Circa 2000 B.C., horses were domesticated and used for transportation. A horse can travel about five miles per hour over moderate distances. For the next 3,800 years, technological innovation for traveling on land was almost nonexistent. In 1800, the horse still was the prime method of transportation on land—and still at about five miles per hour. So 3,800 years of stagnation!

Everything started to change in 1769 when James Watt designed an efficient steam engine. Originally, steam engines were used to power pumps and industrial machinery, but soon inventors saw the potential to use the engines to power boats. While a number of steamboats were built by entrepreneurs as early as 1787, Robert Fulton became known as the father of steam navigation when, on August 7, 1807, his Clermont completed the 150-mile trip from New York to Albany in 32 hours—still an average speed of five miles an hour.

If steam engines could be used to power boats, why couldn’t they be used to power wagons that ride on tracks? They could. In 1814, a self-educated British laborer named George Stephenson designed the world’s first steam locomotive to operate on tracks. The purpose of the locomotive was to haul coal from the mouth of a mine. Eleven years later, the first common carrier railroad, the Stockton & Darlington Railroad Company, was formed. Stephenson designed a locomotive for the Stockton & Darlington that could pull 6 coal cars and 21 small passenger cars over nine miles of track in about one hour—already nearly twice the speed of a horse.

In the United States, the city of Baltimore decided to build a railroad to Ohio to compete with the Erie Canal, which, at the time, was the least expensive transportation route from the farms in the Midwest. On July 4, 1828, with great fanfare, Charles Carroll, the last surviving signer of the Declaration of Independence, symbolically dug a spade of dirt to commence construction of the Baltimore and Ohio Railroad (B&O). In early 1830, the B&O became America’s first commercial railroad when it started operations on 13 miles of completed track. But expansion to Ohio was slow. Bridges had to be built, tunnels dug. In 1868, the B&O finally reached the Ohio River. Then, three years were required to construct a bridge to span the river. At last, in 1871, the railroad reached the farmlands of Ohio.

The almost immediate commercial success of the B&O in the 1830s and 1840s sparked a boom in the construction of railroads. By 1840, about 3,000 miles of track were in operation. By 1860, about 30,000 miles. Between 1815 and 1860, the cost of moving farm produce and manufactured goods over long distances fell by 95 percent. In 1860, trains could average about 20 miles per hour over a 24-hour day. The horse was on the way out for long distance travel. The railroads had revolutionized transportation. They were the Internet of their day.

Soon, the growth and potential profitability of railroads attracted the attention of Cornelius Vanderbilt, Edward Harriman, Jay Gould, and other wealthy businessmen and speculators, who purchased control of many regional railroads in order to create local monopolies that could set high rates (i.e., prices). In 1887, the U.S. Congress, responding to complaints from many farmers and other users of railroads about the high rates, passed the Interstate Commerce Act, which made the railroads subject to federal regulation. The act formed the five-member Interstate Commerce Commission (ICC), which had the power to regulate many aspects of the railroads, including the rates they could charge.

The railroads continued to grow and generally prosper until the 1920s, when the automobile started to erode passenger travel by rail. Then in the 1930s, the railroads suffered heavily during the Great Depression. Their revenues declined by about 50 percent between 1928 and 1933—and by 1937 many were in receivership. At the outset of World War II, the entire railroad industry was in trouble. War-related traffic from 1942 to 1945 brought a reprieve to the industry’s stressed condition, but the reprieve was temporary. Soon after the war ended, the construction of the interstate highway system gave trucks a competitive advantage over railroads for many types of cargo. By the mid-1970s, the railroad’s share of intercity freight traffic had declined to 35 percent, down from 75 percent in 1930. Usage of railroads to carry passengers also declined sharply after the end of the war, partially due to the construction of the interstate highway system and partially to the commercialization of the jet airplane.

The plight of the railroads was further aggravated by poor and excessive regulation. The ICC established rates. In an effort to help farmers, rates for grains and other bulk farm produce generally were kept low, while rates for manufactured goods generally were kept relatively high. The high rates for manufactured goods incentivized many manufacturers to switch to transportation by trucks. Thus, the railroads faced adverse selection. They lost many of their most profitable shippers to trucks, while still being saddled with their marginally profitable agricultural shippers. Furthermore, many of the ICC’s multitude of rules and regulations stifled the introduction of such efficiencies as unit trains.

The railroads were in trouble. In the 1960s and 1970s, many went bankrupt, including, on June 21, 1970, the once venerable Penn Central, one of the largest and most important railroads in the country. Finally, even the U.S. government recognized that change was needed. In 1978, the U.S. Department of Transportation noted that “the current system of railroad regulation … is a hodgepodge of inconsistent and often anachronistic regulations that no longer correspond to the economic conditions of the railroads, the nature of intermodal competition, or the often-conflicting needs of shippers, consumers, and tax payers.” (“A Short History of U.S. Freight Railroads,” paper by the Policy and Economics Department of the Association of American Railroads, April 2013, page 5.)

In 1980, Harley Staggers, chair of the House Interstate and Foreign Commerce Committee, introduced an act that essentially deregulated the railroads. Under the Staggers Act, railroads could establish any rates they chose, unless the ICC determined that there was no effective competition for the service. Other provisions of the act increased the flexibility of the railroads to provide efficient service. For example, procedures for abandoning unprofitable lightly used tracks were streamlined.

The Staggers Act was a breath of fresh air. Railroads immediately started adjusting their rates to make economic sense. Unprofitable routes were dropped. With increased profits and with confidence in their future, railroads started spending more to modernize. New locomotives, freight cars, tracks, automated control systems, and computers reduced costs and increased reliability. The efficiencies allowed the railroads to reduce their rates and become more competitive with trucks and barges. According to the Association of American Railroads, the average inflation adjusted rate charged by the railroads, expressed in 2012 dollars, declined from about $0.07 per ton-mile in 1981 to below $0.03 in 2003—an amazingly large decline.

In the 1980s and 1990s, the railroad industry also enjoyed increased efficiencies through consolidating mergers. In the west, the Burlington Northern merged with the Santa Fe, and the Union Pacific merged with the Southern Pacific. As a result of the mergers, only two large railroads remained to serve the western half of the country. With increased efficiencies of scale and with reduced competition post the mergers, both the Burlington Northern Santa Fe and the Union Pacific could stabilize their earnings at a time when their rates were still declining.

During the 2001–2002 recession, Union Pacific’s management reduced the railroad’s costs in order to mitigate the effects of the slowdown. Employment levels were reduced by 8.1 percent, from about 50,500 in 2000 to about 46,400 in 2003. The reduced costs permitted the railroad to enjoy larger earnings in 2003 than in 2000. However, the reductions left the Union Pacific with insufficient capacity when demand for rail transportation accelerated in 2004. The insufficient capacity led to serious congestion on many routes. An analogy might be traffic jams on a highway during rush hour. Too much traffic on a highway, and the traffic slows and backs-up. Union Pacific’s congestion caused its profits to decline. Its trains took longer to reach their destinations. A railroad’s revenues are a function of ton-miles, but many of its costs, especially hourly labor, are a function of time. The more the delays, the higher the costs. Furthermore, Union Pacific needed to spend large sums to hire and train an abnormally large number of new employees—and also large sums to repair and maintain old locomotives that were brought out of retirement to meet the increased demand.

Investors became aware of the earnings shortfall during the first half of 2004. On the last trading day of 2003, the price of Union Pacific’s shares closed at $17.37.¹ On June 30, 2004, the shares closed at $14.86, down 14.5 percent during the six-month period. During the same period, Burlington Northern Santa Fe’s shares appreciated by 8.4 percent.

It seemed logical to us that Union Pacific’s congestion problems could be and would be solved within a few years. New employees could be hired and trained, new locomotives could be purchased, some trains could be rerouted away from congested lines, and new track could be added in some of the more congested corridors. Decongestion should not be a difficult task. It also seemed logical that the stronger market for rail traffic would lead to at least modest rate increases. We constructed an earnings model for Union Pacific for the year 2006 that assumed that the railroad had solved most of its problems and had increased its rates modestly. Based on these assumptions, our model projected that Union Pacific would earn close to $1.55 per share in 2006. By comparison, we expected the railroad’s 2004 earnings to be only about $0.70 per share. I decided to value Union Pacific’s shares conservatively at 14 times earnings. Therefore, I estimated that the shares would be worth just over $22 in 2006. In addition, the company was paying a $0.30 annual dividend. Including the dividends, I concluded that the shares, which were selling at about $14½ at the time, could provide a 55 percent total return over the two-year period 2004 to 2006. I decided to purchase the shares. They did not appear to be a particularly exciting investment opportunity, but the stock market and our stocks had been strong in 2003, and we did not have alternatives that appeared to be more attractive. Furthermore, and importantly, I believed that the shares had very solid protection from material permanent loss. Their book value was $12. The company had a strong balance sheet for a railroad. And the company and Burlington Northern were a duopoly in the western half of the United States in an industry that remained vital to the country.

There was another reason to own shares of Union Pacific rather than holding cash. Over a period of many decades, the stock market, as measured by the Standard & Poor’s (S&P) 500 Index, has enjoyed an average annual total return of roughly 9½ percent—and our best guess was that returns could average somewhere around 9½ percent over the next few decades.² When we purchase a stock, we believe that it will appreciate by far more than the stock market appreciates. However, what if we are wrong and the stock appreciates only as much as the market? Then, if the stock is typical, it should earn an average annual return of about 9½ percent over time, which is much better than holding cash.

I note that the 9½ percent long-term upward bias of the stock market is one reason that shorting stocks generally is a bad business. To do as well as a market that tends to provide a 9½ percent annual return, a shorter of stocks must find stocks that decline by at least 9½ percent per year or that underperform the market by 19 percent. There are not many investors in the world who can outperform the stock market by 19 percent per year, and thus I imagine that there are not many investors who can find stocks that will underperform the market by 19 percent. Furthermore, when an investor is long a stock, the most he can lose is his original investment—and his upside profit potential is unlimited. When an investor is short a stock, his loss potential is unlimited if the stock happens to appreciate sharply, and his profit potential is only equal to the value of his original investment. Thus, I believe that shorting stocks generally is a miserable business, except for rare investors who have extraordinary abilities to predict when individual stocks or when markets will decline sharply.

Another reason for owning high-quality stocks is that they sometimes benefit from unanticipated surprises. That is why my old boss, Arthur Ross, continually advised me to purchase high-quality stocks and “stay in the game, Ed, stay in the game” (remember, Arthur Ross usually said things twice when he wanted to make a point).

In the case of Union Pacific, we did materially benefit from an unanticipated surprise. The surprise had to do with the price of diesel fuel. At the time we established our position in Union Pacific’s shares, the price of diesel fuel was a hair over $1.50 per gallon. For the next several years, diesel prices increased steadily and sharply, reaching a high of about $4.70 per gallon in 2008. Railroads typically are three to four times more fuel efficient than trucks. Thus, to offset the higher fuel costs, trucking companies needed to increase their rates much more sharply than the railroads needed to. The sharply higher rates charged by trucking companies incentivized some shippers to switch from truck to rail. The increased demand for rail affected Union Pacific in two ways: (1) with demand exceeding supply and with trucks less competitive, Union Pacific could materially increase its rates in excess of the amounts needed to offset the increased cost of fuel; and (2) the increase in traffic delayed the solution to the congestion problem because it increased the amount of capacity that Union Pacific needed to add.

After purchasing our holding in Union Pacific, we constantly monitored the railroad’s progress. From the onset, it appeared that the company was receiving larger rate increases than we had projected, but that the congestion problems were stickier than we had expected. When looking ahead, we believed that the large rate increases likely would continue and, sooner or later, the company would benefit from decongestion, and when that happened the company’s earnings and share price should be buoyant. We were increasingly excited about our investment—so excited that we also purchased shares of other railroads that also should benefit from the anticipated large increases in rates.

Wall Street generally did not agree with our excitement. On July 13, 2005, at a time when Union Pacific’s shares were trading at about $16, an analyst with J. P. Morgan downgraded the shares and recommended that they be underweighted—which, in the language of Wall Street, means that they should be sold. I was intrigued that, in spite of the bearish recommendation, the analyst wrote favorable comments about the company’s longer-term potential. He wrote about Union Pacific’s excellent route structure and advantageous access to customers. He added that these strengths should lead to solid margin, earnings, and cash flow performance. However, the analyst was concerned that current capacity constraints and resulting operating inefficiencies would lead to short-term problems and continued disappointments that likely would take longer than a few more quarters to correct. Simply, the analyst could not wait much longer than a “few quarters” for the operational turnaround. We could.

By mid-2007, it appeared that Union Pacific was beginning to materially benefit from decongestion. In the quarter that ended on June 30, the company’s average train speed was 21.6 miles per hour, up from 21.2 miles per hour during the comparable year-earlier quarter. Average dwell time (the average time a rail car spends in a terminal) was 24.7 hours in the June 2007 quarter, a large improvement from the 27.6 hours during the June 2006 quarter. The company’s earnings in the first half of 2007 were up 19 percent year-over-year, in spite of adverse flash floods in several states that caused track washouts and bridge outages. Earnings estimates for the full year 2007 had increased to close to $1.50 per share—a sharp improvement versus the depressed $0.77 per share earned in 2004. Wall Street analysts had become more positive about the company’s outlook, and the price of the shares had increased to above $30.

I was amused that the same J. P. Morgan analyst who in mid-2005 recommended that the shares be underweighted at $16 now was recommending that the shares be purchased at $31½. In a report dated July 17, 2006, the analyst wrote that Union Pacific’s positive story was firmly intact, that the company had more pricing power than most other railroads, and that it had additional opportunities to improve its efficiencies. Ironically, the inefficiencies that were a reason not to purchase the shares in mid-2005 were a reason to purchase the shares in mid-2007. The analyst concluded that he saw meaningful upside potential for the shares.

I meet regularly with my associates at Greenhaven to discuss our ideas and holdings. By mid-2007, it was becoming clear to us that most other investors also believed that Union Pacific was benefiting from sizable rate increases and from operational efficiencies. Therefore, we concluded that a large percentage of Union Pacific’s potential likely was being discounted into the price of the shares. Thus, at one of our regular meetings, we decided to start gradually selling our holding. By the end of 2007, we had sold most of our shares at an average price of roughly $31. We had held the shares for an average of about 3.7 years. During that period, the shares had slightly more than doubled, and we had earned $1.05 per share in dividends. Our average annualized return was close to 24 percent, including the dividends we had received.

The J. P. Morgan analyst who recommended underweighting Union Pacific’s shares at $16 was wrong. After the recommendation, the price of the shares increased quite steadily through September 2008. My experience is that the recommendations of Wall Street analysts are wrong more often than they are correct. An investor told me that he once had an account with one of Wall Street’s most respected firms. He told the Wall Street firm immediately to purchase any stock that was added to the firm’s overweight list—and immediately to sell the stock when it was removed from the list. After several years of following this approach, the investor closed his account with the most respected firm because his results were particularly poor.

Why do analysts tend to be substandard stock pickers? Most analysts follow only one or a few industries and tend to have deep knowledge about the companies they follow. However, there is a large difference between knowledge and judgment. It is said that knowledge is knowing that a tomato is a fruit, but judgment is not putting it in your fruit salad. To have good judgment, you need to have the knowledge, but, in my opinion, you also need many other qualities, including common sense, stable emotions, confidence, and, quite possibly, an indefinable sixth sense.

Furthermore, in my opinion, most individuals, including securities analysts, feel more comfortable projecting current fundamentals into the future than projecting changes that will occur in the future. Current fundamentals are based on known information. Future fundamentals are based on unknowns. Predicting the future from unknowns requires the efforts of thinking, assigning probabilities, and sticking ones neck out—all efforts that human beings too often prefer to avoid.

Also, I believe it is difficult for securities analysts to embrace companies and industries that currently are suffering from poor results and impaired reputations. Often, securities analysts want to see tangible proof of better results before recommending a stock. My philosophy is that life is not about waiting for a storm to pass. It is about dancing in the rain. One usually can read a weather map and reasonably project when a storm will pass. If one waits for the moment when the sun breaks out, there is a high probability others already will have reacted to the improved prospects and already will have driven up the price of the stock—and thus the opportunity to earn large profits will have been missed.

In my opinion, one of the best articles about not waiting for storms to pass was written by Warren Buffett in a New York Times op-ed piece on October 17, 2008, at the height of the financial crisis. Warren Buffett wrote:

A simple rule dictates my buying: Be fearful when others are greedy, and be greedy when others are fearful. And most certainly, fear is now widespread, gripping even seasoned investors. To be sure, investors are right to be wary of highly leveraged entities or businesses in weak competitive positions. But fears regarding the long-term prosperity of the nation’s many sound companies make no sense. These businesses will indeed suffer earnings hiccups, as they always have. But most major companies will be setting new profit records 5, 10, and 20 years from now. Let me be clear on one point: I can’t predict the short-term movements of the stock market. I haven’t the faintest idea as to whether stocks will be higher or lower a month—or a year—from now. What is likely, however, is that the market will move higher, perhaps substantially so, well before either sentiment or the economy turns up. So if you wait for the robins, spring will be over.