7
$\begingroup$

I would like to examine the effect of degree of competition (market power?) among buyers on price of a product. I have data on product price, number of buyers with nonzero purchases and a couple more variables in multiple (around 400) different locations over multiple (around 70) consecutive times periods. I am not familiar with the literature on the topic and might be ignorant of some relevant keywords. I have found very few seemingly relevant papers, e.g. Digal & Ahmadi–Esfahani (2002), and I wonder if there are newer or more important ones that I have missed.
(Update: I do not think this paper offers any method that could be used given the type of data I have.)

Could you recommend a paper or a textbook chapter to bring me up to date with the relevant methodology?

References

$\endgroup$
4
+50
$\begingroup$

You might be looking for the term monopsony power or perhaps oligopsony power.

A superficial look at the literature indicates that the Herfindahl-Hirschman index (HHI), so commonly used in assessing seller market power, is also used to measure buyer market power. Recall that the HHI is the sum of the squared market shares measured in percent (so a duopoly with two sellers of 10% and 90% market shares has an HHI of $90^2 + 10^2 = 8,200$.

If you have buyer level figures (not clear to me from your question) then your data allows you to construct local and aggregate buyer market shares. You'll have to convince yourself which is the right market share for understanding the relevant market. For example, if they were dry cleaners, they tend to draw resources from a small area, and each store is probably a distinct market. If they are car dealers, they draw customers from all over, and so perhaps the entire set of sales is the relevant market.

In the follow up comments and question, Hardy asks, "I want to see how the market power is exercised in terms of its effect on price." In general, you can't do this with price and quantity data alone because prices are simultaneously determined by supply and demand. Unless you know whether supply or demand or shifting, you can't know why prices are changing.

If you really knew that "an additional buyer comes into the market at a given location", and you knew something about what their demand schedule looked like, and you knew that their arrival was exogenous to changes in supply, then I think you could trace out the supply curve and learn about the supply elasticity (for example). But I think you generally need exogenous variation in supply to trace out the demand curve (and vice versa).

However, I'm concerned that the arrival of a purchasing customer is an equilibrium outcome (they only buy when they like the price) while the arrival of a potential customer is what actually is in the demand curve and influences the degree of monoposony. To see this starkly, recall that there are a number of competitive settings where even potential entrants can fully discipline a monopolist to provide the competitive price and ensure they earn zero profits. For example, Tirole Chapter 8. The monopsonist version probably works the same way: in the presence of free entry, if the monopsonist earns any rents then other buyers will enter until those rents are competed away. Therefore we cannot generally learn much just from observing the market quantities (what we really want is unobserved). This makes me skeptical that your data can identify what you are hoping to identify.

However, if you can argue, using your knowledge of the institutional setting, that the appearance of positive quantity purchasers are independent of pricing (they are movements of the demand curve and not movements along a demand curve), then I think you trace out the supply curve with your data. Then, you can look for something similar to the Monopsony and the Minimum Wage outcome. In that setting, the minimum wage raises price and quantity demanded. In yours, entry would lead to higher prices and higher quantities. Could you do that in your data? After deciding whatever is the exogenous shift in demand, can you look (perhaps in a regression difference in difference setting (diff 1: markets with and without such entry with the same change in quantity, diff 2: before and after entry) to see if prices go up along with quantities after entry?

| improve this answer | |
$\endgroup$
  • $\begingroup$ Thank you! I am aware of the index. My problem is not measuring the "potential" market power by looking at how many buyers there are and what their market shares are. Rather, I want to see how the market power is exercised in terms of its effect on price. I would like to measure something like by how much the price is reduced when an additional buyer comes into the market at a given location. (I might need to assume the effects to be equal across locations.) I am also interested in studies that have used data like mine so that I could follow, or at least get acquainted with, their methodology. $\endgroup$ – Richard Hardy Oct 18 at 15:05
  • $\begingroup$ I do not yet know whether I have buyer level figures, because I am still about to receive the actual data. Let us assume I have the quantities for each buyer per location per time period. (In the worst case, I will just have the number of buyers per location per time period.) $\endgroup$ – Richard Hardy Oct 18 at 15:09
  • $\begingroup$ Please see my revised answer above. $\endgroup$ – BKay Oct 18 at 18:38
  • $\begingroup$ Your edit brought in lots of useful insight. Thank you! $\endgroup$ – Richard Hardy Oct 18 at 19:41
  • $\begingroup$ What I have learned recently after talking to people familiar with the industry is that supply can be considered constant. I suppose that makes my task easier. I have a hypothesis that there was considerable buyer collusion before but it has been diminishing over time, and the market has become more competitive. I hope I could quantify this by looking at the number of buyers or buyer shares vs. prices. I am interested in finding how much lower the prices were because of collusion which hopefully could be measured by number of buyers or buyer shares. $\endgroup$ – Richard Hardy Oct 22 at 10:05

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.