In the information economy, intermediaries will also be economic
decision makers. In the shopbot model, we have investigated a scenario
in which the cost structure
reflects an actual price paid by
buyers to a shopbot, and have explored how a shopbot might manipulate
to maximize its own profit [17].
As a simplification, suppose that the shopbot offers two choices: a
single quote for price
and two quotes for price
. A
competing mechanism for obtaining price information costs c' for one
quote and 2c' for two quotes. For example, manual price comparison
by a human (conducted by visiting multiple merchant web sites) might
well cost an amount of time (and therefore money) proportional to s.
Figure 5: Optimal shopbot prices
and
as
function of alternative search cost c'. (Normalized to v-r=1).
The optimal settings of
and
as a function of c' are
depicted in Fig. 5. Regardless of c', the
shopbot should always just undercut the alternate mechanism on a
single quote. The price of the second quote
has a
more complicated dependence on c'. For low c', the second quote
should also be priced just less than c'. However, for intermediate
values of c', the price of the second quote must be less than that
of the first -- otherwise, too many buyers will be discouraged from
buying two quotes. In this regime,
should be a constant
value 0.0957, which maximizes the product of
times the
fraction of buyers that purchase two quotes. For large values of
c',
must be reduced below 0.0957. Reducing
encourages more buyers to purchase two quotes. Increased comparison
shopping forces sellers' prices lower, making it possible for buyers to
afford a high single-quote price
. At the extreme limit of
, practically all buyers purchase two quotes. If almost
all buyers are comparing prices, the sellers' prices drop to just
above the marginal cost (zero). Thus the sellers get virtually no
surplus. The buyers pay very little for the product itself, but pay
almost their entire valuation to the shopbot for the price
information, so they get no surplus either. Thus, if a shopbot has an
effective monopoly on price information (i.e. the alternative search
cost equals or exceeds the difference between buyer valuation and
marginal production cost), then it can extract practically all of the
surplus from the market.