There's a joke that everyone in customer service at a bank is a VP.
Dan Steinberg wrote of trends to decreasing # IT workers and increasing proportion of managers.
One potential factor that I haven't heard anyone discuss is a change in focus from big projects to small. A feature of the dot-com era, we used to speak of "internet time". Time to market was everything. Now the bubble has burst, but the pendulum hasn't swung all the way back to "quality before timeliness" as a driver.
Why should this matter? One feature of the manager/worker interaction is the need for managers to help resolve issues, focus resources on problems, that sort of thing.
Before the internet, projects were typically larger, and mostly internal to the organisation. Large projects with lots of introspection have fewer unplanned issues, at least in my experience, so there is just less for a manager to do. A project comes up, some folks are assigned, by and large you can sit down with the engineers, get them moving in the right direction, and sit back for a spell. There is more people supervision than problem resolution being done by management.
By contrast, a dynamic environment with lots of small projects being formed and reformed, interesting and difficult questions come up more often. And the extension of these projects into the external (that is, outside the firewall) realm means that more of these issues are of potentially broad organisational import. And thus may require more management attention, more often. And so, more managers.
Note in all this, I've been using the word "manager" to mean "responsible person who can speak for the organisation". "Manager" is vague, which is part of the problem - we just don't have common language to differentiate between "people supervisor" and "responsible team-member". Or job titles. Hence the joke about banks and VPs. They have to be VPs because their word can come to bind the bank. Even if they supervise no-one.
Which leads some to imagine that there is a problem when the ratio of managers increases when such is a spurious result due more to language than to reality. Again, in my experience.