Skip to content

Finding Energy Efficiency in an Unexpected Place – The Cockpit

I suspect that most energy economists think there are more unexploited opportunities for energy efficiency in homes than in firms. Firms are cost-minimizers, after all – they’re in the business of making things with the fewest possible inputs. And, energy is an important input for many firms, particularly airlines. So, not even McKinsey – in their exhaustive catalog of potential energy efficiency measures – identify improved fuel efficiency from airlines.

Surprisingly, a new NBER working paper finds a significant opportunity for fuel savings in the airline industry. In a research coup that makes people like me drool with envy, Greer Gosnell, John List, and Rob Metcalfe convinced Virgin Atlantic Airways to let them run an experiment. And, not just any experiment – one that involved their captains, the head honchos, as in, “This is your captain speaking.”

The cockpit of a Virgin Atlantic Airbus A340
The cockpit of a Virgin Atlantic Airbus A340

The researchers sent three randomly selected groups of Virgin Atlantic captains either (a) information about average fuel efficiency on the flights the captain made in the previous month, (b) the same personalized information as group (a) plus personalized targets for the coming month, or (c) personalized information and targets, plus an offer to donate 10 GBP (which was worth more pre-Brexit…) to a charity for each target achieved. A fourth set of captains was in the control group and knew the experiment was going on, but didn’t receive a monthly mailing.

Pilots who only received information – group (a) – significantly improved their fuel efficiency, while pilots who received personalized targets improved their fuel efficiency even more, achieving about the same gains as the pilots who could donate savings bonuses to charity.

Perhaps most surprisingly, even pilots in the control group improved their fuel efficiency considerably relative to the months before the study began. The authors speculate that this is an example of the so-called Hawthorne Effect, which suggests that people behave differently when they know they’re being studied.

The adjustments yielded substantial savings, netting the airline more than $5.4 million in Virgin-Atlantic-Airways-airbus2fuel savings over the 8-month pilot period. Presumably, this would scale up to about $8 million over a year – not too shabby considering that the company’s profits (before taxes and exceptional items) were around $20 million in the year the study took place. And, since the costs of running the experiment were very low – less than $3,000 to send a couple hundred letters a month – this appears to be an example of a negative cost abatement strategy. The authors calculate approximately -$250 per ton of CO2 (yes, negative, since the company saved money on net).

In my mental model, pilots are super-humans. They’re better than super-computers, and could probably even beat Gary Kasparov at chess. In other words, they always, always make the right decisions. This might be what my psyche requires to get on a plane, plus a function of being a teenager when the movie Top Gun came out.


So, what adjustments did these super-humans make to save fuel? And, do we really want them to be thinking about fuel efficiency and not MY SAFETY? As it turns out, most of the improvements involved simply – ahem – following the rules.

For example, when they’re taxiing, pilots are usually supposed to turn off half of the engines (e.g., one on a two-engine plane and two on a four-engine plane). Before the experiment, pilots did this on 35% of the flights, and after they knew they were being observed, this increased to 51%. Other adjustments involved following advice from air traffic control about more efficient routes or using real-time information about the baggage onboard to adjust fuel levels.

So, if even super-human pilots can be nudged into saving fuel, can we expect to find lots of similar opportunities across other industries? Think of the energy inputs controlled by power plant operators, truck drivers, or building supervisors. I’m personally optimistic, but two points temper my enthusiasm:

  • Nudging people to do things they weren’t doing before likely imposes additional costs. (Hunt Allcott and Judd Kessler explore this point in more detail here.) So, even if the cost of sending letters to the pilots is low, there may be other, unobserved costs. For example, whatever extra time it takes the pilots to update the fuel calculations is time they were previously spending doing something else. And, even if it was just getting a cup of coffee at the airport, it was something they must have enjoyed, because they chose to do it.

Greer, List and Metcalfe thought about this, and surveyed pilots after the experiment. They found that if anything pilots reported higher job satisfaction, especially those who met their personal goals and donated to charity. While we can’t rule out additional costs (unobserved costs are notoriously hard to quantify…), that result at least suggests they are small and possibly negative.

  • While the fuel savings are decent sized and statistically significant, they amount to small tweaks to the way an industry does business. But, climate scientists suggest that we need to reduce GHG emissions by around 80% to avoid dramatic disruptions, which we can’t do with only small tweaks. Every little bit certainly helps, but we can’t stop p19he9v0a7157b1eo01ib518a619rr4looking for more fundamental changes within sectors. (The Wall Street Journal had a recent piece on electric and hybrid planes, which still sound a bit futuristic, but you never know.)

As the researchers point out, providing feedback to the pilots involved large amounts of data (on over 40,000 flights), which were collected, quickly analyzed and sent to the pilots. At the most fundamental level, I see this experiment as an example of how using lots of personalized, high frequency data can give valuable feedback to decision-makers. Let’s hope more companies take similar opportunities to find savings this way – it’s in their best interest, as well as the climate’s.

Catherine Wolfram View All

Catherine Wolfram is the Cora Jane Flood Professor of Business Administration at the Haas School of Business, Co-Director of the Energy Institute at Haas, and a Faculty Director of The E2e Project. Her research analyzes the impact of environmental regulation on energy markets and the effects of electricity industry privatization and restructuring around the world. She is currently implementing several randomized control trials to evaluate energy efficiency programs.

11 thoughts on “Finding Energy Efficiency in an Unexpected Place – The Cockpit Leave a comment

  1. I’m also an aviation buff and I own and fly a small airplane. Fuel typically represents about half of my annual operating costs so I’m pretty careful about seeking out low prices and operating efficiently. However the saying goes that “aviation is unforgiving of the slightest error or omission”, and the consequences of errors and omissions are often fatal. Consequently, fuel efficiency is a ways down the list of my priorities when planning and executing a flight, and it should be the same for airline pilots. One of the WORST things a pilot can do is carry too little fuel to deal with adverse weather at a destination. That particular mistake continues to kill people every year in private aircraft, and it has resulted in several tragic commercial aviation accidents plus even more situations with what we call “high pucker factor” over the years.

    • I rent light piston singles, and my club rents “wet” (that is, they charge by the hour, with fuel included) as do most. I don’t know what the primary reason for that approach is, but I’ve always assumed that it is specifically to decouple the pilot from caring too much about fuel cost. Economically inefficient, but probably a big safety win, especially considering renter pilots unfamiliar with particular aircraft’s idiosyncrasies etc.

      Even so, in the spirit of being a good club member, I’ve overflown fields with expensive gas to continue on to ones where it’s cheaper. I did the calculations and it was all safe, but wouldn’t you know that it made the latter part of that leg somewhat less fun for me — as your “pucker factor” comment eludes.

      As you know, compared to commercial airline ops, the GA safety situation is not pretty, so the consequences of increasing efficiency awareness in that area are probably high enough to think twice.

  2. David J: I had somewhat similar concern. Safety and efficiency can be a trade-off at the margin. For example anything that increases the cognitive load during preparation for takeoff increases the chance of missing a checklist item. Rushing through an engine restart certainly could fit this.
    However, compliance was still nowhere close to 100%. ” Before the experiment, pilots did this on 35% of the flights, and after they knew they were being observed, this increased to 51%. ” So that 49% leaves plenty of room for pilots to say e.g. “It’s snowing and the airport is backed up… I’m going to minimize extra work today.”
    On the other-other hand, keeping track of any out-of-the-usual items increases cognitive load. I bet you also read the accident report from the aircraft that hit the Washington DC 17th St. bridge on takeoff. But we give pilots the ability, authority, and knowledge to make these judgments for good reason, and it’s reasonable for an airline to raise their awareness in this non-judgmental way.

    • Agreed. I should have emphasized that I think all of these scenarios, to the degree that they are realistic at all, are happening at a margin that is already very, very safe. But even excellent pilots still commit errors in judgment from time to time. I do think decisions that might inconvenience passengers or make them less comfortable have higher potential. I say this as someone who has casually noticed that most technological progress in commercial aviation in recent years seems to accrue to operators rather than passengers.

  3. As both an econ and an aviation buff, I appreciate this story.

    I wonder about whether the study captures all of the true costs and benefits, however. Pilots make operational decisions for many reasons and not all of them can be captured in terms of fuel cost ad some are external to the airline. Let’s consider a some scenarios:

    0. (Mentioned in the paper). Taking on less fuel. Pilots are required to take on fuel as determined by various calculations. Some carry more based on their judgment that it gives them more options in the air, but the extra weight is an efficiency penalty. Is cutting down to the absolute minimum required by their calculations really “free?” (Also, the /price/ of fuel varies by location, and airlines will sometimes overfill the tanks where gas is cheap and “tanker” it to their next destination where a fillup is more dear. Result: efficiency reduced, but dollars saved.)

    1. Pilot has personal policy to avoid storm cells by 50 nautical miles, even though her company policy only requires he remain 30 nm clear. Now that she is prioritizing fuel efficiency, she shrinks her margin to the company minimum. Result: small, but non-zero increase in risk to aircraft, also more unpleasant turbulent encounters.

    2. Aircraft operates more efficiently at altitude X, but the ride is smoother a few thousand feet lower. Does he descend for pax comfort or operate “optimally?”

    3. Aircraft is encountering stiff headwind, slowing progress and the flight will be late landing by 30 minutes. She can add power to increase airspeed by a few percent at the expense of efficiency.*** Reducing the delay to 10 minutes. How much is 20 minutes multiplied by an airplane full of people worth. (It turns out not very much, according to the models the FAA uses, but there are debatable assumptions in them.)

    4. Pilot taxis to the hold short line on one engine and gets a clearance to take-off immediately, forcing the pilot to complete his checklist — which now has one more major item: start engine — in a slight rush. Again, very small but non-zero elevated chance of error, with potential consequences ranging from extra wear and tear to catastrophic.

    One could come up with other “apparently” costless interventions for aviation efficiency that do actually have a cost for someone: running the APU less (letting boarding passengers bake in an un-air-conditioned plane), reducing weight by charging for bags by the pound instead of flat fee, designing aircraft that are more aerodynamic at the expense of comfort or speed, etc.

    *** For a given airframe, going a few knots faster in normal cruise is actually quite a bit more “expensive”. The power needed to overcome drag increase with the cube of the velocity. And engines themselves becomes less efficient when operating above their optimal power setting for a given circumstance. (

  4. This provides two refinements of the Hawthorne Effect. First, people behave differently when they’re being observed, ESPECIALLY WHEN THEY WERE PREVIOUSLY BREAKING THE RULES. Second, people often like being observed.

  5. Any follow-up info on:
    1) Has Virgin Atlantic acted on the study and implemented its lessons on a company-wide ungoing basis (and if so, with what results)?
    2) Have other airlines learned from the study as well and implemented its lessons?
    3) Did Gosnell, List, and Metcalfe get any share of the millions of dollars they saved Virgin Atlantic?
    (OK, that last question wasn’t serious).

  6. Very nice. Airlines are notable because they have already been aggressive about fuel conservation. The opportunities mentioned here e.g. using only one engine to taxi were developed through years of work. So it’s interesting that this behavioral aspect still adds a lot of value.

    Second comment. “I suspect that most energy economists think there are more unexploited opportunities for energy efficiency in homes than in firms.” Back in the 80s, my theory was that energy economists studied homes because they were amenable to econometrics. Many firms, on the other hand, are idiosyncratic and correspondingly hard to model en masse. Since then armies of consultants have stepped in to fill some of the knowledge gaps. But it’s doubtless still easier to do fundable and publishable econometric analyses of homes, than of industrial conservation.

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s

%d bloggers like this: