Finding Lowest Cost Flights

Welcome back. I don’t fly much anymore, do you? Do you try all the hacks trying to find the cheapest flight? You know, travel on Tuesday, Wednesday or Saturday? Be flexible on days? Avoid peak travel times of summer and holidays? To fly in one peak season, book in the opposite season?

Flyopedia’s tips for getting the cheapest flights (from www.flyopedia.com/blog/amazing-travel-tips-for-getting-the-cheapest-flight-tickets/).

I wish you luck, but a recent study suggests many of those won’t work. Researchers with Chicago; California, Berkely; Texas at Austin; and Yale universities examined data from a large U.S. airline--daily prices and quantities, capacity choices, fare decisions, internal demand model, demand estimates, flight-level forecasts, pricing design, as well as all consumer interactions (clicks) on the airline’s website. The sample covered 300,000 flights and 470 domestic routes over two years.

What they found, which is representative of airlines around the world, was at odds with what most economists would expect and most consumers assume. Permit me to highlight some key points.

Convenience vs. Price
Travelers generally select flights that balance convenience and price: The price of one flight might lead people to select a less convenient but cheaper flight.

Airlines don’t consider that kind of substitution. Instead, they think about the prices of seats on each flight rather than the total seats sold in a day. This, despite changing the price on one flight, will affect the way people think about all options.

Pre-Set Prices
If one airline cuts prices, you would expect others to do the same. But airlines don’t incorporate competitors’ fares in price-setting. This is the result of the specific pricing heuristic--decision-making shortcut--airlines use.

Airlines have a fixed and relatively small number of prices they assign to tickets on each flight. There are large gaps between possible prices, sometimes $100 or more. They might sell the first 30 economy tickets at the lowest price, the next 30 tickets at the next possible price, and so on.

Tickets are sold through global distribution systems that ensure everyone sees the same price. The system emerged from an industry alliance to facilitate inventory management. Other travel sector businesses (e.g., hotel rooms, car rentals) do the same.

Airlines are thus relatively unresponsive to real-time changes in cost. The next discrete fare is often a significant jump.

Departments Don’t Coordinate
Probably the most interesting (OK, strangest) discovery is the process airlines use to set prices.

Following deregulation in 1978, airlines adopted pricing heuristics that relied on decentralized decision-making by separate departments. The network planning department decides where to fly and assigns initial capacities. The pricing department designs itineraries and chooses a menu of fares. The revenue management department (RM) estimates demand, develops flight forecasts and monitors flight-level performance. These inputs are then given to the observed pricing heuristic that decides how many seats it is willing to sell at each fare level.

Key departments and responsibilities at all airlines; solid arrows show decision flow across departments, dashed arrows show departments that influence the pricing heuristic (from fig 1, academic.oup.com/qje/advance-article-abstract/doi/10.1093/qje/qjad051/7284417).

The researchers document three problems.

First, the pricing heuristic does not solve a dynamic program. Every flight is optimized using the same heuristic, regardless of market structure.

Second, department input decisions are biased and subject to miscoordination. Departments do not incorporate all of the decisions made by other departments, leading to input decisions that are incompatible.

Third, departments manipulate inputs they supply to the pricing heuristic. For example, RM analysts inflate the demand model, causing 93% of flights to be over-forecasted. It’s estimated that over-forecasting demand reduces the percentage of flights that are priced on the inelastic side of the internal demand model by up to 60%.

Wrap Up
One of the researchers expects that, over the next several years, airlines may start to adopt more dynamic pricing platforms, and non-business travelers may benefit. For now, trying all the hacks in search of the lowest fares may be futile. 

Static vs. dynamic airline pricing strategies (from expertassignmenthelp.com/price-analysis-case-study-on-comparative-flight-prices/).

She does suggest not waiting until the last minute as prices go up significantly days before a flight. I haven’t looked at that lately, but thanks for stopping by.

P.S.
Study of an airline’s organizational structure and pricing in The Quarterly Journal of Economics: academic.oup.com/qje/advance-article-abstract/doi/10.1093/qje/qjad051/7284417
Article on study on EurekAlert! website: www.eurekalert.org/news-releases/1004467