In my Managerial Economics class today, we were wrapping up the section on perfect competition. Economic models are often assailed because their assumptions are not always realistic. In perfect competition, we assume every firm is identical — an assumption that is clearly not true in many competitive markets. Farms come in a wide variety of sizes. The assumption of equal size is not a crucial one; it just makes the analysis easier. But more important than the accuracy of the assumptions of a model is the accuracy of its testable implications — does the market behave in ways the model predicts it should? For example, one common attack on Keynesian models is that their aggregate supply assumptions typically assume that when output is high it is because real wages are low and that is why firms are using a lot of workers – however, we usually see high real wages during economic booms. Two  major implications of the model of perfect competition are that prices should be a) relatively stable, and b) rising slower than those in less competitive markets.

Prices must be relatively stable because, if price were trending upward over time persistently, it would indicate positive economic profits in the market and this would invite entry, which would increase supply and bring prices down. Sure, there will be small fluctuations, but they should not be large or prolonged.

Prices should be rising slower than average (slower than the rate of inflation) because of the extreme importance that perfect competition places on cost-cutting. If you want to increase profits, advertising won’t do it — you’re already selling as much output as you can and advertising is pointless when your product is the same as every other firm’s. But every dollar you cut in costs is a dollar of profits. And the large size of competitive markets makes it likely that others will develop innovative technologies that they can sell to all the firms in the market. This is one of the reasons why the relative price of food in this country has fallen so dramatically in the last 150 years — produce a new farm implement, fertilizer, or tool and you have millions of customers, all buying your product so they can have a cost advantage over their competitors (except they all buy it, so the price drops and they still have zero economic profit).

I decided to try to test the implications of the theory to see if this holds. While agriculture is often characterized by perfect competition, most people association agriculture with fruits and vegetables. That adds more complications, as farmers plant crops based on expectations of future prices when the crop is harvested; and plants are subject to changes in weather like droughts and freezing. But agriculture also includes beef, poultry, and eggs. These are not as susceptible to changes in supply due to weather, and it only takes about a month to raise a chicken to 3.5-lb broiler size — a relatively quick time to respond to changes in market conditions.

So I compared the average monthly prices of a few different goods. For competitive markets, I used turkey (a very competitive market, with an HHI of about 50) and chicken. For comparison, I also looked at gasoline (a mildly concentrated market as classified by the Department of Justice, with an HHI of about 1600). The numbers on the horizontal axis indicate months, 180 in total (Jan 1990-Dec 2004). The numbers on the vertical axis are both nominal prices — gas is per gallon, turkey is per pound (I think — it says nominal prices to consumers for frozen birds, but the data does not specify the units) and chicken is a retail price index.

Average monthly prices, January 1990-December 2004

Note that turkey prices rise at less than 1% per year, much lower than the rate of inflation. Chicken starts out at a higher price (it’s an index, not an actual price) and rises at around 2.4% per year. Gas prices rise at around 7% per year, twice the rate of inflation. Notice also that while gas prices have swings of 10-20% per month on a pretty regular basis, chicken and turkey prices have much smaller monthly fluctuations. Chicken prices are incredibly stable. In only two of 180 months is the change in prices greater than 3.5% and the highest change is 7%.  (in absolute value, so we include both increases and decreases) is only 1% per month. In contrast, the largest monthly change in gas prices is a 19% increase, and 40% of the monthly changes exceed 3.5%. The turkey market would be as stable as the chicken market except for those big drops about the same time every year in, you guessed it, November.

If explore the turkey data, you will find that in all 15 years, the month with the lowest average price is November, the month when demand is unquestionably the highest: consumption in November is now about 50% higher than the monthly average. (50 years ago it was 200% higher and 20 years ago it was 100% higher; the dropping increase in November consumption, relative to other months, is due to increased consumption of turkey in these other months, a trend that started in around 1990.) I think the likely explanation is that supermarket competition for such a huge consumer base causes firms to reduce prices, hoping to entice customers to do the rest of their Thanksgiving shopping in their store. It’s useful to have the basic model to help our understanding of markets, but also to acknowledge unique features of each market that can’t be simplified or assumed away, like the fact that turkey sales are heavily holiday-dependent, but chicken sales are not.

These are only three examples, but you can find a lot more markets and I’m pretty certain you would find a similar result: competitive markets are more stable and predictable, and offer consumers lower prices. It’s nice when we can point to observable outcomes in markets and justify our models. They might not be perfect, but sometimes they’re a pretty decent approximation of reality.

(I wanted to include the image below in my comment, but I can’t, so I’m doing it here.)