This is becoming an annual event: somebody shouts “crime wave!”, based on a simple comparison of this year versus last year, and if this year's crime count is higher, then bingo, you've got a crime wave.
Last year it was the Washington Post with a burglary surge. This year, it's the Washington Times, headlining Violent crime in D.C. surges in 2012. “Violent crime so far this year has spiked sharply – a 40 percent increase that includes twice as many robberies at gunpoint than at this time last year”, they write, scarily. Continuing, “all police districts are reporting increases in violent crime, and all but one have had double-digit percentage increases, according to internal Metropolitan Police Department documents”. Other news media follow this lead, e.g., D.C. police: Crime wave continues with more robberies in Northwest today. Even the Guardian Angels are out in force, handing out fear-crime flyers at Metro stops.
Well – is this a real “crime wave”, or is there a less frightening explanation for these increases in crime? Here's one skeptic: violent crime in the first six weeks of 2012 fits neatly into long-term trends. Besides the “long-term trends”, this skeptic observes that “Seasonal effects and seemingly random events will trigger brief crime spikes and brief crime declines”. And, further, “as crime rates return to more typical levels, comparisons with an anomalously low violent-crime period create the illusion of a crime spike”.
That's the thing – if the earlier period was characterized by “anomalously low” crime, then a comparison to this year will result in a “surge”, even if this year's count is merely a return to normal. That was certainly the case for the Washington Post's 2011 burglary wave, comparing that year's wintertime count to the count during the winter of 2010 – notorious as “Snowmaggedon”, when everyone, including career criminals, was paralyzed by the deep snow.
The current “spike” in crime is based on this same mistake: comparing this year to last year, implicitly assuming that last year was a “normal” year. Such an assumption is always chancy, but especially in winter, when weather can make a very large difference.
Here's a chart showing the robbery counts, citywide, for the period January 1 through February 22 (approximately, the period considered by the Washington Times). Yes, the robbery count in 2012 is a lot larger than the count in 2010 and 2011. But on the other hand, it's not much larger than the 2009 count, and is consistent with the trend apparent in the 2005-2009 data.
What happened in 2010 and 2011? Quite simply, those were relatively harsh winters, especially 2010, with a record 40 inches of snow in January and February. In 2011, the snow totalled 8 inches, still enough to paralyze Washington. But this year, 2012, has been a “winter” characterized by nearly zero snow, and by balmy, record-high winter time temperatures. Similarly, 2009 saw almost no snow.
In such mild-weather conditions, people are out on the street, enjoying the unseasonably warm weather, and some people will be seen strolling along with smartphones out, their attention focused on whatever they're reading or texting. That's an invitation for a snatch-and-run robbery, and smartphones seem to be a favored target this year. In colder wintertime weather, people spend as little time as possible outdoors, and when they're out on the street, they're hurrying along to whereever they're going, their smartphones safely tucked away in pockets or purses. Plainly, the number of potential robbery targets is much higher in good weather than in bad.
The current “spike” in crime amounts merely to a comparison of a mild-weather winter to a couple of harsh-weather winters. That's what the Post did last year, for burglaries, and that's what the Times has done this year, for robberies. The situation calls not for fear, and for hurling lots of extra police at the problem, and Guardian Angels posted at Metro stations, but for the caution that should be routine for city dwellers.