Name: Heath Gibson

Student Number: c9416574

Email: randyte@hunterlink.net.au

This paper will attempt to provide a rudimentary analysis of the market for web browsers. After providing a brief background to the 'browser wars' it will be argued that the most appropriate model for analyzing the interaction in this particular case is a game theory approach. A game will then be developed which will endeavor to explain the strategic interaction which occurs in this market.

The World Wide Web (WWW) is perhaps the best known element of the Internet. Through the use of a 'web browser', users are able to access multimedia information stored on computers across the globe.

When it comes to the market for web browsers, the market is essentially dominated by two products. The first of these is the Netscape Navigator, produced by Netscape Communications Corp. The second of is Internet Explorer (IE), produced by Microsoft. In a market which appears to be changing so rapidly, obtaining data on the precise market share of these two products is difficult. Some suggestions for the relative positions are;

Netscape 70%, IE 28 % (USA Today (1997))

Netscape 64.7% IE 26.7% (iWorld (1997)) *This site updated daily.

Clearly, Netscape is for the time being the dominant browser however the strong growth being shown by IE (Microsoft (1997)) and the fact that the next generation of Microsoft Windows will certainly ship with a copy of IE make it reasonable to expect that this situation may level out.

Game theory only represents one approach to analyzing situations of oligopoly. Two other approaches available include the Cournot and Bertrand Models.

The Cournot Model was named after the Auguste Cournot who developed this particular model to analyze the duopoly which he observed in the market for mineral water. Cournot duopolists make decisions about the quantity they will produce. "The central assumption of the Cournot Model is that each duopolist treats the others quantity as a fixed number, one that will not respond to its own production decision" (Frank (1991) p. 444). But , as it has been pointed out (Frank (1991) p. 444) that this is a very weak form of interdependence. Indeed, one of the main reasons for rejecting a Cournot approach was that it did not sufficiently reflect the level of strategic interaction between Netscape and IE.

Furthermore, one of the assumptions of the Cournot model is that the product is identical. Whilst the basic function of both products is to enable the user to browse the WWW, one of the key features of the interaction between the firms is competing innovation. Both firms are competing to improve the capabilities of their browsers. These provide some reasons why game theory was favored over the Cournot model for examining this situation.

Another commonly used approach to analyzing duopoly is the Bertrand Model. This essentially arose as a response to the Cournot Model and suggested that the appropriate variable which firms would change would be price. Again, this model assumes that the rival firm will not respond to the change in price. However, as highlighted before, the main interaction between Netscape and IE lies in the competing innovation, and also in the expansion of usage of the browser. Indeed, both browsers are effectively free to any user once they have connected to the internet. It may be though that this situation is because the firms have reached a Bertrand equilibrium in a situation where marginal cost is zero - which is essentially the case where a copy of a browser is distributed electronically over the internet. However in a Bertrand equilibrium the market share held by each firm is equal. This is presently not the case although the market may be moving towards this position. But even so this still fails to account for innovation and perhaps there is a better model given the degree of interaction of these firms.

Given these considerations a game theory approach perhaps allows the best examination of the strategic interaction in the browser market. Perhaps the foundation literature in this field is von Neumann and Morgenstern's book The Theory of Games and Economic Behavior (von Neumann & Morgenstern (1944)). Game theory has been described as "A set of tools for analyzing decision making in situations where strategic behavior is important"(Katz & Rosen (1994) p. 544). As has been alluded to previously, the browser market is one where a great deal of strategic behavior would be expected. Game theory thus appears the logical choice

Essentially this market is characterized by a duopoly between Netscape Navigator , produced by Netscape Communications Corporation, and Internet Explorer – a product of Microsoft. These two browsers currently dominate the market, with other browser such as Maczilla and Lynx making up a relatively insignificant portion of the total browser market. The model constructed to explore this situation, though constrained by several limiting assumptions, seeks amongst other things to offer some explanation for why these two firms continue to engage in innovation of their products.

We now attempt to construct a game theory model to examine the browser wars. It should be said that perhaps one of the most obvious flaws in the model to be constructed is that it assumes the browsers hold equal market share. As was highlighted earlier, most sources seem to regard Netscape as holding a much greater share of the market than IE. IE however has been growing and a model which analyses these as equals may soon be relevant (or indeed already is if one believes the Microsoft press release cited).

This game starts out with the following rules.

• The two browsers are assumed to be of equal distribution in the market. This is reflected in the game by the initial symmetry of the payoffs.
• The firms have two choices. Expand or innovate, with the payoffs listed in the game square.
• The payoff’s represents distribution/sales of the product. They are relative strengths.
• The firms play pure strategies.
• The game is not a zero-sum game.

Given the payoff matrix constructed in 1a. The dominant strategy for either firms in this situation is to expand. A dominant strategy equilibrium occurs with both firms choosing to expand. The payoff for both firms is 1,1.

Now, let us alter the situation slightly. Let us play the game in a second round where the choice from the previous round can effect the second round. Specifically, let us change the rules such that :

• Where a firm chooses innovate, they receive +1 to their payoff for each successive round.

Innovation leads to an expansion in sales for the innovator but not at the expense of the other competitor as the market expands in response to the improvement in the product.

Since the game is symmetrical let us consider what the payoff would look like where Netscape chooses to innovate in round one. The first round matrix is the same as in 1a. However the second round matrix in this case is 2a.

The dominant strategy for IE remains to expand. The dominant strategy for Netscape in round two is also to expand. Over the two rounds, the total payoff for both players is 2.

Now, what if we extend to a third and fourth round. Assuming that after round one (i.e. from round two onwards) both firms play the dominant strategy of expand. The matrix for each round from round two onwards is the matrix 2a. We can then see that Netscape continues to increase the size of the gap between its total payoff and IE’s total payoff. If both firms continue to play expand then after ‘n’ rounds, Netscape’s total payoff will be:

IE’s total payoff after ‘n’ rounds will simply be

TP = n+1.

The gap between the two will be

n-3 (where n>1)

This would appear to suggest in this particular case that whoever innovates first will, from round four onwards, hold a greater proportion of the total market and will continue to expand their share of the market.

But what if the firms realize that the first to innovate will gain this advantage. Both firms will then engage in innovation in the first round. The payoff for the first round is given by 1a. However since both firms have chosen to innovate in round one the payoff matrix for round two is given by 3a.

The dominant strategy for both firms in the second round is, as in 1a and 2a, to expand. If both firms continue to play the dominant strategy of expand from round two onwards then the relative positions of the firms will not change and their total payoffs over ‘n’ rounds will be given by:

TP = 2(n-1).

Now let us consider the impact if Netscape chooses to innovate again in the second round then continue with expansion from round three onwards. The payoffs for the next two rounds can be represented by 4a and 4b.

In this case the payoff for NE is

The payoff for IE if it continues to choose expand from round two onwards is

TP = 2n-2.

This means that the gap between the browsers is given by

n-4 (where n>2 )

Similarly, if IE were to innovate in round two then expand from round three onwards and Netscape innovate in round three then expand from four onwards , then the size of the gap would be

n-5 (where n >3)

Expanding this mode of analysis we are able to arrive at a general set of results. Where :

• Both firms innovate for i_r rounds ;
• The round of the game is denoted by n ;
• One firm chooses to expand in round r_e whilst the other innovates; and
• From round r_e+1 both firms choose to expand.

Gap in Total Payoffs = | n – ( r_e + 2)| where n > r_e

Earlier it was seen that whoever innovates first will gain an advantageous position. The above condition suggests that whoever ceases innovating first will have their payoff equalized with their opponent in round r_e+2 and will fall behind from r_e+2 onwards.

This has an important implication for the firms. If the game is to played over n rounds then both firms must continue to innovate since once they cease innovating then their opponent need only innovate for that round (r_e) then expand for all proceeding rounds in order to achieve a gap in total payoffs or the above mentioned proportion. This helps greatly to explain the continual innovation which takes place in the 'browser wars'.

Are there any implications for government policy from this model? Government policy in the case of oligopoly is usually directed at ensuring that the parties involved do not involve themselves in collusion as this usually results in sub-optimal outcomes when compared to the perfect competition outcome.

In the case of the 'browser wars' it is difficult to see a basis for government intervention. The firms providing the browsers are utilizing distribution technologies which mean that the marginal cost of distributing extra copies is virtually zero and the price of the browsers to the consumer is effectively free. Furthermore, the game theory model presented suggests that whilst ever the game is to be played over n rounds, rather than a finite number of rounds, the firms have an incentive to continue to innovate. Although clearly there is more scope for analysis of the welfare implications of this situation there does not appear to be any basis for government regulation in this market.

The 'browser war' presents a situation of duopoly in an unusual market. The marginal cost of distributing additional units is almost zero and the price faced by users already connected to the internet is also zero. Duopoly models based around price and quantity may therefore not be the best tool for examining the high level of interdependence in this situation. A game theory approach better explains the interaction, in particular why there is such extensive innovation when the game is played over n rounds.

Although a complete analysis of the implications of this situation for government policy was not conducted, there appears no obvious role for government intervention in this situation.

Frank, H (1991) Microeconomics and Behavior, McGraw Hill, New York.

Katz, M and Rosen, H (1994) Microeconomics, 2nd ed, Irwin, Illinois.

iWorld (1997), BrowserWatch - Stats Station [On-line], Mecklermedia, Westport. Available: http://browserwatch.iworld.com/stats/stats.html

Microsoft Press Pass (1997), End User and Corporate Demand for Microsoft Internet Explorer Reaches Major Milestones [On-line], Microsoft, Redmond. Available: http://www.microsoft.com/corpinfo/press/1997/Apr97/IE15mipr.htm

von Neumann, J and Morgenstern, O (1944) Theory of Games and Economic Behavior, Princeton University Press, New York.

USA Today (1997) Netscape yields $8M profit despite competition [On-line], USA Today Online - High Tech, The Associated Press, Arlington. Available: http://www.usatoday.com/life/cyber/tech/ct709.htm