How Delta is Out-Managing its Rivals: A Detailed Look at RASM

The Q3 results are out, earnings calls are complete, and once again we find the season ending with the same story: Delta is winning while American and United struggle to explain why. Operating margins for the three legacies show a clear story: Delta leads with 17% margins while both American and United saw 11%. With a market cap of $35.9B for Delta, $22.7B for American, and $17.7B for United, Wall Street has clearly shown their preference for the Delta way of doing business. Consider that Delta is roughly the same size as United yet twice the market cap, and you start to understand just how differently Wall Street regards these airlines. 

Consider that Delta is roughly the same size as United yet has twice the market cap...

The metric that has risen to the top when comparing the legacy airlines is RASM, or Revenue per Available Seat Mile. Once again, Q3 showed Delta having operated with the highest RASM, either when looking at just passenger revenues (PRASM) or total revenues (TRASM). The ability for Delta to produce this revenue without increasing capacity (ASMs) is a clear indication of their revenue efficiency, and a strong positive sign to Wall Street.

It is within this context of results that theories of how Delta is producing higher RASM and margins are starting to come from senior management at United and American. Just last week, Doug Parker had this to say on the earnings call regarding the difference in margins:

"Margins by region come and go, margins by hub come and go, but what I can tell you right now is Delta has an airline that flies over 40% of their flights in and out of Atlanta, which is a really, really good hub,and if American flew 40% of its flights in and out of Charlotte, we would have a margin advantage there in the business because Charlotte's a really, really good hub."

"It's not about mismanagement or anything close to it. It's about the networks that are currently in place at the airlines. They have that advantage, by the way, in valuation."

According to Parker, Delta is fundamentally positioned as a more profitable airline due to their Atlanta hub. Setting aside the lack of confidence those comments must have shown to the analysts on the call regarding American's ability to match Delta's profitability, it seems logical. The industry appears to have accepted the reasoning as can be read in TheStreet's article this week. Is the profitability of airlines purely a product of their network, built more from years of historical circumstance than strategic management, or has Delta figured out what American and United have not? The challenge is to understand how Delta is outperforming its peers, and most importantly, why.

The intricacies of RASM

While most people in the industry are familiar with the concept of RASM, its true effects are often overlooked. Sometimes referred to as “revenue efficiency,” RASM is simply the amount of revenue brought in for each seat flown one mile. It is often confused with yield, yet differs in that yield looks at how much revenue was brought in per passenger per mile. RASM also considers the empty seats, which makes it a key indicator of not only how much revenue was brought in, but also how much capacity was deployed to capture it. 

At its core, RASM is a simple fraction with revenue on top and total capacity in the denominator. As you would expect, the easiest way to increase RASM is to increase the numerator of the fraction: total revenues. Charge more per ticket without adding any additional capacity, and your RASM goes up. Sounds easy, right?

Based on this methodology, Delta should be commanding higher yields proportional to their higher RASM. Yet it is not entirely fair to just compare yields and consider them indicators of how well an airline is managed since each airline is subject to unique yield challenges in the regions they are most present. Delta is strongest across the Atlantic, American has a higher exposure to Latin America, and United is dominant across the Pacific. To cancel out these region-specific yield challenges, we will only look at the domestic environment where each of the three compete to comparable levels. If Delta is achieving the highest RASM, they must also be achieving the highest yields, right?

...not quite. In fact, not only is Delta not the leader in domestic yields, American is. Whether we look at pure yields or adjust for stage length, American is able to charge the highest ticket prices per passenger-mile flown. Since United has a higher stage length than the average of the three, they get credit when adjusting for stage length, however it is still not enough to bridge the gap.

Delta is not leading domestic yields, American is.

Legacy networks and LCC presence

Another benefit Delta is said to enjoy from its network is a lack of LCC presence. It is no secret that fares are more difficult to maintain at a high level when a market has a low cost carrier present. Looking at the legacy networks, it may be easy to say Delta sees a lower level of LCC competition, but is that truly the case?

Once again the perception of Delta's network differs from the reality. When looking at the numbers from a passenger option perspective, Delta actually sees a much higher exposure to LCCs than American. This is largely due to their heavy presence in Florida where LCCs are present in every market, as well as Seattle with Alaska and even Atlanta, a hub for Southwest and a focus city for Spirit. Compared with American, who benefit from monopoly markets like Charlotte, Delta has far more exposure to LCCs on their domestic network. United sees the highest exposure, with almost 80% of all tickets sold having an LCC alternative. Since LCC presence tends to drive down yields in the markets where they are present, the next logical place to look is within the individual hubs themselves and how well Delta is managing traffic and yields compared to their rivals.

Is Atlanta responsible for Delta's success or is Delta responsible for Atlanta's success?

Which brings us back to Doug Parker's take on why Delta is performing so much better than his airline, Atlanta. This is a more challenging analysis to perform, since hubs are infinitely complex creatures of local and connecting traffic. Any airport can be a connecting hub, however only those markets with a large enough O&D presence can make a hub economically viable. As such, the differentiator between success or failure of a hub can be largely measured by the local yields of the home airline, as shown below.

The results are surprising. Atlanta performs well as expected, however it is not the highest yielding domestic hub within the U.S. That distinction goes to Charlotte and American Airlines. In fact both United and American have higher yielding hubs than Delta in Washington Dulles and Philadelphia. Even Detroit maintains higher yields for Delta than Atlanta.

Also of interest is how well the markets of Newark, Houston Bush, and Dallas/Ft. Worth perform. These are three fortress hubs with other LCC airports within their catchment areas (Jetblue at JFK, Southwest at Houston Hobby, and Southwest at Dallas Love). Even with this LCC competition at other airports in the same market, the legacy carriers are able to command strong yields at their home market, suggesting they do not have the direct impact previously assumed. This is not the case in markets such as Chicago O'Hare, New York JFK, and Los Angeles where the challenge to the legacies is less the LCCs than it is the presence of each other. In fact, Denver, which arguably has the highest LCC presence of any legacy hub, still performs as well as Chicago O'Hare and better than Los Angeles for United. Both cities share a common characteristic: intense competition from other legacy carriers.

Now What?

So far we've only been able to discount the main reason for Delta's RASM advantage, yet we know the advantage is real. While this RASM strength is certainly helped by the facts that Delta does see strong domestic yields and Atlanta is a respectable hub coming in at #5 in the top yielding hubs, neither of these are unique to Delta. In fact American, not Delta, sees the best pure domestic yields, and two of American's hubs out-perform Atlanta from a local yield basis. Exactly how is Delta producing consistently higher RASMs than American and United?

To answer that, it requires us to revisit the RASM equation and the other half of the equation we've been ignoring: ASMs. Remembering that RASM is a measurement of the efficiency of generated revenues, the capacity generated will have an equally large impact to the RASM equation. Measuring the efficiency of capacity deployment can be complex, however at its core, what we really want to know is how closely the macro capacity matches the macro demand.

To measure that, it requires us to understand what the underlying demand of the domestic market looks like. The graph to the left shows the percentage of markets by number of passengers. In effect, we are looking at the natural demand curve of the U.S. domestic market. As markets see increased frequency or combined traffic through hubs, they move up and down the demand curve, however the demand curve remains intact. If an airline were perfectly revenue efficient, their departures would line up perfectly with this graph. Of course that is not realistic within the limitations of fleet choices or schedule requirements, however the deviation from this perfect efficiency allows us a way to measure the airlines performance.

When looking at the airlines' actual departures compared to optimal, Delta's ability to drive higher RASMs becomes clear. With a combined error of 34% from optimal, Delta has a strong ability to manage their capacity when compared with American (56% error) and United (46% error). This alignment of capacity to demand is largely a function of the fleets chosen by the airlines. Based on the above graph, the peaks immediately become obvious at the 50-seat and 150-seat level. This is easily attributed to the large fleet of small regional jets and 737-800/A320 narrow-bodies at United and American. More importantly though, is the valley between 80 and 120 seats. While not perfect, Delta has done a far better job efficiently filling the need for flights with demand between 80 and 120 seats than its peers. This is predominantly due to Delta's willingness to embrace the 100-seat market with the 717 and eventually the CS100; a market largely discounted by United and American as being too expensive. In fact Scott Kirby at United has said publicly several times that he does not see the 100-seat segment as being economical, even while his management team suggests otherwise.

Delta remains the most efficient legacy airline at aligning capacity to demand...

With United looking only at larger narrow-body aircraft, the RASM gap with Delta is certain to last well into the future. For those markets between 80 and 120 passengers, Delta is not required to offer ultra-low basic economy fares or to depart with empty seats, a RASM-boosting luxury not available to the other legacies with large peaks and valleys in their capacity deployment. The Widget remains the most efficient legacy airline at aligning capacity to demand, and the results show in industry-leading RASM.

Appendix - A Quick Note on the Problem with Yield

You may have noticed I used the term "Pure Yield" several times in this article. Since it is often mistakenly assumed that yield is synonymous pricing power, separating the two became important.

Yield is simply the average revenue brought in per passenger per mile. It differs from RASM in that it doesn’t count empty seats. The challenge with yield is that it is typically calculated top-down by taking all revenues and dividing them by the total miles each passenger has flown (revenues divided by revenue passenger miles or RPMs if you’re keeping score). In an airline network where roughly half of the passengers are on connecting itineraries, this method generates an artificially low yield. 

Consider this example: A passenger buys a ticket to fly from Austin to Fort Lauderdale for $400. At a distance of 1,106 miles, this would bring a yield of 36 cents per mile. Only this passenger doesn’t fly directly from Austin to Fort Lauderdale, they connect through Charlotte; traveling a total distance of 1,664 miles. This results in a reported yield of only 24 cents. As much as we like to think yield is a view into an airline’s pricing power, it is still largely subject to the design of the network, a sort of self-fulfilling prophecy. 

Yet it is much easier to divide total passenger revenues by RPMs to arrive at yield, so it remains the industry norm. Still, this distorts the view of the actual pricing power of an airline. To calculate this measurement of pricing power, or pure yield as I call it, you would need to add all ticket revenues and divide by the market distance, or the actual distance between the passengers origin and destination… i.e. where they truly wanted to fly. It takes detailed data and an author crazy enough to spend the time to aggregate this data. Fortunately we have the massive DOT DB1B database and IATA’s PaxIS to create the detailed data. As for the crazy author…