Lines are overrated, and totally empty restaurants are underrated

Some readers (or journalists) ask me if I have further principles for finding good food which are not outlined in my ethnic dining guide or in An Economist Gets Lunch.  Of course I do, though many of them are not easily articulated in the medium of print (some involve scent, for instance, others are about the intangible feel of a place).

Here is one I should have put in the book: Lines are overrated.

Furthermore totally empty restaurants are (often, not always) underrated.

Natasha and I recently took two friends out to a new Bangladeshi restaurant in Arlington, which by the way was spectacular.  But as they first walked into the restaurant, they seemed taken aback that the place was empty and indeed it felt more than a bit deserted, as if no one had eaten there for days.

I showed no sign of wishing to leave.

Here is the logic.  Let’s say a restaurant allows a line to form outside the door.  Why don’t they just raise their prices?  Well, for one thing the line, and the accompanying difficulty of getting a reservation, is a way of marketing the restaurant to potential customers.  Which means the place needs marketing in some manner, which means its audience is in some way not so well-informed about where they ought to be eating.  They tend to be trendy people who follow…lines.  Conformists, in other words.

A lot of places with lines are quite good but when they fall they fall hard.  In the meantime, the presence of a line indicates the place extracts consumer surplus in some fairly inefficient ways, so why should you go, especially if you are not a conformist?  I recall the wise words of my undergraduate differential equations teacher, Professor Lim, who once averred “I don’t want in line.”

What about a totally empty, deserted restaurant?  Well, it depends on ethnicity.  If it’s an Ethiopian place, it means everyone is coming much later.  Go anyway, and enjoy the personal attention you get.

What about an Afghani or Pakistani place or for that matter a Haitian place?  They may make their livings doing catering or weddings.  In those cases, emptiness is often a sign of quality.  It means they make their food for truly demanding customers who demand the best for ceremonial purposes.  It means they have not learned how to sell out or dumb down their food, and they just don’t have enough compatriots in the neighborhood to put many people in the seats on a regular basis (for these reasons, emptiness is not a good sign in say the Eden Center, where the number of Vietnamese diners is quite high, or say in Mexican restaurants on Kedzie street in Chicago, and so on).  Very often empty restaurants come from cultures where consumption is intensely seasonally cyclical, and that is positively correlated with food quality.

Purveyors of empty restaurants are also Adam Smith’s classic overconfident, delusional entrepreneurs.  That’s who I want cooking for me, as most great food is not in fact that profitable.

Best yet is the restaurant which bars its door and remains locked altogether.


This may be true for US but definitely not true for Russia. There are restaurants that are used mostly for money laundering purposes so the food is horrible and you'll never see anyone entering or leaving the building.

My father once went into an Armenian restaurant in North Hollywod for lunch and was the only customer. Yet, delivery men kept coming and going. He figured it was some kind of front operation for organized crime. Now, I know better. Instead, it must have been a very high quality restaurant!

About a year later, four people were murdered in the restaurant:

Intriguing, but were there any hot women?

The Danish films "Pusher 2" & "Pusher 3" have one such restaurant at the center of their plot. It's run by an Eastern European gangster. Romanian I think.

Great films, both.

Serbian. Are you sure about Pusher 2? If memory serves, the only restaurant featured there was the one that hosted the wedding.

Both Pusher 2 & 3 so far as I recall. Not the one that hosted the wedding but the one where the "bad" wedding food was cooked by the brides father.

In a similar (though less sinister) vein, many restaurants in Pennsylvania exist mainly to sell beer, due to the state's silly liquor laws. I suspect one wouldn't want to eat at them, either.

In Oregon, it's video poker.

What if a restaurant is truly horrible? Wouldn't that be totally empty too?

How do you avoid false positives?

White Spot is rarely empty.

Well, he did say it was not always true. But note that restaurants that are empty because they are horrible won't last long.

The problem here is that Tyler is "overthinking" this ... A good restaurant is often a crowded one, but there are also false positives in this area (trendy but bad, overpriced restaurants can also be quite crowded).

He is certainly not overthinking it.

A crowded restaurant is often bad.

Unless it is more often true than not true, it is a moot point. All we've established is, there are at least some empty restaurants which are also good.

Hardly any use if it means I've to endure 20 crappy meals before I find that one gem.

I read it as, don't let emptiness deter you if you have other reason to think it's good. Which is a potentially useful rule. It also has been true in my experience.

Thank you dan1111, your merits shine.

Try looking at the menu at an empty place. You all know he ritual.

As for restaurants that are positively difficult to get into, this one sounds intriguing:

That was the most interesting post I have read in a while. Definitely intrigued.

I fear that this is just a manifestation of the economist preference for contrarianism. Economists are reared on "actually" results, like "seatbelts actually make driving more dangerous" or "tax hikes actually raise less income overall". But sometimes, common sense is more correct. If restaurants are subject to temporary fads, especially non-cyclical once-off fads, raising prices will extract some extra cash from the new diners, but will alienate the audience after the crowd moves on. The line may be uninformed, but that doesn't mean they are wrong. It could mean they are following the advice of critics, who are often right. The line could be staying away because they have heard horror stories you haven't; food is not a zero-information world. Few people are very well-informed about the entire range of dining options. But if you follow contrarian advice to the letter, you have to be very well-informed.

Ah, but if the line is evidence of a temporary fad, then that's a good reason to avoid it, from Tyler's point of view. The food isn't any better now that it's a fad right now.

And yes, one has to be fairly well-informed in order to decrease the use of existing fads and rather be the kind of person who helps starts fads or "likes things before (and after) they were cool." Tyler is that person, though.

Your summary seems fine, but it doesn't disagree with what Tyler is saying. He's not saying that "a line is definite evidence that food is bad." He's saying that "lines are overrated," that they don't say as much about food quality as (Tyler believes) the average person assumes. Remember that Tyler doesn't argue in absolutes, but in differences in the margin. He very often argues not that "X is true" (or false), but that "X is not quite as true as you think it is."

Tyler saying that lines are overrated is no different than him saying that, e.g., comparative advantage is overrated. (By economists and his peer group; the statement that is utterly untrue if you extend it to claiming that comparative advantage is overrated by the general population.)

The larger point I see is that popularity-based measures are only as good as the information level of the people doing the rating. Depending on where you are, especially tourist areas, this information level can be extremely low, and you can be more informed even with a minimal amount of effort. For example, at several major attractions in European cities, I have observed people waiting in line an hour or more to get in, when there was another ticket option nearby or even on site that allowed bypassing the line. Would I want to follow these people's lead for where I should eat lunch?

Tourist locations are the perfect example of the long line != good rule. But an empty restaurant at lunch time in an unfamiliar place makes me suspect that all the locals know better than to eat there.

"...raising prices will extract some extra cash from the new diners, but will alienate the audience after the crowd moves on."

So, restaurant prices are sticky upwards? If only the Fed would be tighter with monetary policy to cause a little bit of deflation, then we wouldn't have to put up with these long lines and undercapacity.

Fads are awesome small talk to get acquainted with strangers. In those long lines the objective is to see, to be seen, talk and maybe get laid with someone else in the line. Food is secondary.

Seriously? I don't think I have ever observed anyone talking to someone in another party in a restaurant line. Usually there is a short line to put your name down and then a long wait, not in a line, with people grouped very closely in their own parties.

Axa is right. Lines are a great venue for doing pickup. They make starting a conversation extremely easy - you already have something in common to get the ball rolling.

"So, I see that you like overrated, conformist food, too..."

Sounds like a chapter in An Economist Goes to Lunch and Strikes Out.

Or maybe it's just shit.

I had a similar experience to Steve Sailer. I used to take my kids to a wonderful Sicilian fish restaurant on the Brixton Road in London. Never full, incredible prices, the freshest fish beautifully cooked by the charming husband and wife couple, who would fuss adoringly over my children. Some years later the wife was jailed for murdering the husband, who was, it appears, one of London's top cocaine bosses.

Wow, such is the state of the cocaine business that top bosses have to dabble in eateries.

It sounds like maybe it was the other way around--the cocaine trafficking was subsidizing the high-quality food.

There's a pattern appearing here, remembers me how how the new year parade guys stopped and stayed for about half an hour at my favorite Vietnamese restaurant. It was obvious that adorable, nice guy had connexions, maybe I underestimated what kind of connexions.

In some countries like India an empty restaurant may mean you get food which is ages old . I guess this does not apply to the eateries in the U.S where the standards of hygiene are observed in all restaurants , empty or full

Does apply. Try a Chinese buffet place.

"Let’s say a restaurant allows a line to form outside the door. Why don’t they just raise their prices?"

Because there are lines out the door only on weekends and it's not culturally acceptable for restaurants to charge more during periods of high demand (except indirectly with weekday coupons and specials)? Or because they want to build a base of repeat customers who will return even after the trendiness (inevitably) wears off (customers who went and stood in line the first time because it was the hot new place and who come back because it also turned out to be a pretty good deal).

Maybe the lines are ways to drive business to slower days. It works on me, anyway. And I suppose restaurants unlike box stores are businesses that require constant churn.

This reminds me of a recent story about lunch congestion pricing at the Goldman Sachs cafeteria:

Not that paying high peak prices would be much of a burden for most GS employees.

One observation: Restaurants with handwritten menus on a blackboard where things get crossed out as they run out tend to be good.

Any one else see this?

In general, yes, I've found that restaurants that are capable of running out of particular food items have good tasting food for the price. That includes places that serve only a few items but close up shop when it's done, like Pioneer Pit Beef in Catonsville, MD or Oink in Edinburgh, Scotland.

It's a similar kind of effect with cash only places. Selling items until they're gone instead of stocking extra lowers overhead and improves food quality (no days old ingredients you have to get rid of) on the one hand, but it imposes a non-cash, non-food quality cost on diners. In equilibrium, one might expect good food at a good price with the annoyance of them possibly being out of what you want, limited selection, or being closed when you go by there. For some people, the latter annoyances outweigh the former, so it depends on what you can handle.

The locked doors reminds me of a southpark episode:

I disagree. People wait in lines for things that are worth waiting in line for. If there is a line outside the place it means that the customers are estimating the product value as worth the price plus the opportunity cost of the time spent waiting in the line.

Also, Mr. Cowen's valuation of his own time or opportunity cost may be (likely is) higher than other (most other) peoples.

Yes, but his valuation of trendiness and fads, and the desire to be seen in crowds or with the popular people, seems lower than other people. If you've read his book, you'll also note that he places low priority on restaurants that encourage (and are full of) lively conversation as well, preferring places with a lot of fairly grim looking serious eaters, since you get the best food there. Lines *can be* the sign of something that is trendy, and you're paying for that trendiness. (And trendiness *might* mean sacrificing the food to appeal to a wider audience, given Tyler's appreciation of unusual food.)

It doesn't have to mean that-- R&R Tacqueria remains good, despite having very long lines at lunch. Part of that is indeed from lots of intrepid people who have heard of the place-- I am amazed every time at the number of first time diners there. Rodrigo keeps the prices fairly low (they have increased since two or three years ago; two tacos plus lamb soup used to be an outrageously good $5, but is still quite good at $6.99) and the lines long, partially to continue to cater to the large Mexican-American population that eats there.

Assuming rational, fully informed consumers with tastes similar to yours.

People wait in lines for things that are worth waiting in line for.

Like Powerball tickets when the number gets high.

Tyler Cowen didn't say there wasn't a correlation, he said the correlation isn't as strong as you think. Further, he said in specific situations, such as unusual ethnic food, (NOT Chinese or Mexican), this especially hold true, and may even be a correlation in the opposite direction.

I agree with the general principles espoused here, but I'm with Professor Lim: I won't stand in a line for anything. Life is too full of alternatives without lines to spend any time in them. And it's not just actual lines, virtual ones count too. I won't book restaurants that open their reservation book at 9 am for a meal two months hence and which fill it up at 9:15. Lines? No Thank Queue.

I have no trouble making reservations, but I don't like lines. I solve that problem with a number of restaurants by being extremely willing to eat at slightly unusual times. The difference in lines in a popular lunch place at 11:30 or 1 pm to noon or 12:30 is enormous. So too for eating dinner at 6pm or earlier, or 8:30 or later.

It's all about cross-subsidies. If it's always empty, it is probably being cross-subsidized by something else. If you are pretty sure that the "something else" is food-related (catering, a late-night rush, delivery, etc), EAT THERE. But there are often reasons to doubt that. A hotel will keep its restaurant going even if the food sucks. There's the aforementioned money laundering, plus human trafficking, whorehouses and other businesses that can be based out of "restaurants."

But it's probably not a whorehouse if it's also devoid of pretty women, so by following Tyler's other rules, you can at least avoid that problem.

A significantly large proportion of "Indian" restaurants are really owned and operated by Bangladeshi families. The logic is classically Bangladeshi in its self-effacing pragmatism: Many Bangladeshi restauranteurs simply believe that "no one will buy Bangladeshi cuisine," so they call it an "Indian" restaurant instead. Considering that I have never seen a declared Bangladeshi restaurant that had a full house of customers, I'd say the owners' logic is pretty sound.

There is also the fact that most ingredients that are important in Bangladeshi cuisine are completely unavailable in the USA. Where are you going to find shutki? Laal shaak - forget about it. I have seen lau in Canadian grocery stores, but never in the USA. Korola (bitter gourd) is easy enough to find, but nobody's going to order that in a restaurant.

You might be surprised at how many people at authentic Chinese restaurants order bitter melon. I don't have the taste for it of my wife, but it's on the menu in lots of places in this DC area.

I very much prefer to be in an empty restaurant. I like being the only one. That enhances my restaurant experience. There are some restaurants which are frequently empty, but mostly it's the time of day. If you go in the middle of the afternoon, many restaurants are empty. (Some are closed, of course.)

Your comment about Afghani restaurants may explain why Afghani House (Sunnyvale, CA) is so often empty or nearly so. The food is some of the best I've ever had.

Becker's Note on Restaurant Pricing (JPE 1991) seems relevant here. He definitely takes the conformist route (people like eating where other people eat) causing some restaurants to be "in" and others "out" -- and "in" restaurants with lines can't raise prices even a little bit to eliminate the line without ending up with an empty restaurant.

I use Yelp! I work with technology, and am ready to embrace my forthcoming allegiance to the top 15%.

What about having a preference for half-empty (or half full) restaurants?

I disagree with the claims about completely empty restaurants, which in my experience are often horrible. Unless you are hitting the restaurant at an off-off-peak hour, it could be empty for some of the reasons Tyler is mentioned, or it could be empty because everyone in the area knows the food is bad and for some reason the restaurant just hasn't closed yet.

However, I agree that in most circumstances (there are sometimes your plans involve trying one particular restaurant, or you may have eaten and the restaurant before, know its good, and can't schedule your visit for an off-peak period), waiting in line for a restaurant is a pointless additional cost because its easy to substitute a less crowded restaurant of comparable quality.

You want some people to be eating in the restaurant, but having it actually overcrowded is more of a sign of trendiness.

And going at off-peak periods is a good idea, though again this doesn't hold in all places. The worst service I've experienced has been in empty restaurants or nearly empty restaurants at off-peak periods. It was clear that though the restaurant was technically open, the staff had other priorities during that time than serving customers. This includes places that were crowded and peak times and unlikely to be fronts for something else.

Years ago we were in Saratoga, New York in the dead of winter, and had a choice between 2 restaurants for lunch. One was an upscale pub with $7 burgers and a convivial Irish name. There was a line stretching outside, so we crossed the street to an Indian restaurant. We peered inside, hesitated a bit at the sheer emptiness of the place (had the health dept. been by recently or something?) but gave it a chance. For $4.95 we both had the most mouth-watering entrees imaginable (Baingan bharta was one). We took literally all the remaining meals on our short visit at that place.

Bedrock anecdotal proof!

Re-Stated (with the admitted benefit of reading others misinterpretations):

I read it as a true economists warning along the lines of correlation/causality. Just because a place has a line doesn't mean it's good and don't be discouraged by a desolate hole in the wall place.

I recently waited over 30 minutes in line for a breakfast in State College, PA (I was with a group that insisted). The meal was over-priced and mediocre at best.

Did this restaurant name involve "waffle" in it? Most of the food in State College is completely warped by college students who don't have transportation and are drunk or hungover.

Does anyone know how to avoid salty pasta while in Rome?

Eat risotto.

That was easy.

I believe Yogi Berra once offered an observation similar to Tyler's: "Nobody goes there anymore because it's too crowded."

For many years I was a happy customer at a Chinese restaurant that tended to be mostly empty even in the middle of dinnertime. The kitchen was good, and kept busy with take-out and delivery orders.

Conversely, I have given up on trying the Spaghetti Factory after two attempts. Both times, we were quoted long waiting times, and there were large empty sections of restaurant. Either they are deliberately maintaining a line to make people think it's worth waiting for, or they have no idea of how to manage staffing. (I was considering it in the first place because it's in walking distance of my new apartment.)

I'm sure you've covered this somewhere, but it'd be seriously useful for me (and everyone else in corporate America) if you could give us some wisdom/basic rules of thumb for utility maximization on a) choosing food trucks b) Seamless, where you're flying blind c)the ridiculous food economics of New York

What explains the difference in average portion sizes in restaurants between one city and another?

The best Pakistani restaurant in Buffalo (and one of the best anywhere for lunch, not necessarily dinner), Zaiqa's, is usually empty.

In my dreams, the opposite is true: the best buffalo restaurant in Pakistan is always full.

I paid for part of college promoting live music in 80's pre-Grunge Seattle night clubs. There was a clear understanding in the music business there that popular music (any band with a line out the door to get in) wasn't always good and that good music (what the musicians and cognoscenti preferred) wasn't always popular. The night clubbing public tended to follow the crowd and their assessment of a band's quality was often linked to how crowded the club was on that evening. A nice experiment that was repeated often: Same music group on two different nights. One night with little competition from other local events and a packed house that loved the music A second night with other events in town (sports, other bands, big TV show...) and a half empty house that thought the music was, well, okay.

A long way to say that people who wait in a restaurant line are often convinced that the food is better than it is because 1) Everyone else likes it and they must be right. 2) I waited in this #$%& line and I want to feel like that was a good decision. 3) Following the crowd is easier than thinking. 4) There are a lot of people here and that is fun!

Does anyone else now have an image of Tyler as a short, bald Sicilian trying to reason out which glass of wine has the poison?

Is cultural preference part of the price? Pakistani food is highly valued in Pakistan but not in the US?

@prior probability There is NOTHING that Tyler can't overthink.

Many defend Tyler as making a marginalist argument, but it's not at all clear that this defence works. He doesn't know how well lines are rated.

The conclusions are almost trivially true, but the reasoning here is terrible.

1) Of course places that are empty are going to tend to be underrated and places that are full are going to tend to be overrated. Unless someone believes that there is a perfect relationship between popularity and quality, then all restaurants are more likely to be average than anything else. Which means that a meaningful percentage of popular places will be out-performing their quality whereas a meaningful percentage of unpopular places will be under-performing their quality.

2) Most places don't have a line by choice of the restaurateur. They have a line because they're busy and they can only handle so many customers at once. Nicer restaurants may have a line (or, more accurately, a wait) because they have some number of reservations and leave the rest of the tables for walk-ins. This allows them to more efficiently serve customers and have a higher number of turns (numbers of times a restaurant is filled) per night, increasing volume and profits. This in turn can keep labor costs lower relative to gross revenues and also allows for prices to be kept lower. Fast casual places handle customers one at a time and usually have a smaller number of seats and this creates a line. Often these lines only exist during prime dining periods and rapidly build and disperse. Generally fast casual places use counter-service in order to keep labor prices low to off-set the necessity of keeping menu prices low. Fast casual restaurants generally depend on volume to make meaningful profits. This is especially true of national chains that compete even more heavily on price and may only have a paltry 3% net profit. But a national chain may gross $2 million or more per restaurant and so that 3% is still a meaningful profit, even if a fine dining restaurant might make the same profit with 1/4 the business.

3) Building on point (2), restaurants can do significantly better when they were consistently busy than they can when their business is more erratic. While food costs are a proportional cost and will maintain their percentage of gross revenues regardless of volume (though there should be less wastage with more volume), both labor and other major costs (rent, insurance, equipments leases, etc) will stay relatively static regardless of volume. Labor can change somewhat, but often a restaurant has to staff to a minimal level in case they fill up. So profits can often be maximized better by keeping prices reasonable and allow lines than to raise prices until there aren't lines. (As if such balances can be found in the real world.)

4) Restaurant customers react slowly. What a restaurant does today often won't be realized in customer actions for months later. If you raise prices, it doesn't stop them from coming in immediately. It takes time for customers to cycle through and realize the prices are raised and for a reputation of prices to build. There is no doubt that lines to lend a certain reputation that other people like what you do. But a restaurant still has to meet some level of expectations to bring customers back. There are certainly a large number of people that will refuse to go to a restaurant with a line, but also people who even if they get a good meal will be offended that they didn't receive a great meal. But ultimately, it's safer for a restaurant because of all of the above to not do anything to dissuade a line. It's the conservative choice. If it isn't broke.... Hell, if one of my restaurants was so popular people started lining up before it was open everyday, I might stop shaving and wear the same underwear every day just to keep from jinxing it.

I guess I should have summed up: Restaurants primarily maintain lines for bottom-line business reasons (and luck) not as some marketing ploy.

"Best yet is the restaurant which bars its door and remains locked altogether." When I and others tried to go to a highly recommended Salvadoran restaurant, it was closed and empty. Best restaurant ever!

Comments for this post are closed