Transcript Titolo slide

Dottorato cultura e impresa
Classe di studio 09-07-2004
On Networks and Markets by Rauch and
Casella, eds.
Ezra W. Zuckerman
Marco Lamieri
[email protected]
Agenda
• Paper summary;
• Review of the various contributions;
• Conclusion and comments.
“People of the same trade seldom
meet together, even for merriment
and diversion, but the conversation ends
in a conspiracy against the public, or in
some contrivance to raise prices.”
Smith - Wealth of Nations
Paper summary
• This paper illustrates some of the most promising approaches
to the study of economic networks.
• An economy can be depicted as a network or graph that links
economic actors with one another in a flow of exchange. This
definition it would seemingly be irrelevant by neoclassical
theory.
• A social network is defined as a set of nodes and the pattern
of ties among such nodes, an economic network must be
considered a subtype of a social network.
• A network is economic if it has effects on future events that are
considered economic by the analyst.
Elements of economic networks
• Nodes
the nodes may be people or organizations. The node is often
achieved through an exercise in aggregation whereby all
nodes that relate to a common agent are subject to common
effects or have common causes. Alternatively, the nodes may
proxy for unobserved agents.
Types of nodes:
– human beings
– human collectivities that are assumed to have significant capacity
for coordinated action.
• Ties
the ties are links between agents. The link must be made
explicit and justified for analyses of network effects to be clear.
Boundary specification problem
Define the boundary of the nodes.
1. Aggregation: all nodes that relate to a common agent are subject to common
effects or have common causes.
2. Proxy: nodes are proxy for unobserved agents.
–
–
The link must be made explicit and justified.
This specification is particularly important for firms (merge) node can be business
unit or the firm
There are two typical approaches:
•
Nominalist strategy
whereby the analyst imposes a definition of the relevant set based on a priori
criteria;
(gathering data on all firms that are assigned to a particular standard industrial
classification (SIC) code or that use the same set of inputs)
•
Realist strategy
whereby actors are included if they are judged relevant by the actors
themselves.
(snowball sampling, idiosyncratic to a particular time and place)
Approach to analyse economic networks
• Network as concentrated exchange
ties as market exchanges and ascribes importance to the
network because it is more concentrated or patterned than is
expected by orthodox market models.
• Networks as primordial relations
– focus on social networks that pre exist (sociologists);
– seems easier to say which aspect of the relationship is social and
which is economic;
– easy to verify the causal impact of the former.
• Networks as structures of mutual orientations
– general definition that include the other two as subtypes;
– one agent's ties to the others varies in the type, valence, and
strength of the connection felt toward the second party;
– sociometric networks.
Kirman on the Marseille Fish Market
•
Networks as concentrated exchange
•
Refrain from searching beyond their usual source for fish
 pay a higher price
 typically obtain better service
•
•
Prices are highly variable on a daily basis,
Stable when viewed over a month-long interval.
•
Irregular patterns from the standpoint of neoclassical theory:
1. there is significant price dispersion;
2. individual demand curves are not downward sloping.
"the regularity at the aggregate level ... is not due to individuals
behaving in isolation ... (according to) the standard competitive
model ... (thereby breaking) any simple link between individual and
aggregate behavior".
Go too far … argue that among agents who are embedded in such
networks, self interest is somewhat muted due to strong
identification with one's partners or even larger collectivities.
Casella on the Kirman's results
•
Casella argues that Kirman's success calls into question
whether we should care about the underlying structure of
individual interactions.
•
Two reasons to care about such networks:
1. while the market may generally aggregate micro activity to
produce the patterns suggested by orthodox theory, this
may not be true for all aggregate features, and
particularly, the rate and direction of the path towards
equilibrium.
2. we should care about how equitably profits are distributed
and the correlation between high prices and privileged
service.
Feenstra, Hamilton and Huang on East Asian Economic
Organization
FHHE try to motivated by longstanding differences in the
organization of the South Korean and the Taiwanese
economies.
1. South Korea economy: very large, vertically integrated firms
and business groups
2. Taiwanese economy: much smaller firms that sell
intermediate goods and services to a large number of small
and medium size firms.
• They runs opposite to that taken by Kirman,
– begin with "regular" behavior at the agent (firm) level,
– produces "irregularity" at the macro level.
• Their model generates multiple sets of stable equilibria, two
types of which broadly resemble the South Korean and
Taiwanese economies.
… FHH on East Asian Economic Organization
The model is based on some assumptions:
1. the market is monopolistic;
2. firm sell the intermediate inputs to their own firms at marginal
cost and to unaffiliated firms at marginal cost plus a markup
3. firms have incentive to withhold their intermediate inputs from
other groups because they are competitors
4. the elasticity of the intermediate goods (assumed to be uniform
throughout the economy) is different for the Taiwanese economy
and the the South Korean economy.
•
No governarce cost and only too small transaction cost.
•
Consider the double marginalization problem but assume the
only way to reduce it is via integration and possible contract
are ignored.
•
Political institutional and historical differences in the two
economics are ignored.
•
Failure to adopt a theory of the firm.
Rauch on Alternatives to Ethnic Networks
How primordial relations are relevant for the analysis of
economic behavior are particularly noteworthy.
"an ethnic business network can be a tool that allows
entrepreneurs to avoid the effects of discrimination".
• Discrimination:
exclusion of businesses owned by minorities or immigrants
from the flow of information about such matters as where to
find appropriate and reliable vendors for key inputs.
• Ethnic communities
fill this void (Caribbean business community) and argues that
the weakness of African American retail entrepreneurship may
result in part from the absence of such networks.
… Rauch on Alternatives to Ethnic Networks
• Policy proposal:
Independent buying office to better serve African American
retailers. Independent buying offices are commercial
intermediaries
– reduce the costs of sourcing inputs
– provide an array of related services to assist retailers
– loci for information exchange.
• Functional alternatives to ethnic networks,
• The market may fail to support such groups in
cases such as the Afro American community
– incentives to invest in the accumulation of knowledge
required are likely to be weak,
– non-contractible good.
Tienda and Raijman reply to Rauch
• Skeptical of Rauch's proposal.
• Recognize that commercial intermediaries cannot
supplement the role played by ethnic networks.
• This critique higligt the sociologist's tendency to
privilege the informal and social over the formal and
economic.
• They describe results from a unique survey of small
businesses in Little Village, a largely Mexican
neighborhood in Chicago, conducted in 1994.
Padgett on Florentine Banks
•
Very speculative paper.
•
Model organizational genesis as a biological process.
•
Genesis of the alternatives
While existing biological organization theories can deal with
reproduction, or choice within given alternatives, Padget deal with the
genesis of the alternatives themselves.
•
Organizational birth is rooted interaction between the economic
domain and other social domains (particularly the political system)
•
Historical research on the partnership structure among fourteenth
and fifteenth century Florentine banks.
•
Four types of partnership structures:
1. family firms (father-son or brother-brother partnerships), which prevailed
through 1348;
2. guild firms (master-apprentice), which prevailed from 1349 to 1378;
3. social class based firms (father in law-son in law or friend-friend), which
prevailed from 1380 to 1433;
4. Medici era patronage firms (patron-client), which prevailed thereafter.
… Padgett on Florentine Banks
• Change from one era to the next was precipitated by
political crisis, which brought about a realignment of
existing elite coalitions.
"often needed to reach into their … political and familial networks
to stave off disaster. Once created under conditions of stress,
moreover, organizational form reproduced through career
recruitment"
• Primordial ethnic or religious identity should be regarded
as salient only under particular conditions.
• Changes or variation in those conditions may be
expected to reduce the relevance of some identities and
increase the salience of others.
• Networks between economic agents (within the firm) are
reconstructed in response to the demands from other
sectors and from politics.
Burt and the mutual trust
•
Structural hole theory:
actors who occupy brokerage positions, by virtue of having large
networks replete with "structural holes" between contacts, enjoy
better access to information and enhanced negotiating leverage.
•
Bandwidth hypothesis:
dense networks facilitate the information flow about reputation
necessary for collective sanctions to be applied to shirkers.
– This hypothesis is very important because is often touted as the basis for
the "social capital".
– Dense network of strong ties facilitates the emergence of mutual trust
among members of the network.
•
Burt casts doubt on the assumption that dense networks transmit
information about reputation more efficiently than do sparse
networks. in proposing the
•
Echo hypothesis:
"third parties do not enhance ego's information on alter so much as
they \dots reinforce ego's predisposition towards alter".
… Burt and the mutual trust
• Etiquette:
one avoids contradicting the preferences of an interlocutor so
as not to damage one's relationship with that person. As a
result, existing predispositions get reinforced since ego tends
to hear reports that "echo" her initial bias.
• Evidence
from social network survey data of three organizations,which
appears more in line with the echo hypothesis than with the
bandwidth hypothesis.
• Proximity
Typical assumption say that proximity between nodes in a
given network implies a greater tendency to share the same
information (or anything else flowing through the network,
such as a disease).
• Burt's analysis suggests that and may sometimes imply
the opposite.
Sobel on Burt's hypothesis
• Joel Sobel reacts to Burt's analysis by showing how an
economic theorist might model the echo hypothesis.
• He succeeds in producing a game model that roughly
approximates Burt's hypotheses but with which an
economist would feel more comfortable.
• In place of etiquette as the third party's rationale for
sharing her true opinion, Sobel postulates that close third
parties are more likely to know ego's predisposition than
are more distant third parties.
Conclusion and comments
•
Network as bridge
Good bridge between economist and sociologist.
•
Why economic networks
The importance of networks in economics is clear when we accept
that not all transactions take places at harms length.
•
Network tie as "pipe"
The theory proposed by the different authors are so different but
there is a trait d'union: the network tie is viewed as a "pipes" through
which information or resources flow. By this shortcut is possible to
analyse social networks of various kinds.
•
How to model networks
AB models are largely used in many of the work presented, and show
powerful explicative power in network study.
•
The underlying theory
In my view, attention must be paid to the underlying theory behind the
models. Many AB model described produce results coherent with real
economic data (I am thinking to the Marseille fish market and to the
East Asia economic organization) but this does not imply that the
model is a correct description of the phenomenon;
… Conclusion and comments
I strongly believe Milton Friedman was not right saying that:
“accurately accounting for economic agents behaviour is
unimportant so long as our model makes accurate predictions
about the economic outcomes about which we care."
the fact that a specific assumption is not rejected by an
econometric test does not mean that the assumption has
cached the relations underlining the economic phenomenon
as well as results from ABM do not assure theory’s general
validity.