The Numbers Behind the Transit Strike

WSJ.com - The Numbers Guy By CARL BIALIK: "Lots of numbers are being tossed around about the economic impact the New York transit strike will have on the city.

Marcia Van Wagner says such numbers are "more an art than a science" -- a notable acknowledgement, given that the government office she runs is behind one of the most quoted stats. "There are ways to gussy it up and make it look more scientific, but that's really all an illusion," Ms. Van Wagner, who is New York City's deputy comptroller for budget, told me of her office's efforts to estimate the potential impact of a strike by bus and subway workers that began Tuesday.

Ms. Van Wagner, an economist, and her agency's chief economist reported that a strike could cost the city $400 million in economic activity on Tuesday and $300 million on subsequent weekdays this week, as city residents miss work and shoppers can't reach stores. The price tag for the weekend, should the strike last that long, is estimated at $100 million total for both days, because less business gets done on weekends, and the economists expect the Jewish and Christian holidays to slow business even more than usual.

But New York's Economic Development Corporation put its estimate for lost economic activity higher, at $440 million to $660 million per day. These numbers were cited by the city's commissioner of emergency management in court filings opposing a Transport Workers Union's strike, to demonstrate the severe economic harm that the strike could cause. Mayor Michael Bloomberg chose a number in between the two estimates for his radio address Sunday, saying the strike would cost the city $400 million a day. (The mayor appoints the president and chairman of the board of the Economic Development Corporation, which isn't a city agency but which is under contract with the city to promote its economic growth.)

When asked at a press conference Tuesday about the $400 million number and whether there was an update to the city's economic-impact estimate, Mr. Bloomberg said, "There's no way to know. These kinds of numbers are based on modeling, and some people's guess."

It's not clear exactly how the different groups arrived at their estimates. Ms. Van Wagner explained how her office "thought through" the numbers and used "professional judgment," but declined to walk me through the calculations. The mayor's office and the Economic Development Corporation referred calls to the city's law office, where a spokeswoman forwarded me the legal filings but declined to elaborate on how the economic-impact numbers were determined, citing the ongoing legal proceedings. The legal filings didn't explain the numbers, but did cite estimates from the two most recent transit strikes, in 1980 and 1966.

Several economists I spoke to questioned the methods behind the estimates, and some called the numbers inflated. What's more, a lack of day-by-day economic stats from the city, coupled with the difficulty of accounting for economic activity that was shifted to before or after the strike, will make it impossible to verify, after the fact, whether the projections were correct.

Yet Mr. Bloomberg's $400 million-a-day projection was reported by Agence France Presse, the Associated Press, CNN and the New York Times. And the higher estimates of $440 million to $660 million were also widely reported, in Newsday, the New York Times and the New York Sun.

Other estimates -- for example, $10 million per day for extra police officers, from the city's court filing -- are better founded, as it is easier for the city to measure such expenses. Still, media reports tend to focus on the larger numbers, which include estimates for lost productivity and retail sales.

Ms. Van Wagner said my questions about the stats were probing the "dirty underbelly of local economic analysis." She went on to explain that her group started by taking the gross city product -- a measure of the city's annual economic activity, which is about $425 billion to $450 billion annually -- and then estimated that the city product on a typical weekday this time of year is $1.5 billion (arrived at by dividing by 365, but then adjusting upward for weekdays and downward for weekend days). Then the economists decided that productivity loss -- due to cancelled workdays, truncated workdays or distracted workers -- would amount to roughly 25% of economic activity on the first day, and closer to 20% as businesses and workers adjust to the strike.

Brendan O'Flaherty, a Columbia University economist, called the estimate of $400 million "pretty high." He noted that New Yorkers had been bracing for a possible strike for days, and likely shifted some of their work (and retail purchases) accordingly.

I asked Ms. Van Wagner if her numbers could be checked after the fact. "No, sadly," she replied. "Those of us in the prognostication business like to see after the fact how far off we were. This time, there will be no way to tell." She said that local retail and employment figures are compiled monthly, not daily, making it difficult to isolate the effects of the strike. (Even daily tallies might obscure the effect of work-shifting; Prof. O'Flaherty worked long hours at the beginning of last week, ahead of a strike threat for Friday, to ensure he finished preparing exams for this week.)

Ms. Van Wagner pointed out that forecasting economic impact is easier when a local event is less diffuse, like a major sporting event or political event, when organizers know how many guests to expect. Yet even these forecasts are dicey and often more useful to boosters than economists. As I pointed out in two earlier columns, a widely reported estimate of benefits Jacksonville, Fla., would reap by hosting this year's Super Bowl wasn't well founded. Economists Victor Matheson and Robert A. Baade, from College of the Holy Cross in Worcester, Mass., have shown that host cities for college basketball tournaments often lose money, despite rosy projections from boosters. And forecasts by Boston and New York about the benefits of hosting the 2004 political conventions proved too sanguine about the loss of local business activity.

The higher estimate on transit-strike costs, of $440 million to $660 million, is more nebulous. It's preceded in court documents by the disclaimer that "it is difficult to make precise projections," and followed closely by a recounting of woes suffered during the city's prior two transit strikes. According to the legal filings, the New York State Public Employment Relations Board said that the 1980, 11-day strike cost the city's businesses $1.1 billion -- an average of $100 million a day, far short of the city's estimate for today.

Since 1980, the city's population and economy have grown, but New York has also changed in ways that could mitigate this strike's effect. Services jobs -- which include many white-collar jobs that can be done from home or alternate locations -- represented 52.7% of New York employment in 2002, up from 40.6% in 1980 and 31% in 1966, according to the comptroller's office. (Of course, some services sectors, like restaurants, are severely affected by the strike.) And the Internet has eased telecommuting. (I write this column from my home in Brooklyn and file it electronically, which would have been difficult for me to do in 1980, and not only because I was an infant at the time.)

Also, a strike incurs costs on the public sector that are difficult to measure in the rubric of economic activity. For example, schoolchildren may miss class. According to the city's court filing, attendance at public high schools ran below 50% in the 1980 strike. And the increased use of automobiles could create more polluting emissions, as Brookings Institution fellow Robert Puentes, who studies metropolitan policy, pointed out to me. Measuring the impact of transit, and of transit strikes, is "a very understudied area," he says."