0% found this document useful (0 votes)
74 views14 pagesChapter 12 R - D Model
Uploaded by
Sheikh Sahil MobinCopyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF or read online on Scribd
0% found this document useful (0 votes)
74 views14 pagesChapter 12 R - D Model
Uploaded by
Sheikh Sahil MobinCopyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF or read online on Scribd
/ 14
(AN APPLICATION: RESEARCH
POC aaa
‘This chapter will present an application of the backward induction solution proposed
in the previous chapter. The economic problem is that of research and development
(ReD), and the setting for this particular application is that of a patent race. The R&D
problem will be discussed in section 12.1, and a model ofa patent race willbe presented
in section 12.2. In section 12.3 we will derive the backward induction solution of that
model. Section 12.4 will conclude with a discussion of possible generalizations of the
model.
12.1 BACKGROUND: R&D, PATENTS,
AND OLIGOPOLIES
‘The engine of growth for modera economies is technological progress; indeed, this
has been the engine for some 250 years, ever since the early years of the Industrial
Revolution. Some economists have estimated that by far the majority of the growth
that has taken place in the United States economy in the 20th century has been driven
by technological advancement. Even today, many of the sectors that are experiencing
the highest rates of growth—such as computers, telecommunications, pharmaceuticals,
and biotechnology—are, in fac, the sectors where the growth rates of technology are the
highest. It is therefore extremely important to answer such questions as, What drives the
growth of new technologies, and how can government R&D policy futher stimulate this
growth? D
‘A central aspect of an innovation is that itis a public good; a new idea can be
exploited by whoever has access to itand the next person who gets access to it will
not have any less of the idea to work with. For example, if now gene is sequenced, and
the result is announced in a scientific journal, then the same code will be available to
A spte of technalgia beck
throughs uncked the Indust
Revolution in glad inthe md
dl ofthe 18th century. Inprtant
Innowations ofthe S6th and 190
centuie inclced the stern a
ne and th pining joy in te
textile indus, ond the Bese
ner proces in the to and tel
Industry.
2 Forinstanoe, in a om study
publish in 1957, Raber Solow of
MA estate that 675 percent of
the gronth in Utd Sates GNP
ten 1809 and 1949 could
be cttabuted fo techoloial
govt (coe Tenia! Change
and the Aggregate Production
Function, Review of Beonasies
snd Stati, val. 2p 912-320) 460 | Part} | Cuarren 12
® his tates spcily rue
gin th eats of developing 0
ur drag rom stat to finish,
(ne study eticted thatthe
verge otf doing sis about
$850 lin (0 Phamaceatical
RED: Cot Rth, nd Rewari,”
1088, US. Congres, Ofc of
Tehnology Assesment). This
sty bose on drags that were
developed inthe 1880, and it may
actualy understate the cots
of deolping nse drugs today,
nous the lng of te that
hogs undego lial eas
‘ow increased to an avo of 15
ers.
* Bdwand Mansfld ofthe Unt
verity ofPennsyvania estimated
the percentage ofcommerilly
introduced Snovatios that would
ot have ben derelpod without
tent protection (se nelle
tual Property Rights and Capital
Formation i the Not Decades,
University Pose of Americ, 1988)
Thezumbersrange fom 6 percent
{orpharmaceutialsand 9 percent
for camicoe trough 17 percent
for machinery and 1 percent for
primary metal.
* Acompony tht hos «patent
can lense the technology tater
companies or can enter into joint
development ovement with
ters. Te pint remains, though,
that a patent io way of making 2
reesrch beakthrogh es fly
callable to those who dd not
make the brakrough
* Since HOTY signal cary more
inforatio, they need wider
bandits for trasmision
than does conventional tlvision
technology, eden, much of
the information inthis subsection
comes from the iol "The Next
Genertion of Teison" by Dale
Cpe, fatced inthe HDTV
Newsletter and availble atthe
‘nb tte com/hdd
every person who reads the scientific paper. Furthermore, using this information, a new
drug could be developed by any pharmaceutical company that has the requisite drug
development infrastructure.
‘This brings up the question, Ifthe original gene sequencing was a costly enterprise,
why would any drug company do this spadework? They would all rather have some
other company's laboratory do the initial sequencing, acquire the information from the
scientific journal, and then only spend money to actually develop the consequent drug.*
However, ifall drug companies thought along these lines, nobody would do the initial,
but critical, work.
‘That is procisoly where patents come in; they are a way of rewarding the company
that did the original sequencing. A patent on a sequence gives a company the exclusive
right to develop any drugs that emerge from that initial research breakthrough. Patents
can be awarded with different degrees of comprehensiveness: some cover all develop-
ments that emerge from the frst innovation, whereas others only cover developments
that arise directly from it. The important point is that patents give private companies an
incentive to do R&D* and they have a “winner-take-all” element to them.
In most major modem economies, R&D is conducted by private firms in oligop-
olistic industries. For instance, the U.S, pharmaceutical industry, which spends $15.8,
billion annually on R&D, is an industry with a fow major firms and another cluster of
smaller players. Hence there is intense competition among these companies in the de-
velopment of new drugs.
Inan oligopoly, since there are only a few rivals and doing R&D isa costly business,
its conduct is a strategic variable. In choosing how much R&D to do, how many new
projects to fund, when to fund them, when to pull the plug on an R&D project that is mot
proceeding satisfactorily, and the like, an oligopolistic firm takes into account decisions
on similar matters that its rivals make. That factor brings game theory into this important
issue and brings us to the analysis of this chapter.
12.1.1 A PATENT RACE IN PROGRESS: HIGH-DEFINITION
TELEVISION
High-definition telovision (HDTV) is thought by many to be the next frontier in home
entertainment. HDTV, by using a digital technology, will provide a much higher pic-
ture quality than conventional (analog) television images. In particular, picture quality
will not depreciate even on very large ‘screens, and the color quality will be several
times better. On Christmas Eve, 1996, the Federal Communications Commission finally
approved the industry-wide standard for this technology. The FCC is also encouraging
broadcasters to switch to the new technology by granting them the necessary increase in
the broadcasting spectrum.®
The history of HDTV is interesting especially for the insights it gives about the
R&D process, The first steps toward its development were taken in Japan, where at the
urging of the Japanese Broadcasting Corporation a consortium of Japanese television
manufacturers started research on digital television technology in 1964.7 In the early
1980s they carried out the first experimental transmission of HDTV signals, and at an
in
tk
a
oe AW APPLICATION: RESEARCH AND DEVELOPMENT | 181
jnternational meeting in Algiers in 1981, they presented the first set of standards for
tho new technology. About this time, the American Electronics Association docided
that HDTV was indeed going to be the technology of the future and urged the Ameri-
con administration to take steps to help catch up with the Japanese. A consortium of
largely American companies that goes by the (sinister sounding) name of the Grand
Alliance’ was formed. The American consortium has made rapid progress since its in-
ception and is now acknowledged to be the technological leader. But the race is still
on.
12.2 AMODEL OF R&D
‘The setting for our model isa duopolistic industry—say, consumer electronics—in which
each firm is working toward a technological breakthrough—say, HDTV. To the winner go
the spoils; the winner gets a patent for the new innovation. The questions of interest are
those: How much should each firm spend on R&D, and how often? When should it get
into a race, and at what point should it opt out of such a race? What factors determine
the likely winner: is it an advantage to be in a related manufacturing area, is it more
important to have a superior R&D department, and so on?®
Suppose thore are two firms in an industry, RCA and Sony—hereafter firm Rand
firm $—each of which is conducting R&D to produce HDTV.!° There are several stages
that need to be completed successfully before HDTV can be brought to the market. To
‘make the analysis tractable, we will make several simplifying assumptions:
1. The distance from the eventual goal can be measured; we can say, for example, that
firm Sis n steps from completing the project.
2, Bither firm can move 1, 2, or 3 steps closer to the end in any one period.
3, Itcosts $2 (million) to move one step forward, $7 to move two steps forward, and
$15 to move three steps forward."?
4. Whichever firm completes all the stops first gets the patent; the patent is worth $20.
Discussion of the Assumptions
‘The first assumption says, in essence, that there is a one-dimensional index by which the
technology can be measured. For instance, the number of technological problems that
need to be resolved before a company can apply for a patent may be the number of steps
from completion of the project.
The second assumption asserts that infinite progress cannot be made in one period.
‘You can think of a period in this game to be, say, anything from one to three months. The
assumption says that if the project is at an early stage, we cannot expect to complete it
"The consortium Included such
ousshold names in consumer
electronics os Sony, Panasoiée,
Toshib, Hitachi, Mitsubishi, ond
we.
"The Grand Alone i made
1p of David Saf Resch
Laboratories (Dovid Samef os
head of RA was tho pioneer
In color tlevision), General
nstument, Zenith, ATET, Philips,
Mtr and Thomson
°VThe emadel in this chapter
ran from Chitopher Haris and
John Vier "Potent Competition
ma Model of Rac,” Review of
Econom Sues, 1985 vol. 52,
pp. 193-208. Tay donot, however,
saga theapllation o HOTV
cand should not therefor beheld
responsible for ny ovkwardass
fn the fi betwen ther model and
thefts of tha HDTV wold
* Altmatvly, tae are two
consti conducting RED. We
se the terminology “two firms”
rather than “wo consort" spl
because the lotr orm i aise
anwield
4 fiom cum aio deide not to
snake any progress at al hati
smove0 steps. The cost i 3, 182 | Pants | Cuarren 12
"This assumption can be @
consequence of technological
cantons x wel If for example,
fe echnoloial problems need
‘salution andthe project team
‘ean sart onthe fourth problem
only aftr theft thee have been
sotsectrly recive, then tis
or likly ha thy wil not gto
‘eithn the next paid.
"9 Thearpumant of sap 1 would be
des true f there ware uncertainty
inthe RED proves. I tht cas,
the carte might wish to hore oth
firms do RED becouse there ea
rater probity that at est one
ofthe two wil fh the project
vicky. The second step would
ood to bo mdifed the cartel
hadreasontowanttobrig the new
_prodact tothe market os quickly
spose We wl discus these
eneraliztions inthe lst section
over the next three-month period. As the project nears completion—that is, when itis
already within three steps of completion—it can indeed be finished in one period."
‘The third assumption says there is no free lunch. Ifthe project is to move faster, its
managers have to hire more personnel or invest in greater infrastructure, and so on, all of
which costs more money. The actual numbers have been chosen to reflect an underlying
decreasing returns to scale in R&D; costs go up faster when a firm tries to complete 3
rather than 2 steps than they do in going from 1 to 2 steps.
Finally, the fourth assumption is the definition of a patent; it gives the winner
a positive reward (whereas the loser gets zero). This reward can be thought of as the
increased profits from selling a product with an exclusive technology, a technology that
no other competitor possesses. Hence, the total profit to the patent winner is the value
of the patent less the cost incurred in winning it, For the other firm, the total loss is the
R&D cost. ~
‘The numbers that we use are for illustrative purposes. In the Exercises section you
will redo the analysis with other sets of numbers. Also, at the end of this chapter we will
suggest some further generalizations of the assumptions.
Before getting to the game-theoretic analysis, however, let us first ask what would
happen in this simple model if the two firms were able to coordinate their R&D activities
and operate as a cartel (and hence pick the R&D expenditures to maximizé joint profits),
Concept CHECK
Carte, StEP 1
Verify the following statement: Since only one of the two firms is going to get the patent,
itonly pays to have one of them do R&D.
Carret, STEP 2
Vorify the following statement: Whichever firm does R&D does so by spending the
smallest possible amount and moving forward a step at a time. Furthermore the firm
chosen will be the one that is closer to finishing.
(Hint: For the fist par ofthe statement zemember that there are decreasing returns
to doing R&D.)
v
The two steps would be true in a modified form if there is uncertainty and
impatience.’
‘To summarize, a cartel would minimize R&D competition. It would let the tech-
nologically more advanced member firm advance toward the patent in minimal cost
increments. Aw APPLICATION: RESEARCH AND DEVELOPMENT | 183
Figure 124
12.3 BACKWARD INDUCTION: ANALYSIS OF
THE MODEL
What would be the duopoly outcome (with the two firms competing in their pursuit ofthe
patent)? Somewhat surprisingly, it tums out that we get a sharp answer to that question
ifwe make one last assumption:
5, The two firms take turns deciding how much to spend on R&D; if RCA makes an
R&D decision this period, it waits to make any further decisions till it learns of
Sony's next R&D commitment. Furthermore, Sony ‘makes its announcement in the
poriod following RCA’s announcement,
One way to think about this assumption is that each firm's management makes
periodic reviews of the project. These reviews are conducted every fow months, and-at
each such review a decision is taken sbout R&D spending until the next review. The
decision might be to step up spending levels, to hold them at the current level, o to cut
back. Firms alternate in their decisions ifthe review dates are differont, although the
length of the review period is the same for the
‘Assumption 5 tums the patent race into a game of perfect information; let us look
atts extensive form. In Figure 12.1, RCA has the first R&D decision, and RCA and Sony
are, respectively, 3 and 4 steps from completing the project.
‘A somewhat more transparent depiction ofthis same situation can be given in a
location space picture. By that we mean a picturo in which the “coordinates” of the
twvo firms—that is, how far they are from finishing—are graphed. In this location space
picture, the northeast point refers tothe joint finish lin for both firms; thats, successful
completion of R&D by both firms. The finish lino for S is the vertical terminal line,
‘whereas the finish line for Ris the horizontal terminal line; see Figure 12.2, 184 | PaRT3 | Cnarren 22
Noe tat we made no mention
of how much ltr fm may
lead hae spent. Regardless of
how much ha leady been spent,
«frm one stp frm fishing
tondeto ake ans prot of
516 by completing the project. ie
val come back hi ideo—the
Inslorance of unk ets the
end of ths chapter.
mndR
su
(3,4)
End
Ss
‘The following notation will be useful for the location space. If is r steps from
completion while $ is steps away, we will denote their location as (r,s). The game will
be solved by backward induction on the location space. In other words, we will show
that when either irm is near completion there is a best way for them to make their R&D
choice, In turn, these eventual decisions will affect the choices the firms make at more
formative stages of R&D.
‘To illustrate these ideas, we will proceed in a stepwise fashion:
Fiune 122
+ Step1. Suppose that the game is at (1,5), and itis firm A's turn tomove. Its optimal
docision evidently is to finish the game in one move, That will yield a patent of value
$20 (million) and will cost §2 (million). Similarly, ifthe location is (r,1) and itis firm
S's tum to move, S will complete the project in one step.™#
+ Step2. Now suppose thatthe two firms are at either (2,1) orat (3,1), and itis firm
Ris tur to move, It can complete in one move, and if it does R makes a positive profit
at both locations: $20 - $7 in the first case and $20 ~ $15 if itis 3 steps from finishing.
Indeod, if R does not finish the game in one move, it knows that $ will do so at the very
next opportunity (why?), and hence R will either make nothing from that point on or
suffer a loss, For example, if R chooses to make no progress, which is costless, it will
find that S will win the patent in the next period. Ifit makes incomplete progress—1 step
starting from (2, 1) or 2 steps or less from (3, 1)—it will incur a cost but will not win the
patent.
Hence, itis best for R to complete in one step ifit has a move at (2,1) or at (3,1).
Of course, the same result holds if the firms are at (1, 2) or at (1,3) and itis frm S's turn
to move; $ will complete in one step.
In turn this result has the following implication:
Step 3. (a) Usethe previous analysis to show that ifthe game sat (2,2), whichever
firm has the first move should invest for two steps and finish the game. (b) Can you then wer
hen
Ficune 12.3
show that if the game is at (3, 2) and it is firm F’s turn to move, it should finish in one
step. What if the game is at (2, 3) and S has the first move?!®
In fact what we have shown via steps 1 through 3 is the following:
Proposition (TI). Ifthe game is at any location (r,s), r <3 and s <3, whichever fim
has the first move at that point will trigger a completion, that is, wil end the project in
one step.
Call this set of locations Trigger Zone I, as seen in Figure 12.3. Let us use the
information about Trigger Zone Ito analyze what happens at other locations. What we are
really going to do is fold the location space back; that is, we will do backward induction
onit. Since we know what is going to happen at the “end” of the space, we can now ask
‘what will happen at a penultimate zone of that same space.
+ Step 4. For instance, what can we conclude about a location such as (4, 3) when
itis firm F’s tum to move? Note that R cannot finish the game in one step. The most that
it can do is move its project forward by three steps to (1, 3). Or it can move two steps to
(2, 3) or one step to (3, 3). Or it can remain where it is by stopping R&D. In the first three
of these cases, knows that $ will, in fact, finish the game at the next step. (Why?) So
the best response for Ris to pick the fourth option, make no progress. This is equivalent
to dropping out ofthe race.
If firm R finds it in its best interest to drop out of the patent race at (4, 3), what
should firm $ do subsequently? Well, firm S, as the sole survivor, will get the patent
eventually. Since rapid R&D is more costly than slow R&D, the best approach for Sis to
move in the least costly fashion, one step at a time, toward the patent,
* Steps. Show that the same conclusion, R should drop out ofthe race and $ should
then advance slowly, is true also for locations (4, 2) and (4, 1). What about locations (5, 3),
6,2), and , 1)?
An APPLICATION: RESEARCH AND DEVELOPMENT | 185
9 ll ofthe cases, ack youself
what would happen if the fst
move did et complete the project
‘none stp. What, n particular,
would it ial doin tho next
pevod? 186 | Part3 | CuaereR 12
EndR
Safety Zone
for R
(3)
End
Safety [ $
Trigger Zone |
Zone! for SF
FlouRe 124
Iterating, we can, in fact, conclude the following:
Proposition 2. . For all locations (r,s), whenever r > 3 and s < 3, the best thing that firm
Rasa first mover can do is drop out, After this, firm S can take a step forward at a
time,
This sot of locations is therefore called Safety Zone I for S; in these locations, $
can coast and R will drop out. Since the game is symmetric, we can also conclude thet
all locations (r,s), r <3 and s > 3, is Safety Zone I for firm R. The two safety zones are
pictured in Figure 12.4.
Lot us continue with the backward induction argument. We know that in Trigger
Zone | there will be a preemptive move while in Safety Zone I, the “war” is over. The
next question to ask is, Is there an incentive for either firm to try and get into its own.
safety zone even if doing so means doing rapid R&D?
* Step 6. Consider a location such as (4,4). Suppose itis firm F’s tur to move. Firm
Ran, in fact, take the game into its Safety Zone Tin one step—at a cost of 2. Thereafter
it knows that $ will drop out, and hence it can move a step at a time toward eventual
completion; those three steps will cost a further $6. The total costs then will be $8, and
‘that is less than the value of the patent.
More is true; as long as R has a way to get into its safety zone—and thereafter move
a step ata time—while incurring costs that are no more than $20, the value ofthe patent,
it is worth R's while to do so, The argument applies symmetrically, of course, to firm S.
+ Step 7. Show that from locations (r,s) if r,s =4,5, the first mover will find it
profitable to take the game into its Safety Zone I, Show that if R has to move at (6,4), its
consequent net profit is $7.
’ *
, Aw APPLICATION: RESEARCH AND DEVELOPMENT | 187
End R
Safety Zone 1
for R
Trigger
Zone!
Safety Zone II
forR | ma
s
‘Trigger
Safety [
Zone
for $
Zone I
zone | 8
Figure 125
However, it is not worth firm F's while to move into its safety zone if it is at (6, 5).
That would cost $21 in total (why?), which exceeds the value of the patent. Of course,
this result implies that, from that location, the best response for R, if it is the first mover,
isto drop out of the race, (Why?)
‘These arguments give us the following
Proposition 3. There is a second Trigger Zone between (3, 3) and (6,5); the first mover
inthis zone should move the game into its own Safety Zone I, There is also a second set
of safety zones. The one for R is 3 5 (and symmetrically for $). In Fi
R's Safety Zone Il, S should immediately drop out. See Figure 12.5.
Continuing in this fashion, we get the picture shown in Figure 12.6 forthe solution
in location space.
To summarize, the associated strategies are the following: If Sis inf’ safety zone—
whatever the zone number—the best thing it can do is drop out of the race, Firm $ in its
own safety zone spends the minimum amount on R&D, moves a step at a time, and coasts
towin the patent. In Trigger Zone n, each firm spends what it needs to—profitably—to
get an invincible advantage for itself and move the game into safety zone n— 1.19
For these numbers on costs and patent value, there are six trigger zones and five
corresponding safety zones. Different numbers on the cost and patent value variables will
change the size and number of these zones, but will not change the qualitative feature of
the solution. Indeed, you will establish the truth of this assertion in the Exercises,
"Note that excep forthe tigger
zones, the outome looks a Tot
ike th eral elution nether
woud, the depot fght—or
speed money—texebish inital
advange, bt once thotadvantage
‘as boen establiied thee only
«sing fem that continves to do
ed. 488 | Paer3 | CaaPren 12
Again, you wl work through
some exetiestobete understand
thi paint
End R
End
Safety Zone I 1H
eget
for R Zone
Safety Zone Il :
‘Trigger
forR Zone Il
Trigger
Safety Zone Il L
Zone III
: ‘Trigger
Safety V Zone V
‘Trigger
Zone V
‘higger Zone VI
Frou 12.6
12.4 SOME REMARKS
First, our solution teaches us the importance of “sunk costs. ” Even if firm Shas already
spent a large sum up until some point, it will be ready to spend up to $20 additionally »
in order to secure the patent, It will be willing to do so because its net profit from that
point onis the patent value minus additional costs, and this net profit is positive as long
as the additional costs are less than $20. Hence the way that the game gets played from
any location (r,s) is completely independent of how the game got to that point, that is
is independent ofthe investments that have been sunk by the two firms in getting to hat
location,
Second, the symmetry assumptions that we made—each firms costs of doing R&D
are identical, and so is the value of the patent—are completely inessential. The entire
analysis can be ropeated for the case of dissimilar costs or patent valuations. All thet
changes is that the sizes of the trigger and safety zones become firm-specific. Ifa firm has
lower costs its zones are going to be biggei*for example. It still remains true, howeveh
that from any location we will ee a one-step movement toa safety zone or an immedial®
7
73
dropout by the first mover AW APPLICATION: RESEARCH AND DEVELOPMENT | 189
‘Third, the larger the value ofthe patent, the larger are the trigger zones. For instance,
if the value of the patent is $30, then the size of Ttigger Zone Il would be the “square”
between (3,3) and (6,6), and that of Trigger Zone Ill would be between (7,7) and (9,9). Put
differently, the patent race would be much more of a race; a firm would drop out only
if it was more than three steps behind its rival but would remain in the race if it was
any closer. The higher the value of the patent—or the lower the costs of doing R&D—the
snore likely itis that we have a horse race.
Fourth, if there is uncertainty about the outcome to R&D, then that might induce
4 firm to stay in the race longer, because there is some chance that would looks like an
insurmountable lead now will not remain so forever:
Fifth, if a firm has @ preference for quick profits—rather than profits that accrue
further off in the future—it may choose to do R&D more rapidly ven when there is no
competitive threat. For example, consider the cartel doing R&D. If a cer
three steps from finishing the cartel, how much should the cartel spend in the current
period? If the patent is only useful when acquired immediately, the cartel will clearly
choose to complete all three steps in the current period (and pay the higher cost of $15).
More generally, if the future is discounted, the analysis applies except the size of the
trigger and safety zones and the behavior within a safety zone depend on how heavily
the future is discounted.1®
Sixth, public policy toward R&D is extremely important in this setting because it
pays to be ahead oven if it is by a small amount, In the case where the other firm is a
foreign competitor for example, it will pay the domestic government to subsidize the
R&D efforts of the domestic firm so as to give the latter the necessary small advantage,
Furthermore, since all that is required is a small advantage, public policy is also very
cost-effective in this context.’
Finally, what have we learned about HDTV from our model? We know that neither
consortium has dropped out. This fact suggests that profits are expected to be high (or
R&D costs are low), Both these inferences sound reasonable; clearly both consortia (ond
their administration backers) view HDTV as the technology of the future—and hence
expect large eventual profits. (Further, each government has defense or other motivations
that make the nonmonetary payoffs also loom large.) Doing R&D through the consortia
tas lowored the riskiness as well as the capital costs of doing R&D. Finally ifthe firms
are in their (large) trigger zones, they should be doing R&D as rapidly as possible. There
is certainly evidence to suggest that they are.
member is
1. Atechnological breakthrough can be profitably utilized by more than one company,
and hence each company has an incentive to let somebody else make a costly
breakthrough. Patents are a way of solving this incentive problem,
# See Chapter 15 fore discussion
ef scouting
ofcourse, bat govenmects
ore subiicing thee respective
RED ofa, publi palicy may
ot be gut as affective. n pa
iulr, relative postions may be
completely unchanged f the to
govemments spend the same—or
sinilanamouns on subuidi
‘ng Re. Tie may bea bate
Aesrption of opanese and US.
public poly toad ReD In HOTY.
Both governments re ctive—tho
Jopanase though the opanese
Broadening Corporations Si
ence ond Technology Center and
‘the Americas though the Ad
vanced Teleision Cnt thir
forts might hrofore bo wash 490 | Parr | Cuapten 12
2, Inmany industries, firms battle furiously to win patents on new technologies; high-
definition telovision (HDTV) is an example of an ongoing patent race.
3. A two-firm patent race in which the competitors take turns making R&D invest-
rents can be modeled as a game of perfect information, Furthermore, this game
ccan be solved using backward induction pn location space.
4. The backward induction solution has the feature that two firms that ere similar
distances from project completion invest heavily to get an R&D’ advantage. A firm
that falls sufficiently behind is better off dropping out ofthe patent race,
5, Droppingoutis less likely the higher the value of the patent and the lower the costs
of doing R&D. These characteristics describe the current HDTV race.
SECTION 12.4
12.4
Give an example of a research breakthrough that was patented by the company thet
made the breakthrough (you might wish to explore the pharmaceutical industry for this
example).
122 ‘
Give an example of a research breakthrough that was privately developed but not
patented. Explain whether it was nevertheless profitable forthe company in question to
‘make the breakthrough.
12.3
Give an example of a research breakthrough that was achieved in the public domain
either at a university or at a government research laboratory.
SECTION 12.3,
Lotus redo the analysis of the chapter with somewhat different patent values and cost
to doing R&D, ‘AW APPLICATION: RESEARCH AND DEVELOPMENT | 191.
424
Suppose that the patent is worth $25, Everything else is unchanged. Solve the R&D game
by backward induction.
12.5
‘What is the difference between your conclusion and that reached in the text? Explain
your answer, What general conclusion can you draw about the effect of patent valuation?
Explain carefully,
12.6
Suppose instead that the costs of moving 1, 2, and 3 steps are $4, $10, and $15, respec
tively (but the patent value remains at $20). Solve the R&D game by backward induction.
12.7 ’
What is the difference between the conclusion that you arrived at in exercise 12.6 and
that reached in the text? Explain your answer. What general conclusions can you draw
about the effect of costs? Explain carefully
The next few questions explore a game in which two firms—firms A and B—have
different costs and benefits from doing R&D. Let us start with different benefits. Suppose
thet all ofthe data of the chapter remain unchanged except forthe fact that the patent
is worth only $12 to firm B (whereas it is worth §20 to firm A), Let a (respectively, ’b)
denote the distance that firm A (respectively, firm B) is from completing its R&D project.
12.8
Show that Thigger Zone I is made up of all locations in which firm A is within 3 steps
and firm B is within 2 steps, that is, is made up ofall locations (a, b) such that a<3 and
bea
12.9
Show then that Safety Zone I, for firm A, is made up of all locations in which firm A
is within 3 stops and firm B is more than 2 steps ftom finishing, that is, is made up of
all locations (a,b) such thet a<3 and b> 2, $imilarly, show that Safety Zone J, for B,
is made up of all locations in which firm B is within 2 steps and firm A is more than 3
steps from finishing, thet i, is made up of all locations (a, b) such that a> 3 and b <2.
12.0
Show that Trigger Zone Il is made up of all locations in which firm A is between 3 and
5 steps from finishing and firm B is between 2 and 4 steps, that is, is made up of all
locations (a, b) such that 3
