve student commitment in some cases) were notable responses in the questionnaire tally. Thus, students, in the form of co-operative interns or recent graduates, may represent the fundamental bond between IT firms and institutions of higher education in Canada, and it is within this realm that government involvement could be enhanced. In fact, several respondents from the survey implied that their firm could benefit from an increase in co-operative program subsidization. In addition, there may be a need for governments to encourage dialogue between technology-based firms and education institutions in the field of curriculum development so that the skills students develop within the university or college environment continue to match those that are needed within the IT sector.


The author thanks Roger Pitblado for the use of his postal code conversion program, Leo Lariviere for his help with the illustrations, Daniel Lebonte for his work in administering the phone surveys and CATAlliance for allowing access to their database. In addition, thanks are extended to Editor Lawrence Berg and three reviewers for their valuable comments and suggestions.


ABDEL-RAHMAN, H. 2000 'Multi-firm city versus company town: a micro foundation model of localization economies' Journal of Regional Science 40, 755-769

AGRAWAL, A. 2001 'University to industry knowledge transfer: literature review and unanswered questions' International Journal of Management Reviews 3, 285-302

ASSOCIATION OF CANADIAN COMMUNITY COLLEGES. 2003 ACCC Membership List, available at (last accessed 24 July 2003)

ASSOCIATION OF UNIVERSITIES AND COLLEGES. 2003 Our Universities, available at accessed 18 July 2003)

AUDRETSCH, D.B. 2001 'The role of small firms in US biotechnology clusters' Small Business Economics 17, 3-15

AUDRETSCN, D.B., and FELDMAN, M.P. 1996 'Innovative clusters and the industry life cycle' Review of Industrial Organization 11, 253-273

BAILEY, T.C., and GATRELL, A.C. 1995 Interactive Spatial Data Analysis (New York: John Wiley & Sons)

BECKSTEAD, D., BROWN, M., GELLATLY, G. et al. 2003 A Decade of Growth: The Emerging Geography of New Economy Industries in the 1990s' The Canadian Economy in Transition Research Paper Series (Ottawa: Micro-Economic Analysis Division, Statistics Canada)

BEESON, P. 1987 'Total factor productivity growth and agglomeration economies in manufacturing, 1959-1973' Journal of Regional Science 27, 183-199

BELDERBOS, R., and CARREE, M. 2002 'The location of Japanese investments in China: agglomeration effects, keiretsu, and firm heterogeneity' Journal of Japanese and International Economics 16, 194-211

BELLEFLAMME, P., PICARD, P. and THISSE, J.F 2000 'An economic theory of regional clusters' Journal of Regional Science 48, 158-184

BERRY, M. 1998 'Strategic planning in small high tech companies' Long Range Planning 31,455-466

BOOTS, B.N., and GETIS, A. 1988 Point Pattern Analysis (Newbury Park, CA: Sage Publications)

BRITTON, J.N.H. 1996 'Specialization versus diversity in Canadian technological development' Small Business Economics 8, 121-138 2002 'Free trade and the high-tech response: a regional innovation system perspective on Toronto' in Knowledge Clusters and Regional Innovation: Economic Development in Canada, eds J.A. Holbrook and D.A. Wolfe (Montreal and Kingston: McGill-Queen's University Press) 85-111

CARLSSON, B. 2002 'Institutions, entrepreneurship, and growth: biomedicine and polymers in Sweden and Ohio' Small Business Economics 19, 105-121

CATALLIANCE. 2003 Techno-Connect, available at (last accessed 17 June 2003)

CHAMBERLIN, T., and DE LA MOTHE, J. 2003 'Northern light: Ottawa's technology cluster' in Clusters Old and New: The Transition to a Knowledge Economy in Canada's Regions, ed D.A. Wolfe (Montreal and Kingston: McGill-Queen's University Press) 213-234

COHEN, W.M., FLORIDA, R., RANDAZZESE, k. et al. 1998 'Industry and the academy: uneasy partners in the cause of technological advance' in Challenges to Research Universities, ed R.G. Noll (Washington: Brookings Institutional Press) 171-199

COHEN, W.M., NELSON, R.R. and WALSH, J.P. 2002 'Links and impacts: the influence of public research on industrial R&D' Management Science 48, 1-23

COOKE, P. 2001 'Biotechnology clusters in the UK: lessons from localization in the commercialization of science' Small Business Economics 17, 43-59

--2002a 'Regional innovative systems and regional competitiveness' in Innovation and Social Learning: Institutional Adaptation in an Era of Technological Change, eds M.S. Gertler and D.A. Wolfe (New York: Palgrave Publishers) 177-203

--2002b 'Regional innovation systems: general findings and some new evidence from biotechnology clusters' Journal of Technology Transfer 27, 133-145

DAVIS, C.H., and SCHAEFER, N.V. 2003 'Development dynamics of a startup innovation cluster: the ICT sector in New Brunswick' in Clusters Old and New: the Transition to a Knowledge Economy in Canada's Regions, ed D.A. Wolfe (Montreal and Kingston: McGill-Queen's University Press) 121-160

DICKEN, P. 2000 'Places and flows: situating international investment' in The Oxford Handbook of Economic Geography, eds G.L. Clark, M.P. Feldman and M.S. Gertler (Toronto: Oxford University Press) 273-291

DICKEN, P., and MALMBERG, A. 2001 'Firms and territories: a relational perspective' Economic Geography 77, 345-363

EBDON, D. 1987 Statistics in Geography (New York: Basil Blackwell)

FELDMAN, M.P. 2000 'Location and innovation: the new economic geography of innovation, spillovers and agglomeration' in The Oxford Handbook of Economic Geography, eds G.L. Clark, M.P. Feldman and M.S. Gertler (Toronto: Oxford University Press) 373-394

FUJITA, M., and THISSE, J.F. 1996 'Economics of agglomeration' Journal of the Japanese and International Economies 10, 339-378

GAD, G., and MATTHEW, M. 2000 'Central and suburban downtowns' in Canadian Cities in Transition: The Twenty-First Century, eds T. Bunting and P. Filion (Don Mills, ON: Oxford University Press) 248-273

GERTLER, M.S. 1996 'Worlds apart: the changing market geography of the German machinery industry' Small Business Economics 8, 87-106

GORDON, I.R., and MCCANN, P. 2000 'Industrial clusters: complexes, agglomeration and/or social networks?' Urban Studies 37, 513-532

HANINK, D.M. 1997 Principles and Applications of Economic

Geography (Toronto: John Wiley & Sons) HIGANO, Y., and SHIBUSAWA, H. 1999 'Agglomeration Diseconomies of traffic congestion and agglomeration economies of interaction in the information-oriented city' Journal of Regional Science 39, 21-49

HILSON, G. 2003 'Canadian SME e-business growth stalled by a lack of IT skills: report' Computing Canada 29,3

HUGGINS, R., and COOKE, P. 1997 'The economic impact of Cardiff University: innovation, learning and job generation' Geojourna141, 325-337

INDUSTRY CANADA. 2003 Information and Communication Technology Clusters, available at inict_c-g_tic.nsf/vwGeneratedlnterE/home (last accessed 24 July 2003)

JOHNSTONE, H., and HADDOW, R. 2003 'Industrial decline and high technology renewal in Cape Breton: exploring the limits of the possible' in Clusters Old and New: the Transition to a Knowledge Economy in Canada's Regions, ed D.A. Wolfe (Montreal and Kingston: McGill-Queen's University Press) 187-212

KHARBANDA, V.P. 2002 'Learning organizations: the process of innovation and technological change' Al and Society 16, 89-99

KLOOSTERMAN, R.C., and LAMBREGTS, B. 2001 'Clustering of economic activities in polycentric urban regions: the case of the Randstad' Urban Studies 38, 717-732

KRUGMAN, P. 1997 Geography and Trade (London: MIT Press)

LAHEY, L. 2003 'IT spending inches up' Computing Canada 29, 4

LANGFORD, C.H. 2002 'Measuring the impact of University research on innovation' in Knowledge Clusters and Regional Innovation: Economic Development in Canada, eds J.A. Holbrook and D.A. Wolfe (Montreal and Kingston: McGill-Queen's University Press) 113-132

LANGFORD, C.H., WOOD, J.R. and ROSS, T. 2003 Origins and structure of the Calgary wireless cluster' in Clusters Old and New: The Transition to a Knowledge Economy in Canada's Regions, ed D.A. Wolfe (Montreal and Kingston: McGill-Queen's University Press) 161-185

LEE, Y.S. 2000 'The sustainability of University-industry research collaboration: an empirical assessment' Journal of Technology Transfer25, 111-133

LEMARIE, S., MANGEMATIN, V. and TORRE, A. 2001 'Is the creation and development of biotech SMEs localized? Conclusions drawn from the French case' Small Business Economics 17, 61-76

LEVINE, N. 2002 Crimestat: a Spatial Statistics Program for the Analysis of Crime Incident Locations (v 2.0) (Houston: Ned Levine and Associates and Washington, The National Institute of Justice)

MALECKI, E.J. 1997 Technology and Economic Development: the Dynamics of Local, Regional, and National Competitiveness (London: Longman)

MCCANN, P. 1995 'Rethinking the economic of location and agglomeration' Urban Studies 32, 563-577

MEAD, N., BECKMAN, K., LAWRENCE, J. et al. 1999 'Industry/University collaborations: different perspectives heighten mutual opportunities' Journal of Systems and Software 49, 155-162

MEYER, S.P., and HECHT, A. 1996 'University growth poles in Canada: an empirical assessment' Canadian Journal of Regional Science XIX, 263-282

MILLS, L., and BRAIL, S.G. 2002 'New Media in the new millennium: the Toronto cluster in transition' in Knowledge Clusters and Regional Innovation: Economic Development in Canada, eds J.A. Holbrook and D.A. Wolfe (Montreal and Kingston: McGill-Queen's University Press) 135-159

MOOMAW, R.L. 1988 'Agglomeration economies: localization or urbanization?' Urban Studies 25, 150-161

MUN, S., and HUTCHINSON, B.G. 1995 'Empirical analysis of office rent and agglomeration economies: a case of Toronto' Journal of Regional Science 35,437-455

MYRDAL, G. 1957 Economic Theory of Under-Developed Regions (New York: Methuen)

NATIONAL RESEARCH COUNCIL (NRC) CANADA, 2003 Technology Clusters, available at _e.html (last accessed 12 June 2003)

NIOSI, J. 2002 'Regional systems of innovation: market pull and government push' in Knowledge Clusters and Regional Innovation: Economic Development in Canada, eds J.A. Holbrook and D.A. Wolfe (Montreal and Kingston: McGill-Queen's University Press) 39-55

NIOSI, J., and BAS, T.G. 2001 'The competencies of regions--Canada's clusters in biotechnology' Small Business Economics 17, 31-42

OWEN-SMITH, J., RICCABONI, M., PAMMOLLI, F. et al. 2002 'A comparison of US and European University-industry relations in the life sciences' Management Science 48, 24-43

PETRAKOS, G.C. 1992 'Urban concentration and agglomeration economies: reexamining the relationship' Urban Studies 29, 1219-1230

PORTER, M.E. 2000 'Locations, clusters, and company strategy' in The Oxford Handbook of Economic Geography, eds Gi. Clark, M.P. Feldman and M.S. Gertler (Toronto: Oxford University Press) 253-274

POYAGO-THEOTORY, J., BEATH, J. and SIEGEL, D.S. 2002 'Universities and fundamental research: reflections on the growth of university-industry partnerships' Oxford Review of Economic Policy 18, 10-21

PREVEZER, U. 1997 'The dynamics of industrial clustering in biotechnology' Small Business Economics 9, 255-271

ROMUN, H., and ALBALADEJO, M. 2002 'Determinants of innovation capability in small electronics and software firms in southeast England' Research Policy 31, 1053-1067

SANTORO, M.D., and CHARRABARTI, A.K. 2002 'Firm size and technology centrality in industry-university interactions' Research Policy 31, 1163-1180

SAXENIAN, A. 1994 Regional Advantage: Culture and Competition in Silicon Valley and Route 128 (Cambridge, MA: Harvard University Press)

SCHIENSTOCR, G., and TULKRI, P. 2001 'The Fourth Pillar? An assessment of the situation of the Finnish biotechnology' Small Business Economics 17, 105-122

SELTING, A., LOVERAGE, S. and ALLANACH, C. 1995 'The economics of agglomeration and firm location' Great Lakes Geographer 2, 17-29

STATISTICS CANADA. 2000 Postal code conversion file' catalogue: 92FOO27XDB (Ottawa: Statistics Canada)

--.2003 Census of Canada, 2001 (Geo-Suite). Population (2001 and 1996), Dwelling Counts (2001), and Land Area for Census Metropolitan Areas and Census Agglomerations' DRL Study Number: 024.2001 (Ottawa: Statistics Canada)

STEPHAN, P.E. 2001 'Educational implications of University-industry technology transfer' Journal of Technology Transfer 26, 199-205

TREMBLAY, D.G., KLEIN, J.L., FONTAN, J.M. et al. 2002 'Modalities for the establishment of the relational firm: a case study of the cite du multimedia, Montreal' in Knowledge Clusters and Regional Innovation: Economic Development in Canada, eds J.A. Holbrook and D.A. Wolfe (Montreal and Kingston: McGill-Queen's University Press) 161-185

VARGA, A. 2000 'Local academic knowledge transfers and the concentration of economic activity' Journal of Regional Science 40, 289-309

WHEELER, J.O., MULLER, P.O., THRALL, G.I. et al. 1998 Economic Geography (Toronto: John Wiley & Sons)

WOLFE, D.A. 2002 'Social capital and cluster development in learning regions' in Knowledge Clusters and Regional Innovation: Economic Development in Canada, eds J.A. Holbrook and D.A.Wolfe (Montreal and Kingston: McGill-Queen's University Press) 11-38

WOLEE, D.A., and GERTLER, M.S. 2003 'Clusters old and new: lessons from the ISRN study of cluster development' in Clusters Old and New: The Transition to a Knowledge Economy in Canada's Regions, ed D.A. Wolfe (Montreal and Kingston: McGill-Queen's University Press) 1-36

YAMAWARI, H. 2002 'The evolution and structure of industrial clusters in Japan' Small Business Economics 18, 121-140

ZELLER, C. 2001 'Clustering biotech: a recipe for success? Spatial patterns of growth of biotechnology in Munich, Rhineland and Hamburg' Small Business Economics 17, 123-141

ZHENG, X.P. 1998 'Measuring optimal population distribution by agglomeration economies and diseconomies: a case of Tokyo' Urban Studies 35, 95-112

(1) If one assumes that additional points lie just beyond the borders of the study area, the Crimestat software package will correct for 'edge effects'. 'If a point is closer to the border than to the measured nearest neighbour distance, then the distance to the border is taken as the adjusted nearest neighbour distance' (Levine 2002, 29). However, given the strong metropolitan bias of IT companies in Canada, it is not reasonable to assume that there is always another IT company outside the border. As a result, an edge correction was not used. For a complete discussion on how boundaries and edge effects can be modelled, consult Bailey and Gatrell (1995) or Boots and Getis (1988).

STEPHEN P. MEYER Department of Geography, Laurentian University, Sudbury, ON, Canada P3E 2C6 (e-mail:
Table 1
Information technology firms in Canada: by activity

                                            Number of    Percent of
Information technology activity             activities       total

Software developer                                 662         38.3
Consulting                                         383         22.2
Value-added reseller                               208         12.0
Multimedia content development                     132          7.6
Hardware manufacturing/original equipment          116          6.7
Software publisher                                  98          5.7
Distributor                                         89          5.2
Internet service provider                           39          2.3
Total                                            1,727        100.0

Table 2
Information technology firms in Canada: by urban area

                                               Number    Percent of
Urban area (CMA/CA)                          of firms        total

Toronto (ON)                                       704         33.9
Ottawa-Hull (ON/QB)                                291         14.0
Vancouver(BC)                                      176          8.5
Calgary (AB)                                       148          7.1
Montreal (QB)                                      137          6.6
Kitchener (ON)                                      81          3.9
Edmonton (AB)                                       61          2.9
Hamilton (ON)                                       36          1.7
Halifax (NS)                                        34          1.6
Winnipeg (MB)                                       33          1.6
St. John's (NF)                                     24          1.2
Quebec City (QB)                                    23          1.1
Victoria (BQ                                        23          1.1
Fredericton (NB)                                    20          1.0
London (ON)                                         20          1.0
Remaining urban (CMA/CA)                           199          9.6
Rural                                               64          3.1
Total sample                                     2,074         99.9

CMA, census metropolitan area.

Table 3
Nearest neighbour (NN) analysis of information technology firms in
seven Canadian census metropolitan areas (CMAs)

CMA             Mean NN    random     NN      Standard
                distance   distance   index   error
  Actual area     514.26   1,503.79    0.34      29.63
  Points area     514.26   1,447.81    0.36      28.52
  Actual area     469.49   2,137.53    0.22      65.50
  Points area     469.49   1,281.51    0.37      39.27
  Actual area     843.83   2,022.08    0.42      79.67
  Points area     843.83   1,666.22    0.51      65.65
  Actual area     449.76   2,930.21    0.15     125.90
  Points area     449.76     830.91    0.54      35.70
  Actual area   1,153.69   2,717.66    0.42     121.37
  Points area   1,153.69   2,339.18    0.49     104.47
  Actual area     456.18   1,597.63    0.29      92.79
  Points area     456.18   1,352.84    0.34      78.57
  Actual area   2,228.22   6,212.96    0.36     415.82
  Points area   2,228.22   3,951.98    0.56     264.50

                Test        P-value
CMA             statistic    (one-
                (Z)         tailed)
  Actual area      -33.40    0.0001
  Points area      -32.73    0.0001
  Actual area      -25.47    0.0001
  Points area      -20.68    0.0001
  Actual area      -14.79    0.0001
  Points area      -12.53    0.0001
  Actual area      -19.70    0.0001
  Points area      -10.68    0.0001
  Actual area      -12.89    0.0001
  Points area      -11.35    0.0001
  Actual area      -12.30    0.0001
  Points area      -11.41    0.0001
  Actual area       -9.58    0.0001
  Points area       -6.52    0.0001

NOTE: The mean random distance was calculated using 'actual area'
And 'points area'. Actual area uses the CMA's area in square kilometres
(Statistics Canada 2003); points area is the rectangular area defined
by the minimum and maximum Xand Yco-ordinates.

Table 4
First-order information technology (IT) clusters: summary of results

                              proportion     Number of
               Number of     of IT firms      IT firms
              first-order   in first-order   in largest
CMA            clusters      clusters (%)      cluster

Toronto           12             54.7            101
Ottawa-Hull        6             S2.2            63
Vancouver          2             33.0            44
Montreal           2             43.1            40
Kitchener          2             45.7            23
Edmonton           2             68.9            23
Calgary            1             37.1            55

               Mean distance     Shortest distance
                of clusters         between a
                 to nearest       cluster and a
              higher education   higher education
CMA           institution (km)   institution (km)

Toronto             4.38               0.25
Ottawa-Hull         5.14               1.17
Vancouver           4.25               3.32
Montreal            3.75               0.70
Kitchener           2.02               0.76
Edmonton            3.30               0.77
Calgary             1.35               1.35

CMA, census metropolitan area.

NOTE: Distances are measured from the centre of a cluster's ellipse to
the nearest higher education institution.

Table 5
Information technology (IT) firm-university/college connections in
Canada (shown in percent of total)

                                  All          Local
                               sampled    head office
Contact type                    firms    and local R&D

No contact                        42.5            43.0
Co-operative student programs     38.4            38.3
General networking with            7.0             7.4
Hire/recruit graduates             5.0             4.1
Sell product/service to            2.9             3.3
Work with faculty on specific      2.0             2.2
Use university/college             1.4             1.1
Teach a course                     0.9             0.6
Total contacts                     558             460

                                       Non-local         IT firms in
                                     head office         first-order
Contact type                   and non-local R&D            clusters

No contact                                  39.8                43.6
Co-operative student programs               38.8                36.3
General networking with                      5.1                 7.4
Hire/recruit graduates                       9.2                 6.4
Sell product/service to                      1.0                 3.4
Work with faculty on specific                1.0                 1.5
Use university/college                       3.1                 1.0
Teach a course                               2.0                 0.5
Total contacts                                98                 204
COPYRIGHT 2006 Canadian Association of Geographers
No portion of this article can be reproduced without the express written permission from the copyright holder.
Copyright 2006 Gale, Cengage Learning. All rights reserved.

Article Details
Printer friendly Cite/link Email Feedback
Author:Meyer, Stephen P.
Publication:The Canadian Geographer
Geographic Code:1CANA
Date:Mar 22, 2006
Previous Article:Measuring the effectiveness of Parks Canada's environmental management system: a case study of Riding Mountain National Park.
Next Article:Political Ecology: A Critical Introduction.

Terms of use | Privacy policy | Copyright © 2019 Farlex, Inc. | Feedback | For webmasters