Tufte in The Visual Display of Quantitative Information points out that the Wall Street Journal is one of the least likely newspapers to display anything in a graph. Indeed, it would be a lot easier for visually minded people like me to check the Journal's arguments if they were displayed in bivariate plots more often.
http://lawprofessors.typepad.com/securities/2008/02/januarys... : "January [2008] was a slow month for IPOs, with only five that raised a total of $892 million. This compares with ten in January 2007 that raised a total of $2.25 billion." I quickly googled, and there's probably more to look at here, since either 5 of the 6 IPOs did it in January, or there's different counting methodology in action here. Still, it's down even in months that weren't bad. It's not just statistical gaming, even if that's in play. Asking what changed is still fair.
"According to the National Venture Capital Association, in all of 2008 there have been just six companies that have gone public. Compare that with 269 IPOs in 1999, 272 in 1996, and 365 in 1986."
It's probably more complex than that, or he'd have hammered the point a lot more. I'd love to see a graph, for instance.