This article, which features two UF marketing professors, discusses the difference between “precise” prices like 19.95 and “imprecise” prices like 20.00. Still, the article doesn’t really address the question of this post.
To be more specific, let’s consider retailers and grocers selling relatively low cost goods in high quantities. These are the industries that we usually consider to be most competitive. First, are goods actually priced as claimed? I haven’t seen any data regarding typical prices at these stores, but I believe that anecdotal evidence is enough to say that sellers typically do not price at whole numbers.
There are several potential explanations I have heard for this phenomena:
- To force employees to open the cash register and thus record the transaction. (supposedly this practice was first instituted by James Cash Penney.)
- To accommodate a newly added sales tax (of around 1 to 5 cents).
- 19.95 looks psychologically less than 20.00, thus increasing quantity demanded by more than one would expect from a very small change in price.
The article linked above discusses bargaining as a potential explanation–although this doesn’t seem reasonable when considering retailers. One typically does not bargain with such firms (I am thinking of Target, Albertson’s, etc.), although if we consider smaller owner managed firms, this explanation seems more likely (e.g., produce stands at farmer’s markets).
The article primarily discusses the possibility that a ‘more precise’ price signals that the cost is closer to the asking price and thus buyers will bid closer to the asking price because they know that the seller cannot profitably stray to far from the asking price. They consider some data from the real estate market to test this hypothesis.
Anyway, back to the original question. Explanation 3 doesn’t seem valid. Many people are innumerate, but do so many people really make such bad calculation mistakes that this practice is profitable? Explanations 1 and 2 make the most sense to me, and we need some further historical evidence to evaluate these claims.
Some other correlational evidence that might support these explanations: If we see a correlation between the price distance (20.00 - 19.95) and the local sales tax; and if we see a positive correlation between the percentage of customers paying with cash and the existence of “19.99″ pricing.
Suppose that the latter correlation does not exist. That is, even in firms with a high percentage of customers making non-cash payments, we still see “19.99″ pricing. If we also suppose that “19.99″ pricing is simply a relic of the past, then why does this practice remain? Perhaps some industries are simply so competitive that, given a baseling of 19.95, a 5 cent price increase would lower profits?
One final note: Wal-Mart typically does not price in whole numbers. Their prices, anecdotally, appear to be evenly distributed across the spectrum. They have some goods with whole number prices, some goods ending in .86, some ending in .37, .54, .93, etc. Perhaps this is a signal that they are always trying to provide the lowest price, even if it’s just by a few cents.
I actually always thought it was 3! Stuff like bargaining and especially tax doesn’t really explain the universality of this pricing scheme. I’m thinking currently of car prices, which ALWAYS seem to be advertised as “$14,990, drive away!” (or similar). People always know they can bargain down car prices, and I’d be surprised if the tendency to do so diminishes when the price is reduced. The fact that this pricing is so universal probably diminishes this effect of ‘oh, it’s precise, must be close to the best price they can give’. I might agree with it more if it only happened occasionally.
I dare say you’re giving the average consumer too much credit. I am so often frustrated by the irrationality of the average consumer, but then that’s what I get for being an economist. Often when you look at a price, the first thing you notice is the first couple of numbers. Or at least I do. And I think that something being in the ‘teens’ in price rather than the ‘twenties’ (for the 19.95 price example) is not something that would convince someone on a conscious level, but subconsciously it might just be enough to grab their attention that little bit more and, at the margin, make slightly more of a difference than that simply attributable to a 5 cent price difference.
Honestly, I am baffled by your insouciant dismissal of explanation #3 as well. It seems like it would be very convincing, especially given the recent research into consumer perception of prices and how much it does, in fact, strongly affect them.
If anything, the first explanation that should be dismissed is #2. That explanation would only accommodate a new sales tax if the good was in the $1-$2 range, not for the vaaaaaast majority of goods which are not in that subset of priced goods. I think explanation #1 makes sense and I think that the explanation for Wal-Mart’s pricing sounds about right. Instead of imprecise pricing, in which they would lose profit, they go for the exact lowest they can go while maximizing profit.
First: I don’t believe current taxes explain the scheme, but rather past taxes. Now the scheme would be mostly a relic of the past. However, you two do argue a good case: taxes probably aren’t the explanation.
I also agree that we’d need to investigate the bargaining explanation in the auto market and the real estate market a little further. What Mark makes a lot of sense: If too many firms incorrectly signal precision, then consumers will figure things out and the gain will be eliminated.
Second: We can discuss “rationality” and related topics in more depth later, but briefly…. “rational” is not well defined in economics. Everyone seems to have a different definition of what it means for consumers to be rational. For me, I say a consumer is rational if they attempt to do what’s best for them given their constraints. That is, a rational consumer will not knowingly do something that’s not in their best interest.
This definition allows for all kinds of behavior that others term “irrational” to be rational. This kind of behavior can arise from calculation costs, complex preferences, or whatever. The main point is that a rational consumer does what’s best for them, given their limitations.
I strongly believe that all consumers are always rational under this broad definition. Furthermore, I think it is likely impossible to prove that consumers are not rational under this definition, and it must be taken as an axiom. I have no problem with this. Anyway….
Back to the topic though, above and beyond this broad definition, I do believe calculation costs are relatively low for most consumers, although I haven’t seen any data. I really just don’t think consumers are fooled by this pricing scheme. Besides, we really need to see the data. Well…I was about to say that I think most goods are not priced right below a whole number. But then I went to my local grocer’s weekly ad…lol. Anyway.
I suppose my argument is quite weak and I’ll have to work on it, but my intuition just says that the psychological explanation is wrong: people can round 21.99 up to 22. They know that 21.99 is almost 22 dollars. They don’t think it’s 21 dollars or 21.50, or 21.95. But again, maybe it comes down to calculation costs? This leads into a general equilibrium problem I think, but perhaps I should end my rambling now and gather my thoughts for a bit.
The “19.99” pricing practice may relate to impulse buying (which is not irrational, but I don’t want to get too much into my thoughts on that here).
Admiral, you cite “recent research into consumer perception of prices,” do you have any specific papers in mind? Are you referring to the tax salience research? I’m not sure that research informs our question about “19.99” pricing too much.
P.S., Megan McArdle recently mentioned the psychological explanation for this practice, although only in passing, and she doesn’t give any evidence either way.
First, your response to my point isn’t a response. I mean, okay, great, so it applies to past taxes. That doesn’t change the fact that your whole reason would *only make sense* for the $1-$2 range. Maybe there was a time in human history when most prices in the United States were in that range, and it might have seemed convenient to price them so, but the pricing anomaly you cite applies to cars and all kinds of far more expensive goods as well. Moreover, there’s no reason that the vast majority of firms across all manner of goods would keep that scheme just by inertia that I can see unless it is related to the other explanation. In short, your favored reason is as unlikely as it is unpersuasive. It doesn’t matter if it applied to prices in 1900, 1950, 2000, or today.
Second, the psychological explanation is much, much, much stronger and it is certainly related to what Mark highlighted. [[ MATERIAL OMITTED. ]]
Third, I think that the recent tax salience papers (but mostly Chetty, et al (2007)) combined with that neurolinguistics research would indeed make for a pretty paper that would inform our discussion quite a bit. But that’s me.
So Admiral and I argued about this for about 2 hours the other night and we came a little bit closer to agreeing (read: I retreated a bit more).
First, regardless of whether there is a perception effect or not, let me discuss an interesting point that arose in our conversation.
Veblen effects occur when consumers care explicitly about prices of goods. The main examples are luxury goods. With Veblen goods you get an upward sloping demand curve—consumers prefer to have a good that is scarcer to a good that is more available. I haven’t studied the specifics of this effect much, but I believe this paper is the main reference.
In my micro class last fall, Dr. Slutsky noted that the primary analysis of Veblen effects assumes that consumers care about relative prices.
In standard demand analysis, absolute effects don’t matter. Double all prices and everything stays the same and no one cares. –This is no longer true if we allow for right digit effects (as “19.99” pricing is apparently called in the literature). If we increase all prices by 5 cents then consumption is going to change significantly. Hence absolute effects matter.
I’m not sure what the implications of this are, but I thought this relationship between the analysis of Veblen goods and the analysis of right digit effects was interesting; they essentially fit into the same framework—consumer preferences depend on prices.
As a side note, I drew up what I think a demand curve would look like given right digit perception effects:
I’m not sure if people actually estimate the shape of demand curves to such an extent but this would be one test of the perception theory.
——————————————————
Admiral’s first post on Stigler’s text discussed the evolution of preferences over time. Consider an even smaller scale: the evolution of a single consumer’s preferences. Perhaps such perception effects exist only in the early stages of a consumer’s preferences, that is, when they are still young. Or perhaps they appear as consumers age? Allowing perception effects to appear and disappear, rather than arguing that they are always present, is a much stronger argument.
So in the aggregate, we may see 19.99 pricing remain at a relatively constant level, but it is targeting a small, changing, subset of consumers.
So how would preferences change to remove or allow such perception effects? It follows from decision costs. If the brain starts out with a bias towards towards the first few digits on the left, consumers can perhaps (indirectly) train themselves to remove this bias as they gain more experience in the marketplace. Similarly, as one grows older and mental capacity diminishes, the bias may resurge.
——————————————————
Here is some research I dug up related to this debate. I couldn’t find any research in strictly economics journals though, all of it comes from marketing research. I haven’t read these studies yet so I can’t comment on their quality. For now I want to mention them for reference only.
Keith S. Coulter and Robin A. Coulter. “Distortion of Price Discount Perceptions: The Right Digit Effect” Journal of Consumer Research: August 2007.
Robert M. Schindler and Patrick N. Kirby. “Patterns of Rightmost Digits Used in Advertised Prices: Implications for Nine-Ending Effects” The Journal of Consumer Research, Vol. 24, No. 2 (Sep. 1997), pp. 192-201 (article consists of 10 pages) Published by: The University of Chicago Press
Hooman and Estelami. “The computational effect of price endings in multi-dimensional price advertising” Journal of Product & Brand Management (1999). Volume: 8, Issue: 3, Page: 244 – 256. Publisher: MCB UP Ltd