Imperfect Competition

Link post

Pre­vi­ously in Se­quence: Moloch Hasn’t Won, Perfect Competition

This post looks at a few ex­am­ples of im­perfect com­pe­ti­tion to illus­trate var­i­ous ways in which perfect com­pe­ti­tion is kept at bay and value is pre­served. Con­crete ex­am­ples seem more likely to en­lighten than ab­stract prin­ci­ples.

Let’s start with the most literal of mar­kets, the mar­ket for food, and then talk about the mar­ket for cars.

As I noted last time, I no­ticed af­ter writ­ing this that a Lo­cal Farmer’s Mar­ket was Google’s top re­sponse to ask­ing for an ex­am­ple of perfect com­pe­ti­tion. Which makes it kind of perfect as the cen­tral ex­am­ple of im­perfect com­pe­ti­tion.

For length, I’ll stick to these ex­am­ples. I am happy to give quick mod­els for other mar­kets in the com­ments if any­one wants to work through differ­ent ex­am­ples.

1. The Lo­cal Farmer’s Mar­ket

A few times a week, at var­i­ous places around the city, lo­cal pro­duc­ers open stands to sell their goods to the pub­lic. Many peo­ple choose to shop here, rather than at the su­per­mar­kets a few blocks away in ev­ery di­rec­tion. They typ­i­cally pay higher prices, do a lot of in­ves­ti­ga­tion and sam­pling of goods, and even­tu­ally of­ten be­come loyal to pro­duc­ers and prod­ucts that provide strong value. When I was grow­ing up, we of­ten got our cheese and ap­ple cider from one such mar­ket (along with more healthy foods that I was in­ca­pable of eat­ing, and there­fore don’t re­mem­ber as well).

What’s go­ing on here? A lot of differ­ent things are com­bin­ing at once – and this list likely isn’t even com­plete. Each of these is a vi­o­la­tion of perfect com­pe­ti­tion.

Prod­ucts and pro­duc­ers are not ho­mog­e­nized, and in­for­ma­tion about them is costly. Even the rel­a­tively ho­mog­e­nized farm­ers’ mar­ket prod­ucts, like the ap­ple cider, differ in qual­ity from batch to batch, from sea­son to sea­son and har­vest to har­vest. There is no re­li­able way to differ­en­ti­ate high qual­ity from low qual­ity prod­ucts or pro­duc­ers, or any way to di­v­ine the product de­tails you might pre­fer or dis­like, with­out sam­pling the goods, which is costly out­side of the oc­ca­sional free sam­ple. When buy­ing pro­duce, one needs to pick out the items one likes and things get even more vari­able. Rep­u­ta­tion and ex­pe­rience mat­ter a lot.

Mar­ket in­for­ma­tion is costly for con­sumers. Differ­ent pro­duc­ers bring differ­ent goods to mar­ket at differ­ent times, with no known sched­ule. Often the goods sell out. The only sure way to know is to go to the mar­ket and look around. If one wants to be sure, one would even walk around the full mar­ket and ex­am­ine things be­fore mak­ing any de­ci­sions, but that takes a while. Once you’re there, it of­ten makes sense to buy some­thing even if you wouldn’t have both­ered go­ing out there if you’d known what was available to­day.

Mar­ket in­for­ma­tion is costly for pro­duc­ers. Farm­ers and bak­ers and cheese­mon­gers tend not to be those with a deep in­ter­est in mar­ket dy­nam­ics. They don’t know what con­sumers will want or what prices they’ll be will­ing to pay, and find out by try­ing to sell goods and see­ing what peo­ple buy. They are quite bad at max­i­miz­ing prof­its from what they pro­duce. We like it that way. Some­one who was good at max­i­miz­ing prof­its would be bad at pro­duc­ing qual­ity goods. Each pro­ducer has a unique situ­a­tion, so they can’t rely on gen­eral sur­veys to price goods cor­rectly. Good­will is vi­tal to the busi­ness, mak­ing pric­ing ex­per­i­ments gen­er­ally a bad idea. That leads into a key ad­di­tional con­cept, which is com­plete skill stacks are rare. This al­lows be­ing very good at one as­pect to com­pen­sate for be­ing rel­a­tively poor in an­other, whether or not this is by choice, or pro­vides slack.

Un­known un­known in­for­ma­tion ex­ists. We know lit­tle about nu­tri­tion and what is dan­ger­ous for us. Some of what we are buy­ing, when we buy from small lo­cal pro­duc­ers, is their in­abil­ity and un­will­ing­ness to use var­i­ous in­dus­trial meth­ods we don’t know about or un­der­stand, that may have un­known and hard to mea­sure long term effects on our health. We are un­cer­tain about the or­der of mag­ni­tude of this effect, but such effects are highly un­likely to be good. This pushes us away from ho­mog­e­nized prod­ucts.

Fixed costs ex­ist and pro­duc­tion costs are asym­met­ric. Skills differ. Each farmer and pro­ducer both spe­cial­izes in a differ­ent set of prod­ucts, and faces the quirky re­source con­straints of their own land, skills, tools and back­ground, and what they know how to and like to grow and make. Economies of scale ex­ist due to many fixed costs. Even if you did know how to grow crops, do­ing so on too lit­tle land would be pro­hibitively time in­effi­cient. Even if it wasn’t, you would still have no prac­ti­cal way to bring those goods to mar­ket, as that too re­quires fixed costs. You can’t par­ti­ci­pate in these mar­kets un­less you’re big enough to be a full mem­ber and com­mit­ted to sus­tain­ing re­la­tion­ships and rep­u­ta­tion.

Pro­ducer prefer­ences differ. This is an im­por­tant vari­a­tion on vari­able pro­duc­tion costs. Farm­ing in the mod­ern age is of­ten a la­bor of love. If a farmer likes ex­tra sharp ched­dar or plum toma­toes, they’re go­ing to pro­duce them even if it isn’t the best thing for the bot­tom line, then bring them to mar­ket any­way. They’ll also pro­duce more va­ri­ety than the mar­ket would re­ward on its own. Seek­ing this out is one way to find the best qual­ity.

Lo­ca­tion mat­ters. The farmer’s mar­ket ex­ists largely be­cause it’s bet­ter to buy lo­cal prod­ucts with lower money and time spent in trans­porta­tion, re­sult­ing in cheaper and higher qual­ity goods. The goods in su­per­mar­kets of­ten were en­g­ineered to travel bet­ter and keep longer, which re­duces qual­ity in other ways. When you go to the mar­ket for lo­cal goods, those closer to the mar­ket have a huge ad­van­tage. Those farther away will choose to go to a differ­ent mar­ket, even if its cus­tomer base isn’t quite as good.

Con­sumers care about the in­di­vi­d­ual pro­duc­ers. Lo­cal farm­ers also of­ten get ex­tra busi­ness from those who like to sup­port lo­cal pro­duc­tion for var­i­ous pro-so­cial rea­sons, or even de­velop per­sonal re­la­tion­ships over time. Most peo­ple at the farmer’s mar­ket would pay a non-zero amount ex­tra to buy from the farmer’s mar­ket over an iden­ti­cal product at the su­per­mar­ket.

Skin in the game. Skin in the game doesn’t mix with perfect com­pe­ti­tion, be­cause in perfect com­pe­ti­tion there’s no skin available to have. When own­ers op­er­ate and stick around for the long term, and face the con­se­quences of their de­ci­sions in terms of skill de­vel­op­ment, rep­u­ta­tion, ca­pac­ity, and so on, they have to care about more than just cre­at­ing a ho­mog­e­nized product, or even cre­at­ing a product. Every­thing mat­ters, and those who let longer term things slide do not sur­vive for long.

No boss. No one is tel­ling the pro­duc­ers they’ll be fired if they make trade-offs. No share­hold­ers are putting on ex­treme pres­sure to max­i­mize prof­its. Peo­ple get to do the things they want to do, and not the things they don’t want to do, and ac­cept the con­se­quences.

What hap­pens in the long term?

We used to mostly have these types of mar­kets, and now we mostly have su­per­mar­kets.

2. Supermarkets

Su­per­mar­kets are a rough busi­ness to be in. They are much closer to perfect com­pe­ti­tion than farmer’s mar­kets. The prod­ucts are more com­modi­tized and ho­mog­e­nized, and are more pre­dictable in qual­ity. In­for­ma­tion is eas­ier to get. Often goods are iden­ti­cal across sel­l­ers. The re­la­tion­ships are mostly purely trans­ac­tional. The store own­ers are mostly in it for the money.

Iden­ti­cal pack­ages of Ker­ry­gold but­ter are $3.19 at Trader Joe’s on 3rd Av­enue, $3.79 at Whole Foods a block farther west, and some­thing like $4.49 at West­side Mar­ket two blocks to the south. Some­times I see them for $5 or more.

That’s weird. Sounds re­mark­ably un­com­pet­i­tive. What hap­pened to the law of one price? What’s go­ing on?

Here’s my model. Peo­ple care a lot about dis­tance when pick­ing a su­per­mar­ket, and care a lot about batch­ing their pur­chases so they don’t have to go too of­ten. Cus­tomers are choos­ing be­tween a hand­ful of po­ten­tially rea­son­able bun­dles of op­tions. Each has some unique offer­ings in terms of product va­ri­ety and qual­ity, a unique lo­ca­tion, a unique brand iden­tity and rep­u­ta­tion, and a his­tory with each po­ten­tial cus­tomer. It takes time to learn what each one has to offer, to re­mem­ber where ev­ery­thing is lo­cated so things can be found quickly, to know which things are be­ing over­priced at each store. Stores of­ten price some items lower than nor­mal and/​or be­low cost to tempt cus­tomers in (hence the term ‘loss leader’) and then jack up the prices of other items (like the Ker­ry­gold but­ter) where cus­tomers are likely to be price in­sen­si­tive.

Thus, when I en­ter a new su­per­mar­ket, I price check the Ker­ry­gold but­ter be­fore buy­ing items whose fair price I do not know.

Su­per­mar­kets are effec­tively in oli­gopolis­tic com­pe­ti­tion with lo­cal al­ter­na­tives and a hand­ful of ser­vices like FreshDirect. They’re not in any­thing like perfect com­pe­ti­tion with each other.

The same is true for differ­ent brands within the su­per­mar­ket. We have what al­most any other cul­ture in his­tory would con­sider amaz­ing product va­ri­ety and choice. It is very rare for there to mostly be mul­ti­ple brands that sell es­sen­tially iden­ti­cal prod­ucts at the same price point, in the same store. It’s rare for there to be even two best op­tions that are es­sen­tially iden­ti­cal. Often, the choice will be close be­tween mul­ti­ple brands, but that’s be­cause they’re offer­ing differ­ent trade offs. At­tempts to copy gold stan­dard prod­ucts like Heinz ketchup some­how don’t work. Most con­sumers don’t know the qual­ity or de­tails of most op­tions of things they’re con­sid­er­ing buy­ing. And there’s a ton of brand loy­alty. There is in­tense com­pe­ti­tion. The sys­tem works well. But it’s definitely some­thing very differ­ent than perfect com­pe­ti­tion.

On­line gro­cery ser­vices like FreshDirect also face suffi­cient scale and dis­tance re­quire­ments that there are only a hand­ful available in a given area, and they tend to not have that much product va­ri­ety ei­ther, and to mar­ket them­selves as unique op­tions com­plete with many unique items (e.g. for FreshDirect I recom­mend the take and bake Ci­a­batta rolls).

One can imag­ine a world in which trans­porta­tion costs con­tinued to fall, and skill in preser­va­tion con­tinued to im­prove, as did the abil­ity to com­mu­ni­cate what you want on­line and get what you asked for (this is much less re­li­able than my pri­ors would have pre­dicted), to the point where lo­ca­tion stopped be­ing im­por­tant and it made sense to break up or­ders into in­di­vi­d­ual prod­ucts. One can imag­ine bet­ter mar­ket knowl­edge be­com­ing available and known to more cus­tomers. Or the food is made in repli­ca­tors, or by robots us­ing stan­dard­ized open-source pro­ce­dures. That day may come. For now, a lot of things stand be­tween here and there.

3. Res­tau­rants (in brief)

Res­tau­rants are a strange case.

I am told that no one makes money and com­pe­ti­tion is su­per rough, largely be­cause peo­ple are of­ten open­ing and run­ning places for non-mon­e­tary rea­sons. A lot of peo­ple think they can do it when they can’t be­cause they don’t un­der­stand what it takes (and be­cause their friends and fam­ily of­ten lie to them about their cook­ing). Which can all cause su­per-perfect com­pe­ti­tion.

De­spite that, we clearly see highly im­perfect mar­kets. We reg­u­larly see places that can pro­duce suffi­ciently su­pe­rior prod­ucts to com­mand strong pric­ing power, which they rarely fully use. There is lit­tle doubt that many places have huge ad­van­tages over the marginal com­pe­ti­tion. Places have sus­tained rep­u­ta­tions and re­la­tion­ships with reg­u­lars, earned by Play­ing in Hard Mode. See­ing peo­ple play in hard mode is a sure sign com­pe­ti­tion is far from perfect. And those who do these jobs well earn the abil­ity to charge more and of­ten get more busi­ness than they can han­dle.

My model says the cen­tral sto­ries here are that lo­ca­tion mat­ters a lot, skill and ex­pe­rience mat­ter a lot, prod­ucts are highly differ­en­ti­ated, in­for­ma­tion is ex­pen­sive on many fronts, and peo­ple have differ­ent val­ues with skin in the game. Too many par­ti­ci­pants en­ter, and ex­it­ing is ex­pen­sive, so mak­ing zero prof­its is re­ally hard. But there are huge marginal re­turns to be­ing great, so huge suc­cess re­mains pos­si­ble.

4. Buy­ing a Car

I don’t drive, so I’ve never bought a car. But buy­ing a car is tra­di­tion­ally so frus­trat­ing to so many peo­ple that I know a rea­son­able amount about the pro­cess.

In the first phase, differ­ent com­pa­nies at­tempt to con­vince you to buy their car in­stead of an­other car.

This is done through some com­bi­na­tion of brag­ging about small fea­tures or ar­bi­trary in­dus­try awards, se­lec­tive com­par­i­sons (which my brain always reads as, ‘that other car you’re men­tion­ing is prob­a­bly worth check­ing out’), cre­ation of emo­tional res­o­nance and iden­tity af­fect as­sign­ment, load­ing up more ob­serv­able fea­tures, pe­ri­odic ‘good deals’ and ‘sales’ with su­perfi­cially cheap struc­tures, offer­ing test drives, and pro­vid­ing in the back­ground real in­for­ma­tion re­sources.

Cus­tomers then com­bine this with per­sonal anec­dotes and ideally some in­for­ma­tion from plau­si­bly ob­jec­tive third par­ties (e.g. my par­ents used Con­sumer Re­ports mag­a­z­ine and the NPR show Car Talk) to eval­u­ate salient choices along nu­mer­ous axes, nar­row things down to choices that feel ap­pro­pri­ate and rea­son­able, then con­sider what ‘deals’ are available and make a fi­nal choice how­ever big fi­nal choices are made. One must also then choose ex­pen­sive new car ver­sus cheaper new car ver­sus pre­mium used car ver­sus cheap used car.

Economies of scale, both in pro­duc­tion and dis­tri­bu­tion and then in mar­ket­ing and rep­u­ta­tion, keep the mar­ket limited to a hand­ful of pro­duc­ers, each of whom pro­duces a hand­ful of mod­els each year. Even ma­jor par­ti­ci­pants can only effi­ciently high­light a hand­ful of offer­ings, or the mes­sages get mud­dled. Differ­ent cus­tomers care about wildly differ­ent things. One will care about safety and re­li­a­bil­ity, an­other how the car looks or whether it is seen as sexy, an­other pri­mar­ily cost to own, a fourth comes from a fam­ily that drives Fords, a fifth wants ev­ery­one to know how good and re­spon­si­ble and for­ward think­ing they are with their elec­tric model, and the last is a six­teen year old who just got their li­cense and wants re­ally good air con­di­tion­ing and sweet cup hold­ers.

You want to differ­en­ti­ate your­self from the com­pe­ti­tion, not match them, even if you’re bet­ter on ob­jec­tive qual­ity and value, since those are hard to mea­sure and com­mu­ni­cate.

The prod­ucts do their best to not be ho­moge­nous.

It’s not that one el­e­ment of perfect com­pe­ti­tion is miss­ing. It’s that all the el­e­ments are each on their own miss­ing.

En­try and exit are costly.

Economies of scale are cen­tral.

Mar­ket knowl­edge is ex­pen­sive and difficult.

Mo­bil­ity of goods is not only not free, it is what is be­ing sold.

There are mas­sive re­stric­tions on trade.

Past this point, I have no per­sonal ex­pe­rience, so this could be un­fair, ob­so­lete, in­ac­cu­rate or all three. And of course, the whole point is that these mar­kets are not perfectly com­pet­i­tive, so ex­pe­riences wildly differ from per­son to per­son and lo­ca­tion to lo­ca­tion. Great ex­pe­riences do ex­ist, and only some of them in­volve not re­al­iz­ing you’re be­ing had.

Once a car has been se­lected, it’s time to head to the dealer. To ne­go­ti­ate and make a deal.

This pro­cess is not less ab­surd than it sounds. It’s also man­dated by the gov­ern­ment in the United States, which bans sys­tem­atic au­to­mo­bile sales that don’t go through phys­i­cal deal­er­ships.

Cus­tomers come in and look around the shiny cars. They are then likely to be ac­costed by sales­man who share se­lec­tive in­for­ma­tion, hide other in­for­ma­tion, frame shift, an­chor on ab­surd re­tail prices, ma­nipu­late, out­right lie, ‘check with their man­ager’, ca­jole, play good cop bad cop and other similar games, cre­ate ex­pec­ta­tions of re­ciproc­ity, bait and switch and so forth. They do this so brazenly and re­li­ably that ‘ev­ery­body knows’ that car sales­man are out to get you, and peo­ple study up on how to nav­i­gate this in rea­son­able time and not get taken for a ride.

Most cus­tomers right­fully dread this. Many find it so un­pleas­ant that they know­ingly ac­cept los­ing hun­dreds or even thou­sands of dol­lars in value to make the pro­cess less painful. They are of­ten un­able to bring them­selves to use tac­tics like “call mul­ti­ple deal­ers and ask for quotes” or “take a day to think about it or see if the deal im­proves,” and even those who know rarely hold out for the most pow­er­ful, “wait for the end of the month,” whose mechanism will be ex­plained in a bit.

And that’s for new cars. Used car sales­man car­ries even worse con­no­ta­tions.

Im­por­tantly, it’s not max­i­mally bad con­no­ta­tions. It’s noth­ing close to perfectly com­pet­i­tive fraud. It’s wildly vary­ing mag­ni­tudes of fraud.

The com­mon ques­tion, ‘would you buy a used car from this man?’ high­lights that there are peo­ple you un­der no cir­cum­stances should buy a used car from, and oth­ers, even pro­fes­sional used car sales­man, that you should be cau­tious with but who it would be rea­son­able to pur­chase a used ve­hi­cle from.

New car sales­man pur­sue a wide range of strate­gies at a wide range of skill lev­els. They even rad­i­cally ad­just tac­tics based on the time of the month. They have tar­gets to meet and will do what­ever it takes to meet those tar­gets if they’re about to miss, but find do­ing so un­pleas­ant on many lev­els, as well as ex­pen­sive. It not only burns slack and isn’t fun, it de­grades the long term and bot­tom line to push too hard. So most sales­man avoid do­ing this when it isn’t nec­es­sary, and pick a trade-off be­tween short term prof­its, long term prof­its, in­ten­sity of work, and their own sense of moral­ity.

When we do see the worst be­hav­iors, my model says it is usu­ally be­cause of pres­sure from bosses and quo­tas from higher up the cor­po­rate lad­der, rather than any need to be max­i­mally com­pet­i­tive.

It is done hap­haz­ardly and in­effi­ciently to save peo­ple’s hides. That you can typ­i­cally liter­ally save hun­dreds to thou­sands of dol­lars by buy­ing your car on the last day of the dealer’s sales in­cen­tive pe­riod speaks to how dis­torted all of this has be­come.

We are see­ing the re­sults of im­moral mazes.