David R. Henderson  

Taking Consumer Surplus Seriously

PRINT
New from the NBER... What I'm Watching...

In a comment on my December 21 post, "John Papola on Behavioral Economics," John Papola wrote:

[W]hat is consumer surplus really and why does anyone take such an idea seriously?

As someone who takes it seriously--I don't know of any economist who doesn't--I want to explain (1) what consumer surplus is and (2) why I take it seriously.

Consumer surplus is the difference between the maximum amount the consumer is willing to pay and the amount he actually pays. As such, it's simply a measure of the consumer's gain from trade. That's what it is.

Now, why do we take it seriously? There are many reasons. One is that we realize that we can't simply use GDP as a measure of well-being. Let's say someone comes along and radically changes technology so something that was priced positively is now given away. That component of GDP will fall because a zero price times a positive quantity, no matter how large, is zero. But consumer surplus will rise.

Another reason--this is one of my reasons and not all economists will share it--is that it's a way of celebrating our gains from trade. When I teach the concept in class, I tell a consumer surplus story from my life. It's too long to tell here and so I'll tell it as a separate post later this week.

A final reason (this list is by no means complete) is that it explains why so many of us who hate what the TSA is doing put up with it. Think of an important trip I want to make that's over 1000 miles. Now consider the consumer surplus I get from flying rather than driving. It's immense. So I don't want to mess with TSA: I don't want to get on a no-fly list--that would be a huge loss. I don't even want to miss the flight because the consumer surplus loss from being delayed even 4 hours is immense. Indeed, with my time value, everything I said applies to any trip that's over 400 miles each way.


Comments and Sharing


CATEGORIES: Revealed Preference



COMMENTS (27 to date)
Seth writes:

Can consumer surplus be deduced by a third party?

David R. Henderson writes:

@Seth,
Good question. Not typically, because of what I think you're getting at: Values are subjective.
But there are cases where one can deduce another's CS. Imagine an auction in which the top bidder pays what the second-from-top bidder bid. There, your incentive to bid accurately is quite strong. So if you're the top bidder and you bid $100 and end up paying the second-from-top bidder's bid of $80, I, as an outsider, can be reasonably confident that your CS is $20.

Robert writes:

You can conclude that the CS is at least $20, not exactly $20; the winning bidder might have happily paid $120, or more.

jsalvatier writes:

@Robert

In well designed, competitive 2nd bid auctions (http://en.wikipedia.org/wiki/Vickrey_auction) with rational actors the bids should reflect each person's actual willingness to pay.

Eric Hosemann writes:

Re: the TSA

Look at it this way: the surplus gained from consuming jet transport is so great, some people are willing to alienate heretofore inalienable rights to get it.

So far the TSA has claimed only a small portion of that surplus. Imagine what it might do to get the rest.

Philo writes:

I have a headache, so I go to the drug store to buy a package of aspirin. I buy one for $1. What is the maximum amount I would have been willing to pay for _it_, if necessary? Answer: $1; there is no consumer's surplus, even though I am suffering greatly from the headache. If I had had to pay more for _this package_, I would instead have bought one of the many other packages in the store, all of which are being offered for $1. To repeat: the consumer's surplus from my purchase of _this package of aspirin_ is zero.

How about the consumer's surplus from my purchase of _a package of aspirin_ (no specific one)? Perhaps I considered buying ibuprofen instead of aspirin, but at the offered prices I slightly preferred aspirin. Perhaps this preference would have held up through $1.50; if I had had to pay more than that for the aspirin, I would have switched to ibuprofen. So my consumer's surplus for an unspecified package of aspirin is 50 cents.

Then there's the consumer's surplus from my purchase of a headache remedy. Now we're talking big money (I _am_ suffering mightily). If necessary I would have paid, say, $20 for a headache remedy; my consumer's surplus is $19.

Lots of surpluses here.

Philo writes:

"Let's say someone comes along and radically changes technology so something that was priced positively is now given away. That component of GDP will fall because a zero price times a positive quantity, no matter how large, is zero. But consumer surplus will rise." Fair enough; but then you should not say, of consumer's surplus, that "it's simply a measure of the consumer's gain from trade." Consumer's surplus is broader than that, since when you get something for free there is no _trade_.

Philo writes:

Today I am getting auto transportation for free: I don't have to pay anything, because I own a car, it is licensed, in good repair, the gas tank is full, etc. But if I had to pay for auto transportation today, I would be willing to pay, let us say, $50. That is my consumer's surplus for _auto transportation today_.

I am also getting the services today of a steering wheel on my car. Again, I am paying nothing. But if I had to pay for _steering wheel services today_ (everything else as is, the only change from actuality being my having to pay for the services of a steering wheel), I would pay . . . $50! My consumer's surplus for steering wheel services is the same as for auto transportation.

(This is intended not as an objection to the notion of consumer's surplus, but merely as an illuminating observation.)

Joe Cushing writes:

Seth,

Sales tax deducts some of the consumer and seller (could be producer, retailer etc.) surplus. If there were no surplus, there could be no sales tax. Actually, there could be no taxes at all--of the kind we have today.

rpl writes:
Today I am getting auto transportation for free: I don't have to pay anything, because I own a car, it is licensed, in good repair, the gas tank is full, etc.
No, you're not. You may not be spending anything today, but that doesn't mean there's no cost. You are consuming goods that you purchased in the past, namely the car itself, fuel, insurance, and so on. Moreover, you will eventually have to incur those costs again if you want to continue using your car. Those costs must be imputed to the trips you take with your car in between gassing up, maintenance, etc. (Note that if we adopt your form of cost accounting, almost nothing "costs" anything, since we rarely consume goods at the exact moment we buy them.)
I am also getting the services today of a steering wheel on my car. Again, I am paying nothing. But if I had to pay for _steering wheel services today_ (everything else as is, the only change from actuality being my having to pay for the services of a steering wheel), I would pay . . . $50! My consumer's surplus for steering wheel services is the same as for auto transportation.
Nope. You're double-counting the steering wheel, since it was part of the stuff you paid for when you bought your car. You could try to break out costs associated with the steering wheel and other parts of the car individually to compute a "consumer surplus" for each one, but in this case it makes no sense to do so, since the steering wheel is worthless to you on its own. Only as a part of an automobile is it valuable, and there the value is in the automobile as a functioning system, rather than in any specific part. With other complementary goods, where each good has some value on its own, but they are more valuable together, the accounting is less clear, but you still have to do it in a way that avoids double counting.
Dave Tufte writes:

I face this argument in the classroom sometimes.

I tell my students that since economic profits earned by businesses are analogous to producer surplus ... that consumer surplus is personal profit.

I then make the case that - to be fair - we ought to start taxing consumer surplus or stop taxing producer surplus (which we do indirectly when the government takes some of the taxable income a business reports on their income statement).

Robert Johnson writes:

Great post!

I worry about measuring consumer surplus in dollars. A dollar isn't valued the same way by all people. A dollar of consumer surplus to Bill Gates obviously doesn't mean much, but it means a great deal to anyone living on less than a dollar a day.

If I'm starving, then I might be willing to spend everything I have for a loaf of bread. But if everything I have is one dollar, and the bread costs $2, then it's definitely the wrong conclusion to say that my consumer surplus from buying the bread is minus one dollar. If the bread costs $1 then it's still the wrong conclusion to say that my consumer surplus from buying the bread is $0.

Suggestions for how to solve this problem?

Tracy W writes:

To repeat: the consumer's surplus from my purchase of _this package of aspirin_ is zero.

But your consumer surplus from asprin in general is possibly a lot higher. How much would the price of all of the boxes of asprin had to have risen, before you would have not bothered buying for it?

it's definitely the wrong conclusion to say that my consumer surplus from buying the bread is minus one dollar

Yes, because in that hypothetical case you haven't bought the bread, because you can't afford it. Consumer surplus only applies when people buy something, if they don't, they don't gain from trade in that particular point.

If the bread costs $1 then it's still the wrong conclusion to say that my consumer surplus from buying the bread is $0.

Why is it the wrong conclusion? Obviously in that case, consumer surplus doesn't capture all the gains you get from the bread, but it does capture that you don't have any spare money left over.

Philo writes:

@ rpl:

"You may not be spending anything today, but that doesn't mean there's no cost." My answer: Sunk costs are sunk.

"Moreover, you will eventually have to incur those costs again if you want to continue using your car. Those costs must be imputed to the trips you take with your car [today] . . . ." You're right about that, though note that I may not actually take any trips today, even though _ex_ante_ I would have been willing to pay $50 to have auto transportation available today (in case I needed to use it). But even so we must count depreciation as a current cost, and if I do take trips there will be a cost of the gas used, etc. Suppose that comes altogether to $10. Then my consumer's surplus was only $40.

". . . the steering wheel is worthless to you on its own." What does "on its own" mean? The auto is useless to me without roads (bridges, parking spaces, other places I might want to go to). Consumer's surplus for a particular item is measured by determining how much the consumer would pay for that item if necessary, _leaving_everything_else_as_is_, varying only the amount he has to pay to get _that_item_. (It is not determined by asking how much the consumer would pay for that item "on its own.") I don't have to pay for the services of a steering wheel (except for imputed depreciation for the day, which is a very small amount--say, 10 cents), but if I did I would pay $50, because (given everything else as is) steering-wheel services are necessary and sufficient for my enjoying auto transportation today, which is worth $50 to me. So my consumer's surplus for steering-wheel services today is $49.90--actually greater than my consumer's surplus for auto transportation today.

ajb writes:

But the real problem is comparing consumer's surplus across people. There are a variety of hand-waving reasons why it's not too bad to compare surpluses for average people with similar incomes. But it really fails for someone with unusual utilities or when comparing nations with very different incomes and strong budget constraints. This really emphasizes the difference between consumer's surplus and utility.

Robert Johnson writes:

ajb,

Yeah, you're right. Comparisons across people are problematic.

For me, the relevance of trying to measure or estimate consumer surplus comes in relating it to human welfare. Like in the post, Prof. Henderson's reasons for taking consumer surplus seriously mostly point toward being aware of increases in well-being. And so my point is that, even without trying to compare different people, consumer surplus measured in dollars is not a very good measure for impact on well-being.

Noah Yetter writes:

[W]hat is consumer surplus really and why does anyone take such an idea seriously?

This literally knocked me back in my seat. I don't know how anyone could say Word One about economics without a full comprehension of consumer surplus.

rpl writes:

Philo,

Your sunk costs may affect your decision about whether to drive or not going forward, but they don't change the fact that your up-front costs must be amortized over the individual driving trips in order to get a true picture of the cost of driving. As I said, under your proposed cost accounting nothing "costs" anything because by the time you get around to consuming it the purchase price is already sunk. That is, of course, nonsense.

You're also confused about the steering wheel. You say that if your consumer surplus from driving the car is $50, then your consumer surplus from having a working steering wheel is also $50. I gathered that you were planning to add those together to get a total of $100 in consumer surplus, which is double-counting. If, on the other hand, you merely meant to observe that the $50 of consumer surplus could be viewed as being "lodged" in any of a number of components of the car, then I suppose that's true, but I don't see the significance of it.

Finally, you make another mistake when you impute the entire value of driving your car to both the car itself and to the steering wheel, but then impute only part of the cost of the car to each of them. This is how you arrive at the absurd result that your consumer surplus from the steering wheel is greater than that from the car the wheel is attached to.

Much of what is confusing you here is that the utility of the car is very different from the sum of the utilities of its parts, so imputing utility to any individual part is problematic.

Seth writes:

@David - Thanks and thanks for answering my question.

Am I wrong or is consumer surplus maximization a key argument against government stimulus spending?

@Joe - No disagreement there. That people are willing to pay sales tax shows their consumer surplus is higher than the all-in cost, but it doesn't tell us by how much, does it?

Tracy W writes:

Robert Johnson, the problem is coming up with a better way of measuring impact on well-being.

Economists have come up with other measures for well-being, such as contingent valuation, hedonic regression, etc, but they're not exactly perfect either.

Robert Johnson writes:

Tracy W, I think you're right.

I think one simple improvement would be to contextualize willingness to pay, by simply dividing the dollars that one is willing and able to pay by the total value, in dollars, of all of that person's resources. So, if I'm willing to pay everything I have for bread, then the extremity of my desire for bread is revealed. It shows that I need to eat more than I need other desirable things like comfort, entertainment, etc.

Tracy W writes:

I think one simple improvement would be to contextualize willingness to pay, by simply dividing the dollars that one is willing and able to pay by the total value, in dollars, of all of that person's resources.

Simple conceptually, large data problems in practice. I don't even know what all of my resources are worth, let alone anyone else's (eg, if I started selling my books, how much could I get for them? How about that I could hit my parents up for a loan if I was at risk of starving to death? I don't even know how much money my parents have, but their willingness to do me favours is a resource of mine, btw, my resources are also at the disposal of my parents, so we could easily wind up in the situation where the resources at each person's disposal when summed, are larger than the sum of the resources everyone owns.)

So, if I'm willing to pay everything I have for bread, then the extremity of my desire for bread is revealed. It shows that I need to eat more than I need other desirable things like comfort, entertainment, etc.

I think we already knew that.

Robert Johnson writes:

"Simple conceptually, large data problems in practice."

Yeah, I meant simple conceptually. Also simpler conceptually (and practically?) than contingent valuation and hedonic regression.

"I think we already knew that."

Sorry for stating the obvious. My point in stating it is that you *don't* know how badly I want/need the bread just from the price I pay for it. Or even from the highest price I bid for it in an auction.

But you're hip to that.

Tracy W writes:

Contingent valuation is based on stated preferences, not behaviour. Hedonic regression has all the problems of regressions, particularly unobserved variables.

James Oswald writes:

My willingness to pay for a good is the price at which I am indifferent to buying that good and spending the same amount on another alternative good. Opportunity cost is at the core of supply and demand. For this reason, I do not think that consumer surplus is a good measure of utility, since people with fewer alternatives have much higher consumer surplus than those with more alternatives.

Philo writes:

@ rpl

". . . I don't see the significance of it." The significance is that in general it makes no sense to aggregate consumer's surpluses.

I'll stand by my "result that [my] consumer surplus from the steering wheel is greater than that from the car the wheel is attached to." If this seems absurd to you, there is something wrong with your grasp of the notion of consumer's surplus.

RS writes:

@ philo

Having spent many happy hours behind the steering wheel of a car I wasn't permitted to drive, I cannot help but accept that not only was my consumer surplus from the steering wheel higher than from the car, so was my total utility.

Unfortunately, I cannot imagine a situation when I am not just able to sit behind the wheel, but actually drive the car and yet state that the steering wheel contributed to my surplus in any measure distinct from the car. It just seems to be a case of going around in circles (pun intended). If acquiring a steering wheel reduced all your other costs to zero, somehow managing to move you over a distance without consuming fuel or subjecting the car to wear, then yes you would probably derive more consumer surplus from the wheel. Which makes me wonder, why would you pay $50 to be able to walk with a steering wheel, knowing that you could drive a complete car for the same amount?

The problem seems to be one of erroneous externalities more than anything else. What you say is only possible if you believe that all other costs associated with using the car are external to your use of the steering wheel. Kind of like the cost of electricity being a cost you would consider when using your TV, but not when you are using the power switch on it. The fact is that the power switch is useless without the TV, just as your steering wheel is useless without a car, so when you use either the power switch or the steering wheel, all costs associated with using the TV or the car respectively are direct costs and hence, have to be reduced from the utility you derive from using the power switch or steering wheel, which cannot exceed the utility you derive from the TV or the car. So, at best, your surplus from the steering wheel can be equal to your surplus from the car. Which would hold true for the wheels, tyres, engine (broken up into the casing, gaskets, pistons, piston rings etc.) etc.

I really hope this makes sense to you, because I really doubt my willingness to type anything similar to the above passage again.

Cheers

Comments for this entry have been closed
Return to top