Response to Mark

Mark’s criticisms are either irrelevant or incorrect. The main thrust of this critique seems to be that I had the temerity to suggest that All Good Things did not flow entirely from the Coal Mine Safety Act (CMSA) of 1969. The act probably did improve safety—although perhaps not in the way Mark believes, for one article I cite suggests its main effect may have been to drive out small mines rather than to enhance safety in those that remained. Mark claims that I am “reluctant to credit” the CMSA, but whatever that act may have done is irrelevant to my article that deals with mine safety before 1970. I am always reluctant to attribute prior events to later causes.

Mark dislikes my statement that after 1965 fatality rates began a “long decline,” apparently because the slide ended about 1980 and he prefers the term “precipitous drop.”

OK, but such a terminological squabble seems argumentative.

My critic asserts that because I claimed that the Bureau “did not know how to study roof falls,” before the introduction of bolting, I reveal a “misconception of how roof support systems actually work.” This is a non sequitur. In the 1920s a Bureau engineer admitted they did not know how to study roof falls. They did, however, know how to prevent them, using what I term “proper mining practice.” Unlike older approaches, however, roof bolting did provide a research agenda.⁵

Mark claims “roof bolts were more expensive than timber posts.” In fact, they did reduce costs per ton of coal. I detail a whole set of benefits they generated, all of which raised productivity and/or reduced costs. One Indiana mine reported that productivity jumped 50 percent when it shifted to bolts. While other companies reported similar gains, a review of 1951 suggested that improvements of 10 to 20 percent would be more typical, although better equipment and learning surely increased these numbers. North American Coal reported that bolting raised production and reduced costs $.20 a ton. This is why their use spread, after all. However, the relevant issue is how bolts affected incentives to provide safety at the coal face—and the answer is they increased safety incentives because unlike timbering they did not interfere with machine mining.⁶

Why did bolting have so little impact on safety for so long? Mark claims that companies aimed for “acceptable level of risk,” but this does not get us very far: what, for example, influenced this level? An economist might suggest that it was the costs and benefits of safety, and so if bolting reduced the costs of protection near the coal face, why didn’t it have any impact for so long? I provide several explanations. One reason was that up to about 1965, the share of small mines increased and they did not use it. Mark points out that their role had fallen by 1968, which simply supports my claim.

I also argue that the new technology might increase both safety and production and so companies might trade safety for output. Bolting also required a lot of learning, for initially no one knew how these might trade off and indeed, the package was changing as bolting improved. Mark seems to dismiss this, but contemporary literature provides abundant documentation. In the article, I discussed what they were learning, but the literature also reveals that the learning was often quite systematic and experimental. Consider how an engineer from Island Creek Coal describes the company’s experience after it introduced bolting in early 1949: “extensive tests were underway … bolting alone has not been successful … in our early experimenting we tried various methods of roof drilling.” The emphasis is mine. Another article about a Kaiser Steel mine was entitled “Roof Bolt Testing Ground” while one on Carbon Fuel Company was “Better Bolting Methods Improve Safety and Output.” Tennessee Coal, Iron and Railroad systematically investigated failures and modified practice as a result. Emphasizing learning is not just a matter of semantics for it implies historical change—trial, error, mistakes, and improvement.⁷

As Mark suggests, this learning often involved “calculated risks,” and sometimes the calculations...