Activities

In an attempt to better-predict the effects of proposed mergers in concentrated industries, competition authorities and other analysts are showing increasing interest in simulation models.  These models combine certain assumptions about the nature of competition in the market with some premerger market data to predict the impact of the merger on important variables such as price, consumers' surplus and total welfare.

We are pleased to make available this simple simulation program for teaching, research and policy purposes.  The program has been developed by Professor Werner Antweiler.  

About this simulation program

To simulate a merger, the analyst needs to first make some informed assumptions regarding the

general nature of competition in an industry.  In this program we assume that the products sold by the firms in the market are reasonably homogeneous, so that there is only one price in the market at any one time.  In addition, it is assumed that competition is of the "Cournot" type -- that is, individual firms choose their outputs based upon conjectures regarding the output choices of their rivals, but without believing they can affect those choices.  We assume that the firms find "equilibrium" (i.e. that their conjectures are correct).

An interesting alternative approach, more appropriate for many markets, would be to assume that the products sold are differentiated and that individual firms are free to pick their own prices.  A differentiated-products model, however, requires that the analyst provides more information about the substitutability of every product for every other product in the market.  For a simulation model with these characteristics, please see the reference to more sophisticated simulation models below.   

The Merger Simulation Tool includes some assumptions about the functional forms of demand and costs: it assumes that demand curves are linear and that marginal costs are constant (but not equal across firms).  

What data you need to run a simulation?

The analyst interested in simulating a merger with this program needs the following data: 

     (i) the premerger market price (if there are a variety of products in the market a weighted average price might work as a proxy);

     (ii) the number of firms in the market and their market shares -- these shares must sum to 100%.  These must be shares of sales, not capacity.  If there is a fringe of minor firms about which little is known, there are two approaches to consider.  If the sum of their outputs is still a small fraction of total output, they could simply be ignored for the purposes of the simulation.  That is, the market shares are based on output not provided by the fringe and so sum to 100%.  The alternative is to treat the fringe as one firm.

     (iii) total market sales (in dollars, not units)

     (iv) the market elasticity of demand -- if this is not well known, the analyst can experiment with several candidate elasticities.  There is also a feature of the program that allows one to graph the results on price and surplus as a function of different market demand elasticities. One limitation here -- the program will not accept a market elasticity number that is not larger than the largest market share.  For example, if the largest firm in a market had a share of 50%, the program will not accept an elasticity of 0.5 or less.

In addition, this program allows the analyst to incorporate his/her own assumptions about the effects on fixed and marginal costs of the merger.  The analyst is allowed to choose whether the program considers the merged firm to have marginal costs equal to the lowest marginal costs of the merging firms, the highest marginal costs of the merging firms, a weighted (by premerger sales) average of the merging firms' marginal costs, or equal to some number specified by the analyst.  In addition, any anticipated savings in fixed costs can be incorporated.

How the program works

Entering the required data is a simple exercise that should take only a couple of minutes.  The program uses the data on market sales and the elasticity of market demand to solve for the market demand function.  Knowing this function and the premerger market shares, the program can infer what individual firms' marginal costs must be, using the first-order conditions from each firm's profit maximization problem.  Once this is done, we have a full model of the market and the program can determine the effect of the merger on price, concentration, consumers' surplus, profits and total surplus.  One caution: the total surplus number does not incorporate any fixed costs, though the change in total surplus does include any allowance for reduced fixed costs provided by the analyst.

For a sample of what the program can do, call up Widgets from the Previous Models file.

To use the Merger Simulation Tool, click here.

More sophisticated merger simulation programs can be found at:  

There is an "antitrust" package at CRAN (maintained by US DoJ):

https://cran.r-project.org/web/packages/antitrust/index.html 

with links to the reference manual.  From an R session, using the package is as simple as

R> install.packages("antitrust")

R> library(antitrust)

which automatically downloads the latest version from CRAN and unpacks it.  R itself can be downloaded (at no charge) from CRAN at https://www.r-project.org  and by following the link for your operating system (win/mac/linux).