OPPORTUNITY RECOGNITION AND DISRUPTIVE TECHNOLOGY: THE U.S. LOCOMOTIVE INDUSTRY FROM 1920 TO 1940

William R. Sandberg, University of South Carolina
H. Dixon Wilcox, University of South Carolina

ABSTRACT

We examine the U.S. locomotive industry from 1920 through 1940, the first half of the steam-to-diesel transition, and compare conditions and events to those of two earlier technological challenges to steam locomotion, the electric locomotive and the gasoline-powered railcar. We conclude that only the diesel-electric locomotive was disruptive as described by Christensen (1997) and discuss the implications of this finding for entrepreneurship research.

WAS IT INDUSTRIAL MYOPIA?

The transition from steam to diesel locomotives in the United States railroad industry has come to be seen, in light of the diesel’s economic superiority and eradication of its rival, as obviously inevitable. Indeed journalists today invoke steam locomotive builders as icons of industrial myopia: “And let’s not forget that utterance from a Baldwin Locomotive Co. executive, just before diesel-electrics swept over the business: ‘They’ll never replace the steam locomotive’ ” (Mack & Summers, 1999: 89). This tale has not been widely challenged by scholars. According to a leading railroad historian, the diesel locomotive as a replacement technology has received limited attention while the managerial decisions of steam and diesel locomotive builders are the focus of only one monograph (Klein, 2000).

Through Cooper and Schendel’s balanced but brief account of steam locomotive manufacturers’ responses to the diesel locomotive, management scholars learned of the chancy judgments and difficult decisions that face such companies—conditions in which failure is not automatic proof of myopia. Cooper and Schendel (1976: 66) treated General Motors’ introduction in 1934 of “the first mainline diesel-electric locomotive” as more noteworthy than the appearance a decade earlier of the first diesel-electric switcher, for it was the capability to haul trains at track speed that attracted the interest of railroad executives. G.M. sold only ten locomotives in 1934, but by 1938 more diesel locomotives than steam were sold in the U.S. (Cooper & Schendel, 1976). World War II interrupted the transition, but in 1947 G.M. would begin a skein of eleven years in which it never held less than 53.8 percent of a domestic market that in each year exceeded 1,000 diesel-electric units (Marx, 1976). By contrast, U.S. railroads ordered only 80 steam locomotives in 1947—and none after 1951.

For entrepreneurship scholars, the early years of the steam-to-diesel transition raise several questions. Locomotive builders circa 1934 “could look back upon two earlier threats which they had survived” in the form of electric locomotives and gasoline-powered railcars (Cooper & Schendel, 1976: 66). Compared to those threats, were there differences in the diesel-electric locomotive that undid the steam locomotive builders? If so, could a different response have saved them? Should their managements have recognized an opportunity in the new technology? In this paper we explore these questions by using the lens of “disruptive technology” to examine locomotive builders’ managerial decisions during the advent of the diesel-electric locomotive.

CONSIDERING A “DISRUPTIVE TECHNOLOGY” EXPLANATION

A particular menace to an industry’s dominant firms is an emerging technology that disrupts rather than enhances the existing order. Clayton Christensen (1997) has described the disruptive technology as “the innovator’s dilemma” on the basis of its destructive impact on “great firms” that invested aggressively in new technologies yet lost their positions of leadership.

What Christensen styled “the failure framework” rests on three findings of his research. First, most new technologies are “sustaining technologies” in that “they improve the performance of existing products, along the dimensions of performance that mainstream customers in major markets have historically valued.” Disruptive technologies, by contrast, “result in worse product performance, at least in the near-term.” They generally “underperform established products in mainstream markets . . . [but] have other features that a few fringe (and generally new) customers value” (Christensen, 1997: xv).

Second, technologies may progress so rapidly that today’s underperforming, disruptive technology “may be fully performance-competitive in that same market tomorrow” (1997: xvi). Third, established companies often decide against aggressive investment in disruptive technologies. Christensen (1997: xvii) identified three bases for this decision: the technologies’ products, being simpler and cheaper, offer lower margins than existing products; the markets in which they first compete typically are “emerging or insignificant”; and established companies’ most profitable customers rarely want or use the products.

Was the Diesel-Electric Locomotive a Disruptive Technology?

Forbes offered the steam-to-diesel transition as an example of opportunity overlooked in a disruptive technology but supported this thought only with a chart showing the proportions of steam and diesel locomotives on the rosters of U.S. railroads from 1940 to 1960 (Mack & Summers, 1999). As expected, steam accounted for nearly all the locomotives in 1940 and virtually none in 1960—but this fact alone does not establish the transition as the result of a disruptive technology.

In this paper we examine the steam-to-diesel transition in light of three descriptive claims (Christensen, 1997) about disruptive technologies and the dynamics they unleash: (1) disruptive technologies change the basis of competition within an industry; (2) their weaknesses (relative to the original basis of competition) are their strengths relative to the new one; and (3) they are typically simpler, cheaper, and more reliable and convenient than the established technologies they displace. Our aim is to determine the suitability of Christensen’s “disruptive technology” and “failure framework” to describe the locomotive industry’s dynamics.

METHODOLOGY

We analyzed primary and secondary accounts of the U.S. locomotive industry, of companies that competed in it, of railroads that were among its major customers, and of various steam and diesel-electric locomotive types. Of special importance was the extensive archival research of Albert J. Churella, presented most notably in his From Steam to Diesel (1998).

Our account is divided chronologically into two periods (1920–1933 and 1934–1940) marked by General Motors’ introduction of the diesel-electric passenger locomotive in 1934. We recount the salient events and each locomotive builder’s responses and interpret them according to the “disruptive technology” explanation. For historical background, we begin by describing the locomotive industry in 1920 and the two challenges it survived prior to the diesel-electric locomotive.

THE U.S. LOCOMOTIVE INDUSTRY IN 1920

The U.S. locomotive industry in 1920 was an oligopoly characterized by incremental technological change, custom and small-batch production, and uncommonly great exposure to the business cycle. Its sole customers were railroads; the great bulk of its orders came from domestic roads.

Three companies accounted for nearly all the commercial production of locomotives in the U.S., with various railroads also building some locomotives for their own use. Baldwin Locomotive Works traced its origins to 1831 and by 1900 “was driving all other builders to the wall” (Withuhn, 2001: 30). Baldwin was an innovative, highly respected company and “more than any other locomotive builder, experimented with nonstandard locomotive products” (Churella, 1998: 64). Its innovation extended to manufacturing methods and plant management.  Baldwin president Samuel Vauclain had risen since 1883 from shop foreman to general superintendent and then vice president; he was elected president in 1919 and turned 64 in the following year. (Vauclain himself in 1891 had developed a major advance in locomotive design.)

Baldwin’s dominance led in 1901 to the birth of the American Locomotive Company (Alco) through the consolidation of eight struggling competitors. By 1920 Alco competed on roughly even terms with Baldwin, each with approximately 40 percent of the market. The remainder of the market went mostly to Lima Locomotive Works, a specialized builder of geared locomotives for use by loggers and quarrying companies.

Technological change long had been incremental. Higher-pressure boilers, larger fireboxes, improved cylinders, refinements of driving wheels, etc., allowed modest but steady increases in steam locomotives’ power while improved brakes and couplers allowed somewhat heavier and longer trains (Bruce, 1952). Even so, the country’s railroads had proved inadequate to the demands of wartime traffic as their slow “drag freights” pulled by aged locomotives over obsolete physical plants (e.g. bridges, track, signals) could not satisfy increased freight volume and urgency. The United States Railroad Administration, taking control of the railroads, created standard specifications for twelve freight and passenger designs, representing eight types (wheel arrangements) of steam locomotives. Baldwin, Alco, and Lima each built these standard designs (Alexander, 1950).

Standardization rapidly gave way following the return of railroads to private control in 1920 (Bruce, 1952). The demands of individual railroads were the principal cause of customization. They differed in operational characteristics such as terrain, importance of passenger traffic, emphasis on fast freight, and quality of their physical plants, and more—each a rationale for differentiation of their locomotive fleets. Builders catered to their best customers’ idiosyncratic demands, working closely with the railroads’ motive power officials to design locomotives tailored to their preferences. The personal and institutional ties developed through this process were enduring; often the locomotives of a particular builder would dominate a railroad’s fleet.

In the face of extensive product customization, steam locomotive builders, “notably Baldwin, succeeded admirably in their efforts to standardize production techniques” and the components used within each locomotive design, but very few components were shared across designs. Components were crafted to precise specifications but typically with “somewhat ‘loose’ ” manufacturing tolerances (Churella, 1998: 11).

Locomotive builders were exceptionally vulnerable to the business cycle. Their executives displayed “patient resignation” during slow years and from long experience “did not equate slumps in demand with the technological obsolescence of their products” (Churella, 1998: 13). They did not necessarily see cause for radical changes in product or management even in years that brought no sales.

The first challenges to steam locomotion on American railroads came during the first two decades of the century. As Cooper and Schendel (1976) observed, the steam locomotive builders survived these challenges; they may also have drawn lessons from their experience

The Challenges of Electric Locomotives . . .

Unlike their smaller sisters, the streetcar and the interurban train, electric locomotives were designed for service on steam railroads. A principal limitation on their employment was the heavy investment in power supply and distribution, including overhead catenary wire or a “third rail” for every track on which they would operate.

Electrification began as a solution to critical problems that rendered steam locomotion unsuitable in specific locations. Whether beneath Manhattan’s urban blocks or through mountains, tunnels threatened the safety of passengers and crews of steam-powered trains. In such sites the electric locomotive was an option, albeit one that required heavy capital investment. Another potential application of electric locomotion was on main lines through mountainous terrain where steam locomotives struggled up steep grades with heavy trains. Electric locomotives provided greater horsepower than steam, especially in starting or at low speeds; the presence of long tunnels on some of these routes only added to electrification’s appeal.

Baldwin played an active role in the electric locomotive market. It built an experimental unit in 1895 and in 1896 built two electric locomotives for use in mining operations. The latter locomotives were built in cooperation with Westinghouse Electric. Subsequent use of Baldwin-Westinghouse electric locomotives in mainline service included the tunnel beneath the St. Clair River linking Canada and the U.S., where in 1908 the locomotives replaced four tank locomotives that in 1891 had been the largest locomotives ever built by Baldwin, and the New Haven Railroad’s line between New York City and New Haven, where by 1912 one hundred units of various types were used in freight, passenger, and switching service (History of the Baldwin Locomotive Works, 1923).

In 1915 the Pennsylvania Railroad electrified its mainline suburban service west of Philadelphia to Paoli, Penna. The trains consisted of multiple passenger cars some of which included electric traction motors. In 1928 the railroad announced its plan to electrify its main lines from New York to Washington and westward from Philadelphia (eventually reaching Harrisburg). To handle mainline freight and through passenger trains, separate electric locomotives would be used. Working with Westinghouse and other electrical engineers, Baldwin built most of the initial fleet as well as the famous GG-1 that followed in 1934. Continuing a joint-venturing history that began in 1903 to supply electric locomotives to the New York Central (Churella, 2001), Alco and General Electric combined to build the other units to the Pennsylvania’s standard plans.

 . . . and of Railcars . . .

Railcars were self-propelled passenger cars that marked the first use of internal-combustion engines in railroading. The Union Pacific Railroad and General Electric separately pioneered this technology. The Union Pacific’s motive was to reduce operating expenses by shifting low-density passenger service from steam locomotives to railcars. William R. McKeen Jr., the company official charged in 1905 with the project, later founded his own railcar company. Between the two companies McKeen produced more than 200 railcars but “enjoyed scant success . . . because their engines were too heavy and unreliable for railroad service” and, “more important,” their mechanical transmissions proved “difficult to control” and experienced “frequent catastrophic breakdowns” (Churella, 1998: 24).

General Electric had built its first electric locomotive in 1895 and produced equipment for streetcars and electric interurbans. When demand for electric locomotives failed to develop as expected, GE sought a related market in which to use its capabilities. GE established a gasoline-engine research department in 1904 and sold its first railcar in 1906. (Alco supplied the car body.) GE eventually sold 88 gasoline-electric railcars but abandoned the business in 1919 with a cumulative loss of $1.5 million. Churella (2000) attributed the abandonment to the railcars’ technical problems and the heavy demands of World War I on GE’s capacity.

Westinghouse entered the business in 1925 during a period of renewed interest in railcars. Its aims were essentially the same as General Electric’s twenty years earlier: to add a market for its traction motors and other railway electrical equipment. Westinghouse sold only 15 railcars then quit the business following the market’s collapse in the Great Depression (Churella, 1998).

Of greater ultimate import was the emergence of the Electro-Motive Company. Its founder, Harold L. Hamilton, had managerial experience in both railroading and the automobile industry. Resigning in 1922 as western wholesale manager of White Automobile Company, he began EMC and soon recruited four men who had worked in General Electric’s railcar program.

From its inception EMC pursued a different strategy than did its peers. It was essentially a design and marketing company. Unlike its competitors in the railcar business, EMC, “strictly speaking, did not manufacture anything” (Churella, 1998: 32) but instead subcontracted the production of railcars its employees designed. General Electric supplied most of EMC’s electrical equipment and remaining veterans of its railcar venture provided extraordinary assistance to the newcomer. EMC subcontracted the manufacture of car bodies to established builders of railroad passenger cars and acquired all its engines from the Winton Engine Company. A subsequent program to develop a suitable diesel engine, begun in 1928, failed on account of Winton’s weak technical knowledge and inability to manufacture to close tolerances (Churella, 1998). (EMC would continue to use Winton’s gasoline engines until 1934.)

By 1930 EMC had sold approximately 500 railcars and accounted for 84 percent of the market since the company’s founding. Churella (1998) credits EMC’s control of the railcar design process with much of its success. More than its rivals (or the steam locomotive manufacturers, for that matter), EMC standardized its railcar designs and reaped the benefits of lower costs by setting lower prices and gaining volume. Patterning itself after the automobile industry’s customer service procedures, EMC provided training, product warranties, replacement parts usually within 24 hours (even using air delivery), and other elements of service that were unprecedented in its industry. Moreover, Hamilton and his sales team dealt directly with railroads’ top executives, who typically had not risen from the ranks of motive power officials but were more likely to have financial backgrounds. The top executives were less steeped in the culture of steam locomotion and more attuned to the savings promised by railcars—and they wielded ultimate authority.

 . . . Survived

At the peak of their market in the late 1920s, railcars had grown powerful enough that railroads used them to haul short trains of passenger cars or a few freight cars, in spite of manufacturers’ instructions not to do so. This was the limit of their substitution for steam locomotives. Their limited power and speed prevented their use in long-distance passenger service, and the lightly patronized runs on secondary lines fell victim to improved highways. Railroads increasingly discontinued uneconomical passenger service rather than convert it to railcars.

Like the gasoline-powered railcar, the electric locomotive gained its toehold in uses that were unsuited to steam locomotives. Unlike the railcar, however, electric locomotives came to outperform steam locomotives on dimensions that were at the heart of mainline railroading: power and speed. Their principal buyers were not marginal competitors or newcomers, as is expected of disruptive technologies (Christensen, 1997), but were such major customers of the steam locomotive builders as the Pennsylvania (known as “The Standard Railway of the World”), New York Central, Great Northern, and Milwaukee roads. Even in its earliest applications as a tunnel motor or a solution to smoke abatement ordinances, the electric locomotive’s buyers were major railroads and its applications were in mainline settings.

The Pennsylvania’s Depression-era project was the last large-scale electrification in American railroading. For years the allure of electrification prompted feasibility studies and speculative writing, but no railroad was sure enough of the heavy traffic necessary to justify electrification’s huge infrastructure costs. Well into the diesel era, some mountainous and tunnel lines were de-electrified as dwindling traffic and more powerful diesel-electric locomotives combined to discourage the replacement of worn-out electrical locomotives and aged catenary power lines. The New York-Washington-Harrisburg main lines continue in 2002 as the electrified exception.

1920–1933: THE DIESEL-ELECTRIC LOCOMOTIVE FINDS A NICHE

The 1920s brought the steam locomotive builders a dramatic upturn in volume and profits. From a low of 214 locomotives in 1919 their orders jumped to 1,998 in 1920 and reached 2,600 in 1922. Their orders drifted downward through 1926 then fell sharply in 1927 and a bit further in 1928, to 603 locomotives. Demonstrating its longstanding volatility, the market rebounded to 1,212 locomotives in 1929 (Churella, 1998).

Lima dramatically changed its business strategy in 1922 by beginning the development of its “Superpower” steam locomotive, a “radically improved” design offering much improved thermal efficiency and lower fuel consumption. The design debuted in 1925 and in short order Baldwin and Alco adopted many of its dozens of improvements (Churella, 1998: 118). Whereas Lima’s gambit advanced the industry’s technology and product performance, it did not overturn the competitive structure: from 1920 through 1928, Alco held 47 percent of the market, followed by Baldwin (39 percent) and Lima (14 percent) (Churella, 1998: 10–11).

Tonnage per train and locomotives’ average tractive force continued to increase in the 1920s. Locomotive performance improved only slowly after the mid-1920s; the builders did not recognize “that by the late 1920s steam locomotives had reached a technological dead end” (Churella, 1998: 13). Builders satisfied railroads’ demands for more powerful locomotives primarily by building larger, heavier models. Locomotives’ size and weight reached the clearance limits of railroads’ tunnels and structures and the load-bearing capacity of their bridges.

The steam locomotive builders pursued financial policies that apparently focused on maintaining investors’ confidence and their share prices. Baldwin paid more than 60 percent of its net profits in dividends during the 1920s, maintaining its high dividend even after steam locomotive orders began their mid-decade decline. Alco went even further, paying $8.00 dividends annually from 1926 through 1929 even though earnings per share averaged less than $5.00 (Churella, 1998).

In spite of its high dividend payout, Baldwin also invested substantially in locomotive building. In 1928 the company completed the transfer of all operations and administration to its new, 590-acre site alongside the Pennsylvania Railroad’s main line at Eddystone, Penna. There $40 million in structures and equipment gave Baldwin the industry’s premier plant (Churella, 2000; Westing, 1966) but also a heavy bonded debt.

The First Diesel-Electric Locomotives

Churella (1998: 23) has described the railcar business as the “springboard to participation in the diesel locomotive industry” because three railcar manufacturers—General Electric, Westinghouse, and EMC—sought to design and build diesel-electric locomotives. GE had begun diesel-engine research in 1911 and in 1917–18 produced four locomotives that failed to perform acceptably. The company ceased the manufacture of gasoline and diesel engines for railroad use when it dropped its railcar business in 1919.

Urban switchyards and industrial sidings, often combining intensive, start-and-stop operation with proximity to business districts or residential neighborhoods, became the target of smoke abatement ordinances. Steam locomotives were restricted or even banned New York, Baltimore, Chicago, and other cities. The New York Central asked Ingersoll-Rand, a builder of diesel engines, to build a prototype diesel switcher; GE joined Ingersoll-Rand in the venture. After the two companies built a successful prototype, Alco joined their consortium in 1924 as the builder of locomotive bodies; GE and I-R supplied all other components, which GE assembled and I-R marketed.

Churella (1998: 26–27) has explained how the performance criteria for this locomotive departed from those of most steam locomotives. In descending rank they were “reliability, high potential speed, low maintenance costs, minimal noise and smoke, good fuel economy, and ‘reasonable first cost’[;] . . . horsepower was not even mentioned.” Alco and Baldwin, he added, knew that railroads accepted these performance characteristics when forced by law to do so but “were not likely to buy diesels . . . where cost and power were the sole considerations.” Diesel engines were overweight and underpowered for most railroad uses.

The consortium produced 33 diesel-electric locomotives from 1925 to 1931, adding to its 300-h.p. switcher a 600-h.p. passenger and a 750-h.p. freight locomotive—all “for specialized niche markets where steam locomotives, the preferred form of motive power,” were uneconomical or unsafe (Churella, 1998: 27; italics added). In 1928 GE decided to produce the locomotive bodies; Alco left the consortium and acquired McIntosh & Seymour Co., a maker of marine diesel engines, in order to control more of the critical components of diesel locomotives. In Churella’s (1998) judgment, Alco’s involvement with GE and I-R was insufficient to support its later claims to a pioneering role in diesel locomotives.

Baldwin also responded to the smoke abatement legislation. It built two 1,000-h.p. diesel-electric locomotives. The first, in 1925, cost more than $250,000 with all castings and engine components obtained from Baldwin subsidiaries and the electrical items from Westinghouse. The second locomotive, built in 1929, included a Krupp engine that cost $100,000. Neither locomotive performed acceptably, nor did industry observers think that Baldwin marketed them with any enthusiasm. Samuel Vauclain personally declined the Pennsylvania Railroad’s repeated requests to test the second locomotive on its lines, notwithstanding the two companies’ close ties. He also failed to reply when the Cummins Engine Company offered to supply a better and less expensive diesel engine (Churella, 1998).

As it had in railcars, Westinghouse followed General Electric into the diesel-electric locomotive industry. In 1928 it built two 660-h.p. switchers using Baldwin bodies. The two companies entered an agreement in 1929 to produce 400-h.p. and 800-h.p. switchers, but Baldwin was merely the independent supplier of bodies, underframes, and running gear; Westinghouse designed and marketed the locomotives (Churella, 1998: 29–30). Compared to its competitors’ diesel-electrics, Westinghouse’s locomotives were “exceptionally light” owing to “extensive use of aluminum alloys” and superior in body design. Westinghouse sold 13 diesel-electric locomotives by 1931, then another 13 during the Depression before quitting the business (as well as the sale of diesel engines) in 1936.

The Electro-Motive Company neither built nor sold diesel-electric railcars through 1933. Its frustration in relying on the Winton Engine Company to develop a diesel engine, recounted above, indicates that EMC’s management sought the new technology to no avail. In 1930 EMC “faced [railcar] market saturation and financial catastrophe” (Churella, 1998: 37). Then, more by chance than design, EMC fell into the General Motors network when G.M. acquired Winton to gain the latter’s research capabilities. Improvements in Winton’s diesels are attributed variously to the two-cycle design devised by G.M. (Sloan, 1963) or to G.M.’s superior skills in close-tolerance manufacturing and engine technology as well as its metallurgical and fuel research (Churella, 1998).

Believing that EMC could not survive as an independent company, Harold Hamilton sought the company’s purchase by GM. He estimated a cost of $10 million to develop a diesel-electric switcher and prepare for volume production, and EMC had neither the money nor the technological expertise for the project. According to Churella (1998), GM made the acquisition without thoughts of a program to develop diesel locomotives. Instead GM acted because EMC was Winton’s principal customer and on account of personal friendships and working relationships, not least between Kettering and Hamilton but also among the three companies’ engineers and technicians.

Space does not permit an account of EMC’s coup in securing Ralph Budd’s order for a diesel-electric locomotive to power the lightweight, streamlined passenger train being built for his Chicago, Burlington, and Quincy Railroad. Kettering and Hamilton overcame Sloan’s initial skepticism regarding the idea and his company’s ability to develop a diesel engine that would stand up under the stress of railroad operation. By the end of 1933, EMC was in the diesel-electric locomotive business.

1934–1940: DIESEL-ELECTRICS ON THE RISE

The debut of the Burlington’s Zephyr was a national event in 1934. Running non-stop from Denver to Chicago, the streamliner averaged 78 m.p.h. and reached a top speed of 112. Placed in regular service in November between Kansas City and Lincoln, Neb., the train ran faster, at lower cost per mile, and attracted more passengers than the steam-powered train it replaced (Reutter, 2000). The Burlington ordered four more diesel-powered streamliners from EMC in 1936. Meanwhile the Union Pacific followed its own 1935 inauguration of a gasoline-powered streamliner with the addition of a fleet of similar trains.

EMC’s demonstration diesel locomotive set, boxy and utilitarian in design, toured Eastern railroads in 1935–36 and impressed railroad executives by its speed, power, smooth starting and stopping, economical operation, and operating range without stops for fuel. Not satisfied with selling this model, EMC revamped the locomotive’s shell and inner workings; the “E” locomotive, introduced in 1937, “changed the face of American railroading” (Reutter, 2000: 50). In 1938 the improved 567 diesel engine enabled EMC to outfit The Orange Blossom Special with a 6,000 h.p. multiple-unit locomotive—ten times the horsepower of the Zephyr just four years earlier. In November 1939 EMC introduced the prototype of the FT, a four-unit, 5,400-h.p. freight locomotive that produced more tractive effort than any existing steam locomotive; the Santa Fe ordered 80 FTs and by March 1941 three other roads had followed suit (Churella, 1998).

An unprecedented speed-up of American passenger trains accompanied the spread of diesel locomotives and streamlined trains. Significant technological advances in steam locomotives during the 1930s contributed to the improvement: roller bearings and the use of lighter, alloyed steels in side rods and other moving parts allowed safer, smoother operation at high speeds. Even so, at high speeds steam locomotives were harder on track than diesel locomotives. Diesel locomotives also proved to operate with undiminished efficiency in cold climates, unlike steam locomotives (Reutter, 2000).

Once G.M. tasted success, it moved toward comprehensive involvement. Churella (1998) estimated that G.M. spent as much as $22 million during the 1930s to enter the locomotive industry. The parent company’s commitment to EMC included a new, $6 million plant outside Chicago and heavy investment in marketing and service, such as training programs for railroad operating and maintenance employees and attractive financing of locomotive purchases through General Motors Acceptance Corporation. The streamlined passenger locomotives garnered the glory, but most orders were for switchers. Railroads that used their new diesel switchers 16 hours per day, a common utilization in large freight yards, saved more per locomotive per month than the monthly payment. EMC also delivered switchers from stock, usually within a few weeks of an order, an availability that steam locomotive builders could not approach. In 1938 EMC sold 10 diesel switchers for every one of its competitors’ sales, and outsold steam switchers by 100-to-one (Churella, 1998). The same year appears to have been the first in which diesel locomotives outsold steam (Cooper & Schendel, 1976).

Events hit the steam locomotive builders hard in the 1930s. Baldwin’s sales of locomotive products fell from $31 million in 1930 to a little over $1 million in 1933. Five consecutive years of losses culminated in bankruptcy in 1935 when Baldwin defaulted on its bonds. In spite of diesel locomotives’ dominance in the switcher market and spectacular emergence in passenger service, Baldwin’s 1937 study of motive power concluded that “modern steam power must continue to be the mainstay of railroad operations for the indefinite future” (quoted in Marx, 1976: 16).  In 1938 top management and the board of directors were ousted save Samuel Vauclain, whose retention as director was a token appointment. Efforts to sell diesel locomotives had increased following bankruptcy and now intensified under new management.

Baldwin had purchased a diesel engine manufacturer in 1931 but had not integrated its operation into the locomotive division. In 1935 Baldwin began to build a prototype locomotive using a diesel engine that had been designed for stationary service. A 660-h.p. switcher introduced in 1936 was prone to serious mechanical problems and to cracked frames; new 660- and 900-h.p. models introduced in 1939 used the same engines and experienced the same problems. Baldwin sold only 23 diesel locomotives between 1936 and the end of 1940 (Churella, 1998). The company’s product was inferior to EMC’s, more expensive to produce, and unenthusiastically supported by many Baldwin executives and salesmen. At the same time Baldwin was coming out of bankruptcy in 1938 and returning to profitability in 1939, trying to break into the diesel locomotive market, and fighting to retain its steam locomotive business.

Alco fared better than Baldwin in the 1930s but fell far short of EMC’s progress and G.M.’s investment. Alco captured more than 80 percent of the U.S. diesel locomotive market in 1935 with switchers that introduced many design features that would become industry practice. That market share was short-lived, but Alco’s new diesel engine and two switcher models introduced in 1940 became staples of its product line. In spite of these efforts, Churella (2001) concludes that Alco’s top management remained committed to steam locomotives, citing their statements in Railway Age. The company thought of diesel locomotives as supplementary to steam power for mainline operation; their main purpose was in specialized applications where steam was not viable. Alco’s executives, employees, plant, and organizational routines all were oriented to steam, and “they believed they were being innovative with steam [even though] changes were often restricted to marginal, incremental improvements, not radically different technologies” (Churella, 2001: 37).

Marx (1976) portrays an Alco organization that was fully aware of the diesel’s ascendancy even as it strove to preserve its steam locomotive business. He argues that Alco management stayed in touch with technical developments and, when the FT’s road tests in 1939 demonstrated its effectiveness, prepared to enter the freight market. Indeed, by the end of 1940 “Alco had completed the design and engineering of a complete line of new diesel passenger and freight locomotives” (Marx, 1976: 14). To Marx, Alco demonstrated a shrewd defensive strategy of being prepared to enter new markets for diesel locomotives as soon as their attractiveness was demonstrated, and well ahead of demand, while maintaining its strong position in steam locomotives

Lima chose not to enter the diesel locomotive business in the 1930s, focusing instead on continued improvement of its steam locomotives. To Churella (1998), that decision was wise in view of Lima’s lack of capital and access to diesel technology.

A Reprieve, Then the Demise of Steam Locomotives (and Their Builders)

World War II brought increased, urgent demand for locomotives; critical military needs for alloys, technical skills, and manufacturing capacity; and federal allocation of scarce materials and assignment of particular roles to specific locomotive builders. The war meant a reinvigorated market for steam locomotives as the only option for railroads that could not obtain diesels. It also ensured buyers for any available diesels. Electro-Motive was assigned exclusive production of diesel freight locomotives on account of its FT. Alco and Baldwin were allowed to build switchers and 2,000-h.p. freight locomotives; having no diesel locomotives in its line, Lima was restricted to steam locomotives. Shortly after the war, orders for steam locomotives collapsed; their builders variously sought to compete in the diesel locomotive business and to survive as corporations, but all failed in both endeavors.

CONCLUSIONS

Our conclusions, based on ours and others’ interpretation of this industry’s history, must be brief: In the diesel-electric locomotive, steam locomotive builders faced a disruptive technology. Its early manifestation was as an inferior replacement where legal or safety requirements ruled out steam locomotives. Its rapid penetration and dominance of the switch-locomotive market sprang from EMC’s revolution of the broader technology (Christiansen, 1997): the entire system of production, marketing, and service associated with the locomotive. From that base, sustaining improvement of its design and materials pushed the diesel locomotive into the passenger locomotive segment and proved its road-service superiority to steam. By 1940 the diesel locomotive’s superiority was beginning to become apparent in freight service too.

Steam locomotive builders participated slowly and ineffectively in diesel technology. They lagged in metallurgy, production technique, marketing, and more—in virtually all aspects of the new technology’s implementation. As Churella (1998) concluded, the diesel locomotive ran counter to steam locomotive builders’ entire experience and their organizations’ competences. Its early inroads in uninteresting or inaccessible market segments solidified steam locomotive builders’ impression of its marginal or supplemental nature, and of their own technology’s impregnability. When builders, notably Alco, ventured seriously into the new technology, they stumbled over their unsuited routines and competences. Their failure to respond effectively is unsurprising, given that steam locomotive builders had never faced a disruptive technology.

Earlier technological challenges were not disruptive. The electric locomotive met or surpassed steam on key performance measures but its infrastructure costs were justified only in the heaviest mainline corridors; in a sense, it was a superior technology that did well in its niche but could not downgrade into the performance and cost range demanded by the market. Steam locomotive builders’ cautious joint venturing was effective in satisfying prime customers’ special needs and foreclosing the segment to unfriendly entrants.

The gasoline-powered railcar’s beginnings were indistinguishable from those of a disruptive technology, but the failure to upgrade its capabilities kept internal combustion out of either passenger or freight locomotion. Moreover, the railcar’s fortunes rested on the health of the segment in which its inferiority was tolerated—low-volume passenger service, especially on branch lines— and faded with the withering of that marginal business.

We also conclude that distinction between disruptive and unthreatening technologies is made more easily after the fact. Opportunity recognition remains messy and fog-shrouded (Hench & Sandberg, 2000) in spite of lucid post hoc explanation. We are continuing our research on the later phases of the steam-to-diesel transition and on the actions of both incumbents and newcomers.

CONTACT: William R. Sandberg, Moore School of Business, University of South Carolina, Columbia, SC 29208; (T) 803-777-5980; (F) 803-777-6876; sandberg@moore.sc.edu

REFERENCES

Alexander, Edwin P. (1950) American Locomotives: A Pictorial Record of Steam Power, 1900–1950. New York: Bonanza Books.

Bruce, Alfred W. (1952) The Steam Locomotive in America: Its Development in the Twentieth Century. New York: W.W. Norton & Company.

Christensen, Clayton M. (1997) The Innovator’s Dilemma: When New Technologies Cause Great Firms to Fail. Boston: Harvard Business School Press.

Churella, Albert. (2001) “Death by Diesel.” Trains, 61(9): 34–45.

Churella, Albert J. (2000) “Business Strategies and Diesel Development.” Railroad History (Millennium Special: The Diesel Revolution): 22–37.

Churella, Albert J. (1998) From Steam to Diesel: Managerial Customs and Organizational Capabilities in the Twentieth-Century American Locomotive Industry. Princeton, NJ: Princeton University Press.

Cooper, Arnold C., & Dan Schendel. (1976) “Strategic Responses to Technological Threats.” Business Horizons 19 (1) (February): 61–69.

Hench, Thomas J., & William R. Sandberg. (2000) “‘As the Fog Cleared, Something Changed’: Opportunity Recognition as a Dynamic, Self-Organizing Process.” Paper presented at the Babson Kauffman Entrepreneurship Research Conference, Wellesley, MA, June 8–10.

History of the Baldwin Locomotive Works. (1923) Philadelphia: Baldwin Locomotive Works. Facsimile reprinted in Westing (1966).

Klein, Maury. (2000) “The Continued Neglect of the Diesel Locomotive.” Railroad History (Millennium Special: The Diesel Revolution): 6–15.

Mack, Toni, & Mary Summers. (1999) “Danger: Stealth Attack.” Forbes, January 25, pp. 88–93.

Marx, Thomas G. (1976) “Technological Change and the Theory of the Firm: The American Locomotive Industry, 1920–1955.” Business History Review 50(1): 1–24.

Reutter, Mark. (2000) “Building a Better Iron Horse.” Railroad History (Millennium Special: The Diesel Revolution): 38–61.

Sloan, Alfred P. Jr. (1963) My Years with General Motors. New York: Doubleday & Company.

Westing, Fred. (1966) The Locomotives That Baldwin Built.  New York: Bonanza Books.

Withuhn, Bill. (2001) “American Experience.” Trains 61(9): 30–33.

2002 Babson College. All Rights Reserved. Last Updated March 2003.