In the empirical analysis, stock market and operating data are obtained from
Compustat. The market value of equity is the market price of common stock
(Compustat variable A#24) times the number of shares outstanding (A#25) as of
December 31 for each year. Size effects are normalized by deflating size-related
variables (market value and the number of patents) by the book value of total
asset (A#6).
Table 7.2 shows sample descriptive statistics quantity and scientific quality
of inventive output from Tech-Line
®
. Japanese companies are more prolific than
their U.S. counterparts in the number of U.S. patents granted. Among firms that
received 10 or more U.S. patents per year over the 1989–95 period, the typical
Japanese high-tech firm was issued 88.14 patents compared to 82.56 patents for
similar U.S. firms. However, many U.S. patents issued to Japanese firms appear
to be of relatively lower quality and economic value. Japanese companies had an
average CI of 0.91, meaning that Japanese patents receive only 91% of the typi-
cal number of citations in subsequent patent applications. U.S. patents issued to
domestic firms receive 114% of the typical number of such citations. U.S. patents
issued to Japanese companies are also characterized by relatively lower NPR.
Japanese firms have only 45% of the typical number of “other references cited”
on the front page of each patent, compared with 107% for U.S. firms. Finally, the
TCT indicator value of 7.85 years for Japanese firms, defined as the median age
of earlier U.S. patents referenced on the front page of a U.S. patent, is below the
10.02-year average for U.S. firms. This means that U.S. patenting activity by
Japanese firms is most prolific in emerging technologies where innovation
is rapid.
In estimation, it will be interesting to note the extent to which the stock
market perceives any size-based differences in the effectiveness of patent activity.
Similarly, it will be interesting to note the extent that patent quality appears more
important for Japanese and U.S. firms with differing growth opportunities.
VII. METHODOLOGY
A simple predictive equation for stock-price changes is considered where:
P
i,t+1
=
Φ
(P
i,t
, Patents
i,t
, Patent Quality
i,t
(CI
i,t
,SL
i,t
,TCT
i,t
), (7.1)
and P
i,t+1
is the next-year stock price, and P
i,t
is the present-year stock price. In
this relation, notice that the stock-price effects of ordinary physical assets and
other such influences are reflected in P
i,t
. Following Hall, Jaffe, and Trajtenberg
(2000) and a long tradition of research on the economic implications of patenting
activity, a simple linear specification of this model is estimated.
The focus of interest is on how investors might use scientific indicators of
patent quality as useful evidence concerning the future economic worth of the
firm’s patent activity. Under the main hypothesis, indicators of the scientific merit
METHODOLOGY 169

Get Tech Stock Valuation now with O’Reilly online learning.

O’Reilly members experience live online training, plus books, videos, and digital content from 200+ publishers.