Mississippi and other parts of the deep South have long held the unwanted distinction of home to America’s deepest poverty. Footage and newspaper photos of New Orleans during Hurricane Katrina captured Southern state deprivation for all to see.  Even after decades of progress, Mississippi’s income per capita is still the country’s lowest—fully 23% below the national average.  But now the Census Bureau’s new improved Supplemental Poverty Measure (SPM) shows that the poverty rate is lower in Mississippi than in much richer New York and California.

More specifically, under the official poverty rate, 23.2% of Mississippi’s population was poor in 2009, when 15.9% of New Yorkers and 15.5% of Californians were.  But, by the new SPM, the poverty rate falls to 17% in Mississippi and jumps to 17.6% in New York and to 22.4% in California.  And keep in mind here that New York’s income per capita is more than 1.5 times Mississippi’s.

So, what’s going on?  Is there really less deprivation in Mississippi than in New York or California?   This and similar reversals of fortune in other states reflect differences in housing costs.  Since spending on housing can eat up as much as half of all household income for those at the poverty threshold, geographic adjustments for housing costs—in the Census Bureau’s calculations, rent for a two-bedroom apartment—can dramatically change both that threshold and poverty rates.

A problem with this approach is that housing quality isn’t taken into account.  A two-bedroom apartment in New York or California might be a lot sturdier and more attractive than one in Mississippi.  And it might be located in dicier or more isolated neighborhoods with fewer amenities. Sure, housing costs less (and thus drives poverty rates down under the new measure) in rural than in metropolitan areas.  That means that some rent differences can reflect differences in living standards, not simply price variations.

To put the SPM measure on the stand for a minute, let’s look at how the official and SPM measures relate to material hardship.  Census Bureau economist Trudi Renwick reports that the official measure is more highly correlated with food insecurity, health, and education while the SPM measure is more reflective of such shelter issues as rent burdens, crowding, homelessness, and foreclosures.

To further complicate this comparison, housing costs and poverty rates can vary within states too. In fact, housing prices vary at least as much within  as between states.   The SPM factors these within- state housing cost  adjustments into poverty rates, but, for example, low rents in outer-ring suburbs can be offset by less access to cultural opportunities and high transport costs or unmanageably long commutes.  Again, in housing you probably don’t get what you don’t pay for.

The SPM’s geographic adjustments for poverty may illustrate a variant of H. L. Mencken’s dictum that “for every complex problem there is [a metric] that is clear, simple, and wrong.” At any rate, trying to summarize poverty in one number is at best awkward and, however resistant to sound bites and quick takes, multidimensional indicators probably serve us better.