day 106
![]()
alice in wired world 10: inverse econ 101
Now, here is the non-Keynesian part. Classical learning curves predict price
declines (or an increase in quality) as the
volume of produced goods doubles. This is the
so-called theory of diminishing returns. But in
the new software economy, the opposite is true --
volume production leads to increasing
returns, according to new-age economists such
as W. Brian Arthur ("The Theory That Made
Microsoft", FORTUNE, April 29, 1996,
p. 65-66). I call it the theory of inverse
economics, Arthur calls it the theory of
increasing returns.
Here is how it works. Assuming true pricing
(the price you pay is equivalent to the quality
you get), the learning curve model predicts
that the commodity-priced product will
eventually overtake the premium-priced product
simply because of higher production volumes.
This is non-intuitive, because it says that if
Toyota makes enough cars, eventually Toyotas
will be both better and cheaper than
Mercedes-Benz cars!
Figure 2 illustrates the impact of inverse
economics on premium/commodity market
segments. For example, suppose a commodity
producer manufactures 20 times as much product
as a premium producer. This is indicated as a
20 on the vertical axis of Figure 2. If both
manufacturers learn at the same industry rate,
say B = 0.5, then (0.5, 20) represents the
point at which both products are of equal
quality. If the premium product initially sells
for 5 times the price of a commodity product,
it will eventually be no better in quality than
a commodity product that sells 10 times as may
units. The commodity product eventually catches
up with the premium product.
In a way, volume production allows the
commodity-product to accelerate the learning
curve needed to make a better and cheaper
product. It is the superiority of mediocrity
over elitism.
Figure 2. Ratio of Volume Production of Commodity-Priced product to ![]()
Volume Production of Premium-Priced product versus the learning rate
parameter B. Recall that small values of B represent high learning rates,
and high values represent low learning rates, 0<B<1.
alice in wired world 1
2
3
4
5
6
7
8
9
Daily Dose Index