A mighty rough road: The deterioration of work safety on American railroads, 1955–75

Abstract

While nineteenth-century American railroads were fabled for their dangerous working conditions, after 1900 a new safety regime grew up. Threatened by union and public pressures that resulted in numerous safety regulations and rising accident costs, the carriers innovated safety programs intended to stave off further regulation. The carriers’ formula for successful safety work involved worker participation, management commitment, and technological modifications. Finally, the new regime included expert oversight from both government and the trade press but comparatively little direct regulation. While these resulted in dramatic improvements in work safety down through World War II, there were weaknesses in the system and after 1955 work safety worsened for two decades. While this resulted in part from the continued shift from passenger to more dangerous freight work, some new technologies also worsened risks. In addition worker involvement had eroded, many safety departments had become moribund, and declining profits undermined safety. By the 1960s, as derailments skyrocketed, railroad safety again became a public issue and a new regime resulted. Congress transferred responsibility from the Interstate Commerce Commission to the Federal Railroad Administration and in 1970 extended federal control over nearly all aspects of rail safety.

Notes

All data on railroad fatalities are from the Interstate Commerce Commission's Accident Bulletin in various years, although the data on man-hours prior to 1922 are from Aldrich, Safety First, appendix 1. In the following I rely exclusively on fatality rates, for they are less likely to be under-reported than are injuries. However, because of differences in data-gathering techniques neither railroad injury nor fatality rates can be compared with those of other industries. On this see United States Railroad Retirement Board, Reporting of Accidents & Casualties.

For an assessment that stresses declining profitability see Savage, The Economics of Railroad Safety, chapter 2.

On the coming of safety appliances see Clark, ‘The Development of the Semi-Automatic Freight Car Coupler’ and ‘The Railroad Safety Movement;’ Usselman, ‘Airbrakes for Freight Trains.’

For early employers’ liability law see Witt, ‘The Transformation of Work.’

On the early railroad safety movement see Aldrich, ‘Safety First Comes to the Railroads.’ Rice is quoted on p. 12.

Richards, ‘The Safety First Movement;’ Railway Age, ‘The Campaign against Accidents;’ United States Railroad Administration, ‘Suggestions and Recommendations.’

‘Simply another method’ is from Railroad Trainman, ‘An Object Lesson in Safety First,’ 985. For the meeting between Hale and the union leaders see Railroad Trainman, ‘Safety First.’ See also Lee (President, Brotherhood of Railroad Trainmen), ‘Benefits Accruing to the Railroad Trainmen,’ National Safety Council (NSC) Railroad Section, Proceedings 14 (1925): 1017–21; Hale, ‘Discussion,’ 1022–23.

A good introduction to the role of institutional and organizational innovation is North, Institutions, Institutional Change and Economic Performance.

H. R. Lake, ‘Some Pro's and Con's on Railroad Safety Methods,’ ARA Safety Section Proceedings 10 (1930): 101–16. For the St Louis & San Francisco see ‘Frisco's Safety Program Gets Results,’ MR 6 (September 1950): 33, 35–36, quotation on p. 35; John Banks, ‘Remarks,’ AAR Safety Section Proceedings 28 (1948): 76; Howard E. Banta, Chair Local 1,000 Brotherhood of Railroad Trainmen, to R. C. Diamond, Superintendent B&O, 12 February 1955; Diamond to Banta, 11 March 1955 and Banta to Diamond, 29 June 1959. All in box 16, Brotherhood of Railroad Trainmen Collection, 5149, Keehl Center, Cornell University (KCCU).

For the New Haven see ‘Minutes of Safety and Efficiency Committees, 1923–1930,’ folder 3677, box 25, New Haven Railroad Historical and Technical Association Records, RG-8, Dodd Library, University of Connecticut; North, Institutions, Institutional Change and Economic Performance, chap. 8; George Bradshaw, ‘The Safety First Movement,’ Western Railway Club Proceedings, 18 February 1913, 240–89, quotations on p. 242; ‘Meeting of the Regional Safety Committee,’ 2 September 1927, box 635, Pennsylvania Railroad Collection, Hagley Museum and Library (hereafter PRRC, HML).

The quotation is from an untitled editorial, Railroad Gazette, 10 September 1901, 542. Where two rails cross (a ‘frog’ in railroad parlance), the resulting ‘V’ works like a trap to catch a foot. Some states required carriers to block them, but Vermont did not. ‘In re Fatal Accident to Maynard Ryan which Occurred on the Vermont Railway at St Albans, 12 July 1903,’ box 00599445, Vermont Board of Railroad Commissioners, Docket, Complaints and Petitions, 19 December 1902 to 24 March 1904, Vermont State Archives.

On the importance of outside criticism see the discussion of Charles Francis Adams in McCraw, Prophets of Regulation, chap. 1.

On safety work in manufacturing see Aldrich, Safety First, chaps 3 and 4.

For a critique of the rail unions’ approach to safety see US Railroad Retirement Board, Safety in the Railroad Industry, chapter 3. For the experience with running boards see Railroad Trainman, ‘Metal Running Boards’ and ‘Economy and Safety First.’

For examples of shop improvements see Railway Locomotives and Cars, ‘Baltimore & Ohio Streamlines its Shops’ and ‘What Is Light Worth?’ Improvements in equipment and testing methods are described in numerous articles in Railway Age and Railway Locomotives and Cars

For the safety engineer see ‘Report of the Committee on Direction,’ AAR Safety Section Proceedings 28 (1948): 150–51; 29 (1949): 26.

The power brake rules could only be changed for the purpose of improving safety. See US Railroad Retirement Board, Safety in the Railroad Industry, chap. 3. See also US Congress, Office of Technology Assessment, An Evaluation of Railroad Safety, 102.

For improvements in locomotive design see Railway Age, ‘A Crown Sheet Failure;’ Railway Review, ‘Recent Types of Freight and Passenger Locomotives;’ Railway Age, ‘Can Boiler Explosions Due to Low Water be Prevented?’ For an example of blaming the engineer see Hall (Bureau of Locomotive Inspection), ‘Beware of the Empty Water Glass.’ For similar critiques of the ICC's safety work see US Railroad Retirement Board, Safety in the Railroad Industry, chap. 4; US 89th Congress, 2nd session, House Committee on Government Operations, Interstate Commerce Commission Operations (Railroad Safety), H.R. 1452 (Washington, 1966); Bernstein, Regulating Business by Independent Commission, 88. That the quotation on p. 88 applies to the ICC is my judgment.

I estimated the following equation relating (the natural log of) worker on-duty fatality rates, 1922 to 1965, to passenger and freight train miles in millions (Ptm and Ftm), a war dummy (for 1940–45), a postwar dummy (1957 on), and an interaction between the trend and the postwar dummy. Figures in parentheses are t-ratios. The results indicate that fatality rates declined at 4.9 percent a year before the 1950s but rose after 1957.

Had total train miles in 1965 been unchanged but had they been 62 percent freight as they were in 1957 instead of 71 percent freight, there would have been 56 (million) more passenger and fewer freight train miles. Using the coefficients for freight and passenger train miles in the equation above yields exp(− 0.0023 × 56) − 1 = − 0.12 and exp(0.0023 × − 56) = − 0.12 and so the shift would have reduced worker fatality rates by 24 percent.

‘Santa Fe Accident Prevention Plan,’ NSC Railroad Section Proceedings 47 (1959): 9–12. For the ACL see Railway Age, ‘Safety Men Aim for Perfection’ and ‘On the Frisco Safety Is a Family Affair.’ The Milwaukee is from L. J. Benson, ‘Remarks,’ AAR Safety Section Proceedings 28 (1948): 54–56.

For a survey and critique of railroad safety programs see US Railroad Retirement Board, Safety in the Railroad Industry, chap. 2; ‘Report of the Safety Activities on the B&O and the NYC Compared with the Pennsylvania Railroad,’ n.d. [c. 1947], box 342, PRRC, HML; ‘Report of Interdepartmental Committee on Personal Injuries,’ 15 June 1951, box 342, PRRC, HML; James M. Symes to Edward W. Smith, 5 May 1949, box 342, PRRC, HML. Manufacturing corporations were also reducing worker involvement in safety during these years with similar results. See Farris, ‘Institutional Change in Shopfloor Governance.’

For the Norfolk & Western see Modern Railroads, ‘Accidents under Control;’ Railway Age, ‘New SP Cabooses Stress Safety;’ L. A. Villella, ‘Track Maintenance,’ NSC Railroad Section Proceedings 45 (1957): 11–13. For the Lackawanna see J. P. Hiltz, ‘Remarks,’ AAR Safety Section, Proceedings 32 (1952): 169–72.

‘The Standardization of Hand Brakes,’ AAR Safety Section Proceedings 30 (1950): 101–7. The dangers of old-style brakes are from L. E. Hoffman, ‘Remarks, AAR Safety Section Proceedings 29 (1949): 88–91.

For a broader discussion of the effects of mechanization on employment and work see Haber, Maintenance of Way Employment on U.S. Railroads; C. E. Hightower, ‘Remarks,’ AAR Safety Section Proceedings 30 (1950): 107–8.

Walter Kintz, ‘Statement,’ 15 August 1952, Misc. Accident Correspondence, box 17, MG 199, Lehigh & Hudson River Railroad Collection, Railroad Museum of Pennsylvania; J. P. Hiltz, ‘Remarks,’ AAR Safety Section Proceedings 32 (1952): 169–72.

Hill's quote is from ‘Non-Train Accidents,’ NSC Railroad Section Proceedings 31 (1943): 620. See also Railway Track and Structures, ‘What Radio on M/W Equipment Is Doing.’

On the lack of set-offs see W. E. Rader to R. H. Burkett, 11 October 11, 1939, file 3, box 2, William Doble Papers, Collection 5182, KCCU. An accident that resulted because the track car would not trip a crossing signal is reported in L. R. Bloss to F. M. Brown, 23 May 1950, file 13, box 2.

Railway Signaling, ‘New Automatic Block on the Katy;’ Railway Age, ‘AAR Studying Track Car Safety.’ Refusal to provide a line-up is from R. M. Smith to W. E. Green, 17 January 1945, file 7, box 2, Doble Papers, Collection 5182, KCCU. US 85th Congress, 2nd session, Senate Committee on Interstate and Foreign Commerce, Safety Regulation of Railroad Track Motorcars, Hearings on S. 1729, 26 May 1958 (Washington, 1958).

The equation for 1964 is

where C is the casualty rate, MH is man-hours, a measure of company size, and ROA is the return on assets. Figures in parentheses are heteroskedastic-consistent t-ratios. The standard deviation of ROA is 0.0257; a two-standard-deviation increase implies a decrease in casualties of exp(− 6.57 × 0.0514) − 1 or − 29 percent. Other equations are available from the author upon request.

ICC Bureau of Valuation, Railroad Maintenance Study, Railway Track and Structures, ‘Where Do We Stand on Deferred Maintenance?’

The survey is from Modern Railroads, ‘Where Do We Stand on M of W?’ Complaints about maintenance are voiced in United States 91st Congress, 1st session, Senate Commerce Committee, Hearings on Federal Railroad Safety Act of 1969 (Washington, 1969).

The ICC sharply increased the cost threshold for reporting train accidents several times between 1948 and 1957 and then held it constant until 1975. These changes first overcompensated and then undercompensated for inflation, leading to spurious changes in reported accidents. Note, however, that such difficulties do not affect the divergence in the performance of strong and weak carriers depicted in Figure 4. In a statistical analysis (not shown) I have attempted to correct for these difficulties as well as changes in car miles. After such corrections derailments rise after 1957. For similar conclusions see United States Congress Office of Technology Assessment, An Evaluation of Railroad Safety.

For a statistical study that finds train accidents to be inversely related to company profitability see Golbe, ‘Product Safety in a Regulated Industry.’

National Transportation Safety Board, Broken Rails. Many technical articles for this period refer to the problems created by the new large cars. See for example Railway Locomotives and Cars, ‘Wheel, Rail Design Involves Dynamics’ and ‘Larger, Heavier Equipment Poses Problems.’ The figures for the Penn Central are for that company and for prior years they are the sum of its predecessor companies. See US 92nd Congress, 2nd session, Senate Commerce Committee, The Penn Central and Other Railroads (Washington, 1973), Table C-10.

The quotation is from C. D. Young to Martin W. Clement, 21 April 1947, box 342, PRRC, HML; Mansfield, ‘Innovation and Technical Change.’ For the importance of unit trains see Savage, Economics of Railroad Safety, chap. 2.

Carrier returns on investment are from Railway Age, ‘Review of Railway Operations’ and Railway Track and Structures, ‘Diesels and Rail Lubricators.’ Equipment installations are from Railway Signaling ‘Signaling Construction’ and ‘Signaling and Communications.’

New York Times, ‘Rail Tank Car Explodes,’ ‘Derailed Train Afire,’ and ‘Railroad Toll Rises to Five;’ Turner, The Train Wreck Disaster.

US 88th Congress, 2nd session, House Committee on Government Opions, Hearings on Interstate Commerce Commission Operations (Railroad Safety) (Washington, 1965). For a review of the legislation and an argument that economic deregulation improved railroad safety see Savage, ‘Trends in Transportation Employee Injuries.’

For the automobile companies’ approach to safety see Gusfield, The Culture of Public Problems, chap 2. On the role of electrification in manufacturing and mining see Aldrich, Safety First, chaps 4 and 6.