On my other cars ('89 Mitsu and a '99 Celica) the city estimates were really accurate for the years in which they were produced. In other words, we actually do get what the estimate was in that model year. But not with this car. The city estimate is 25, but we get only 22.keep in mind too the epa "city" mpg determination is not 100% to realistic city driving, also it probably doen't include the same amount of sitting time as living near a big city like new york/long island if that is the city you are talking about.
I suspect that in recent years the EPA ratings have become inflated. In 2009 there was a push by the government to produce and market smaller, more efficient cars. Hence, they came up with the 'cash for clunkers' incentive not only to get older, larger vehicles off the road, but also to boost the bad economy by luring people into buying new vehicles. Considering those policy goals, the government would have an interest in inflating the figures, since official fuel economy data that is attractive entices people to buy new. If you think the idea of their inflating the numbers is far-fetched, ask yourself who now owns the majority controlling interest of General Motors? Yep, Uncle Sam. And Chevy certainly wouldn't object either, as long as it sells more cars.
As for 'sitting' in traffic, believe it or not, idling burns relatively little fuel and plays a very minor role in reducing fuel economy. In fact my wife's daily commute is on local roads and involves virtually no idling in traffic except for at red lights (which undeniably are a part of real city driving).