Economics of Cities – Summary
Chapter 2: Agglomeration and clustering
2.1 Introduction
- Most industrial activities tend to be clustered together in space -> why?
- Certain activities take place in certain clusters -> some are more dispersed (more
transportation), others more concentrated.
-> Generally one largest city cluster with almost all types of activities -> larger number of
smaller clusters.
-> Urban hierarchy.
- This chapter: why cluster activities together in space?
2.2 Industrial clustering: returns to scale and geography
- Constant returns to scale: Clustering will create demand for land and labour -> land rents
and wages go up -> profitability of firms will be reduced, so no incentive for firms to cluster.
- External economies of scale: Clustering still leads to higher land rents and higher wages,
could be offset by increase in efficiency for each firm.
-> Clustering increases profitability -> other firms will join too: cumulative process of local
growth.
-> Are limits though as it is not continuously cumulative: not all activities are at one
location.
2.3 Agglomeration economies
-> Location-specific economies of scale.
- Sources of agglomeration economies (Marshall):
- Knowledge spillovers: Tacit information (incomplete knowledge shared on a non-market
basis) flows easily in clusters through informal contacts.
-> All firms able to get a clear picture of market environment -> more competitive in the
market.
-> Information advantage for firms within the cluster over outsiders.
- Non-traded local inputs: If many firms in the same industry group together, they form a
large demand for specialized inputs to be provided.
- Local skilled labour pool: Easier for the firm to find new labour when market demand
conditions improve + little costs to retraining activities (they are skilled).
-> Duranton and Puga: Processes of learning, sharing and matching: together the three
sources allow firms to experience economies of scale external to a single firm, internal to the
group of firms together.
-> Agglomeration economies for firms clustered within the same industry.
- Types of agglomeration economies (Hoover):
- Internal returns to scale: Economies of scale due to size of one firm: large factory requires
a lot of capital and a large labour force at the location.
-> Production factors clustered together within one firm -> internal economies of scale are
,location specific.
-> Firm-specific economies of agglomeration.
- Economies of localization: Advantages due to a group of firms within the same industrial
sector located at the same place (Marshall’s sources of agglomeration).
-> Industry-specific economies of agglomeration.
- Economies of urbanization: Advantages of firms of different sectors clustering together:
activities not related to dominant sector providing services for firms and employees of that
sector (education, health care, leisure, legal).
-> City-specific economies of agglomeration.
2.4 Clusters, firm types, and the nature of transactions
- Hard to empirically distinguish between different types and sources of agglomeration.
- Three different typologies of spatial industrial clusters:
- Pure agglomeration: Firms are atomistic (perfect competition) -> intense local competition
(no loyalty between firms).
-> Everyone is able to enter the cluster: benefit is local presence, cost local real-estate
rents -> the last one also an indicator of the cluster’s performance.
-> Marshall model, also localization and urbanization economies.
- Industrial complex: Long-term and stable relations between firms in the cluster.
-> Input-output linkages, high entry and exit costs -> proximity is required to reduce inter-
firm transport costs.
-> Internal returns to scale.
- Social network: Common culture of mutual trust -> no opportunistic behaviour.
-> Spatial proximity fosters trust relations: local business environment of confidence, risk-
taking and cooperation.
-> These types differ quite a lot -> each cluster will have a dominant typology and helps
determining the characteristics and behaviour of any particular cluster.
-> Does not show why firms wish to cluster in space due to potential unintended knowledge
outflows (privately developed ideas or techniques through R&D) -> is controlled for by
contracts (industrial complex) and shared values (social network), but not in a pure
agglomeration -> discourages investments in R&D.
-> Decision of clustering issue of comparing knowledge advantages and disadvantages.
-> Clusters are related to innovation and economic growth, ways to observe these relations
empirically remain very unclear.
2.5 People clustering: creativity and urban consumption
- Other advantages of urban clustering:
- Creative class (Florida): Culture of tolerance will lead to an environment of creativity
(fostering innovation in multiple areas) and commercial dynamism.
-> Creative personalities will move to tolerant places -> increasing innovation in these
places.
- Consumer city: Modern city a place of work and leisure.
-> High-income consumers benefit of high-quality creative services and amenities provided
by creative people.
,-> Clustering not only about firms, also related to interactions of people.
2.6 Limited information, uncertainty, and the evolution of clusters
- In environments of uncertainty and limited information the process of agglomerations
needs to be reconsidered: three themes:
- Bounded rationality: Firms due to limited information, limited in their ability to be
rational.
-> Firms location decision bounded by limited information on locations -> unable to
choose profit maximizing location -> needs to focus on other objectives.
- Conflicting goals: Firms can choose to optimize short run profits, sales, minimize costs.
-> All best in different location -> due to limited behaviour, firms focus on different
objectives and might miss the best possible location.
- Relocation costs: Due to high relocation costs, firms will not move due to small changes in
factor prices or market revenues.
-> Relocation is a last resort strategy if other competitive weapons did not work.
-> Behavioural theory explains why observed spatial patterns do not reflect optimum
location behaviour, but behavioural theory does not indicate why a firm chooses a particular
location.
-> Necessary to reconsider agglomeration theories: no.
- Alchian (evolutionary approach):
- Adoptive environment: All firms are identical, no information advantages or a priori
knowledge of whether their strategy is successful.
-> In the end the firms best suited to the needs of the economy will survive.
- Adaptive environment: Large firms able to analyse the market and acquire information,
small firms are not, but they copy behaviour of larger firms.
-> Large firms overcome uncertainty by analysis, small firms by imitation.
-> Small firms (most cases spin-offs) will locate near large firms -> explains why localized
clusters in related technological fields arise.
-> Moving close to large firms good strategy for risk averse firms: market share is not
lower than equivalent firm + well defended against price movements.
2.7 Conclusions
- Generation, acquisition and transfer of knowledge key component of industrial clustering.
- Spatial economy characterized by geographical concentration and dispersion, the balance
produces a system of cities: an urban hierarchy.
Chapter 3: The spatial distribution of activities
3.1 Introduction to industrial dispersal
- Not all activities are located together, some with large land use for instance are dispersed.
-> Industrial dispersal creates a spatial pattern or system of cities: urban hierarchy.
3.2 Firm dispersion: price discrimination, market areas, and distance costs
- In absence of agglomeration mechanisms, different firms with different output for different
, markets will tend to be dispersed.
- Firms might use monopoly power over local consumers (less transport costs) to charge
them more than distant consumers: price discrimination.
-> Increases spatial market of the firm + increases the tendency towards firm dispersion.
- Market area: Reilly: Ability of market to attract customers directly related to size of market
and indirectly to distance to the market.
-> Large market increases attractiveness of market, distance increases costs of overcoming
that distance.
-> Greater urbanization economies associated with larger cities, implies larger hinterland;
more local market areas are dominated by purchases of lower variety of goods.
- High-value products tend to have a high value-weight ratio, meaning that transport costs
will remain small portion of the price; low-value products tend to have low value-weight
ratio, meaning transport costs quickly become a significant portion of price.
-> High-value products tend to be in larger urban centres.
-> The higher the value-weight ratio, the greater average distance of shipment.
- Problems: Method is useless for the weightless service economy, some high-value
products also produced outside major centres + major centres also production location for
low-value products.
-> This paragraph contrasts with chapter 2: countervailing forces of agglomeration and
dispersal create a spatial pattern of cities and regions.
3.3 Urban hierarchies and classical central place theory
- Observations: Nations dominated by one or two primal cities, followed by smaller cities, as
city size falls, the number of those cities increases.
-> Spatial distribution of urban centres follow a hierarchical pyramidal pattern.
- Christaller approach to central places: (inductive)
- There is a hierarchy of different goods, market areas and urban centres.
- Higher order goods, need larger spatial market areas.
- Outcome: Spatial pattern with hexagonal market areas, which ensures that all locations
are supplied with all goods from a minimum amount of supply points.
- Lösch approach to central places: (deductive)
- Variety of firms, producing a variety of goods.
- Aim to find ideal economic landscape with the most efficient spatial allocation of activity
(arises naturally in a competitive economy).
- Quantity demanded decreases with distance at some point price + transport costs is so
high, that quantity demanded is zero.
-> Market area of firm is a circle.
- Competition: all land is occupied by identical firms: spatial economy will follow a pattern
of hexagons, firms locate at centre of their hexagon.
-> Ensures that the average price + transport costs is minimized as there is a maximum
number of competing suppliers in the economy (the opposite of Christaller).
- High-value goods price inelastic -> larger market area as demand does fall relatively slowly
as distance (transport costs) increase, low-value goods have smaller market areas.
-> Firms producing price inelastic goods located at fewer points (larger hexagonal areas),
Chapter 2: Agglomeration and clustering
2.1 Introduction
- Most industrial activities tend to be clustered together in space -> why?
- Certain activities take place in certain clusters -> some are more dispersed (more
transportation), others more concentrated.
-> Generally one largest city cluster with almost all types of activities -> larger number of
smaller clusters.
-> Urban hierarchy.
- This chapter: why cluster activities together in space?
2.2 Industrial clustering: returns to scale and geography
- Constant returns to scale: Clustering will create demand for land and labour -> land rents
and wages go up -> profitability of firms will be reduced, so no incentive for firms to cluster.
- External economies of scale: Clustering still leads to higher land rents and higher wages,
could be offset by increase in efficiency for each firm.
-> Clustering increases profitability -> other firms will join too: cumulative process of local
growth.
-> Are limits though as it is not continuously cumulative: not all activities are at one
location.
2.3 Agglomeration economies
-> Location-specific economies of scale.
- Sources of agglomeration economies (Marshall):
- Knowledge spillovers: Tacit information (incomplete knowledge shared on a non-market
basis) flows easily in clusters through informal contacts.
-> All firms able to get a clear picture of market environment -> more competitive in the
market.
-> Information advantage for firms within the cluster over outsiders.
- Non-traded local inputs: If many firms in the same industry group together, they form a
large demand for specialized inputs to be provided.
- Local skilled labour pool: Easier for the firm to find new labour when market demand
conditions improve + little costs to retraining activities (they are skilled).
-> Duranton and Puga: Processes of learning, sharing and matching: together the three
sources allow firms to experience economies of scale external to a single firm, internal to the
group of firms together.
-> Agglomeration economies for firms clustered within the same industry.
- Types of agglomeration economies (Hoover):
- Internal returns to scale: Economies of scale due to size of one firm: large factory requires
a lot of capital and a large labour force at the location.
-> Production factors clustered together within one firm -> internal economies of scale are
,location specific.
-> Firm-specific economies of agglomeration.
- Economies of localization: Advantages due to a group of firms within the same industrial
sector located at the same place (Marshall’s sources of agglomeration).
-> Industry-specific economies of agglomeration.
- Economies of urbanization: Advantages of firms of different sectors clustering together:
activities not related to dominant sector providing services for firms and employees of that
sector (education, health care, leisure, legal).
-> City-specific economies of agglomeration.
2.4 Clusters, firm types, and the nature of transactions
- Hard to empirically distinguish between different types and sources of agglomeration.
- Three different typologies of spatial industrial clusters:
- Pure agglomeration: Firms are atomistic (perfect competition) -> intense local competition
(no loyalty between firms).
-> Everyone is able to enter the cluster: benefit is local presence, cost local real-estate
rents -> the last one also an indicator of the cluster’s performance.
-> Marshall model, also localization and urbanization economies.
- Industrial complex: Long-term and stable relations between firms in the cluster.
-> Input-output linkages, high entry and exit costs -> proximity is required to reduce inter-
firm transport costs.
-> Internal returns to scale.
- Social network: Common culture of mutual trust -> no opportunistic behaviour.
-> Spatial proximity fosters trust relations: local business environment of confidence, risk-
taking and cooperation.
-> These types differ quite a lot -> each cluster will have a dominant typology and helps
determining the characteristics and behaviour of any particular cluster.
-> Does not show why firms wish to cluster in space due to potential unintended knowledge
outflows (privately developed ideas or techniques through R&D) -> is controlled for by
contracts (industrial complex) and shared values (social network), but not in a pure
agglomeration -> discourages investments in R&D.
-> Decision of clustering issue of comparing knowledge advantages and disadvantages.
-> Clusters are related to innovation and economic growth, ways to observe these relations
empirically remain very unclear.
2.5 People clustering: creativity and urban consumption
- Other advantages of urban clustering:
- Creative class (Florida): Culture of tolerance will lead to an environment of creativity
(fostering innovation in multiple areas) and commercial dynamism.
-> Creative personalities will move to tolerant places -> increasing innovation in these
places.
- Consumer city: Modern city a place of work and leisure.
-> High-income consumers benefit of high-quality creative services and amenities provided
by creative people.
,-> Clustering not only about firms, also related to interactions of people.
2.6 Limited information, uncertainty, and the evolution of clusters
- In environments of uncertainty and limited information the process of agglomerations
needs to be reconsidered: three themes:
- Bounded rationality: Firms due to limited information, limited in their ability to be
rational.
-> Firms location decision bounded by limited information on locations -> unable to
choose profit maximizing location -> needs to focus on other objectives.
- Conflicting goals: Firms can choose to optimize short run profits, sales, minimize costs.
-> All best in different location -> due to limited behaviour, firms focus on different
objectives and might miss the best possible location.
- Relocation costs: Due to high relocation costs, firms will not move due to small changes in
factor prices or market revenues.
-> Relocation is a last resort strategy if other competitive weapons did not work.
-> Behavioural theory explains why observed spatial patterns do not reflect optimum
location behaviour, but behavioural theory does not indicate why a firm chooses a particular
location.
-> Necessary to reconsider agglomeration theories: no.
- Alchian (evolutionary approach):
- Adoptive environment: All firms are identical, no information advantages or a priori
knowledge of whether their strategy is successful.
-> In the end the firms best suited to the needs of the economy will survive.
- Adaptive environment: Large firms able to analyse the market and acquire information,
small firms are not, but they copy behaviour of larger firms.
-> Large firms overcome uncertainty by analysis, small firms by imitation.
-> Small firms (most cases spin-offs) will locate near large firms -> explains why localized
clusters in related technological fields arise.
-> Moving close to large firms good strategy for risk averse firms: market share is not
lower than equivalent firm + well defended against price movements.
2.7 Conclusions
- Generation, acquisition and transfer of knowledge key component of industrial clustering.
- Spatial economy characterized by geographical concentration and dispersion, the balance
produces a system of cities: an urban hierarchy.
Chapter 3: The spatial distribution of activities
3.1 Introduction to industrial dispersal
- Not all activities are located together, some with large land use for instance are dispersed.
-> Industrial dispersal creates a spatial pattern or system of cities: urban hierarchy.
3.2 Firm dispersion: price discrimination, market areas, and distance costs
- In absence of agglomeration mechanisms, different firms with different output for different
, markets will tend to be dispersed.
- Firms might use monopoly power over local consumers (less transport costs) to charge
them more than distant consumers: price discrimination.
-> Increases spatial market of the firm + increases the tendency towards firm dispersion.
- Market area: Reilly: Ability of market to attract customers directly related to size of market
and indirectly to distance to the market.
-> Large market increases attractiveness of market, distance increases costs of overcoming
that distance.
-> Greater urbanization economies associated with larger cities, implies larger hinterland;
more local market areas are dominated by purchases of lower variety of goods.
- High-value products tend to have a high value-weight ratio, meaning that transport costs
will remain small portion of the price; low-value products tend to have low value-weight
ratio, meaning transport costs quickly become a significant portion of price.
-> High-value products tend to be in larger urban centres.
-> The higher the value-weight ratio, the greater average distance of shipment.
- Problems: Method is useless for the weightless service economy, some high-value
products also produced outside major centres + major centres also production location for
low-value products.
-> This paragraph contrasts with chapter 2: countervailing forces of agglomeration and
dispersal create a spatial pattern of cities and regions.
3.3 Urban hierarchies and classical central place theory
- Observations: Nations dominated by one or two primal cities, followed by smaller cities, as
city size falls, the number of those cities increases.
-> Spatial distribution of urban centres follow a hierarchical pyramidal pattern.
- Christaller approach to central places: (inductive)
- There is a hierarchy of different goods, market areas and urban centres.
- Higher order goods, need larger spatial market areas.
- Outcome: Spatial pattern with hexagonal market areas, which ensures that all locations
are supplied with all goods from a minimum amount of supply points.
- Lösch approach to central places: (deductive)
- Variety of firms, producing a variety of goods.
- Aim to find ideal economic landscape with the most efficient spatial allocation of activity
(arises naturally in a competitive economy).
- Quantity demanded decreases with distance at some point price + transport costs is so
high, that quantity demanded is zero.
-> Market area of firm is a circle.
- Competition: all land is occupied by identical firms: spatial economy will follow a pattern
of hexagons, firms locate at centre of their hexagon.
-> Ensures that the average price + transport costs is minimized as there is a maximum
number of competing suppliers in the economy (the opposite of Christaller).
- High-value goods price inelastic -> larger market area as demand does fall relatively slowly
as distance (transport costs) increase, low-value goods have smaller market areas.
-> Firms producing price inelastic goods located at fewer points (larger hexagonal areas),
