Seventeen Months on the Chemical Safety Board

By Beth Rosenberg

On August 6, 2012, a pipe in the Chevron refinery in Richmond, CA, ruptured and leaked flammable fluid. The fluid partially vaporized into a cloud that engulfed 19 employees and then ignited. Miraculously, the workers narrowly escaped and were not seriously burned. The large plume of particulates and vapor travelled across Richmond, sending more than 15,000 people to the hospital with respiratory and eye irritation. Dozens of citizens also developed hearing problems as a result of the loud explosion.

Such an incident triggers a number of investigations. The Occupational Health and Safety Administration (OSHA) and the US Environmental Protection Agency (EPA) investigate to see if there are legal violations. The company itself, Chevron in this case, investigates. If there is a union, it will investigate or participate in the company’s and OSHA’s investigations. The US Chemical Safety and Hazard Investigation Board (CSB) can also investigate, if the CSB decides that the accident is momentous enough to represent a significant threat to public well-being, or if there are valuable lessons that the accident can offer for improving the safety of the chemical industry.

The CSB is a small federal agency that investigates incidents—mainly explosions and leaks—in the petrochemical industry. It has a budget of $11 million and about 50 employees. Based on the model of the National Transportation Safety Board, it has no regulatory authority; it makes recommendations, based on root cause analysis, to the party or parties that will be most effective in preventing similar accidents from recurring. Recommendations are usually made to companies, federal agencies such as EPA and OSHA, municipalities, local public health and environmental agencies, unions such as the United Steelworkers of America, or trade associations such as the American Chemistry Council. CSB was created as part of the 1990 Clean Air Act amendments.

An invitation to serve

In 2011, the chair of the CSB, a former professor of mine, asked me to apply for a job as a board member of this small agency. It seemed like a perfect fit for me. On family vacations when I was a child, we would occasionally tour factories. While my family was absorbed in the smell of melted chocolate at the Hershey factory, or the glow of a red-hot slab of copper from which pennies would be stamped at the US Mint, I was noticing the workers. By the time I was 10, I wanted to do something to address the injustice that some people had jobs that were soul-witheringly boring, dangerous, painful, or even fatal, and I eventually made it my life’s work.

With a doctorate in work environment policy and 30 years in the field of worker health, I was nominated to be a board member of the CSB by President Obama on September 20, 2011. The lengthy vetting process involved written recommendations from colleagues, as well as Federal Bureau of Investigation interviews with me, numerous friends, coworkers, and neighbors about my spending habits, drug and alcohol use, travel history, and character. I was finally confirmed by the Senate on January 1, 2013.

Board members serve for five years, and there are five slots, only three of which were filled during my tenure. We presidential appointees were supposed to direct policy and the priorities of the agency. We were to decide what events were to be investigated, and to work with CSB’s professional staff investigators to determine the focus of the investigations. The agency’s dozen investigators have a range of backgrounds in chemistry, public health, industrial engineering, chemical plant safety, human factors, refining operations, and law.

At the end of an investigation, the investigators provide a report of their findings to the board’s recommendations staff, which writes policy recommendations to the pertinent parties, based on the evidence revealed by the investigation. A public hearing is then held at which the investigators present the findings and recommendations to the affected community, and board members vote to approve or reject the report of the investigation along with the recommendations. During my tenure, senior leadership of the CSB consisted of a general counsel and a managing director. The managing director oversaw the investigations, but was also responsible for the board’s public relations, which included helping to produce training videos that are very popular with both labor and industry.

As a scientist and academic, I was unprepared for the politics of the job. I expected to be pressured by stakeholders outside the agency—from oil company executives and environmental activists—to approve or reject certain recommendations. That didn’t happen at all. Instead, the real problem turned out to be the internal politics of the CSB itself, and the related problem of lack of accountability to anyone outside the board. My time at the CSB raised some important and troubling questions about the role of science and democracy in ensuring occupational, public, and environmental health and safety in the petrochemical industry.

Soon after I arrived at the CSB, I and my fellow board members were presented with a draft report on the Richmond explosion. The draft revealed the agency’s dysfunction that came to dominate my tenure as a board member.

An explosive culture

CSB investigators learned that in the 10 years before the explosion, Chevron Richmond management was told at least six times, by their own employees, as well as by an outside group of Chevron technical experts, that the pipes that ultimately were the source of the leak should be inspected for corrosion. Sulfur, found in crude oil, reacts with iron-containing compounds, and at high temperatures (450-800 degrees F), the reaction will corrode steel pipes. This sulfidation corrosion had been documented in the industry literature for two decades. Alerts repeatedly went out to all Chevron plants from both the company headquarters and the industry to inspect the pipes. The glaring question that rose from our investigation was why did management at the Richmond plant not inspect and replace the vulnerable pipes?

The petrochemical industry in the United States is, to put it mildly, mature. Ninety-five percent of the 144 refineries in the United States were built before 1985. The average refinery is about 40 years old and some are almost 90. People who work at these plants sadly joke that they are held together with duct tape. Often, equipment is not maintained and instead is “run to failure.” In 2012, the CSB tracked 125 significant process safety incidents at US petroleum refineries.

Financial incentives are lacking to ensure that workplaces, including refineries, are safe and well-maintained. Even for workplace fatalities caused by willful negligence, the fines levied by federal OSHA are low; in 2014, the median penalty in fatality cases was $5,050. Criminal penalties are weak, too; a willful violation resulting in death is prosecuted as a misdemeanor. Killing workers is cheap.

Financial incentives are lacking to ensure that workplaces, including refineries, are safe and well-maintained.

Injuring them is more expensive, because injured workers have medical bills, but companies bear little of the cost. The workers’ compensation system is supposed to pay for medical care and some salary replacement, but the system is deeply flawed, and a minority of injured workers bothers to use it. In the 1920s, before workers’ compensation came into being, jury awards to injured workers were wildly unpredictable and sometimes very high. The workers’ compensation system provided employers with predictable expenses in the form of premiums. Workers gave up the right to sue, and lost any compensation for pain and suffering (which was part of the award in pre-compensation days) in exchange for a “no fault system” that would pay for salary replacement and medical costs.

Compensation payments have not kept up with the cost of living over the past 40 years, and recent changes in state compensation systems have made it more difficult for workers to obtain coverage for medical expenses that are rightfully theirs. And now some states are opting out of the formally required comp system, so employers are no longer even paying premiums. A recent OSHA report noted: “Employers now provide only a small percentage (about 20%) of the overall financial cost of workplace injuries and illnesses through workers’ compensation. This cost-shift has forced injured workers, their families, and taxpayers to subsidize the vast majority of the lost income and medical care costs…” The financial incentives for companies to maintain a safe workplace are minimal.

Nor do companies seem to be very concerned about reputational damage, or the loss of refining capacity that occurs when a plant is out of service due to an accident. My conversations with insurance specialists revealed that most companies have business interruption insurance or are self-insured. Not only does the insurance adequately protect against such losses, it also covers the cost of replacing the equipment—another reason to run to failure. Moreover, the company as a whole may benefit after an accident, if the lost refining capacity leads to increased gas prices, as it did on the West Coast after the Richmond explosion. Overall, the incentives line up against repairing or replacing aging equipment.

The human element is important for plant safety as well. In particular, the tenure of plant managers varies by company and may influence maintenance practices. Some companies move the managers every two or three years, while others move them every five years or more. In my discussions with environmentalists and industry people alike, I was told that plant managers are incentivized to not spend money on maintenance, and the managers who know they will soon be moving on are therefore not inclined to do repairs, gambling that the accident will happen on the next guy’s watch. If a manager is there for seven years, he is more likely to do the repairs because he doesn’t want an accident to happen while he is at the helm. There is no evil intent here—it’s a matter of perverse incentives. A research study on the effect of the tenure of plant managers on maintenance practices could be very enlightening.

This is the social context of the industry that the CSB advises and in which the Chevron incident occurred. Because so many people were affected by the explosion, the CSB held a public hearing in Richmond, CA, in April 2013 to present the board’s initial findings to the community. CSB investigators presented excellent information about the ignored warnings of corroding pipes. All three board members agreed that the incident could be explained in terms of poor management decision making. What was going on in Chevron’s senior management that allowed them to ignore the literature and the warnings of its own employees? Congressman George Miller (D-CA), in remarks at this meeting in his district, echoed our questions about the organizational culture of Chevron. Some board members also raised questions about the role of California’s state-level OSHA, whose severe personnel and resource shortages may have affected its ability to sufficiently inspect plants in the state. We also wondered how and why Contra Costa County’s much-lauded Industrial Safety Ordinance, which requires chemical plants and refineries to submit a risk management plan to the US EPA and the County Health Service, missed this problem.

The Board Members are supposed to have some influence on the focus of investigations, so we three board members expected some investigation of management decision making and organizational culture at Chevron. In April 2013, the CSB staff presented the first Chevron Interim Investigation Report. Rather than assessing why known corroding pipes were not inspected and, if needed, replaced, the report provided a thorough technical analysis of sulfidation corrosion and why the pipe broke. Our concerns as board members about organizational management issues were simply ignored in the investigation.

The “safety case”

The way you define the problem shapes how you solve it. If you define the problem as sulfidation corrosion, then you end up recommending adding chrome to the pipes to make them more corrosion resistant. If you define it as a Chevron management problem, it is messier, more controversial, and requires more thought about how you’re going to fix it. CSB spent many thousands of dollars on technical experts for all sorts of investigations, but I could not get any momentum to hire an organizational management consultant to help with the Chevron investigation. To me it seemed clear that protecting the industry, workers, and the public alike demanded an investigation into management as well as technical failures. And indeed, some Chevron managers who were not from the Richmond plant later told me they believed that Chevron management had “dropped the ball.”

But as a recent member of the board, I failed to appreciate that the board chair had abdicated power over the investigations to the managing director, who tightly controlled the staff. Although I repeatedly brought up the need for an organizational aspect to the investigation, the staff, directed by the managing director, was trained to ignore the will of the board. Indeed, that was the pattern of my tenure at CSB.

The second CSB report on the Chevron accident came out in January of 2014. It again essentially ignored the issue of management failure, although it did recommend a major change in regulatory regime: the adoption of what is known as the “safety case” regime in California refineries. The safety case is a regulatory regime in which the company presents information (a case) to an informed regulator, affirming that the company can operate safely. The government then gives the company permission to operate if the government deems that the company has sufficient safeguards in place. This approach is used in the United Kingdom, Norway, and Australia.

Despite the fact that CSB has no regulatory power, its authority as a federal agency with a presidentially appointed board means that this recommendation for regulatory regime change for California refineries would have had significant weight. And given the deficiencies of the current regulatory regime, I understood the impetus to recommend a new approach. Nonetheless, I was reluctant, as was one of the other two board members, to support this recommendation. The recommendations staff was also uncomfortable with this approach because it did not flow from the evidence; instead, they said, it was being foisted on the staff by a senior staffer in the agency.

My reservations about the safety case reflect its complete dependence on a well-funded, well-informed regulator to approve or reject the industry’s case. The number of OSHA inspectors in the nation is dismally low: 2,200 inspectors for 130 million workers in 8 million worksites, or one inspector per 59,000 workers. There is no reason to think that this will change radically anytime soon. Labor is relied on to provide a check and balance to industry’s proposed case, which is a strategy that may work in countries where labor, industry, and government all have somewhat equal power, but that is not the case in the United States. I feared that a safety case regime would simply duplicate the same power relations of weak labor, weak government, and strong industry that underlie the current failed regulatory regime. Additionally, the safety case regulatory regime lacks transparency. Safety cases are seen by regulators but they are not public. Although we may disagree with the levels of air pollutants that a given plant is permitted to emit, the permits under the Clean Air and Clean Water acts, as well as federal environmental statutes, are public documents. By contrast, communities are shut out of the safety case process.

The CSB presented the second regulatory report on the Chevron accident for a vote at the public hearing in Richmond in January 2014. It provided a literature review of the proposed safety case regulatory regime, but it did not include any of the downsides. It was an advocacy piece. Two of the three board members wanted to settle some of the questions about the safety case before we voted to approve the report, and so instead we voted for a postponement, and offered criteria that we wanted to be met before we approved the report. The meeting was contentious, and many in the city were understandably desperate for something to rein in the company that dominated their city. We advocated an immediate remedy of increasing resources for California state OSHA inspectors—which eventually did indeed happen.

After the meeting, to answer our questions, we proposed gathering an expert panel. This was dismissed by senior leadership as too complicated, because it would require compliance with the Federal Advisory Committee Act, which mandates that such meetings be open to the public. We proposed a National Academy of Sciences investigation of the safety case, which was dismissed by senior leadership as too expensive. We wanted to hold a conference, to hear all pros and cons of how a safety case regime would work in the United States. That did not happen because neither the chair nor senior staff leadership wanted it. Once more, the will of the majority of the board—two of its three members—was ignored.

Instead, the other board member and I were vilified in an email from the chair to the whole agency for agreeing with industry in opposing the safety case recommendation. Although the Richmond local of the United Steelworkers union did support the safety case, the chair’s email did not bother to mention that the health and safety leadership of the national union opposed it. There was never an open debate about the safety case, either within or outside the CSB, despite the fact that a majority of the board members had serious reservations about it. Would the proposed regime change improve a bad situation or make it worse? We had no idea, and there was no discussion, or even a mention, of the downsides of safety case regimes in the still-unapproved January 2014 report.

The whole safety case debate at the CSB raised questions about how recommendations were formed. The recommendations staff was justifiably adamant that recommendations must flow from the evidence; staff members were uncomfortable, and powerless, when recommendations from investigators, or senior staff leadership, were inserted into reports without their endorsement. In the end, CSB’s investigation of the Chevron case was undermined by board governance that allowed senior staff and the board chair to impose their will on other board members and staff, skewing agency process aimed at serving the public, workers, and industry.

A better board

During my tenure as a board member, the CSB’s agenda seemed to be captured by its professional staff, whose members’ interests largely lay in preserving their power and positions, not in ensuring that the board served the public good. Indeed, many at the CSB thought their jobs were primarily about doing public relations for the board, because good PR affects the budget and the budget is power. CSB’s senior management was savvy about the ways of government, whereas the board members—presidential appointees with time-limited terms, who come in from industry, nongovernmental agencies, labor, or academia—often are not. The well-intentioned but easy-to-manipulate board chair was apparently convinced by senior staff that his job was to serve the interests of the agency, which means protecting its image so that it continues to get funded.

The consolidation of power by senior management through alliance with the chair began before my arrival, and it continued throughout my tenure. Because the board senior staff members served the chair, not the whole board, they were able to manipulate the chair to prevent other board members from bringing up votes at public meetings. The board was disempowered and marginalized to such an extent that when I requested additional information from staff, this was framed as “disrespecting the staff.” This framing was used by senior staff to vilify board members other than the chair. We were treated as temporary pests and portrayed to other staff as enemies of the agency. Some staff members claimed we were denying the chair “his legacy” by not supporting regulatory regime change.

The job of a presidential appointee is to serve the public, not the agency or one’s ego, and that means being accountable to the public. One obvious way to be accountable to the public is to have public business meetings, where investigations that are under way can be discussed before reports are drafted and voted on. Many of the staff members and the majority of the board believe that safe workplaces require an organizational culture where constructive criticism is welcomed rather than viewed as an attack. Yet we lacked this culture in our own workplace. Despite efforts by the majority of the board, there were no public business meetings because they were seen as “airing dirty laundry.”

A second way to be accountable to the public is to release draft reports of investigations and recommendations with enough lead time before public meetings—at least a month—to allow for meaningful public comments and for those comments to be considered and incorporated into the final report. Even better would be interim public meetings to discuss the findings and the recommendations of draft reports with the public and stakeholders, where no voting would occur. The vote to approve the report could then occur at a subsequent public meeting, after comments had been considered and debated. Yet when I was on the board, reports were presented and voted on at the same meeting. The public comments at these meetings could have no impact on either the report or the recommendations; they merely provided the illusion of public input.

A third way to be accountable to the public is to make the evidence of the investigations public as the report is being written. The National Transportation Safety Board, for example, has a public docket that contains all the evidence of an investigation, to be transparent and to ensure that recommendations flow from the evidence.

CSB’s apparent capture by senior staff penetrated and poisoned the entire agency and reduced its effectiveness. A backlog of investigations, with no plan for how to clear it up, led to the feeling that although everyone was working hard, things were not getting done. According to the annual federal survey of 2014, employee morale was at an all-time low, as was faith in senior staff leadership. The atmosphere was one of intimidation and distrust.

I was told that in the past there had been an open and free exchange of ideas among everyone in this small agency, but this had now changed. The year before my arrival, in an email, the staff had been instructed by the managing director to not communicate with board members without his knowledge “as a courtesy,” but the real message was clear to everyone. Fear of reprisals was rampant. Employees were leaving the agency to work at other federal jobs. One staff member was denied an annual bonus for making a policy inquiry to another federal agency without “permission.” People disagreeing with the chair or the managing director said they “had targets on their backs.” Office doors were mostly closed. Staff was so nervous about talking to board members that I had meetings off-site or in the ladies room. Yes, a presidential appointee was conducting business in the ladies room.

I resigned after 17 months, at the end of May 2014, because I wasn’t able to accomplish anything useful. The problems—agency mismanagement, low employee morale, intimidation of the staff, the toxic work environment, and marginalization of board members by the chair and senior management—had been brewing for three years before I got to the agency. Complaints to Congress and the EPA’s Inspector General (which oversees CSB) from the CSB staff and board members, former and current, finally culminated the next month in a very dramatic hearing of the House Committee on Oversight and Government Reform entitled “Whistleblower reprisal and management failures at the Chemical Safety Board,” at which I testified.

A second hearing of the same committee, in March 2015, resulted in a bipartisan call for the chair’s resignation. That’s right: amid all the partisan rancor in Congress, the mismanagement and abuse of power at CSB was enough to foster a bipartisan moment. The White House acted, and in April 2015, two months before his term expired, the chair resigned. Senior staff members were subsequently put on administrative leave, and the managing director was soon recommended for dismissal. My former colleagues report that the atmosphere is no longer intimidating, but relaxed and open. A new chair and new board members have put the agency on an even keel. Yet key positions—the general counsel and head of recommendations—remain vacant. The status of the managing director has not been settled. As of May 2016, he is still listed on the CSB website as the managing director, although he has not been in the office since June 2015.

In its third and final report on the Chevron accident, released in January 2015, the CSB did examine some of the organizational problems at the Richmond plant, and provided useful diagrams about the flow of information in the company. Still, it never got to the crux of the issue: how decisions were made; what happened when technical experts disagreed with managers; why the many recommendations whose implementation could have prevented the accident were ignored; who was responsible for ignoring them; and what corporate incentives were in place to allow all the warnings to be dismissed? The organizational management study that should have been done transparently and openly by the CSB was done instead by Chevron, behind closed doors. Unless Chevron publicizes its findings, neither any other company nor the public will benefit from its research.

The 2014 report advocating the safety case was never approved by the board, but as of May 2016 it remained on the CSB website.

My experience raises the more general question of how the government can improve the potential for scientists to contribute effectively when they are appointed to agency boards aimed at protecting the health and safety of the public. Before coming to Washington to work on the CSB, I had been warned that people in government often forget they are serving the public and not their agencies. But when I arrived at the CSB, I had no instruction, guidance, or orientation at all. New board members need to be trained about their rights and responsibilities, as well as about how an agency governed by a board with a chair—as opposed to an agency run by a single presidentially appointed director (such as the EPA)—operates.

Congress, in its wisdom, deemed that CSB should be guided by a board; much of the conflict during my tenure stemmed from disagreements about power distribution among the chair and the board members. Board members, including the chair, must be trained that the business of the agency should be done in a transparent and collegial manner. Perhaps the training could be done by the National Academy of Public Administration. It certainly should not be done by anyone inside the agency at which the new board members will serve. Another useful reform might be to adopt the NSTB’s practice of having two-year terms for its chair, extendable by the President, who typically seeks the consent of all the other board members. Such a policy would have helped the CSB enormously. A third critical need is to explicitly specify the process for adjudicating differing expert opinions among board members. Better mechanisms of public accountability, as I discussed above, should help prevent abuse of power by staff and particular board members, but a culture of respectful deliberation among board members is necessary, too, followed by public votes that make clear where members stand on divisive issues, and why. Such changes will flow from effective board leadership.

The problem of how to guarantee good leadership, however, remains. Scientists serving on government boards must remember that the agenda of politically savvy staff does not always mesh with the public mandate of the agency. Good leadership requires wisdom, integrity, good judgment, and a sense of fairness and focus. The ability to listen to opposing views openly without defensiveness and the ability to discern truth from manipulative lies are crucial for good leadership. Occupational physician Tee Guidotti points out that leaders who are worried about leaving their mark on an agency are driven by power and ego, not service. Good leaders must be comfortable with power, but they shouldn’t need it. A White House and congressional vetting process that focuses as much on character and leadership skills as it does on resumes would benefit everyone with an interest in chemical safety: the agency, the industry, workers, and the public.

Correction

The originally-published version of this article contained several errors: the hearing said to have taken place in October 2014 actually took place in January 2014; and the hearing said to have taken place in February 2013 actually took place in April 2013 (there was no February 2013 public meeting). A complete transcript of the April 2013 meeting is available on the Chemical Safety Board’s website. Also, the article misstates how National Transportation Safety Board (NTSB) leaders are selected. These errors have been corrected in the online version of the article. In addition, one of the editors of Issues, Dr. Daniel Sarewitz, is the brother-in-law of the author of the article. The article meets the standards for publication in Issues.