Chapter 7 focused on a set of issues that affect all firms that are trying to construct a locational strategy that spans international borders.

We argued that firms should select from amongst many possible international production strategies the one that best resolves the tradeoffs

between the four elements: factor advantages, trade costs, market sizes,

and scale economies. My view is that the four elements should be the

most important influences on where to deploy a firm's resources. If

building a factor in, say, Vietnam, makes no sense based on the four

elements, then it probably is a bad idea. Nevertheless, there is another

consideration that many managers invoke to justify the location and

timing of their outward investments: the actions of rival firms. In this

chapter we will consider how interactions with competitors affect location strategy of the multinational. We will examine some simple situations in which the location and timing of investment decisions depend

critically on the actions of a firm's competitors.

9.1 Motivating Examples

To set the stage for the analysis that follows, we will briefly review a

few episodes that may illustrate competitive interactions in practice.

9.1.1 SUVs in the USA

In 1990 both BMW and Mercedes served the world's largest auto market using their factories in Germany. During the 1990s this changed,

with both firms establishing factories in states in the Southeast of the

US. The following time-line shows when the factories were established

and also how BMW altered its production plans. The time-line also

shows later investments by Hyundai and Nissan in the same region of

the US.

1992 BMW picks Greer, South Carolina, as the site for a car production

plant.

1993 Mercedes announces it will build sport utility vehicles (SUVs) in

new factory in Vance, Alabama.

1994 BMW plant begins production of the 318i at Greer.

1995 BMW plants starts to produce Z3 roadster at Greer.

1997 Mercedes begins M-class production at Vance.

1998 BMW decides to expand the Greer plant and use it to produce

the X5 (a new SUV).

2002 Hyundai announces plant in Montgomery, Alabama. Production

to begin in 2005.

2003 Nissan begins car production in Canton, Mississippi in May.

2003 Mercedes announces $600 million expansion of the Vance plant,

with 2000 employees to be added by early 2005.

Mercedes seemed to follow its rival BMW in producing vehicles in

the Southeast of the US. However, BMW followed Mercedes in producing an SUV there (although BMW called the X5 a "sports activity

vehicle").

Retail in China

With the world's largest population and decades of rapid growth in per

capita income, China has become one of the largest retail markets in

the world. Now Wal-Mart (the world's largest retailer, based in the US),

Carrefour (the second largest, based in France), and Metro (the third,

based in Germany) have all entered China. The timing and location of

their entries are highly suggestive.

Carrefour was the first to arrive, setting up its first "hyper-store"

in the nation's capital, Beijing, in 1995. The following year Wal-Mart

established its first super-center in Shenzhen, a rapidly growing exportoriented city near Hong Kong. Also in 1996, Metro opened its first store

in Shanghai. Since then all three firms expanded considerably. As of

early 2006, the first mover, Carrefour had the largest presence, with

70 hypermarkets and 225 discount stores. Wal-mart had 56 stores but

later in 2006 announced plans to expand via an acquisition. Metro was

the smallest, with 30 discount (cash-only) outlets.

KFC and McDonalds established their networks of franchise restaurants around the same time. KFC moved first and established a restaurant in Beijing in 1987. McDonalds moved three years later and opened

its first restaurant in Shenzhen. The company did not cede Beijing to

KFC and instead established its largest restaurant in the world there

in 1992. By 2006 there were 680 McDonalds restaurants in China, considerably less than the over 1400 KFCs.1

9.2 Key Concepts

Analyzing the situations where actor's payoffs are interdependent requires the use of some game theory. There are two key concepts that

need to be introduced right away: strategic complementarity and firstmover advantages.

Strategic complementarity means that if you do more of something,

then the payoff to me of doing that thing increases. The alternative

is strategic substitutability. In that case you doing something makes

that thing less attractive to me. To take an example from outside of

location strategy, consider the purchase of advertising time during the

Superbowl. If Pepsi's decision to advertise during the game makes it

more attractive for Coca Cola to advertise then (and vice-versa), we

say the firm's advertising decisions are strategic complements. Location

decisions would be strategic complements for Japanese auto-makers if

Honda's decision to manufacture in some state or country made that

place more attractive to Toyota.

First-mover advantages (FMA) is a notion that is invoked frequently,

but rarely defined unambiguously. The basic idea is that the first one to

do something, obtains a lasting advantage over subsequent imitators.

The "something" in question is often the introduction of a new product.

Thus one could debate whether the Apple Ipod's strong market share

(as of 2006) and ample profit margins derive from being the first mover

in the digital audio player market. This is a topic for general strategy

books. Our interest here is in the location aspect of the FMA. Does the

first firm to establish operations in China retain a sustained advantage

over subsequent entrants? If so, we say there is an FMA.

Strategic complementarity is important for deciding whether to invest in a region or not. The presence of first-mover advantages determines the optimal timing of entry into region. In the next two subsections we explore the underlying forces that lead to strategic comple-

1 The information on the entry of Carrefour, Wal-mart, KFC and McDonalds into

China was collected by Ran Jing from the Chinese language websites of each

corporation. Roberts (2005) provides a good overview of the recent "retail wars"

in China.

mentarity and first-mover advantages. This is important if one is to be

able to apply the concepts in practice.

9.2.1 Sources of Strategic Complementarity

There are very different mechanisms behind strategic complementarity

in MNE location decisions that have been considered by economists

and management professors.

? Agglomeration economies or Cluster advantages: Marshall (1920)

and Porter (1990) described how groups of related firms often perform better (higher productivity, more competitive products) when

they choose geographically proximate locations.

? Informational herding: Choices made by others may reveal information they gathered on the attractiveness of a location.

? Oligopolistic reaction: Knickerbocker (1973) argued that in industries with a small number of firms, the follower matches the leader's

move to maintain competitive stability.

The most well-documented cause of strategic complementarities are

mainly referred to by economists as agglomeration economies. They are

similar in some respect to plant-level scale economies except that they

occur at the level of a region, rather than a factory. The idea is that

the greater the scale of an activity at a regional level, the lower will be

the average costs of undertaking that activity.

Marshall (1920) described the three classic causes of agglomerations

economies. The first is what we now call "knowledge spillovers." In an

often-quoted passage, Marshall wrote that when people in the same

line of business locate near each other,

"The mysteries of the trade become no mysteries; but are as it

were in the air, and children learn many of them unconsciously.

Good work is rightly appreciated, inventions and improvements

in machinery, in processes and the general organization of the

business have their merits promptly discussed: if one man starts

a new idea, it is taken up by others and combined with suggestions of their own; and thus it becomes the source of further

new ideas."

While his language style seems dated, his ideas seem so current that

one almost thinks he must have been talking about Silicon Valley.

The second point has to do with what Marshall called "subsidiary"

industries and Porter calls "related and supporting" industries. Marshall's idea was that the bigger is the local cluster in the downstream

industry, the more suppliers will be willing to sink capital into specialized machinery. Porter tells a similar story: a cluster of downstream

rivals will stimulate the formation of a cluster of proximate upstream

suppliers. These firms will create dvantages for the downstream firms

because they "deliver the most cost-effective inputs in an efficient, early,

rapid, and sometimes preferential way."

Finally, Marshall argued that groups of firms attract skilled workers

and vice-versa.

"The owner of an isolated factory, even if he has access to a

plentiful supply of general labour, is often put to great shifts

for want of some special skilled labour; and a skilled workman,

when thrown out of employment in it, has no easy refuge."

Porter (1990) argues that on top of these benefits, by locating near

your strongest rivals, you set into motion forces that will ultimately

make your firm more competitive. The presence of local rivals creates

"pressure on companies to innovate and improve." The rivals vie for the

"bragging rights" of being the best in the cluster and they can offer "no

excuses" for relatively poor performance. In the short run, then, local

rivalry makes life rather unpleasant for managers. But Porter assures

us that in the long run, "dynamic improvement" will create sustainable

competitive advantages.

The Marshall and Porter analysis of industry clusters has implications for multinational firms. When other MNCs in the same industry

establish operations in a country, then this will create cluster advantages for the remaining firms in the industry. For example, Ireland has

established a strong cluster in the information communications technology (ICT) sector. The Irish Development Agency (IDA) reports that in

2006 there were 109,000 employees in the sector of which 45,000 worked

for foreign firms such as IBM, Intel, Dell, Apple, Hewlett-Packard, and

Microsoft. There are 1300 firms in all and seven out of the ten largest

MNCs in the ICT sector have a "substantial base" there. The large

number of suppliers and the IDA's claim that Ireland has the highest proportion of science graduates in Europe both support the Marshall/Porter arguments for clusters based on specialized inputs and

highly skilled workers.

Informational herding leads to a very different type of strategic complementarity. The best example is only loosely related to international

business but it is something that most travellers experience repeatedly.

After checking into a hotel in an unfamiliar city, you set out to find

a restaurant for dinner. Someone suggests a restaurant that looks fine

and has affordable prices. But you decide not to enter because there

is no one else inside. If your reluctance to eat there is simply because

you like the noise and bustle of a crowded restaurant, then this would

just be another application of the agglomeration economies described

above. However, suppose you actually prefer a quiet and uncrowded

restaurant. You might nevertheless avoid this restaurant because you

fear that the lack of customers is a sign of poor quality. Then you would

continue to search until you found a place that was sufficiently full so

as to inspire confidence in the quality.

Multinational location decisions have some of the same features as

choosing a restaurant in a new city. The MNE is often uncertain about

a variety of things. Is there a good market for the product we make?

How productive are the workers? Is the infrastructure reliable? Are the

politicians, bureaucrats, and courts corrupt? One can learn some general features from publicly available data but it is often very difficult

to infer the attractiveness of a location for a specific industry. After

conducting some initial research, the MNE formulates a guess?but it

is just one guess. Alternatively, one can rely more on the information

gathered by other firms. If all the firms in an industry have chosen Ireland as their base for serving the European market, that may be seen as

a persuasive vote of confidence. Even without direct (Marshall/Porter)

cluster advantages, the MNE might decide to locate in the cluster because of the collective wisdom embodied in the other firms' decisions.

The third type of complementarity, oligopolistic reaction, is also related to uncertainty about production and market conditions in a foreign country. Knickerbocker (1973) studied data on US multinational's

overseas investments and concluded that the tendency of firms to invest

in the same countries at approximately the same time was strongest in

oligopolies (industries with a just a few firms). His story emphasized

the damage that one firm could suffer if its rival was the sole producer

in a country that turned out to be a great production location. Knickerbocker's basic thesis is best described in the following quote explaining

how a firm lowers risk by following a rival into a foreign market:

"To illustrate, if firm B [the follower] matched, move for move,

the acts of its rival, firm A [the leader], B would have roughly the

same chance as A to exploit each foreign market opportunity.

Thus for each new market penetrated by both A and B, B's

gains, either in terms of earnings or in terms of the acquisition

of new capabilities, would parallel those of A. If some of A's

moves turned out to be failures, B's losses would be in the range

of those of A. Neither firm would be better or worse off. From

the point of view of firm B, this matching strategy guaranteed

that its competitive capabilities would be roughly in balance155with those of firm A." (page 24-25)

Knickerbocker never wrote down a mathematical version of his ideas so

we can only speculate on what exactly he was assuming. Reading the

story now, it appears to have been influenced by the Cold War notions

of "balance of power" between the USA and the USSR.

One way to think about oligopolistic reaction is that it is a story

in which managers care only about relative performance and are seeking to minimize the probability of being the under-performer. This

might seem superficially appealing until one realizes that a firm that

consistently engages in oligopolistic reaction will pass up opportunities

to earn higher expected profits by taking a different course from that

taken by rivals. As a result, the strategy of matching a rival's moves is

not likely to yield high returns for shareholders.

Thierry Mayer, John Ries and I wrote a paper where we reinterpreted Knickerbocker's hypothesis using a standard model of oligopoly.

We found that firms usually do better by choosing different production

locations: what we called reverse oligopolistic reaction. The reason for

avoiding each other is something called the market crowding effect.

The basic idea is that when a company faces a nearby competitor, it

is forced to charge lower prices and cede a larger portion of the market than when its competitor reaches the market via exports from a

remote production site. The reason is that trade costs insulate a firm

from vigourous competition from a distant rival. This makes it desirable

to put some "space" between us.

The market crowding effect is stronger for goods that are undifferentiated and therefore close substitutes for each other. For example

consider a highway exit where there is currently only one gas station.

If a second one opens at the same exit, they will have to share the

market. They will probably compete harder for the same customers,

pushing down the price. At the national level, consider a country like

Brazil that had high tariffs on imported autos. The first foreign maker

to producer there (Volkswagen) had something of a captive market.

The second firm will tend to "crowd" the market, lowering the price

and per-firm volume.

Choosing distinct locations can also help to mitigate crowding effects in factor markets. If my competitor is seeking the same key managers or production sites as me, we will bid up salaries and land prices.

Unless, there is a vigorous supply response that more than offsets the

demand increase?possible in the case of skilled managers (according

to Marshall) but not conceivable for land?firms would be better off

leaving each other alone.

If market crowding effects are strong relative to agglomeration effects, then firms' location decisions can be strategic substitutes instead

of strategic complements. Since clusters and imitative FDI decisions

seem much more common than firms that choose isolated locations,

we may be inclined to think that location decisions are never strategic

substitutes. That is probably a mistake. One cause of the mistake is a

thing called "common cause."

Why do we see clusters of gas stations near highway exits, wineries

in the Southern Okanagan valley, offshore financing corporations on

Caribbean islands? One reason is that in each of these cases there is

a local attraction that appeals to all potential entrants. In the case

of gas stations, it is the cars emerging from the exit ramp in need

of gas. For the wineries, it is the sunny and relatively warm climate,

combined with the night-time cooling effect of the lake. For the offshore

financing companies it is the low taxes and lenient financial reporting

requirements offered by the tax haven governments (see Chapter 12).

The existence of a common cause driving location decisions can mask

the existence of strategic substitution in a case like the gas stations. It

can also give the appearance of strong strategic complementarity even

when those effects are weak or absent.

Consider the following case. The dean of XXXX School of Business announces a new MBA to be offered in Shanghai, China. When

confronted with skepticism, he announces that other big-name US business schools have launched MBA programs in China and "we need to

get on the boat." The question is whether the presence of other MBA

programs raises the payoff to XXXX of establishing its own program.

The fact that other MBAs have chosen to set up programs in Shanghai

may reflect common cause. The growth of outward-oriented business

in Shanghai has created a pool of managers seeking advanced business

education programs that existing universities in China do not know

how to offer. However, since XXXX and its US rivals will compete for

the same pool of students and even some of the same skilled lecturers, market-crowding effects may be strong enough to imply a strategic

substitutes setting. The XXXX School might be better off avoiding

Shanghai and establishing in another city with similar demand conditions that is not yet served by other programs.

9.2.2 Sources of First-Mover Advantages

We now turn to the second key concept: when is there an advantage

due to moving first? And if you do not get to move first for one reason

or another is it better to be the last mover or should one try to move

as early as possible?

There are two main sources of first-mover advantages.

? scarce resource preemption: In a "first-come, first-serve" situation,

the early firm seizes the most prized resource and later movers have

to settle for less attractive alternatives.

? consumer switching costs: the amount a customer who has been

buying from a given seller must give up (in time, money, or expected

benefit) in order to switch to a different seller's product.

Resources subject to preemption include uniquely advantaged retail

sites and production sites. Sometimes foreign investors seek to obtain

a whole set of such resources by acquiring an existing domestic firm

that has managed to assemble an attractive resource portfolio. The

second best acquisition target may possess a substantially inferior set

of assets. Similarly, in countries where the preferred mode of entry is

a joint venture, there will often be potential partners that are much

more attractive than others.

In a well-functioning market economy, the resource preemption effect may not confer much of an advantage to the first-mover. The reason

is that the market will tend to price resources according to their values. Hence, the most attractive acquisition target is also likely to be

sold at the highest price, lowering the net gains of moving first. Even

a non-market asset, such as a site permit for building a factory, may

be available only at high "price"?in the form of offers the firm must

make to government officials.

The form and magnitude of switching costs depends on the nature

of the good. One distinction is between search and experience goods.

In the case of the former, attributes of the goods are known before the

first purchase?all that is required is a search to find the product. If

a new firm enters a market, consumers may be willing to switch easily

as long as they are aware of the new entrant's product and therefore

do not need to search for it. For experience goods, the consumer learns

the quality of the product after buying it and trying it and sometimes

the evaluation period is protracted (e.g. cars). Experience goods are

likely to engender greater consumer loyalty because once the buyer

has identified a product that works, she may be reluctant to take the

risk of trying one that will not work. A third type of good requires the

consumer to invest time in learning how to operate it. Examples include

software, video games, and some types of machine tools. This learning

is an irreversible investment and it will take a large inducement in terms

of superior quality or price to persuade some customers to repeat their

investment in a new product.

When switching costs are large the firm that enters first in effect

takes "ownership" of a large set of customers. These customers can be

wooed away by new entrants but only at the cost of selling at a low

(or even negative) profit margin for an extended period. This leads to

first-mover advantages, which extend to a lesser degree to other early

movers. The worst situation is to be a late mover.

There is a powerful force offsetting first-mover advantages: the investments of early movers generate spillover benefits for followers. For

example, a firm that introduces a new product to a country often has

to teach consumers about the desirability and uses of the product.

When subsequent firms enter with competing products, they find an

educated consumer base already in place. The sort of educational process operates for workers and managers. The first mover trains potential

employees for the followers. Furthermore, legal issues involving production permits and government product approval can be resolved at great

cost by the pioneering firm. Followers may find a well-functioning legal

framework already in place. One of the most important spillovers is

probably knowledge about which business practices work well in a new

business environment. If followers pay close attention to the leader's experience, they will be able to copy effective strategies and avoid failed

approaches.

In the history of new products, an early mover rather than the first

mover, usually ends up on top. The Apple iPod was not the first digital

audio player (it had been introduced three years earlier by a Korean

manufacturer named SaeHan). Google was not the first internet search

engine. Excel and Word were not the first spreadsheet or word processor

programs. But all three entered fairly early in the life cycle of these

products and now it would be very hard for a new entrant to dislodge

them as the dominant players.

The history of country investments by multinationals provides mixed

evidence for first-mover advantages. The two first foreign companies to

manufacture in China were Volkswagen (1984, Shanghai) and Peugeot

(1985, Guangzhou). Fourteen years later, many other firms had entered

but Volkswagen was the dominant player, taking 50% of the market.2

On the other hand Peugeot had bailed out, selling its factory to Honda.

2 New York Times, December 18, 1998.

Another early entrant, Jeep (part of AMC at the time, but later part

of Chrysler), also failed to succeed in China. On the other hand, GM

entered China relatively late, producing its first car there in 1999. Nevertheless China became a profitable market for GM and its market

share there moved ahead of VW in 2005 (11% versus 9%). This is of

course just one case and the Chinese auto market is very dynamic,

with late entrant Toyota expected to see its current market share of

about 3.5% increase substantially as a result of recent investments.3

The point of this case is to illustrate that while we cannot dismiss the

importance of first-mover advantages, we should not exaggerate them

either.

9.3 Multinational Location Games

We now use some simple game theory to explore how the concepts

from the previous section influence competitive interactions. To make

things concrete, let us consider the case of Mercedes and BMW trying

to decide where to place factories to assemble their new Sport Utility

Vehicles (SUVs). We will assume throughout this example that plantlevel economies of scale are so important that each firm will only have

a single factory. The German location has the advantage of placing

the SUV factory close to parts suppliers and their main design engineers. The SUV line might even be added within an existing factory.

Despite their high wages, German workers might offer factor advantages based on their high skill in making luxury cars. Offsetting these

considerations is the savings in downstream trade costs achievable by

locating production within the major market for high-end SUVs: the

USA. There are lower transport costs, better feedback from customers,

and avoiding possible tariffs on luxury vehicles that might emerge in

a future trade conflict. Thus, the basic attractiveness of Germany and

the USA depend in large part on the elements of multinational strategy

that we brought together in Chapter 7. Each of these elements applies

even if the rival company were not planning to produce an SUV at all.

How then do the BMW and Mercedes decisions relate to each other?

First consider the payoffs in Table 9.1.

It is not difficult to see that these payoffs strongly argue for the

two companies to make different decisions. To see this, we should look

for the Nash Equilibria. A Nash Equilibria is an outcome of the game

that neither side would want to deviate unilaterally from. This gamehas two Nash Equilibria: (BMW in USA, Mercedes in Germany) and

(BMW in Germany, Mercedes in USA) In both situations, only one

firm produce in the US while the other make its SUV in Germany.

Underlying these payoffs are market-crowding effects. If Mercedes is

the only firm in the USA it can sell more SUVs at higher prices than

if it faces local competition from BMW. Thus, strategic substitution

implies that following your rival makes no sense. If Mercedes chooses

to produce in the USA, BMW will be better off producing in Germany

and concentrating on sales to the European market.

The payoffs in Table 9.1 were chosen to neutralize the effects of factor advantages and market sizes that would normally bias the location

decision in favour of one country or the other. This situation artificially

made each firm indifferent as to the two countries for any reason other

than the rival's decision.

Suppose instead that the USA market is much larger and its factor

costs are quite competitive as well. Then both firms would rather jointly

locate in the USA than jointly locate in Germany. Table 9.2 provides

some numbers in which the USA is intrinsically preferred. This case

also features very large market-crowding effects. One way to see this is

to sum the profits of the two firms. The two firms could earn a total

of 7 by choosing different countries. However, if they matched location

choices, combined profits could be as low as 1 + 1 = 2 if they chose

Germany.

What is the equilibrium of this game? If the firms chose locations

simultaneously, BMW in the USA and Mercedes in Germany is a Nash

Equilibrium but so is Mercedes in the USA and BMW in Germany.

If one firm could move first, it would select the USA. The rational

response of the rival would be to choose Germany. This would give

it more profits than co-locating in the USA (3 vs 2) but less profit

9.3 Multinational Location Games 161

than the first mover (3 vs 4). Thus we have a case with a first-mover

advantage caused by a combination of location differences and marketcrowding. Even though the USA is considered the better location, and

it is the place the first mover selects, it is unwise, given these payoffs

for the other firm to follow it.

The payoffs in Table 9.2 do not help us to understand what actually

happened since they predict that only one of the two firms would have

selected the USA. There are two distinct ways to alter the payoffs to

yield a game that predicts both firms choosing the USA.

First, we could strengthen the inherent advantage of the USA. To see

this, increase the payoffs of choosing the US by two, regardless of what

the other firm is doing. As shown in Table 9.3, the market-crowding

effects (which would remain since each firm would still improve its profit

if the other firm chose a different country) are no longer strong enough

to separate the firms. Regardless of what the other firm chooses, each

firm prefers the US location.4 Since the follower's location can not be

influenced, being the first mover is of no value. Both firms obtain a

payoff of 4 in either case. What would be useful is for the firms to

get together and agree to choose different locations. Then they could

achieve a combined profit of 9 instead of 8. However, this outcome

would require the firm in the USA to compensate the firm in Germany,

a transaction that might be difficult to execute in practice.

Table 9.4 illustrates a second, and more interesting, way to explain

why both Mercedes and BMW chose to produce their SUVs in the

USA. As in the preceding game, we make the US intrinsically more

attractive. The big difference is that the new payoffs exhibit strategic

complementarity. We can see this by starting from the position of both

firms in Germany. Neither firm would want to unilaterally move to the

US because its profit would fall from 3 to 2. But if both firms were

to choose the USA, their individual and combined profits would be

maximized. This new set of pay-offs yields two Nash equilibria for simultaneous

location choices: Both can produce in Germany or both can produce in

the USA. This set of pay-offs is called a "coordination game." There is

no conflict between the interests of the two firms. Indeed, as long as they

can communicate their intentions to each other, it is easy to coordinate

on the choice that maximizes both firms' profits. What underlies this

payoff structure? First of all, market-crowding effects must be small

enough that they are overwhelmed by other determinants of payoffs.

Market crowding might be weak because the two German companies would be primarily stealing customers from American and Japanese

SUV makers rather than from each other. Alternatively, if transport

costs are low relative to the value of the product, then the market could

be essentially global. This will tend to lead to small market-crowding

effects because far-away competitors are just as important as nearby

ones.

The payoffs in Table 9.4 do not just exhibit a lack of marketcrowding effects; they suggest the presence of agglomeration effects.

Why in practice would two auto plants obtain an advantage from locating near each other? One important reason is that they will share

information about how to produce efficiently in whatever location they

choose. Some information would be related to SUV manufacture; other

information would be related to the location itself, such as tips for

finding qualified workers. More importantly, for the auto industry, the

co-location in the same area of the US will encourage parts suppliers to

set up in the same area, giving both firms access to cheaper and more

reliable sources of components. In summary, a combination of strong

agglomeration economies and weak market crowding effects generates a

coordination game in location choice.

Note that if one firm can choose its location first, this leader would

certainly choose the USA and the second-mover would match locations.

Note the first mover obtains no advantage?and the second mover no

disadvantage?because both firms agree that the USA is their preferred

location.

Table 9.5 modifies the situation slightly to create more complex

situation. As in Table 9.4, there are agglomeration effects. This can be

seen in that collective profit when the firms choose common location

(4+ 3 = 7) is higher than when they choose different locations (2+ 1 =

3). The key difference between the games in Tables 9.4 and 9.5 is that

now Mercedes prefers that they choose Germany (which will give it a

pay-off of 4) while BMW prefers the USA.

Games of this form are often called a "battle of the sexes."5 This

game has a first-mover advantage. If BMW could choose first it would

choose the USA and Mercedes would (reluctantly) follow. BMW would

earn higher profits than Mercedes. BMW's payoff is also higher than it

would have been as the second mover (in which case Mercedes would

have selected the equilibrium in which both firms stayed in Germany).The value of the first mover advantage can be thought of as the amount

of money a firm would be willing to pay prior to starting the game to

be able to move first when the game began. In this case the value is

4 - 3 = 1.

The game illustrated in Table 9.5 makes a general point about location choice. With agglomeration effects (causing firms to want to

locate together) and asymmetric preferences (one firm's payoffs tend

to be higher in a particular location) there is an advantage in being

the leader, because the firm that chooses first ends up selecting where

both will locate.In this section we have seen how to build agglomeration effects and

market-crowding effects into the payoffs for settings in which two firms

are deciding where to produce. We have analyzed five "games" these

firms play using the "normal" (tabular) form of the game. This form

is useful because it allows us to consider simultaneous moves as well as

when one firm or the other has the opportunity to move first. However,

most people find it easier to understand the sequential-move games

when they are expressed in extensive form, i.e. as decision "trees."

Figure 9.1 uses the tree representation for a location game that

is designed to capture some aspects of the Carrefour and Wal-mart

story from sub-section 9.1.2. The tree is set up such that Carrefour

chooses its base first. The upper "branch" corresponds to Beijing and

the lower branch is Shenzhen. For simplicity, we imagine the other

possible locations are substantially less attractive and can therefore be

omitted from the figure. After Carrefour chooses, Wal-mart can copy its

location?the upper "twig" coming out of each upper branch?or select

the alternative twig. The payoffs are shown at the end of the twigs, withCarrefour's profit shown in bold on the left, and Wal-mart's shown on

the right.

If Carrefour is smart, it should figure out Wal-mart's best responses.

Inspecting the payoffs on the right, we see that Wal-mart is slightly better off choosing Shenzhen when Carrefour has chosen Beijing and it is

much more profitable if Carrefour selects Shenzhen and Wal-mart can

have Beijing to itself. The payoffs shown here combine two important

features. First, we have a situation of strategic substitutes: the second

mover wants to avoid the location selected by the first mover. The

likely explanation for such payoffs is market crowding. Second, there is

a basic asymmetry between locations: Beijing is preferable. The aggregate profit when both choose Beijing, 14, is higher than the aggregate

profit when both choose Shenzhen, 10. Of course the market crowding

effect implies that the firms will not want to choose the same location. Thus, the important thing is that when Carrefour selects Beijing

and Wal-mart goes to Shenzhen the profit for Carrefour is higher, 10,

166 9 Competitive Interactions

than in the opposite configuration, 8. By now the astute reader will

have figured out that this game is a sequential version of the game

shown in Table 9.2. The representation and payoffs are different, but

the underlying assumptions and message are the same. The combination of a better location (Beijing's inherent advantages) and strategic

substitutes (due to market crowding) leads to a first-mover advantage.

References

Head, Keith, Thierry Mayer, and John Ries, 2002, "Revisiting Oligopolistic Reaction: Are FDI Decisions Strategic Complements?" Journal of Economics, Management, and Strategy, 11(3).

Knickerbocker, Frederic, 1973, Oligopolistic Reaction and Multinational Enterprise, Cambridge, MA: Harvard University Press.

Marshall, Alfred, 1920, Principles of Economics 8th ed., Book IV,

Chapter X, London: MacMillan.

Porter, Michael, 1990, The Competitive Advantage of Nations New

York: The Free Press.

Roberts, Dexter, 2005, "Let China's Retail Wars Begin," Business

Week January 17 online edition.