NEW YORK — As concern about economic inequality rises to the top of the issue agenda, it is instructive to note that the upturn in poverty of recent years has not been accompanied by a rise in violent crime. To the contrary, since 2008, unemployment and homicides have been inversely related.

Is this a puzzling anomaly? Most people assume that hard times cause crime spikes. They reason, plausibly enough, that financial pressures – as a consequence of, say, becoming unemployed – lead to stress, anger and violence.

Many criminologists agree. The “strain theory” of crime holds that high or rising unemployment and poverty rates may be indicators of increasingly unequal opportunities, and that periods of sharply unequal opportunity are likely to produce more crime.

But U.S. history does not support economic explanations for the rise or fall of violent crime. Such crime (defined as murder, rape, robbery and aggravated assault) has stayed high or risen during boom periods, as it did in the 1920s and the 1960s. And it has continued low or declined during recessions, as it did in the 1890s and the 1930s. It has also done the reverse: rising or staying high during recessions, as in the 1970s, for example, and declining or remaining low in good times, as happened in the 1950s.

The Depression of the 1930s illustrates the inconsistent relationship between economics and crime. This was a time of unprecedented economic catastrophe for the United States.

By the early ’30s, more than 3,600 banks had gone under, more than 40 percent of home mortgages were in default and 1,300 local governments had failed to make obligatory payments. National income had fallen to 54 percent of 1929 levels. By 1932, an estimated 28 percent of the nation’s households (34 million people) did not have a single employed wage earner.

With all this misery, did violent crime rise? At first, yes. In 1933, the nationwide homicide mortality rate hit a high for the century up to that point of 9.7 per 100,000 people. In fact, the homicide rate exceeded 9.0 per 100,000 in every year from 1930 to 1934. Only in the 1970s would we again begin seeing rates that alarming.

On the other hand, as the economic recovery proceeded from 1934 to 1937, the homicide rate declined by 20 percent. Prohibition ended in 1933, and immigration, along with the great black migration to Northern cities, plunged dramatically at about the same time, driving down urban crime.

But whatever the reasons, crime continued to fall. The economy stalled again in 1937 and 1938, but homicide rates kept going down, reaching 6.4 per 100,000 by the end of the decade. In fact, the mid-’30s launched a multi-decade crime trough that ran, despite occasional spikes, right up to the 1960s.

Lest anyone think that reductions in economic inequality triggered the crime downturn, the following staggering figure should be kept in mind: Even as crime plunged in the late 1930s, the percentage of Americans at or below the income poverty line hit an astonishing 68 percent.

While the 1930s provide one of the most dramatic illustrations of the unpredictable relationship between crime and the economy, other eras are also revealing. The 1950s economic boom, for instance, was accompanied by some of the lowest crime rates of the 20th century. But even though the prosperity continued and even accelerated into the next decade, the 1960s produced a great crime tsunami that lasted more than 25 years.

What explains this puzzling lack of consistency? The answer seems to be that crimes of violence are unique; unlike property crimes, they are not economically motivated. Murder and its junior partner, assault, for instance, are mainly precipitated by such things as anger, sexual jealousy, perceived insults or threats, and long-standing personal quarrels, and they are frequently facilitated by disinhibitors such as alcohol or drugs. These factors, it appears, are unaffected by economic downturns or upswings.

Robbery, however, is different, as it is a mix of theft and violence. But the relationship between robbery and economic trends is complex. On the one hand, economic declines create incentives to steal, but on the other hand, so do economic booms, which tend to produce more valuable goods.

The lesson of all this? Social analysis requires long time horizons. Examining patterns over one, five or even 50 years may be terribly misleading.

— McClatchy-Tribune Information Services