The railroad with better profit margins than Google

Like many of Buffett’s investments, BNSF has been wildly successful, more than doubling its earnings since 2009. But Buffett might have been even more successful had he stayed closer to home and bought the railroad with headquarters

A few blocks away from Berkshire in Buffett’s hometown of Omaha: Union Pacific, No. 123 on this year’s Fortune 500 list. Union Pacific (UNP) and BNSF are far and away the dominant carriers in the Western U.S., and stand nose to nose in revenues. But as Buffett himself noted in Berkshire’s most recent annual report, his railroad’s service slipped badly last year, and Union Pacific both performed a lot better for customers and made a lot more money. “We lost market share as a result,” wrote Buffett. “Moreover, [Union Pacific’s] earnings beat ours by a record amount.”

Photograph by Andrew Hetherington for Fortune Magazine

Indeed, the 153-year-old Union Pacific—often referred to as “UP” within the rail industry—is not only outpacing Buffett’s railroad but also beating almost every other industrial company in the Fortune 500. The old-economy warhorse generates profits at a rate that rivals those of the best tech, pharmaceutical, and financial services companies. In 2014, Union Pacific logged $5.18 billion in net profits on sales of $24 billion, for a return-on-revenues ratio of 21.6%. By that measure, the railroad company ties Apple (21.6%) (AAPL) and beats J.P. Morgan (21.3%) (JPM), Goldman Sachs (21.1%) (GS), Intel (20.9%) (INTC), Google (20.2%) (GOOG), and Pfizer (18.4%) (PFE). The only non-oil industrial enterprise that ranks higher is forest products producer Weyerhaeuser (22.9%) (WY), whose sales are one-third those of Union Pacific’s.

The key to UP’s remarkable profitability is a highly disciplined investment approach overseen by a department called network planning. Developed over the past decade, the system is implemented by a group of analysts whom employees refer to as the “smart guys” (though many are women). The central tenet of network planning is that every outlay for new track, locomotives, or terminals must yield a return of at least 15%—and the “smart guys” brook no excuse for failure.

Given the railroad’s strong profitability, it’s no surprise that Union Pacific has richly rewarded shareholders. From the start of 2005 through the end of 2014, it delivered a total return of 746%, or 23.8% per year, compared with a 20.6% annual return for the S&P railroad index and 7.7% for the broader market over that time. With a recent market cap of $91 billion, Union Pacific is tied with UPS as one of the two most valuable transportation companies on the planet. It’s worth more than American Airlines, Delta, and United Continental combined.

Despite that performance, Union Pacific is suddenly facing big challenges, because two of its major growth engines, coal and shale oil, are under pressure. Even its vaunted efficiency has taken a hit as the sudden pullback in those businesses—following a big buildup in 2014—has saddled Union Pacific with too many people and too much equipment.

Now a new CEO must take on the challenge of getting Union Pacific back on track. In February the board appointed Lance Fritz to succeed the retiring Jack Koraleski, a 43-year company veteran. After years of flush times, it’s unclear whether railroads will keep thriving, or whether Union Pacific will retain its status as the best managed of its peers. Fritz, who most recently headed operations at UP, is well aware of the task ahead. “We were the best of the competition, but we lost some of the fluidity of our network,” he says. “We have a lot of work to do.”

The current downturn for the rail industry feels especially sharp because 2014 was such a blockbuster year. An extremely cold winter and high natural-gas prices led to a big increase in coal shipments to utilities, especially from Wyoming’s Powder River Basin. A terrific harvest swelled carloads of grain. As the economy rebounded, the number of trainloads of finished cars and auto parts surged. Business from shale oil production, especially transporting “frac sand” from mines in Wisconsin to prime regions such as the Bakken in North Dakota and the Eagle Ford in Texas where oil is extracted by hydraulic fracturing, expanded around 25% from 2013, building on three years of fast growth. All told, Union Pacific’s volumes soared 7% in 2014, around four times the 1.5% or so that’s normal in a good economy.

Even in late 2014, Union Pacific’s management was forecasting—and adding manpower and equipment for—a fantastic 2015. But so far, this year has proved to be a disappointment. The biggest surprise has come in coal. Electric utilities shifted en masse from coal to natural gas as gas prices dropped sharply in late 2014 and early 2015. At the same time the sudden decline in oil prices caused a slowdown in fracking activity. To make matters worse, labor strife at the ports in Los Angeles and Long Beach caused a big drop in shipments of imports from Asia. Union Pacific’s freight revenues actually fell 1% in the first quarter compared with the first three months of 2014, the first such decline in six years. After gaining 42% in 2014, the railroad’s stock had dropped 13% through late May, erasing around $14 billion in market value.

Investors are overreacting. The railroads, and especially Union Pacific, should quickly regain momentum—for three main reasons. First, the U.S. rail market essentially consists of two duopolies: CSX and Norfolk Southern in the Eastern U.S., and Union Pacific and BNSF in the West. (Two other railroads, Kansas City Southern and Canadian National, operate limited routes in the U.S.)

The best territory is the vast Western region. The Union Pacific network stretches across 23 states and 32,000 miles of track, from Los Angeles and Seattle to Chicago and New Orleans. The market it shares with BNSF offers gateways to the nation’s busiest ports, and ultra-long routes from origin to destination greatly lower the cost per mile of transporting boxcars of auto parts or tank cars of chemicals. UP and BNSF do compete for business, but not super-aggressively. “In the mid-2000s the railroads decided they’d been killing themselves competing extremely hard on price,” says Larry Gross of FTR Transportation Intelligence. “They learned price discipline, and they’ve become pretty darned disciplined.”

Photograph by Geoff Johnson for Fortune Magazine

Customers grudgingly accept price increases that consistently exceed inflation partly because service has substantially improved, but also because they have no other good choice. That’s led in the past to accusations of collusion, which the railroads, including Union Pacific, have strongly denied. But the industry’s price discipline—reminiscent of muted competition between airlines—shows no signs of going away. Hence, UP will continue to benefit from the biggest factor powering its profits: the ability to consistently raise prices faster than costs, and by a generous margin.

The second reason to be bullish is that Union Pacific’s revenue mix is highly diversified, giving it good growth prospects despite the recent reversals in coal and fracking. Its coal volumes have been falling for several years, and the combination of tougher environmental regulations and, in all probability, continued low natural-gas prices make it likely that the decline will persist. The fracking-enabled boom in production of shale oil has filled the gap from coal since 2010. But it’s highly uncertain if that will rebound or fade.

The fracking phenomenon has also created new problems. There have been a number of high-profile derailments of trains—including one by UP—carrying shale oil, much of which is produced in new drilling areas without established pipeline networks and must be moved by rail. That has drawn a lot of scrutiny to the railroads. The industry counters that accident rates have actually been falling consistently since 2004.

Click graphic to enlarge

In any case, Union Pacific’s business is hardly dependent on shale oil. Transporting sand, drilling pipe, and crude oil furnished only 4.5% of UP’s volumes at the peak in 2014. Meanwhile three other major franchises—chemicals, automotive, and agricultural products—are all showing strong gains. More than a dozen big chemical companies are building tens of billions of dollars in petrochemical plants along the Gulf Coast, and Union Pacific, boasting by far the region’s best routes, is destined to get most of their rail business. Best of all, the railroads enjoy a big, and growing, cost advantage over trucks for long-haul shipments. The shift of cargoes from trucks to trains will be Union Pacific’s principal growth locomotive in the future.

Third, Union Pacific is an expert at constantly, relentlessly improving its efficiency. In a hugely capital-intensive business, that means increasing its volumes of freight far faster than it adds new employees, locomotives, and boxcars. Don’t think that Union Pacific skimps on capital investment. Far from it: Its capital expenditure has almost doubled, from $2.2 billion in 2006 to a planned $4.2 billion in 2015. It simply gets more profit from the dollars invested than its competitors, and serves customers better as a result.

Still, the surge in business last year caused logjams in Union Pacific’s normally nimble network. The heavy traffic slowed its trains, and railcars spent far more time in storage between trips. To move all the new freight through an increasingly congested system—think of rush-hour traffic all day long—Union Pacific added 3,600 employees and purchased 260 new locomotives. The extra trains moved more cargo, but delivery times slowed considerably. For example, it took as long as 200 hours for a coal train to run from the Powder River Basin to Missouri and back, compared with 120 hours in 2013.

As a result, UP has recently begun furloughing employees and moving locomotives back into storage. As congestion eases, the speed of its trains—a crucial measure of efficiency called “velocity”—is starting to improve. Now its average speed is 24.5 miles an hour; that’s about 1 mph better than its velocity through most of 2014, but well below the average of 26 mph the previous year—a number it plans to exceed in the next 12 months. That improvement will be crucial: Each 1 mph increase eliminates the need for as many as 200 locomotives.

Union Pacific’s last three CEOs spent their entire, multidecade careers at the company. By contrast, the new boss, Fritz, 52, worked for a wide variety of manufacturers before joining Union Pacific in 2000. The son of an aircraft engineer, Fritz grew up in suburban Philadelphia and attended Bucknell University. After stints at General Electric (GE), conglomerate Cooper Industries, and Fiskars, a Finnish gardening products manufacturer, Fritz in 2000 accepted an offer to run Union Pacific’s coal division. “I went from running a $100 million business at Fiskars to a $2.5 billion business at UP,” says Fritz, who, after all the businesses he’d seen sold, appreciated the tradition of long service at UP.

As chief of coal, Fritz reworked long-term, below-market contracts with utilities, becoming an expert at trading better service for higher pricing. Fritz then cycled through a variety of jobs. In rapid succession he ran the railroad’s Northern region, then its Southern network, headed labor relations, and, before becoming CEO, oversaw operations for the entire railroad for four years.

Fritz says his most pressing challenge is to “right-size” equipment and personnel so that the network regains its old fluidity, but at the same time to keep sufficient railcars, locomotives, and personnel in reserve to handle a new surge in business. He recently furloughed 600 workers. But he’s also giving those workers eight days of work, and pay, each month, and continuing full benefits. The hope is that instead of spending months training fresh recruits, Union Pacific may be able to hire furloughed folks back quickly if needed. Fritz is extremely proud of the program. “We used the program in 2009, and 90% of the furloughed people came back, compared with 33% before,” he says. “I expect that to happen again.”

One of Fritz’s key lieutenants is chief financial officer Rob Knight, the architect of Union Pacific’s capital-management system. When Knight took the CFO job at Union Pacific in 2004, UP was the least profitable major railroad. Its operating margin stood at a paltry 10.6%. The following year Knight set a goal of 25% by 2010. “People thought I was crazy,” he recalls. “The attitude was, ‘All business is good business.’ People didn’t care about pricing.”

Photograph by Andrew Hetherington for Fortune

Knight knew that Union Pacific couldn’t much change its business mix or growth rate. Those factors depended on the economy. Instead, the crucial levers would be productivity (keeping the cost of shipping each carload in check through smart investments to increase the average speed, length, and reliability of trains) and pricing (ensuring that rates consistently rose far faster than costs). That was Knight’s formula for growing margins. To make the numbers, Knight figured that managers would need to deliver 15% annual returns on all new business and capital outlays.

Today the network planning group of 70 analysts oversees this process from cubicles on the 11th floor of Union Pacific’s office tower in Omaha. The “smart guys” are anything but wonks. Many are managers from the field who spend a year or two in the department and blend excellent math skills with rail yard know-how. A case in point is Danny Torres, who spent most of his career working in repair facilities and depots, and now runs a network of 10 terminals in Iowa. “We work with a financial model that says, How much profit will adding this siding or extra track add? Will it slow or increase efficiency in other parts of the network? When it’s all taken together, will the total return reach 15%?”

Knight also built a second financial function that might be called “green, yellow, red.” In each of the big operating businesses—coal, industrial products, chemicals, and so on—Knight installed financial managers to evaluate new business. They enter the proposed pricing on all new contracts, as well as the extra costs in fuel, manpower, and everything else the business will require, into an online operating system that projects the rate of return. If the number is well over 15%, the system flashes green. If it’s on the margin, the signal is yellow. “If it’s red,” says Knight, “and it’s the best pricing we can offer, we let it go.”

He gradually won converts. By 2008, UP achieved the 25% goal. But Knight didn’t stop there. Today UP’s operating margin stands at 36.5%, and Knight pledges to hit 40% by 2019.

On a late April morning, a female conductor at the Council Bluffs facility is moving train cars. She stands on the side of a single track filled with a long line of refrigerated boxcars carrying oranges, kiwis, and assorted fruits from California. A locomotive sits at the end of the train, ready to push the railcars forward. Around the conductor’s waist hangs a yellow control box about the size of a small backpack, what Union Pacific calls the “big belt pack.”

One after another, she uncouples the lead car, then hits a switch in the belt pack. By remote control, the device powers the driverless locomotive forward, just far enough to push the car onto one of 18 tracks that fan out from the single track. A second employee controls the switches that align the cars with the correct track. Within an hour, she has broken down 150 cars into groups of around 50 cars each that will be joined with more cars, then head to Minneapolis, Des Moines, and Topeka.

Before the big belt pack’s debut in the early 2000s, the conductor had to relay instructions on when to push the next car forward via walkie-talkie to a third employee—an engineer running the locomotive. Now that engineer is freed up for the long hauls that make Union Pacific so profitable.

The big belt pack—which UP developed and now licenses to other railroads—is just one example of how the carrier attacks costs. And cost savings from heightened efficiency allow the company to plow more money into extending its capacity.

Today Union Pacific is investing heavily to expand its best growth business: intermodal freight. So-called intermodal cargoes are typically finished products such as electronics, clothing, or appliances that are shipped in corrugated steel containers or trailers. The containers often travel long distances by rail from one hub to another. Then they’re loaded onto trucks—more than one mode of transportation is the “intermodal” part—and hauled a relatively short distance—say, 50 miles—to a Walmart (WMT) or Target (TGT) warehouse.

Intermodal divides into two categories driven by different forces: international and domestic freight. The two markets are around equal in size, and each offers excellent prospects for growth. The international shipments tilt strongly to imports. That business suffered badly last year and in early 2015 because of labor strife at West Coast ports from Seattle to San Diego. The bottlenecks did lasting damage. Fearing future strikes and slowdowns, big shipping companies switched business to the Eastern Seaboard. Still, the strong dollar is poised to overwhelm that shift. According to economist John Husing, an expert on California’s economy, shipments from the L.A. ports will jump to well over 8 million containers in 2015, a new record. “Keep in mind that clothing and furniture from abroad is a lot cheaper than a year ago because of the dollar’s rise,” notes Husing.

In recent years it’s the domestic side that’s been powering intermodal’s growth. And it’s destined to remain the main driver. U.S. intermodal is the territory where railroads compete most directly with trucks. For long-distance shipments, rail is winning. A major advantage is fuel efficiency. Trucks ship a ton of freight an average of 120 miles on a gallon of diesel fuel; railroads can move the same cargo 600 miles on that gallon.

But what’s bound to hasten the shift is the sudden shortage of manpower plaguing the trucking industry and swelling its costs. In mid-2013, federal regulations imposed stiff new limits on the hours truck drivers are permitted to spend behind the wheel, rules designed to improve highway safety. As a result, the industry can’t recruit enough drivers even as the economy rebounds. Since last year about 1.4 million commercial trucks, over 4% of the total, have disappeared from America’s roads.

Photograph by Geoff Johnson for Fortune Magazine

Today intermodal accounts for 20% of Union Pacific’s revenues, and it’s been growing at around 6% a year. Given the strong dollar and stress in trucking, the pace will quicken. Research firm Trefis forecasts that intermodal will become Union Pacific’s biggest moneymaker by 2021, growing cash flow 77%, to $3.76 billion.

To efficiently handle the growth, Union Pacific is investing heavily in its principal artery for intermodal. That’s the Sunset Route, a 760-mile corridor running from Los Angeles to El Paso. UP acquired the Sunset in its merger with Southern Pacific in 1996. Then, as now, the Sunset Route connected the L.A. ports to the biggest intermodal destination: Chicago. But that route competed directly with BNSF’s far better intermodal service, and BNSF was getting a lot more of the business. The reason was basic: BNSF’s L.A.–Chicago route had two side-by-side tracks along almost its entire 2,200-mile distance. By contrast, the Sunset Route was mostly single track. Instead of running trains in both directions, Union Pacific had to park a westbound train on a siding, often for hours, to avoid an eastbound train using the same track.

Over the past decade or so, UP has spent $1 billion to equip 80% of the Sunset Route with double tracks, and plans to reach 100% in the next few years. As a result 55 trains a day now run on the Sunset, compared with around 40 in the mid 2000s. It now carries one-fifth of Union Pacific’s total traffic. UP still lags BNSF in intermodal revenues, but it is gaining market share. The double-tracking has helped raise the average length of Union Pacific’s intermodal trains—which are now less limited by the need to fit on sidings—from 158 cars in 2007 to 170 today, meaning that UP is handling far bigger volumes.

When the double-tracking is completed, Union Pacific plans to run 90 trains a day on the Sunset Route. To prepare for all that traffic, it has just opened a giant new terminal in the New Mexico desert, near the Mexican border just west of El Paso. The $415 million Santa Teresa facility is the gateway to three major routes, branching to the Midwest and Chicago, Dallas–Fort Worth, and the Gulf Coast. Intermodal trains will refuel and change crews there before fanning out.

At Santa Teresa and other terminals near the Mexican border, Union Pacific is also capitalizing on Mexico’s fast-growing manufacturing base. A big portion of that growth is in automobiles. In recent years virtually all the world’s major automakers—Nissan, Ford, Chrysler, BMW, and many others—have announced or completed major manufacturing hubs in Mexico, in projects totaling $12 billion. The nation’s car production is forecast to rise from 3 million vehicles in 2014 to 4.7 million by 2019. Union Pacific owns a big chunk of Ferromex, or FXE, one of Mexico’s leading railroads. FXE carries trainloads of cars to Union Pacific’s gateways, where the cargoes switch to UP locomotives. Union Pacific now transports two out of three of the new automobiles that Mexico exports to the U.S.

The railroad’s superior strength in Mexico isn’t lost on Buffett. During his presentation alongside Berkshire vice chairman Charlie Munger at the company’s annual meeting in April, Buffett remarked, “Union Pacific’s rail network is much better positioned for Mexico than BNSF.” Buffett might have picked the wrong railroad to buy, but he knows a great business when he sees it. 

A version of this article appears in the June 15, 2015 issue of Fortune magazine with the headline ‘Profit Engine on the Rails.’