This is why I left it vague...it really is a complicated issue with several angles. That...and my Bachelors only goes so far.
That said, I think you are conflating a couple of economic principles.
> It's necessary to maximize total surplus as well. If someone wanted to get on the plane, and they took off with a seat empty, that's a deadweight loss.
I think you are confusing utility [1] with surplus somewhat here. If someone wanted to get on and they took off with an empty seat that is a loss in Utility. If the airline turns around and sells that seat to the new person that increases Producer Surplus (and conversely decreasing Consumer Surplus) to increase Utility.
I don't think this situation (empty seat) is DWL in the traditional sense [2]. The traditional DWL scenario is when producers could supply more but don't because of pricing constraints or other artificial ceilings/floors. It is definitely a marketing allocation inefficiency which results in Utility loss and I am kind of interested to see if there are better quantifications of this loss...but it isn't a DWL in surplus technically.
> Everyone then is better off. In fact, on a lot of flights, you can probably find multiple people to bump with just a token payment -- they'd rather stay another day.
It's totally possible that everyone can be better off. The issue isn't about when everyone is better off...it's when the system breaks down and you have situations where no one wants to make that trade.
> The bumped guy is happy to have the money, the last minute guy is happy to be on the plane, and the airline is happy to make $4700 extra.
This is an example of maximizing producer surplus. In my mind it's a variation of Price Discimination [3]. And while this overall beneficial situation is plausible, it isn't really overbooking.
For me, overbooking is basically selling more goods than you actually have and playing on the statistics trying to bank on people not making the flight. If you sell 30 seats and 5 people don't show, you take off with 25 people on board. If you sell 35 seats and 5 don't show, you take off with a full flight and pocket some additional revenue from the 5 no shows.
Now there is a lot of gray area since most no shows are from missed connections and so some of that additional revenue can go back into the system to allow for cheaper cancellation fees as well as accommodating rescheduling costs for these missed connections. But in general I suspect that producer surplus increases more than the gain in consumer utility.
> This works even better with a name-your-price bump question on checkin for flights that are oversold. Even if it's nonbinding, the airline now knows that it has willing passengers to bump and can keep selling tickets.
This is again, Price Discrimination but now adding information imbalance in order to help the airline more effectively Price Discriminate thereby maximizing their Producer Surplus.
Technically the United situation isn't the result of overbooking anyway...but I still think the practice warrants scrutiny. United's recent announcement to increase payouts and mileage compensation is testament to the fact that there was Producer Surplus to spare.
Like I said, all of this is speculative discussion and there is a lot of gray area. It is also plausible that overbooking results in maximizing Utility and an even handed redistribution of benefits to both producers and consumers. But I believe Producers are driven by incentives...and they don't have strong incentives to share the benefits of overbooking altruistically back to the consumer.
I fly from SFO - PDX every other weekend and consistently, no matter how early I book, Virgin America flights are cheaper than United and Virgin doesn't overbook whereas United does. Generally speaking, if United were even slightly returning the added revenue from overbooking to the consumer shouldn't their flights be cheaper than non-overbooked competitors?
Again this observation and subsequent question is anecdotal and broad...but I like it for FFT.
That said, I think you are conflating a couple of economic principles.
> It's necessary to maximize total surplus as well. If someone wanted to get on the plane, and they took off with a seat empty, that's a deadweight loss.
I think you are confusing utility [1] with surplus somewhat here. If someone wanted to get on and they took off with an empty seat that is a loss in Utility. If the airline turns around and sells that seat to the new person that increases Producer Surplus (and conversely decreasing Consumer Surplus) to increase Utility.
I don't think this situation (empty seat) is DWL in the traditional sense [2]. The traditional DWL scenario is when producers could supply more but don't because of pricing constraints or other artificial ceilings/floors. It is definitely a marketing allocation inefficiency which results in Utility loss and I am kind of interested to see if there are better quantifications of this loss...but it isn't a DWL in surplus technically.
> Everyone then is better off. In fact, on a lot of flights, you can probably find multiple people to bump with just a token payment -- they'd rather stay another day.
It's totally possible that everyone can be better off. The issue isn't about when everyone is better off...it's when the system breaks down and you have situations where no one wants to make that trade.
> The bumped guy is happy to have the money, the last minute guy is happy to be on the plane, and the airline is happy to make $4700 extra.
This is an example of maximizing producer surplus. In my mind it's a variation of Price Discimination [3]. And while this overall beneficial situation is plausible, it isn't really overbooking.
For me, overbooking is basically selling more goods than you actually have and playing on the statistics trying to bank on people not making the flight. If you sell 30 seats and 5 people don't show, you take off with 25 people on board. If you sell 35 seats and 5 don't show, you take off with a full flight and pocket some additional revenue from the 5 no shows.
Now there is a lot of gray area since most no shows are from missed connections and so some of that additional revenue can go back into the system to allow for cheaper cancellation fees as well as accommodating rescheduling costs for these missed connections. But in general I suspect that producer surplus increases more than the gain in consumer utility.
> This works even better with a name-your-price bump question on checkin for flights that are oversold. Even if it's nonbinding, the airline now knows that it has willing passengers to bump and can keep selling tickets.
This is again, Price Discrimination but now adding information imbalance in order to help the airline more effectively Price Discriminate thereby maximizing their Producer Surplus.
Technically the United situation isn't the result of overbooking anyway...but I still think the practice warrants scrutiny. United's recent announcement to increase payouts and mileage compensation is testament to the fact that there was Producer Surplus to spare.
Like I said, all of this is speculative discussion and there is a lot of gray area. It is also plausible that overbooking results in maximizing Utility and an even handed redistribution of benefits to both producers and consumers. But I believe Producers are driven by incentives...and they don't have strong incentives to share the benefits of overbooking altruistically back to the consumer.
I fly from SFO - PDX every other weekend and consistently, no matter how early I book, Virgin America flights are cheaper than United and Virgin doesn't overbook whereas United does. Generally speaking, if United were even slightly returning the added revenue from overbooking to the consumer shouldn't their flights be cheaper than non-overbooked competitors?
Again this observation and subsequent question is anecdotal and broad...but I like it for FFT.
[1] https://en.wikibooks.org/wiki/Principles_of_Economics/Utilit...
[2] https://en.wikipedia.org/wiki/Deadweight_loss
[3] http://thismatter.com/economics/price-discrimination.htm