Spatial Analysis of Local Competitiveness: Relationship of Economic Dynamism of Cities and Municipalities in Major Regional Metropolitan Areas in the Philippines

Abstract

The measurement of cities and municipalities competitiveness in the Philippines has been put in place by the Department of Trade and Industry since 2013. However, its use as a spatial planning parameter is lacking in the literature. This paper reviewed the factors that drive competitiveness. The research revealed that “economic dynamism” appeared at the top of the factors that contribute to competitiveness and influence regional development. Given urban and regional planning theories, metropolitan areas were chosen as the most appropriate case study sites that exhibit economic dynamism. The study revealed that the “Economic Dynamism Index”, or EDi, of cities and municipalities are spatially correlated, indicative of their clustering pattern in the economic space. The clustering pattern was determined by treating the EDi as a spatial attribute in the major metropolitan case study areas. Using Moran’s I global spatial autocorrelation analysis, the clustering pattern of cities and municipalities observed through the GIS map was validated by the 99% significance in the spatial statistics of the EDi dataset. This suggests that “complementation” among cities and municipalities exists rather than competition. Thus, sustainable regional spatial/economic development strategies can be reformulated, given the spatial interactions of areas with higher EDi with the less endowed cities/municipalities at the periphery.

1. Introduction

The Cities and Municipalities Competitiveness Index (CMCI) framework of the Department of Trade and Industry (DTI) in the Philippines can be categorized under the Theory of Creative Destruction [1], where constant innovation was introduced in the economic geography of cities, towns, and regions to maintain competitiveness. This framework viewed competitiveness alongside Porter’s [2] concept of location-based productivity, where available resources in a city or municipality are converted to productive assets. One of the metrics used by the CMCI in determining productivity within these administrative units is the “economic dynamism index” (EDi). However, the application and usage of EDi as a regional and urban planning analytical tool were not yet explored and covered in the literature. There is also no scientific proof of a causal relationship between various factors and levels of regional competitiveness. Further, the theoretical or methodological basis for increased regional competitiveness is lacking. In this context, the use of the competitiveness index among cities and municipalities, as a planning tool for urban and regional development strategy formulation, needs to be explored.

In this research, EDi explains the nature of competitiveness arising from the spatial hierarchies of local government units (LGUs) that refer to municipalities and cities. The varying levels of competitiveness across space emphasizes Perroux’ observation that “economic growth does not happen everywhere and all at once. Growth manifests itself into patches of “growth poles” across space at varying intensities’’ [3]. Consequently, the impact of concentration of growth in existing centers would often result in backwash/polarization [4,5].

The general objective of the study is to examine the spatial relationship in the EDi of cities and municipalities in metropolitan areas as an approach to regional development planning. Specifically, the study intends to: (a) examine the hierarchy and spatial association in the EDi of cities and municipalities in major metropolitan areas; (b) asses the nature (competing or complementing) of economic competitiveness in regions, based on the spatial association of EDi of cities and municipalities; and (c) evaluate the influence of the prevailing regional spatial development strategy and economic development policy. These objectives are hinged on the assumption that spatial correlation exists in the economic dynamism indices of neighboring cities and municipalities within a metropolitan area.

2. Literature Review and Conceptual Framework

The theory on agglomeration of firms in space is relevant in the analysis of spatial and hierarchical competition, whereby businesses tend to cluster towards each other [6,7,8]. This phenomenon is observed in areas of high population, such as large metropolitan areas. Clustering establishes competitive advantage, characterized with geographical concentrations of interconnected companies, specialist suppliers, service providers and firms in related industries and associated institutions [9].

An economic cluster refers to geographically contiguous spatial units that have similar, interconnected and complementary firms that share infrastructure/services and a common institutional environment similar to a metropolitan authority set up [10]. The contribution in economic development and urban growth of clusters and metropolitan areas has been studied extensively from different conceptualizations in North America, Europe and East Asia [11,12,13,14]. It was observed that different types of industry participants were more concentrated inside centers than outside centers, and the degree of clustering was higher for the larger, relatively dense centers [15]. These empirical studies resonate with Porter’s concept that the path from these cluster growths to local economic competitiveness is through strengthening the connectedness of productive and non-productive factors driving economic development [9].

In the case of the Philippines, economic clusters are typically formed among contiguous political administrative units that are spatially organized into metropolitan districts. This spatial organization is reflected in the Philippine Development Plan (PDP) 2016–2022, where spreading growth links metropolitan areas with lagging localities to address socio-economic inequalities [16]. This is done by inducing spatial interactions between economic clusters following inter-metropolitan complementation.

Literature also recognizes that regional development and competitiveness are driven by hierarchy and the grouping of several factors. Development theorists recognized geographical diversification as growth reached localization stage [4,5]. Indeed, cities and, more generally, economic agglomerations, are considered as the main institutions where both technological and social innovations are developed through market and nonmarket interactions.

Regional developments and competitiveness are characterized by dynamic and static factors. Dynamic factors represent the pyramid of factors of regional competitiveness to include the basic factors, supporting factors and dynamic factors (Figure 1). Arranged hierarchically, beginning at the top of the pyramid, are dynamic factors, treated as the most active and flexible [17]. Business or economic activity is considered the core of the dynamic factors, with innovation and international integration as complementary factors. The middle of the pyramid accounts for the supporting factors that create a basis for regional development. The bottom of the pyramid is composed of basic factors that form initial peculiarities of a given region.

Static factors are factors acting as policies/mandates. This includes regional development directions mandated in the form of a National Framework on Physical Planning (NFPP), which sets the parameters for spatial development in a country. The NFPP establishes a countrywide direction for the judicious, sustainable and strategic utilization of land, alongside the medium-term and long-term economic development agenda. It serves as the reference document for physical planning at the regional, provincial and local levels, to ensure wholistic and consistent actions across decision-makers within the government bureaucracy. In the Philippines, the NFPP laid out spatial plans from the national down to the local level over a span of time, and periodically updated. The main direction is towards fostering spatial agglomeration, interconnectivity and resiliency through a National Spatial Strategy (NSS). The NSS is an established hierarchy in the network of settlements, based on the overarching roles performed by an administrative unit (

Table 1. Role, function and population size of an administrative unit per level of hierarchy.

Level of Hierarchy	Role	Indicator Functions	Population
Regional center	Major center that services national and international companies	Presence of international port, airport, commerce (shopping malls; luxury brand auto dealership; and business process outsourcing)	1,200,000 and above
Sub-regional center	Provides support services that serve beyond local boundaries	National airport, Level 3 hospitals, Commercial establishments (fast food chains; malls; deluxe hotels; and auto dealership), National real estate developers	120,000–1,200,000
Provincial center	Provincial services and administration	Presence of tertiary level education, Level 2 hospital, commercial banks, commerce (other hotels; hardware; grocery; convenience store; IT store), other residential subdivision, auto repair shop, service forwarders i.e., DHL, FEDEX, LBC, etc.	50,000–120,000
Local center	Rural and agricultural services	Presence of food retail (carinderia or snack house), secondary school, primary health care (local health center), bus terminal, lodging inn	50,000 and below

Source: [18].

).

In operationalizing the spatial development directions of the Philippines, the NEDA has carved out three (3) major metropolitan areas (

Table 2. List of metropolitan areas in the NEDA planning document.

Year 2007 Planning Document	Philippine Development Plan (2017–2022)
(1) Metro Manila (NCR) (2) Metro Cebu (Region VII) (3) Metro Davao (Region XI) (4) Metro Angeles (Region III) (5) Metro Bacolod (Region VI) (6) Metro Baguio (CAR) (7) Metro Batangas (Region 4A-CALABARZON) (8) Metro Cagayan de Oro (Region X) (9) Metro Dagupan (Region I) (10) Metro Iloilo-Guimaras (Region VI) (11) Metro Naga (Region V) (12) Metro Olongapo (Region III)	(1) Metro Manila (NCR) (2) Metro Cebu (Region VII) (3) Metro Davao (Region XI) (4) Metro Cagayan de Oro (Region X) 1

¹ Projected to reach metropolitan level in terms of population by 2025.

), one for each major island group, with the end view of achieving efficiency and economies of scale.

Given the exhaustive review of existing literature in regional development and competitiveness, the research espouses a conceptual framework that establishes the relationship between the drivers of regional competitiveness and the structure of clustering cities and municipalities in metropolitan areas toward an enhanced approach for regional spatial/economic development strategy (Figure 2). The independent variables are the static and dynamic drivers of regional competitiveness. The former refers to the established term-based (based on a planning horizon) spatial strategies or the policy guidelines for spatial development imposed by the national government. Comparatively, the latter represents the productivity indicators or the dynamic factors that are confined to the CMCI EDi. This index is considered relatively the most important productivity metric on the hierarchy of factors that influence regional competitiveness.

3. Materials and Methods

3.1. Research Methodological Framework

The research followed a quantitative approach in dealing with the primary data sourced from the Philippines DTI–CMCI, particularly the pillar on Economic Dynamism (EDi). The other pillars of local competitiveness include government efficiency, infrastructure, resiliency, and innovation. These other pillars were excluded from the study, instead the preference was focused on EDi, taking into account the results of the literature review which ranks business activity on top of the hierarchy of factors influencing regional competitiveness [17]. This is also based on the fundamental concept of the global competitiveness index, which recognizes that the start to existence of competitiveness is where there is business activity producing goods and services. Further, it denotes that the other pillars of competitiveness are but complementary in the creation and attainment of higher added value in pursuit of economic growth [19]. These underlying considerations led the study to focus on the EDi in explaining the competitiveness of cities and municipalities within the coverage of the identified major metropolitan case study areas, particularly Greater Metro Manila, Metropolitan Cebu and Metropolitan Davao.

The schematic diagram of the research methodological framework was sequentially organized (Figure 3) to test the assumption of the study and, consequently, establish the spatial clustering in the EDi values of cities and municipalities within each case study metropolitan area.

This is in sync with the conceptual framework that takes into lens the theoretical constructs that mean competition results in uneven spatial development [4,5], and the concepts of agglomeration in the economic space manifested in growth poles [3], which eventually leads in the clustering of firms to achieve competitive advantage [2]. These key considerations are built within the three major parts of the methodological framework discussed as follows: (a) analysis of regional competitiveness drivers; (b) analysis of cities and municipalities spatial arrangement; and (c) reformulation of regional spatial/economic development strategy.

Analysis of the drivers of regional competitiveness. This involves data collection of the static factor, descriptive review of these data and translating them to a spatial structure through mapping. This part involves the gathering of relevant spatial development strategies/plans at the national and administrative region levels of each metropolitan area. The spatial strategies that were indicative of hierarchical organization in the network of urban settlements were identified, described, compared, and mapped, featuring cities and municipalities using Geographic Information System (GIS).

Another component of assessing the drivers of regional competitiveness is the characterization of the dynamic factors, reprocessing, and mapping of the reprocessed EDi, called REDi. The characterization involves database preparation of EDi indicators and sub-indicators for the cities and municipalities within each metropolitan case study area. This is followed by the reprocessing of the EDi, which entailed the recalculation of the economic dynamism index (REDi) from the database of indicators and sub-indicator values. The design of calculating the competitiveness pillar “indices” (economic dynamism, government efficiency, infrastructure, resiliency, and innovation) from the CMCI operations manual follows the concept of competitive ranking, where the individual index is expressed as a function of the part (individual city or municipality) to the whole (entire group of cities and municipalities). This is done by considering the actual value against the minimum and maximum values within the dataset per indicator or sub-indicator being analyzed [20]. In visualizing the spatial clustering, corresponding REDi values assigned to each city and municipality were mapped through ArcGIS.

Analysis of the Spatial Arrangement of Cities and Municipalities in Metropolitan Areas. This phase of the research involved the analysis of the spatial structure, observed from the overlay of the REDi values within the urban hierarchy map generated in the analysis of the drivers of regional competitiveness.

The Moran’s I global spatial autocorrelation analysis tool in ArcGIS was applied to primarily test the occurrence of clustering in the REDi dataset. The resulting Moran’s I coefficient (index) also yields for ArcGIS a “test statistics” that further validates clustering through spatial and attribute similarity in the REDi dataset. A clustering pattern is confirmed if the similarity in REDi values between observations (cities and municipalities) in the dataset for the whole metropolitan case study area is related to the similarity in the locations of such observations. This confirmation of spatial association in the REDi dataset established the spatial relationships against the hierarchical functions assumed by each city and municipality in the network of settlements within the case study area.

The matching of the spatial statistics with the established settlement hierarchies tested the assumption that complementation influenced spatial interaction, which is manifested in the clustering of the REDi values across cities and municipalities in the case study areas.

Reformulation of an enhanced approach in the regional spatial/economic development strategies. This part of the methodological framework is based on the visual (mapping) and numerical (spatial autocorrelation statistics) confirmation of clustering in the REDi dataset established in the spatial arrangement of cities and municipalities. The research formulated enhancements in terms of spatial/economic development strategies that can be adopted prospectively in planning land use, formulating regulatory policy to create and expand metropolitan areas and broaden spatial complementation.

3.2. Selection of Case Study Areas

Metropolitan areas are considered representations of variability in the levels of competitiveness between the center and its surrounding political and administrative units. The growth pole theory by Perroux demonstrates spatial concentration and expansion of the central city [3], causing “polarization effects”, and leading to uneven development [4,5] between the center and the periphery.

The Philippines central planning agency–NEDA—has established the same conceptualizations in achieving long-term aspirations (coined as “Ambisyon Natin” 2040) through the PDP and in operationalizing the National Spatial Strategy (NSS). Both the PDP and NSS imply that the economic, social, and political interactions should originate in the major metropolitan areas and trickle down to the regional, sub-regional, provincial and local centers in a network of hierarchically arranged interconnected settlements. This interconnection network outlines the fabric for spatial interaction within the market of goods and services to be “complementary” between the larger order settlements and the subsequent rank settlement or administrative area.

In light of these conceptualizations, the research has aligned the selection of the case study areas alongside NEDA’s designation of Davao (part of southern Philippines), Cebu (part of central Philippines) and the greater Metro Manila (part of northern Philippines) as major metropolitan areas (Figure 4) appropriate for explaining “spatial complementation” through the clustering of the economic dynamism index.

3.3. Measurement of Variables and Indicators

For consistency with the sphere of factors mentioned in the conceptual and methodological frameworks, “variables” shall refer to the static and dynamic factors that are considered drivers of regional competitiveness.

The static factors analyzed in this study include the regional spatial development strategies, or spatial conceptualization, for the case study areas. The analysis of these factors aimed at fostering regional socio-economic interactions within the established hierarchy in the network of urban settlements, usually spanning over a 30-year period. Comparatively, the dynamic factors centered on EDi from the various observations in literature show that productivity and competitiveness is primarily a function of business and economic activities [17]. The research treats the EDi as a collective measure indicative of the business and economic output of a particular administrative area following the DTI-CMCI manual of operations [21]. The methodology of computing and ranking the index considered the EDi as a full representation of the economic dynamism pillar of a city or municipality within the case study area. This pillar is comprised of the following sets of indicators and sub-indicators (

Table 3. DTI-CMCI list of indicators and sub-indicators on economic dynamism pillar.

No.	Indicator and Sub-Indicator(s)	No.	Indicator and Sub-Indicator(s)
1.	Size of Economy Gross sales of registered firms Total capitalization of new businesses	7.	Cost of Doing Business d. Daily minimum wage rate Non-Agricultural iii. Non-agricultural establishments with more than 10 workers iv. Non-agricultural establishments with 10 workers or below e. Cost of land in a central business district f. Cost of rent
2.	Growth of Economy Growth of gross sales of registered firms Growth of total capitalization of new businesses
3.	Structure of Local Economy Total number of approved business permits for new business applications Total number of approved business renewals
4.	Safety Compliant Number of occupancy permits approved Number of approved fire safety inspection	8.	Financial Deepening Number of universal/commercial banks Number of thrift and savings banks Number of rural banks Number of finance cooperatives Number of savings and loans association with quasi-banking functions Number of pawnshops Number of money changers/foreign exchange dealer Number of remittance center
5.	Increase Employment Number of declared employees for new business applications Number of declared employees for business renewals
6.	Cost of Living Local inflation rate
7.	Cost of Doing Business a. Cost of electricity Commercial users Industrial firms/customers b. Cost of Water Commercial users Industrial firms/customers c. Price of diesel (on 31 Dec of the preceding year) d. Daily Minimum Wage Rate Agricultural i. Agricultural plantation ii. Agricultural non-plantation	9.	Productivity Gross sales of registered firms Number of declared employees for business renewals
		10.	Total Number of Business and Professional Organization Total number of LGU-recognized business groups Total number of other business groups

Source: [20].

) and measured consistent with the methodology prescribed by the DTI-CMCI [20].

3.4. Data Collection, Sampling Methods, Statistical Tests and Parameters, and Analysis

This study recognizes the need to further articulate the data analysis and their corresponding major output to build the visual and conceptual connections that are needed to understand the technical aspect of the research and the findings of this study. The following discussions were sequenced alongside the output of the three major parts of the methodological framework.

3.4.1. Analysis of the Drivers of Regional Competitiveness

• Mapping of urban settlement hierarchy

Data Collection, Sampling Methods and Parameters. The published information about the NSS, the spatial development strategies for the major islands of Luzon (northern Philippines), Visayas (central Philippines) and Mindanao (southern Philippines), the regional framework on physical planning (RFPP), and the legislative approvals creating the metropolitan areas were gathered from either the websites or the public domain of NEDA, regional development councils and the cities and municipalities. Specific information about the spatial strategies/framework documents that the research took note of are in the areas of: (a) NSS thrust and directions, e.g., agglomeration, interconnection and resiliency; (b) RFPP established hierarchy in the network of settlements (metropolitan, regional, sub-regional, provincial and local centers); and (c) Regional development spatial strategies (i.e., integrated area development, balanced agro-industrial development, growth pole, hub and spoke, etc.). The review of the hierarchies in the network of urban settlements aided in establishing consistency or possible differentiation between the NSS and RFPP. These functional hierarchies assigned for a particular city/municipality are visually reflected and anchored in a GIS map.

Data Input and Analysis. The locations of the city/municipal halls and the shapefiles of the administrative boundaries of the city/municipality obtained from the Open Street Map of the United Nations Office for the Coordination of Humanitarian Affairs (UN OCHA) was used in creating the GIS map on the hierarchy in the network of settlements. A descriptive review of the functional hierarchies in the network of settlements was discussed by means of comparing the current role of the city/municipality indicated in the RFPP against the default/standard function prescribed in the NSS. A similar but broad-based comparison among the significant metropolitan case study area was also highlighted.

Data Output. GIS map of the hierarchical organization in the network of urban settlements for each case study area. This settlement hierarchy map is considered representative of the prevailing spatial strategies in the metropolitan region.

b.

Mapping of cities and municipalities REDi values

The descriptive information about the EDi of cities and municipalities within the case study areas was obtained from the DTI-CMCI online portal [20]. The characterization includes raw data collection, database organization and the reprocessing of EDi. The reprocessing of the EDi primarily entailed the recalculation of the index, using the economic dynamism pillar indicators and sub-indicators raw data of all cities and municipalities that belong to the metropolitan case study area, irrespective of their income classification. Through GIS mapping, the spatial pattern/arrangement formed by the reprocessed EDi was surfaced and visualized.

Data Collection, Statistical Tests and Parameters. The year 2020 raw data on the indicators and sub-indicators under the economic dynamism pillar were processed, consistent with the scoring procedure indicated in the year 2021 DTI-CMCI operations manual [21]. From the CMCI formulas applicable for each indicator and sub-indicators, a reprocessed EDi, termed in this paper as “REDi”, value was generated for each city/municipality within the case study area. This reprocessing in the entire metropolitan study area dataset is necessary to surface out the presence of clustering. The current approach in the DTI-CMCI operations manual of grouping and ranking the EDi according to the city/municipality income classification cannot surface such a spatial pattern out, apart from being disconnected with the concept on competitive clusters or the spatial interaction of firms/economic agents introduced by Porter [9].

The grouping of cities/municipalities according to income classification is also counterintuitive to Tobler’s first law in geography that everything is related to everything else, but nearer things are more connected than distant ones. The general concept applied in economic geography assumes that, regardless of the income class, these administrative units referring to the cities/municipalities (spatial representations) are more likely to engage in spatial interaction (socio-economic and political) as neighbors than those distant ones. A neighbor in spatial analysis is conceptualized to be contiguous or shares a boundary, or one that belongs to a specific distance bandwidth within an established unit of analysis (case study area). Such conceptualization tested the research hypothesis alongside Porter’s theoretical notion of competitiveness through clustering of firms outlined in the conceptual framework of this study [9].

Sampling Methods. The indicators and sub-indicators dataset collected accounts for 100% of the total number of cities and municipalities (collectively referred in the Philippines as cities and municipalities) in each case study area. For metropolitan Davao dataset there are 49 cities and municipalities; the metropolitan Cebu dataset has 53 cities and municipalities under the administrative province; while the greater metropolitan Manila area dataset featured 49 cities and municipalities. These numbers satisfy the minimum prescribed 30 values or attributes fit for spatial autocorrelation analysis.

Data Input and Analysis. Data inputs for the mapping include: (a) shapefiles of the administrative boundaries of the cities and municipalities in each case study area, sourced from the United Nations Office for the Coordination of Humanitarian Affairs (UN OCHA) open street map; and (b) attribute value referring to the REDi, based on the 2020 economic dynamism pillar raw data of indicators and sub-indicators.

The REDi was organized in a spreadsheet database and loaded in ArcGIS/ArcMap tool as an “attribute table”. The attribute table contained the names of the cities and municipalities for each case study area with their corresponding REDi values. The mapping process entailed selecting and assigning color symbology to a specific range of REDi values. This was accomplished by sorting the REDi dataset from highest to lowest to mark or establish the full range in the REDi values.

Across the case study areas, the lowest common REDi value is above 2, while the maximum is measured around 15 for the metropolitan Davao area. This REDi range of values was standardized across the case study areas by grouping into buckets starting from the 2–4 bandwidths. It gradually increases in equal spread/range until it reaches the maximum REDi value within the dataset. By grouping the REDi values into buckets, the intensity levels or differentiation across space can be visualized and shown in the map legend for the figures presented in Section 4 (Results).

The map legend applied color symbology (differentiation) according to the scale of grouping REDi values using the ArcMap tool in ArcGIS. The output is a GIS map of cities and municipalities (administrative area) with higher REDi values, shaded in a darker color and tapering off gradually in terms of shade for cities and municipalities with lower REDi values. The GIS map facilitates visualization and analysis in terms of the occurrence of clustering, the size or extent of the spatial structure, and the indicative spatial association of cities and municipalities with the metropolitan area. This visualization enhancement applies the fundamental mapping technique that translates the REDi values into spatial arrangements, reflective of either differentiation (dispersion), similarities (clustering), or the absence of any pattern.

Data Output. A GIS map of the case study areas bearing the color scheme was generated from processing the input data. The resulting GIS maps for each metropolitan case study area provided a visual indication of the occurrence of clustering pattern.

3.4.2. Analysis of Cities and Municipalities Spatial Arrangement

A spatial autocorrelation analysis based on the REDi values was carried out to validate the spatial clustering of cities and municipalities observed from the GIS map. The output of the global spatial autocorrelation analysis in ArcGIS is a Moran’s I (index/coefficient) that establishes the nature of the spatial pattern and a spatial statistic that enabled the validation of the pattern. This spatial analytical process guided the confirmation of clustering pattern in line with the functions assumed by the city/municipality, based on the prevailing regional spatial development strategy.

• Spatial autocorrelation analysis of REDi dataset

The global Moran’s I spatial autocorrelation tool in ArcGIS was used to evaluate whether the pattern expressed by the REDi attribute indicates clustering, dispersion or complete spatial randomness for each case study area. Preference for the analytical technique was considered due to its suitability for polygon shapes, such as that of an administrative unit [22].

The variables analyzed are the attributes relative to their geospatial coordinates, thus, making the analysis different from the bivariate Pearson correlation and similar correlation statistics. The statistical concept of spatial autocorrelation in literature treats the attribute in line with proposition that everything is related to everything else, which lends the concept of independence and randomness from traditional statistics irrelevant [23].

Data Inputs and Analysis. The data inputs were: (a) shapefiles of the administrative boundaries of the cities and municipalities in each case study area from the open street map of the United Nations Office for the Coordination of Humanitarian Affairs (UN OCHA); and (b) attribute value or the REDi value. These data inputs were processed consistent with the following standard sequence of carrying out spatial autocorrelation:

• Moran’s I tool is selected from ArcMap in ArcGIS to initialize loading of shapefiles of the case study area GIS map and REDi dataset;
• Contiguity edges, or rook contiguity, is applied as the mode of conceptualization in treating cities and municipalities sharing administrative boundaries as neighbors. This conceptualization in literature is a “first order contiguity principle”, consistent with Tobler’s law that near things are more relevant than distant things. A shared or contiguous administrative boundary relates to spatial similarity, while REDi values relative to a specific range of “high and high values” or a “low and low values” describe attribute similarity;
• Euclidean measurement is selected as the measurement unit for the distance between administrative areas (polygons); and
• Processing of the Moran’s I using the ArcMap tool, based on the following standard mathematical expression of the global index.

$$\mathrm{I}=\frac{\mathrm{N}{{\displaystyle \sum }}_{\mathrm{i}=1}^{ \mathrm{n}}{{\displaystyle \sum }}_{\mathrm{j}=1}^{ \mathrm{n}}{\mathrm{w}}_{\mathrm{ij}}\left(x_i-\overline{x}\right)\left(x_j-\overline{x}\right)}{({{\displaystyle \sum }}_{\mathrm{i}=1}^{ \mathrm{n}}{{\displaystyle \sum }}_{\mathrm{j}=1}^{ \mathrm{n}}{\mathrm{w}}_{\mathrm{ij}}){{\displaystyle \sum }}_{i=1}^{ n}{(x_{i }-\overline{x})}^2}$$

(1)

where, $(x_i-\overline{x})$ is the deviation (z_i) in the REDi value at location i from its mean, ${\mathrm{w}}_{\mathrm{ij}}$ is the spatial weight between location i and j, and N is the total number of observations. The ArcMap tool calculates the mean and variance of the REDi being evaluated. For each REDi value, the mean is subtracted to generate the deviation from the mean. Deviation in REDi values for all neighboring locations are multiplied together to create a cross-product.

Data Output. The ArcGIS/ArcMap processing tool generates the global Moran’s I value of the REDi dataset for each case study area. A Moran’s “I” coefficient value of “0” indicates a random pattern, while a negative one (−1) value represents checkered or dispersed pattern, and a positive one (+1) value suggests a clustered pattern (Figure 5). The coefficient is interpreted as the global result for the case of the metropolitan area under study.

b.

Validation of spatial association in the REDi dataset.

ArcGIS is also used to generate the spatial statistics of the REDi dataset together with the graph of the statistics.

The report of the test statistics features the Moran’s “I” coefficient, together with the corresponding “p” value and “z” value. The “p” value is the probability that the observed spatial pattern is due to some random process. The “z” value reflects the standard deviations of the observed spatial pattern. The basic literature on “z” value offers scenarios with very high (positive) or very low (negative) z-scores, associated with very small p-values, that are found at the tails of the normal distribution curve [22]. This enabled the study to generate comparison and generalization among the metropolitan case study areas in relation to the null hypothesis of complete spatial randomness. The calculation of the “z” value for Moran’s I is based on the normal frequency distribution of each REDi dataset expressed in the equation below:

$$Z=\frac{I-E\left(I\right)}{S_{error\left(I\right)}}$$

(2)

where I is the calculated value for Moran’s “I” from the REDi dataset, E (I) is the expected value for a random distribution, and S is the standard error. The expected value is representative of a random distribution and is determined based on the following standard expression:

$$E\left(I\right)=\frac{-1}{n-1}$$

(3)

These statistics are used to confirm whether the clustering pattern observed is significant, or due to some random form. The standard level for accepting significance is 90% confidence level in the “p” value, with the corresponding critical standard “z values” shown in

Table 4. Standard Range of Spatial Autocorrelation Statistics.

Significance Level (in p-Value)	Critical Value (z-Score)
0.01	Less than −2.58
0.50	−2.58 to −1.96
0.10	−1.96 to −1.65
-	−1.65 to 1.65
0.10	1.65 to 1.96
0.05	1.96 to 2.58
0.01	Greater than 2.58

Source: [22].

.

The spatial statistics derived for the metropolitan case study areas enabled the research to validate the consistency between the numerically measured spatial pattern versus the spatial pattern envisioned in the NSS and/or RFPP. This validation is important in pointing out the influence of the national and regional spatial development plan in the development directions of the case study areas. The validation was also able to emphasize the degree of spatial similarly among cities and municipalities in the case study areas. This was achieved by denoting the count of clustered cities and municipalities or similar REDi values with high-and-high or low-and-low similarities, based on the REDi bandwidth scale.

3.4.3. Formulation of Enhanced Approach in the Regional Spatial/Economic Development Strategies

The research reviewed possible refinements in the existing spatial development strategy approved by NEDA, based on the results of the spatial analysis and the validation of the spatial statistics. The examination looked into the locations where clustering exists, based on the validation in the REDi spatial statistics. Such locations are considered locus or points of spatial interactions that can reshape spatial development models (Integrated Area Development, Balanced Agro-Industrial Development, Growth Pole, Hub and Spoke), implemented in the past by the national and regional planning authorities.

4. Results and Discussion

4.1. Hierarchy of Cities and Municipalities within the Network of Settlements

All cities and municipalities (LGUs) in the identified metropolitan case study areas follow the network of settlements hierarchy in the NSS and physical framework plans (NFPP) approved for the metropolitan regions. The LGU classification in the network of settlements is based on the population movements over time and the strategic roles these settlements assume in support of regional development. On top of the hierarchy are the metropolitan centers, which collectively account for 44 LGUs across the case study areas (

Table 5. Count of cities and municipalities in the study areas, based on settlement hierarchy classification.

Urban Settlement Hierarchical Classification 1	Indicative Population Size 2 (In ‘000)	Greater Metro Manila	Metro Cebu	Metro Davao	Total
Metropolitan Centers	~4000	17	13	14	44
2. Regional Centers	>1200	3	0	0	3
3. Sub-regional Centers	<120–1200	15	4	1	20
4. Provincial Centers	50–120	9	5	13	27
5. Local Centers	<50	5	31	21	57
Total		49	53	49	151

¹ [18,25,26,27], ² [28].

).

The succeeding tiers in the settlement hierarchy have relatively smaller population sizes and limited market functions than the larger ones in servicing the requirements of the surrounding LGUs. The use of the Urban Service Center and Special Service Center was noted in the regional spatial framework plan for the Metropolitan Cebu study area. These specific classification names are different against the standard terms indicated in the NSS. In reconciling these differences, the population size was used in line with the NSS/NEDA criteria that classifies functional roles in the network of growth areas.

The general equivalent of an Urban Service Center is a Provincial Center (city and/or municipality), while a Special Service Center refers to a Local Center (municipality), located in smaller islands detached from the mainland province. For a metropolitan center, these are the LGUs (cities and municipalities) forming highly urbanized centers in the emerging urban areas. The Ekistics Logarithmic Scale (ELS) introduced by Doxiadis [28] for a Metropolis was adopted in Table 5 above, in view of the absence of a specific population size standards from the central planning agency (NEDA).

Case Study Area 1: Metropolitan Davao. The unit of analysis for Metro Davao covers the entire Davao administrative region or Region XII. Comprising the study area are the five (5) administrative provinces of Davao del Norte, Davao del Sur, Davao Oriental, Davao Occidental and Davao de Oro. All 49 cities and municipalities that belong to these provinces were grouped according to their assigned hierarchy shown in

Table 6. Hierarchy and Administrative Composition of Metro Davao.

Metropolitan Centers	Sub-Regional/ Provincial Centers	Local Centers
1. Carmen (DN) 2. Davao City (DS) 3. Digos City (DS) 4. Hagonoy (DS) 5. City of Samal (DN) 6. Maco (DDO)	15. Mati (DO) 1 16. Asuncion (DN) 17. Baganga (DOR) 18. Compostela (DDO) 19. Gov. Generoso (DOR) 20. Jose Abad Santos (DO)	29. Bansalan (DS) 30. Boston (DOR) 31. Braulio Dujali (DN) 32. Caraga (DOR) 33. Cateel (DOR) 34. Kiblawan (DS)	41. Matanao (DS) 42. Mawab (DDO) 43. Montevista (DDO) 44. Manay (DOR) 45. San Isidro (DN) 46. San Isidro (DO)
7. Malalag (DS)	21. Kapalong (DN)	35. Mabini (DDO)	47. Sarangani (DO)
8. Malita (DO)	22. Laak (DDO)	36. Magsaysay (DS)	48.Talaingod (DN)
9. Padada (DS) 10. Panabo City (DN)	23. Lupon (DOR) 24. Monkayo (DDO)	37. Maragusan (DDO) 38. Banaybanay (DOR)	49. Tarragona (DOR)
11. Sta. Cruz (DS) 12. Santa Maria (DS) 13. Sulop (DS) 14. Tagum City (DN)	25. Nabunturan (DDO) 26. New Corella (DN) 27. Pantukan (DDO) 28. Santo Tomas (DN)	39. Don Marcelino (DO) 40. New Bataan (DDO)
Total: 14	Total: 14	Total: 21

¹ Designated as sub-regional center in the 2015–2045 Davao Regional Physical Framework Plan: DN–Davao del Norte; DS–Davao del Sur; DO–Davao Occidental; DDO–Davao de Oro; DOR–Davao Oriental. Source of Basic Data: [18,25].

. From this count, 14 cities and municipalities were created by law to be part of Metro Davao [29].

Case Study Area 2: Metropolitan Cebu. The entire island province of Cebu, including Camotes Islands and Bantayan Island, was treated as the unit of analysis. Roads and bridges generally connect the mainland province, while the two (2) smaller islets are detached and are only accessible via sea/marine transport from the mainland. These island municipalities include Madrilejos, Bantayan, Santa Fe, San Francisco, Puro, Tudela and Pilar. The Metro Cebu area geospatially covers the 13 LGUs within the 25-km radial distance from Cebu City. Seven (7) are municipalities, and the rest are cities. The Regional Spatial Development Framework for Central Visayas (2016–2045) has expanded the existing metropolitan area to a greater metropolitan area, with the addition of 23 LGUs radiating beyond the existing 25-km band by stretching farther toward the 50-km distance [26]. A total of 53 LGUs were mapped together with their assigned functional hierarchy in the network of settlements and are listed in

Table 7. Hierarchy and Administrative Composition of Metro Cebu.

Metropolitan Centers	Sub-Regional/Urban Service Centers	Local Centers/Special Service Centers
1. Carcar City 2. Cebu City	14. Balamban 1 15. Bogo City 1	23. Alcantara 24. Alcoy	36. Ginatilan 37. Malabuyoc	49. Pilar 2 50. Poro 2
3. Compostela	16. San Remigio 1	25. Alegria	38. Medellin	51. San Francisco 2
4. Consolacion	17. Toledo City 1	26. Aloguinsan	39. Oslob	52. Santa Fe 2
5. Cordova	18. Argao	27. Asturias	40. Pinamungahan	53. Tudela 2
6. Danao City	19. Carmen	28. Badian	41. Ronda
7. Lapu-Lapu City	20. Moalboal	29. Barili	42. Samboan
8. Liloan 9. Mandaue City 10. Minglanilia 11. Naga City 12. San Fernando 13. Talisay City	21. Santander 22. Tabuelan	30. Boljoon 31. Borbon 32. Catmon 33. Daan Bantayan 34. Dalaguete 35. Dumanjug	43. Sibonga 44. Sogod 45. Tabogon 46. Tuburan 47. Bantayan 2 48. Madrilejos 2
Total: 13	Total: 9	Total: 31

¹ Designated urban service center and ² Designated special service center in the 2015–2045 regional physical framework plan. Source of Basic Data: [18,26].

for the whole study area.

Case Study Area 3: Greater Metropolitan Manila. The case study area extends past the present LGU composition of the National Capital Region (NCR), stretching up to a minimum 20-kilometer radial distance reckoned from “km 0” reference point located at the City of Manila. The 20-km minimum radial distance is the typical size of the urban extent of metropolitan districts in China and the United States [30]. Larger sizes of metropolitan districts are observed in literature in areas that have prior headway in terms of economic maturity. Locally, the 20-km distance coincides with the indicative acceptable distance for long travels in the NCR, using a public jeepney as a mode of transportation vehicle for inter-city commute [31]. Such distance bandwidth corroborates the spatial coverage envisioned for the NCR in the physical planning document prepared by the NEDA Luzon Regional Development Committee (RDCOM) for the year 2015 up to year 2045 [27].

Within this greater Metropolitan Manila case study area are 49 LGUs composed of the 17 cities and municipalities of Metro Manila (which is the National Capital Region); 10 LGUs each from the provinces of Bulacan, Cavite and Rizal, with San Pedro City and Biñan City completing the remaining two LGUs (

Table 8. Hierarchy and Administrative Composition of Greater Metro Manila.

Metropolitan Centers	Regional/Sub-Regional Centers	Provincial Centers	Local Centers
1. Caloocan City 2. Las Piñas City 3. Makati City 4. Malabon City 5. Mandaluyong City 6. Manila City 7. Marikina City	18. Antipolo City 1 19. Dasmariñas City 1 20. Malolos City 1 21. Bacoor City 22. Biñan City 23. Cainta City 24. General Trias City	36. Angono 37. Balagtas 38. Bocaue 39. Bulakan 40. Guiguinto 41. Kawit 42. Morong	45. Binangonan 46. Cardona 47. Cavite City 48. Noveleta 49. Teresa
8. Muntinlupa City 9. Navotas City	25. General Mariano Alvarez 26. Imus City	43. Obando 44. Rosario
10. Parañaque City 11. Pasay City 12. Pasig City 13. Pateros 14. Quezon City 15. San Juan City 16. Taguig City 17. Muntinlupa City	27. Marilao 28. Meycauyan City 29. Rodriguez 30. San Jose del Monte City 31. San Mateo 32. San Pedro City 33. Santa Maria 34. Tanza 35. Taytay
Total: 17	Total: 18	Total: 9	Total: 5

¹ Designated as regional center in the 2015–2045 Luzon Spatial Development Framework. Source of Basic Data: [18,27,32,33].

). These LGUs were included based on their administrative boundary contiguity as nearest/adjacent neighbors of NCR.

4.2. Spatial Association of Cities and Municipalities Based on the REDi

The result of grouping and bucketing the REDi dataset according to the range of values is shown in

Table 9. Top 10 Cities and Municipalities (LGUs) in the Metropolitan Case Study Areas, based on REDi.

REDi Range	Greater Metropolitan Manila		Metropolitan Cebu		Metropolitan Davao
	LGU Name	REDi Value	LGU Name	REDi Value	LGU Name	REDi Value
>14–16	-	-	-	-	Davao	14.65
>12–14	-	-	Mandaue	11.65	-	-
>10–12	Pasay	11.11	Cebu	10.16	Tagum	8.73
>8–10	Quezon	9.87	-	-	-	-
>6–8	Makati Pasig Manila Taguig	7.67 7.48 7.26 6.59	Lapu-Lapu Toledo	6.86 6.24	Panabo	6.75
>4–6	Mandaluyong Pateros Parañaque Muntinlupa	5.96 5.80 5.51 5.43	Argao Tuburan San Francisco Pinamungajan Tudela Talisay	5.64 5.44 5.43 5.22 5.10 5.00	Island Garden City of Samal Carmen Braulio Dujali Sta. Cruz New Corella Bansalan Kapalong	5.80 5.68 5.48 5.47 5.43 5.36 5.32

. This table features the top 10 LGUs, ranked following the established bucket in REDi range. The level of economic dynamism is highly concentrated in the Metropolitan Davao area, compared to Metropolitan Cebu and Greater Metropolitan Manila area. A stark gap in Davao City’s REDi values (14.65), as opposed the next rank Tagum City (8.73), and the third rank Panabo City (6.75), validates this primacy in economic dynamism.

Such dominance in the level of economic dynamism is typically characterized in the economic space by a significant gap in the variation and count of higher-order goods and services offered at the city center as opposed to its neighbors—a scenario that the research found to be prevailing in Davao City, in terms of financial deepening (from the raw data on the count of financial institutions) and the size of the local economy (from the raw data on capitalization of registered business establishments) posted for 2020. This disaggregation in the number and types of business registrations have implications in describing the sectors or economic activities driving spatial interactions and causing the clustering of LGUs across space.

The level of primacy observed in Table 9 for the Metropolitan Davao study area is different from the case of the greater Metropolitan Manila, based on REDi range comparison. The gaps in the REDi values are narrower between the first rank Pasay City (11.11) with the next rank Quezon City (9.87), or against the third rank Makati City (7.67), or the fourth rank Pasig City (7.48) and the fifth rank Manila City (7.26). These relative similarities in REDi values for contiguous LGUs (Makati, Pasig, Manila) generally indicate clustering. Another important observation that differentiates the clustering of LGUs in the Greater Metro Manila from the Metropolitan Davao and Metropolitan Cebu is the relatively smaller proportion (approximately 88%) of LGUs that belongs to the low-low bandwidth, in terms of the REDi range (

Table 10. Number of Cities and Municipalities (LGUs), according to REDi Range.

REDi Range	REDi Spatial Range	Greater Metropolitan Manila Study Area		Metropolitan Cebu Study Area		Metropolitan Davao Study Area
		LGU Count	% Count vs. Total	LGU Count	% Count vs. Total	LGU Count	% Count vs. Total
>14–16	High-High	0	0	0	0	1	2.04
>12–14		0	0	0	0	0	0
>10–12		1	2.04	2	3.77	0	0
>8–10		1	2.04	0	0	1	2.04
>6–8		4	8.16	2	3.77	1	2.04
>4–6	Low-Low	20	40.22	43	81.13	19	38.78
2–4		23	46.94	6	11.32	27	55.10
Total		49	100	53	100	49	100

).

For the Metropolitan Davao case study area, this low-low group accounts for approximately 94%, and 92% for the Metropolitan Cebu case study area. The cities and municipalities that are under this group have narrower differences in REDi values and, for the Greater Manila Area, this could possibly indicate broader spatial diffusion in economic dynamism. Stated in spatial terms, attribute similarities of nearest neighbors or contiguous administrative areas within a broader economic space indicate complementation, rather than competition. The spatial structure for such representation features softer edges in the zone of transition between the high-high area clusters and that of its periphery, as shown in the GIS map for Greater Metropolitan Manila (Figure 6):

One possible reason for the broader complementation (spatial diffusion) in economic dynamism among LGUs within the Greater Metropolitan Manila is the presence of multiple nodes of growth centers, represented by the spatial distribution of commercial business districts. These centers function as the hub and the nearby emerging and lower rank growth centers alongside the settlement hierarchy as spokes that illustrate spatial interaction. The case of clustering among cities and municipalities in Metropolitan Cebu case study area (Figure 7a) has similarities with that of Metropolitan Davao case study area (Figure 7b), with regard to having a dominant center.

The visual presentation has shown that complementation originates centrally from the contiguous area formed by Cebu City and Mandaue City, spreading linearly and following the shape of the mainland province. Having a linear shape can assist in maximizing spatial interaction by minimizing accessibility to reach the next nearest neighbor via the shortest linear path. While there is no means for the research using the CMCI data to reinforce this observation of achieving shortest access, the overlay of the trunk, primary and secondary road network shown in the GIS map could support such a claim.

In summary, clustering of cities and municipalities exists across the metropolitan case study areas, using their REDi values as an organizing spatial attribute. This implies that economic dynamism drives spatial interactions reflective of the spatial association in the REDi values. Such spatial association in economic dynamism can be visualized from the spatial contiguity or clustering of dark-colored administrative units and light-colored areas in the GIS map. These clustering of darker shaded contiguous administrative areas features REDi values in the “high-high” range. Similarly, the contiguity of lighter shaded administrative areas accounts for clustering among LGUs with “low-low” REDi values.

4.3. Nature of Economic Competitiveness in Major Metropolitan Areas

The clustering in the spatial distribution of the economic dynamism (REDi), visualized previously through GIS mapping, was numerically confirmed by the positive values in the Moran’s I for the metropolitan case study areas (

Table 11. Global Moran’s I of Economic Dynamism (REDi) Spatial Statistics.

Spatial Autocorrelation Parameters	Metropolitan Case Study Areas
	Greater Manila	Cebu	Davao
Spatial Association of REDi	Clustered	Clustered	Clustered
Moran’s I	0.497175	0.236037	0.381229
z-score	5.745155	2.844300	4.791787
p-value	0.000000	0.004451	0.000002
Distance Method	Euclidean	Euclidean	Euclidean
Conceptualization	Contiguity Edges	Contiguity Edges	Contiguity Edges
Standardization	Row	Row	Row

). A Moran’s I greater than zero (0) reiterates clustering and the occurrence of complementation. Had competition existed, Moran’s I would be negative (−1), and extreme differences in the REDi values would be seen as markedly separated by alternating intensity in the color shade of cities and municipalities in the GIS map, similar to a checkerboard pattern. A coefficient (index) value of 0 would indicate that the data is randomly distributed and would have an absence of spatial clustering [34].

This spatial clustering in the LGU economic dynamism was further validated by the spatial statistics of the REDi dataset with z-scores and p-values providing the justification in rejecting the null hypothesis. The normal curve is displayed together with global Moran’s I statistics of the datasets expressed in “p-values” and “z-scores”, both of which validated that the clustering pattern is 99% (unlikely to be from random observations) (Figure 8). Thus, findings accepted the research hypothesis that a spatial correlation (in this case, a positive relationship) exists in LGU economic dynamism.

4.4. Influence of Economic Dynamism of Cities and Municipalities in Metropolitan Areas

4.4.1. Influence on Regional Spatial Development Strategies

The clustering of economic dynamism revealed two important findings for planning metropolitan districts and the larger regional economic space. First, economic dynamism is a result of spatial interaction of nearest neighbor cities and municipalities that are spatially structured, similar to cluster corridors. This was the observation from the clustering of high-high REDi values for Davao City, Panabo City and Tagum City (Figure 9a), as well as for Lapu-Lapu City, Mandaue City, Cebu City and Toledo City (Figure 9b), and for Pasay City, Manila City, Makati City, Taguig City, Pasig City and Quezon City (Figure 9c).

Matching this contiguity in high-high REDi values with the regional spatial development strategy shows that these LGUs (cities and municipalities) belong to the top rank metropolitan center or regional center classification following the national spatial settlement hierarchy. Thus, metropolitanization is the likely route to pursue and spread regional development. This can be operationalized by the clustering of contiguous cities and municipalities into growth corridors to foster specialization and drive spatial interactions that will support economic dynamism. The study by Kekezi et al. on the knowledge creation at the local intra-sectoral and the local inter-sectoral spillovers supports this direction and emphasizes the importance of geographical proximity for knowledge production [35]. It will benefit budget or capital constrained economies by taking advantage of the rewards in economies of scale and efficiency from building these clusters.

The second important finding is related to the relatively smaller proportion of administrative areas with low-low REDi values (below 6) that are contiguous, versus the high-high cluster group (REDi of 6 and above). As shown in Table 10, the high-high group accounts for approximately 10% against the 90% count for the low-low REDi group. These proportions pre-suppose a larger area with lagging economic conditions. Planning authorities could then formulate regional economic development plans by encouraging investments that will induce spatial interactions between the LGUs bearing high-high REDi values with LGU settlements (cities and municipalities) that are lagging.

A local spatial autocorrelation test has to be carried out separately to identify specific clusters among the low-low REDi group of cities and municipalities for targeted functional interventions. The research confirmed through the global Moran’s I spatial statistics that the occurrence of clustering in the REDi dataset is indicative of contiguity of cities and municipalities with high-high and low-low values. The local spatial autocorrelation statistic could identify local clusters (refer numerically to either high-high and low-low clusters) or local outliers (areas not part or numerically do not belong to either group) that can establish the contribution of these clusters to the global clustering statistic. Specifically, prioritization and catalyzing smaller clusters can be targeted for economies of scale and efficiency in the delivery of services/interventions.

These numerical confirmations in the clustering of REDi values are valid justifications for decision-makers to consider Integrated Area Development (IAD) Planning, as a spatial and economic development strategy, over other conceptualizations.

4.4.2. Influence on Regional Economic Development Policy

The key spatial points in the preceding section can be summarized into the following operating elements of competitiveness: (a) complementation across administrative space; (b) metropolitanization of economic areas; and (c) clustering of firms.

Competitiveness Operating Element 1: Complementation across administrative space. This is derived from the spatial association in the REDi values. Complementation has a stimulative effect, particularly on the lagging cities and municipalities, and facilitates the rethinking of their spatial interactions by LGU officials and planning authorities. This is beneficial in managing fiscal space through participatory convergence programs.

Competitiveness Operating Element 2: Metropolitanization of economic areas. This spatial point was derived from the GIS map spatial overlay of the REDi values alongside the national urban settlement hierarchy. The layout noted the presence of growth corridors that were generally an amalgamation of cities categorized as metropolitan or regional centers with dominant (high-high) REDi values. Similar spatial observation on higher development near cities surfaced from the spatial autocorrelation study of Jaramillo and Lotero on socio-economic indices and their distance to Columbian capitals [36]. These economic areas functionally serve as growth anchors, due to their fundamental role as consuming centers and their contribution to continuous specialization (innovation). The confirmation in the clustering of administrative areas with spatially similar REDi values has implications in local planning and in the approvals by the higher authorities when creating new metropolitan districts or in rethinking the expansion of existing ones.

Competitiveness Operating Element 3: Clustering of firms. This can be viewed as the clustering of economic agents or firms within the LGU administrative space. This is based on the confirmation of clustering in economic dynamism through Moran’s I spatial statistics of the REDi dataset. Industry-level clustering may be approached beyond establishing the market center and lagging center’s forward and backward linkages but, more importantly, contemplating the delivery/installation of shared services by the government in areas where the scale for private sector entry is not attractive or absent. In this way, the government can anticipate the tapering off in development as economic dependence from cities decreases, usually manifested by the diminishing strength of the spatial correlation [36].

These key operating elements are the mechanisms that fit into the gap between competitiveness policy and regional economic development policy targets. Once these key research findings are set into motion, they could reinforce the attainment of fiscal and developmental targets, and consequently reduce the budget deficit.

5. Conclusions

This study established at the beginning that spatial correlation exists in the economic dynamism indices of neighboring cities and municipalities (also referred to as LGUs) within a metropolitan area. This argument arose from the lack of spatial lens among academic inquiries involving factors driving competitiveness and its contribution to regional development planning in the Philippines. It was through a review of literature on the ranking of economic dynamism on top of the other competitiveness factors that guided the research to focus on economic dynamism. The study reprocessed the economic dynamism index (REDi) data from the CMCI of the DTI. It builds upon Tobler’s first law of geography that everything is related to everything else, but closer things are more related than distant ones. This conceptualization introduced a different way of analyzing regional development by treating the REDi as a spatial attribute, a feature, or a characteristic found in each LGU.

The Moran’s I global spatial autocorrelation that yielded significant clustering in the REDi dataset among LGUs in major metropolitan regions reinforces that the prevailing nature of competitiveness occurring in space denotes complementation among cities and municipalities rather than competition. Had competition existed, the spatial distribution in the REDi would have been dispersed, indicative of a checkerboard pattern, when viewed in the GIS map. The visual translation that resulted to clustering implied the concentration of “high-high” REDi values around metropolitan centers. Similarly, the clustering in the “low-low” REDi range of values were noted in areas (LGUs) that are not nearest administrative neighbor of the metropolitan center. These empirical observations support regional planning theories on uneven development and growth poles and offered important implications in rethinking the spatial planning of regions.

In reformulating spatial strategies, the planning authorities and professionals may encourage metropolitanization in view of its general notion of fostering complementation. It would also benefit capital-constrained developing economies like the Philippines by concentrating developmental interventions in existing urban centers to achieve efficiency and economies of scale. This will also help stimulate policy viewpoints at the local, regional and central planning level towards planning considerations between metropolitan centers and the areas that are considered lagging (LGUs with “low-low” REDi values).

Planning initiatives on enhancing complementation and spread of growth can be re-examined through the geospatial interconnection of companies, service providers, government and institutions that are within, from and across the metropolitan region.

By making efficient these interconnections or linkages between dynamic economic clusters (LGUs with high-high REDi values) found in metropolitan areas and the areas that are lagging, cities and municipalities that are less endowed or with limited resources would eventually benefit from the economic spread effects.

While these results contribute to the knowledge of sustainability of regional metropolitan centers through spatial interactions, this study has limitations in terms of statistically identifying and ranking priority areas for complementation. Thus, the extension of the analysis in our future research agenda is to apply local spatial autocorrelation techniques for each metropolitan case study area. In this way, spatial clusters similar to “hotspots” and “coldspots” can be identified, which is a natural limitation of the Moran’s I global spatial autocorrelation technique used in this study. The contribution of the other competitiveness pillars (i.e., infrastructure, resiliency, innovation and government efficiency), comprising the CMCI of the DTI, may also be explored in the future to reinforce the spatial dimension surfaced in this research.