As an associate editor at The Atlantic, Jordan Weissmann has been following the reduction of driving in the U.S and written about the reduction in younger drivers. He wanted to know if "the U.S. is getting over car culture", or if the driving reduction was mainly a product of the economic downturn. He looked toward Sivak's findings (PDF) for an insight.
Sivak analyzed the period from 1984 through 2011. He examined the registration "of light-duty vehicles (cars, pickup trucks, SUVs, and vans) for each year from 1984 through 2011."
Its key take-away is that the number of cars per household actually began to decline pre-recession, after 2006. Same goes for cars per licensed driver and cars per person.
In fact, the accompanying graph shows it all, with the vehicle registration rate - per person, per licensed driver, and per household, peaking around 2006 - before the recession which is recognized as beginning in December, 2007 per U.S. Bureau of Labor Statistics (PDF).
Yet, Weissmann is not entirely sold, pointing to pre-recession economic factors, e.g. home values peaked in 2006 and suspects that influenced car ownership rates.
He concludes that "(t)he decline of car ownership might well turn out to be a long-term trend with cultural and demographic roots. But if so, the housing bust and recession still seem to have been the tipping point."
He may be partially correct - if you distinguish between absolute numbers and rate of ownership.
From Sivak's abstract (PDF): "...with the improving economy and the expected increase in the U.S. population, it is highly likely that (from a long-term perspective) the absolute number of vehicles has not yet peaked." However, if you believe Sivak, the rate of ownership peaked seven years ago.