ALBUQUERQUE — Along with sponsoring a bill to prevent a reduction in the amount of Medicaid funding New Mexico gets, U.S. Sen. Jeff Bingaman, D-N.M., and other elected officials might also want to make sure New Mexico’s population is counted correctly during the 2010 U.S. Census.

 

A Price Waterhouse study conducted in 2001 predicted that New Mexico would lose over a $100 million dollars in federal funding between 2000 and 2012 due to being undercounted in the 2000 decennial census. A note on the population webpage of the University of New Mexico’s Bureau of Business and Economic Research, or BBER, gives the population discrepancy over time:

 

The July 1, 2006 population estimate for New Mexico, published by the Census Bureau (2007 round of estimates) is 1,942,302. This represents an increase of 6.8% since April 1, 2000. Bureau of Business and Economic Research (BBER) analysis indicates that this growth rate is low in light of other economic and demographic indicators for the state. The BBER estimate of 2,010,570 for July 1, 2006 shows a 10.5 percent increase from April 1, 2000 to July 1, 2006.

 

New Mexico isn’t alone, under-counting by the U.S. Census bureau is chronic throughout the country. Not only do they use a system that doesn’t account for certain populations, they also do what’s called a “rake” to make numbers coming from the local level match up to the national number.

 

So why are the numbers so different? The folks at BBER explained that it was a matter of methodology.

 

The U.S. Census Bureau conducts a decennial census every ten years, in which they actually count people. Most of us have received census forms in the mail and returned them with information about ourselves and those who live in our house. The bureau then uses what is called a “demographic cohort component” between each 10-year census to estimate population changes annually, using the count as a base number. They look at birth and death records, as well as migration to and from the state. Tracking migration is the trickiest part. They use IRS tax return data as well as data from federal programs such as Medicare to track the movement of people between states. These methods are not fail-safe, and lead to undercounts for various reasons. They miss certain communities entirely, such as those that live off the grid without addresses. In New Mexico, there are a number of these types of communities, some of them known as colonias. Also left out of the estimations are people without social security numbers, or young people who move into the state for college but may still be listed as dependents on their parents tax forms.

 

BBER Senior Demographer Dely Alcantara says the Bureau also “controls” the local numbers to correspond to the national estimate. This is called a “rake,” according to Alcantara:

 

“They estimate the U.S. total first, and then do the county estimate. They then compare the county figures to the national. If they don’t match, they do a “rake” so that all the county numbers added up will be consistent to that national figure. The Rake is based on a ratio and it is applied across the board.”

 

BBER uses the component method to inform their demographic estimates, but also uses a “housing unit” approach to capture how housing stock changes over time. They interview self-permitting agencies throughout the state and gather construction industry data. The data collected is added to the cumulative housing stock they’ve been tracking since the Census 2000 count. BBER Demographer Jack Baker said that the Census Bureau’s numbers support the idea that a housing unit method provides a more accurate estimate because if you took the Bureau’s own housing unit data and made a very conservative estimate, the numbers would be higher than the official component method numbers. And, he said, its closer to what BBER produces.

 

The following table shows the under-estimate over the past six years:

 

Census Bureau and BBER: New Mexico Population Estimates, 2000-2006

Fiscal Year (Begins July 1)

Census Bureau

BBER

Under-Estimate

2000 Count

1,819,046

1,819,046

 

2001

1,832,608

1,852,740

20,132

2002

1,855,400

1,876,287

20,887

2003

1,879,252

1,899,846

20,594

2004

1,903,006

1,929,713

26,707

2005

1,928,384

1,968,352

39,968

2006

1,942,302

2,010,570

68,268

Percentage Change

6.8%

10.5%

 

 

Source: New Mexico Business Current Economic Report, BBER, February 2007

 

 

According to Alcantara, for the Census Bureau’s population numbers to be accurate, New Mexico would be way over-built:

 

“Our contention is that we are being under-estimated year in and year out. If we look at occupancy rates and houses, our numbers are higher than census numbers. We believe the input data and the methodology they use underestimates consistently.”

 

Baker put it more plainly:

 

“Their estimates don’t jive with what we know about the housing stock, too many houses would have to be empty. They’re raking us.”

 

The implications of being underestimated are big, and varied.

 

Population estimates are important for redistricting political maps, as well as for calculating the distribution of federal dollars to the states. Population figures are used to calculate per capita income, which is measured by dividing total income by total population. An incorrect, smaller population number would make New Mexico’s official per capita income figure higher than it really is.

 

Yes, this means we’re probably even poorer than we thought.

 

Getting the per capita income figure wrong has enormous funding consequences for New Mexico’s low-income families.

 

Medicaid is at the top of the list of federal programs that develop their matching formulas for distributing federal dollars to the states based on per capita income. And New Mexico, one of the poorest states in the nation, relies heavily on Medicaid to provide health care to its low-income population. In 2007, New Mexico received $2.56 from the federal government for every one dollar the state spent on Medicaid. That’s 71.9% of the cost of Medicaid in New Mexico, which was $2,490,147,301 in 2006. But the numbers show the percentage has been declining over the past few years:

 

 

Percentage of New Mexico Medicaid Services funded by Federal Dollars, FY 2004-2007

 

 

FY 2004

FY 2005

FY 2006

FY 2007

Federal funds spent for each dollar spent in state Medicaid spending, FY07

New Mexico

74.9%

74.3%

71.2%

71.9%

$2.56

 

Source: Medicaid: A Primer, by the Kaiser Commission on Medicaid and the Uninsured, March 2007

 

 

Other programs affected are Foster Care, Rehabilitation Services Basic Support, Child Care and Development Block Grant and Social Services Block Grant programs, Substance Abuse Prevention and Treatment Block Grant, Adoption Assistance, and Vocational Education Basic Grant.

 

Baker and Alcantara both said that the Bureau provides a process for challenging population estimates at the county level, but it’s rarely used. Alcantara explained that since counties aren’t direct beneficiaries of the federal funds flowing to the state, they might not realize the direct implications. Instead, much of the programs’ benefits are distributed to individuals by the state. But, according to Alcantara:

 

While the direct ramifications are at the state level, it trickles down. Block grants are population driven, anything in fact that is population driven is affected. Clearly counties should be concerned but don’t always see the direct implications of census undercounts.

 

So what can New Mexico do?

 

In 2005, the state of New Mexico appropriated funds to BBER’s demography division, which allowed them to expand their team of demographers. According to BBER’s Kevin Kargacin this enabled BBER to launch a full population estimates program in which they were able to conduct intercensal population estimates at the state and county level. They’ll soon be taking their analysis deeper, to the municipal level. The hope is that a more detailed analysis of New Mexico’s population will influence the Census Bureau estimates. The federal government is also just now finalizing an address mapping project that allowed state teams to add to and correct the master address files used for the census count. This is in preparation for the next decennial count, less than two years away.

 

Kargacin noted that the census count conducted once every ten years is only as successful as the civic participation of citizens themselves. Turning in the form is a "civic duty" required by law. He said it’ll be most successful if local governments plus organizations of all types advertise the census and urge their constituencies to return their forms. The bureau, he said, welcomes all the help it can get. Otherwise, New Mexico will probably see another decade of well over a $100 million in lost revenue.