Agglomeration economies

Question 1

Basic theory (economics A level)

Answer 1

Economies of Scale (external)

Question 2

What is the efficient size of a city a trade off between

Answer 2

Agglomeration economies and Urban crowding.

Centrifugal vs Centripetal forces

Question 3

Key theoretical question on what causes cities to be more productive

Answer 3

Darwinian selection vs agglomeration benefits

• more workers leads to better workers or SML
• more firms (survival of the fittest) or actual location benefit

Question 4

SML

Answer 4

Marshalling Externalities - benefits within a single industry of agglomeration
- Sharing
- Matching
- Learning

Duranton and Puga (2004)

Question 5

Jacobian Externalities

Answer 5

Positive impacts on innovation and economic growth that arises from the diversity of industries and knowledge within a region or city

Question 6

Sharing

Answer 6

The gains from specialisation (Adam Smith,1776) - greater scale allows firms to specialise more
… The labour pool - A local industry offers a constant market for skill (marshall, 1890) thus attracting a greater constant labour pool
… Risks - positive and negative shocks in the local labour market or economy

Question 7

Learning

Answer 7

schooling , training and research
Interactions with others due to proximity
Favour the diffusion of innovations and ideas (Marshall, 1890)
Everyday interactions that lead to knowledge creation, diffusion and accumulation (Lucas, 1988)
Endogenous growth theory (Romer, 1986) : accumulation of human capital, leads to tech progress, tech progress and workers do not move as predicted by neoclassical economists - LOCAL not NATIONAL spillovers

Question 8

Matching

Answer 8

Increase in the number of agents trying to match improves the quality of match (Helsley and Strange, 1990)
The crowding costs associated with cities act as a signalling device (Spence, 1973) - thus only workers whose value is high enough to the city will move in because they can afford to pay rent/cost of living

Question 9

Empirical evidence of SML

Answer 9

Rosenthal and Strange, 2004

• input sharing
• Labour market pooling
• Knowledge spillovers (Patents)

Question 10

Diverse vs specialised cities

Answer 10

(Duranton and Puga, 2001)
Diversified and specialised cities coexist, because each firm finds it in its best interest to locate in a diversified city while searching for its ideal process, and later to relocate to a specialised city where all firms are using the same type of process.

Question 11

Empirical evidence of greater productivity (USA *2)

Answer 11

Founding Paper - USA
Sveikauskas (1975), as it was the first influential modern study to use data on outputs and inputs to estimate how productivity varies across space, and found that a doubling of population leads to an increase of output of around 6%.

Built upon - USA
This data was built upon by Rosenthal and Strange (2004), who concluded that doubling city size leads to between a 3-8% increase in productivity. Data focused on America.

Question 12

Empirical evidence of greater productivity (EU)

Answer 12

Modern Support in Europe
Combes et al., (2009) is another example of a study that used empirical evidence, looking at France, and found that the log productivity in large French cities is well described by taking the productivity distribution in small cities and shifting it to the right, as well as dilating it. This led to the conclusion that firms in large cities are around 9% more productive than those in smaller cities.

Question 13

Urban Wage premium (evidence of greater productivity)

Answer 13

Wages and rents are two other commonly used variables, with Puga stating that several studies use wages as evidence of agglomeration economies, as in competitive markets, labour is paid at the value of its marginal product, and thus higher wages mean that there is higher productivity. There are some issues with this, as more able workers may go to cities, and thus higher wages are due to higher skills rather than the urban area. However, Glaeser and Maré (2001) control for this and still find a wage premium, albeit slightly smaller.

Rents are also used, as if firms are willing to pay higher rents in big cities, it must be due to some comparative advantage. However, it can sometimes be hard to find data on commercial rents, so residential rents are often used. Studies by Rosen (1979) and Roback (1982) combine wages and rents to show workers being more productive, with greater consumption amenities compensating higher rents for workers and higher wages for firms, and thus higher rents and wages mean more agglomeration.

Question 14

Survival of the fittest

Answer 14

Productive advantage of cities is due to it forcing only the most productive firms to survive
Melitz and Ottaviano (2008) : More firms = more competition = more exits for less productive firms

Question 15

Empirical support for Agglomeration benefits when controlling for survival of the fittest

Answer 15

Combes et al., 2009 - studies rightward shift of firm productivity vs left-sided truncation
No evidence for left truncation - yet firms in large cities are on average 9% more productive
Large indivisibilities is part justification which is down to economies of scale and sharing - cannot afford in smaller region (Airport, advanced medtech, Luxury brands)

Question 16

Urban Wage premium evidence of greater productivity

Answer 16

in comp markets, labour is paid the value of its MP

Even if not exactly true, higher wages in dense urban areas can be seen as evidence of higher productivity.

Firms would relocate to a cheaper location unless there were some significant productive advantages

Question 17

Worker sorting? Does UWP still stand when controlling for worker fixed-effects

Answer 17

Workers sort themselves into larger cities. Thus, the workers in denser areas are more competitive and more productive.

Glaeser and Mare (2001) control this variable through observable skills, parental background and individual worker fixed-effects and find that even with these corrections there is a significant wage premium associated with dense cities: although smaller than before. Combes et al. (2010) find that the elasticity of the wage premium associated with each city with respect to urban density is cut by about 50% when these effects are introduced.
Another concern when interpreting the urban wage premium seen in empirical studies is that workers may sort themselves into larger cities rather than any intrinsic advantage to urban location.

Question 18

Rents in denser areas? what do they point to..

Answer 18

Firms pay more in big cities there must be some productive advantage.

Commercial rents are harder to find but residential rents provide a close proxy for study (Dekle and Eaton, 1999)

Question 19

Spatial Impossibility Theorem

Answer 19

“In an economy with a finite number of locations, with locally nonsatiated preferences, and a finite number of consumers and firms, in which space is homogenous and transport is costly, no competitive equilibrium exists in which actual transport takes place.” - Starrett 1978
- Whenever economic activities are perfectly divisible, factor mobility is a substitute for trade

Question 20

USE SIT And disprove assumptions

Answer 20

Implication - to understand the spatial concentrations we observe, need to account for

• Heterogenous space
• Existence of externalities
• Marshalian - sharing, matching, learning
• Jacobian - benefits from diversity
• Imperfectly competitive markets

Tradeoffs between transport costs and locally increasing returns are key, and we need to be able to explain sharply decreasing transport costs

Question 21

Agglomeration economies are

Answer 21

the economic benefits that result from the concentration of businesses and activities in a specific geographic area (Glaeser, 2010)

Question 22

Is urban industrial diversity good

Answer 22

Glaeser et al (1992) use a Herfindahl–Hirschman index (HHI) of employment diversity at the MSA (metropolitan statistical area) level to find that diversity encourages growth

Question 23

At what distance do agglomeration economies attenuate

Answer 23

Principle - agglomeration economies may attenuate rapidly across geographic space.

Indirect evidence from
Henderson (2003) - employment activity in a plant’s own county affects plant productivity, but employment activity in neighbouring counties is not found to affect the plant’s productivity

Question 24

Different agglomerative forces operate at different spatial scales

Answer 24

Rosenthal and Strange (2001) – further empirical evidence that agglomeration economies attenuate across geographic space

Some factors drive agglomeration at the state level
- inputs to production sensitive to quick transport (Airport links etc)
- Manufacturing will locate not necessarily next to other firms but near an airport

Some factors drive agglomeration at the local level
- Knowledge spillovers
- Finance in close proximity

Some at all levels - Labour - can take a job within a wider than local commute

Question 25

PATENT CITATIONS: EVIDENCE OF LOCAL KNOWLEDGE SPILLOVERS

Answer 25

Jaffe, Trajtenberg and Henderson (1993) - identify a “paper trail” of knowledge spillovers in the location of patent citations Patent citations are highly spatially concentrated, with citations 5 to 10 times as likely to come from the same SMSA as control patents

Question 26

Consumption possibilities of large cities - School of thought

Answer 26

The charm, aesthetic, diversity of goods and services and ease of social interaction could explain the urban rent premium. Best workers willing to pay more to be in cities, thus firms pay more to be in cities without a productive advantage .

Question 27

Evidence of Consumption possibilities school

Answer 27

Glaeser, Kolko and Saiz (2001) Rise of reverse commuting - grew by 2.79% between 1980 and 1990, and at a roughly similar pace in the two previous decades Suggests that some individuals working at suburban sites are willing to incur higher central city house prices (quality adjusted) for the opportunity to enjoy central city consumption amenities

Question 28

Diverse cities: knowledge spillovers argument

Answer 28

This is led by Jacobs’ (1969) argument that economic interactions across sectors leads to knowledge spillovers which refine firms’ production processes. Diversity is more relevant for high-tech and knowledge-intensive industries that may draw on cross- fertlisations between industries to boost their innovation activities and ideas (Duranton and Puga, 2001

Question 29

Young Firms location patterns

Answer 29

young firms locate in more diverse urban environments because they benefit from a period of experimentation with regards to workers, components and so on. A recent trend has seen young start-ups locate in clusters of larger cities (Foord, 2013). Guzman and Stern (2016) shows that high-quality start-ups in the US have begun to favour being within big cities (San Francisco and Boston) than their specialised tech bubbles just outside (Silicon Valley and Route 128).

Question 30

Specialised cities

Answer 30

Although this is even more significant in low-tech manufacturing (where there may be less to gain from cross-sector knowledge spillovers) (Groot et al., 2016),Arzaghi et al., 2008 argues that specialisation can also be beneficial in knowledge-intensive industries. Once a firm has found its optimal process, it benefits from the lower costs of production in a specialised cluster where external economies of scale are large

Question 31

do firms move from diverse to specialised

Answer 31

Duranton and Puga (2001) find evidence from France to support this theory, finding that among firms that change location 72% move from a diverse to a specialised city, and that such moves are more frequent among mature firms and firms in innovative industries.

Question 32

Venture capital advantage of clusters

Answer 32

Zhang (2005) finds statistically significant differences in rounds of VC per startup and start-up age at first round of VC for Silicon Valley compared to 6 other regions including the rest of SF bay. At the 5% one-tailed t-test, Silicon Valley’s access to venture capital is significantly higher than startups in the wider SF region (Zhang, 2006) Stuart and Sorenson (2005) show that mutual trust and personal connections are very important because they help overcome the asymmetric information problem and accelerate the investment process

Question 33

Impact of doubling a city size on productivity

Answer 33

3-8% (Melo et al, 2009)

Question 34

Attenuation of agglomeration forces and spillovers - raising incomes

Answer 34

Rice et al (2006) find that the effect of proximity to economic mass in raising average incomes ceases to be important beyond approximately 80 minutes if driving time