From CommonWealth: T report misses mark on absenteeism numbers

Originally published as the CommonWealth Magazine Backstory for April 30, 2015. ONE OF THE MOST ALARMING figures to come out of Gov. Charlie Baker’s MBTA panel report earlier this month was 57. That’s the average number of working days that T employees miss each year, according to the panel’s analysis. The analysis also found that the absence rate across all positions at the T was 11 to 12 percent, twice the level reported at peer transit agencies and four times the rate of the transportation industry as a whole. The logical conclusion from the analysis is that the MBTA is staffed by shiftless malingerers, enabled by an ineffective management structure.

The problem with this conclusion is that none of the absenteeism numbers used in the report stand up to scrutiny. An analysis of the source documents for the panel’s report, supplemented by conversations with state transportation officials, including some who worked on the report, revealed errors of interpretation, comparison, and logic.

There does appear to be a problem of absenteeism at the T, but the report does little to clarify how serious it is. Instead, the report showcases alarming numbers without explaining how they were derived. The numbers were also leaked to the press in advance of the report’s release to generate maximum impact.

Let’s start with the 57 days-off figure. That number includes vacation days, holidays, even days spent at employee training sessions, all of which are scheduled in advance. Stripping out these scheduled absences leaves a total of 22.5 “unscheduled” days away from work. These unscheduled days off are the T’s real problem because it’s difficult to run a transit agency effectively with so many workers failing to show for work.

The T panel’s report said the average absence rate at the transit agency is 11 to 12 percent, meaning roughly a tenth of the workers fail to show up on any given day. In calculating that percentage, the panel divided the average number of unscheduled days away from work (22.5) by the average number of days T workers actually do show up at work (204). The calculation is puzzling. Absenteeism is normally calculated by dividing the average number of days workers are absent by the total number of work days in a year. The T’s “employee availability reports” list 261 work days in a year, meaning the calculation would be 22.5 divided by 261, or 8.67 percent. One could debate whether holidays and other types of scheduled days off should be removed from the 261 work day total, but it is clear that removed the unscheduled days off from the work day total is misleading and inflates the absenteeism rate.

As an illustration of why the T panel’s approach is suspect, suppose a worker played hooky for three full days in a period of 10 regular work days. Using a conventional absenteeism calculation, the absenteeism of that worker would be 30 percent (3 divided by 10). Using the T panel’s approach, the number of absent days is first subtracted from the total work days and then the calculation is performed. So instead of 3 days being divided by 10, the panel would divide 3 by 7, for an absenteeism rate of 43 percent. The effect of the T panel’s approach is to make each unscheduled day off more damaging than the previous, increasing the absence rate faster and faster.

MassDOT spokesman Michael Verseckes confirmed via email that the T panel calculated the absence rate by dividing unscheduled absences by the average number of days actually worked. He wrote that “the 11 percent figure is akin to a performance metric – what it represents is the ratio of employees’ unavailability for reasons that could be considered avoidable. As an average number over the course of a year, it amounts to 22.51 days per employee. But as a measurement, it separates out the days off that were specifically unscheduled, and divided by the total average number of days per year that employees are available to work.”

The T panel’s approach makes impossible any comparison to absenteeism figures outside of the MBTA. The T panel report compared the 11-12 percent MBTA absenteeism rate to rates of 5-6 percent at unspecified peer agencies and 3 percent for the transportation industry as a whole.  Neither comparison is accurate.

Absenteeism slide annotated

The 3 percent figure is reported by the US Bureau of Labor Statistics for the entire “transportation and warehousing” industry, which includes transit agencies but also air and water travel, trucking and warehouse operations, even sightseeing tours. A call to the BLS confirmed the agency does not track absenteeism for transit agencies alone. Given how enormous and diverse the larger category is, it’s hard to see what light it sheds when compared to a large urban transit agency such as the MBTA. The BLS method for calculating absenteeism is also different than what is used by the MBTA.

The T panel report said “peer agencies have reported absence rates of 5-6 percent,” or twice the level the panel found at the MBTA. But the T and the peer agencies cited don’t calculate absenteeism the same way, a fact which is not noted in the report.

In phone calls and emails, state transportation officials said absenteeism figures for other peer transit agencies are hard to come by because most transit agencies are reluctant to publicize how many days their workers miss. One official who worked on the T panel report supplied documents or press releases on absenteeism from three other transit agencies. One was a 2006 audit of the Washington, DC, Metro system that relied on 2004 data. The audit cited an absenteeism rate of 5 percent. A 2013 press release from the Chicago Transit Authority (CTA) said the agency had reduced absenteeism from 7 percent to 5 percent. The third document supplied by the T panel, a report on overtime costs at New York City’s Metropolitan Transportation Authority, includes no concrete information on absenteeism.

The Metro in Washington calculates absenteeism using fewer types of days off than the MBTA. Calculating the MBTA rate of absenteeism using the same types of leave as the Washington Metro system yields a rate of 5-6 percent, depending on what number is used for total work days. Officials from the CTA did not respond to requests for information on how their absenteeism rate is calculated, which is also not explained in the CTA press release cited by Massachusetts transportation officials.

In responding to concerns about the T panel’s absenteeism numbers, state transportation officials acknowledged the report was done quickly under a tight deadline. One official called it a “rapid diagnostic” and said its data and conclusions would have to be examined more closely by the proposed MBTA fiscal and management control board, if it is approved by the Legislature.

Perhaps the greatest irony is that the actual absenteeism data does appear alarming. Employees on average miss 22.5 days of work per year due to unscheduled absences. Use of the Family Medical Leave Act varies widely by employee category, suggesting that there may be some instances where it is being abused. These are areas where MBTA management could and should do more, regardless of the outcome of the governor’s reform legislation. The CTA was able to curtail absenteeism through better management, and the same could perhaps be done at the MBTA.

There are decades of academic research exploring why transit workers miss so much work, and many newsarticles from across the country spelling out the impacts of high absenteeism to systems everywhere. This is not a problem unique to the MBTA. The question is, how much of a problem is it for the MBTA, and how much time and effort should be sent addressing it? The report offers little insight into this crucial question.

Steve Koczela is the president of the MassINC Polling Group, which is owned by MassINC, the publisher of CommonWealth magazine.