substantial income disparities betw

substantial income disparities between social groups that might be socially
disruptive. Similarly, the spatial concentration of copying can reinforce spatial
differences in incomes.
Pooling has more potential to be regressive, because it is reciprocal. Networks
tend to include people with similar quantities of knowledge, because people
with a lot of knowledge will find it advantageous to pool with others who also
have a lot of knowledge. Treating the quantity of knowledge as a constraint on
which networks an agent can join, the incentive to join an information network
is greater the greater the amount of information in the network. People with
little knowledge to share are thus confined to networking with others with little
knowledge, and so have less incentive to join a network than people with
large quantities of knowledge. Pooling is therefore regressive in two ways:
more knowledgeable people will have larger networks, and they will gain more
information from each contact in the network. In effect, pooling introduces an
exponent term onto private knowledge. A tendency to exclude the poor is thus
built into private incentives to pool knowledge.
The second externality is the reduction in opportunism, which is achieved
through repeat transactions and reputation. Repeat transactions have the effect
of socially excluding new entrants and so tend to disadvantage the poor.
Reputation magnifies the gain from repeat transactions. Repeat transactions
produce a promise–trust bilateral relationship; reputation enables those who
are in a promise–trust relationship to access many other transactions. Biggs
and Srivastava (1996) show how this phenomenon hurts African-owned firms
relative to Asian-owned firms in Kenya.
The third effect is collective action, which is achieved through norms and
rules. Some of these norms and rules apply across the society. Putnam’s choirs,
for example, build trust not just among choir members but among the population
as a whole. Such social capital is likely to be proportionately more beneficial
to the poor, who are less able to invest in substitutes. For example, an agent
is protected against crime both by private expenditures on security and by the
norms and rules of society. Because the poor are less able to afford private
security expenditures, they are more dependent on norms and rules. Generally,
the victims of crime are disproportionately drawn from the poor (which probably
reflects differences in security expenditures), so that the stronger are norms and
rules, the more the poor gain relative to the better-off.
Other norms and rules apply only to the group. Thus the rules of a club
may make it an effective agency for increasing the incomes of its members. In
general, the poor have more to gain from norms and rules than higher-income
groups and so have more incentive to join clubs and authorities. In credit transactions,
for example, social sanction and collateral assets are alternative means
of reassuring lenders. Wealthy people do not need social sanction, whereas the
poor do. Similarly, the economies of scale that collective action facilitates are