Embrace bad data, don’t shoot the messenger

Since 2006, Trenton has been included in the new American Communities Survey (ACS) which updates important census data yearly rather than every 10 years. The ACS is a boon for economist and policy planners as it provides neutral and consistent data about cities in America.

One of the most important things an organization can do to improve its operations is to measure its success. Yet, measuring city population and income would be prohibitively expensive for any local government, not to mention inherently inconsistent with other localities. Therefore, it is fortunate that the Federal government fills this role for us.

Yet the first words about Trenton’s most recent ACS Census results from Mayor Palmer were to call the results incorrect.

“Mercer County’s growth spurt Report: Population swells in several townships, but not in Trenton”, Trenton, Times, July 31, 2008.

It’s a fair question, to ask how Trenton’s Mayor could possible have done the statistical analysis needed to contradict the US Census Bureau.

Complaints about Census undercounting are as old as the Census.

In 1790, George Washington complained that the first census undercounted people. However, he determined that it was most likely good enough for apportionment (i.e. good enough for government work). A most illuminating paper on undercounting is entitledCoverage Improvement in Census 2000 Enumeration” located on the Census Bureau’s web site, tells the story of how the census methodology has improved over the year’s and especially over the last 20 years in order to better count our population.

In particular the 1990 census was found to have undercounted the urban and minority population causing much uproar, mostly likely leading to the Mayor’s recent dogmatic assumption that the current numbers are wrong. As a result of the 1990 census and the realization that population patterns had once again shifted, the Bureau radically updated to its procedures to specifically target undercounted populations.

The paper lists several changes in methodology including

  • Telephone and In Person follow-ups
  • Double checking housing records with local officials
  • Several marketing programs, and
  • Surveying soup kitchens and other transient living places

As a result of these changes, post-survey follow-up analysis concludes that the 2000 census had either negligible undercounts or slightly over-counted the population.

Over-counting the poor is bad for Trenton

Even if Trentonians believed the city’s population was undercounted, they should keep it to themselves.

The undercounted are typically believed to be poor. Therefore, by over-counting the poor, Trenton’s per capita income would go down. This makes Trenton look especially bad for investment. We’d rather be a small rich city than a big poor one.

Over-counting also leads to increased Federal funding. The type of federal funding we get is generally meant to address the poor. Unfortunately, by spending more on programs for the under-class, Trenton attracts the under-class to the city. This sounds terrible but its true, the more attractive Trenton becomes because of its funding for subsidized housing, drug treatment programs, crime prevention grants and subsidized healthcare the more it becomes a magnet for people needing those services.

The more Federal funding Trenton receives the bigger the problem becomes and the more Federal funding it needs. It’s a vicious cycle that many urban cities have been in for years. The cycle is driven largely by those in the “Under-class” industry who lobby for more Federal funds and use Census data to justify their claims.

The Census Bureau has nothing to gain by presenting inaccurate results

The Census bureau is made up of professional statisticians with reputations. They take their work seriously as evidenced by the written methodologies they provide. They publish the entire methodology on the bureau’s web site. If you were a demographer wouldn’t you want to do a good job on the most important demography project in the word?

Given that the bureau has no reason to make up results and every reason to try and get them right, I chose to trust them. It would take one of the grandest conspiracies of all time to consistently, year and year, undercount cities no matter who’s in the White House.

Groups with something to gain still complain about the Census

Federal funding pays for all sorts of programs supporting the poor and depends on Census data for allocation. In stands to reason that the people who receive their salaries from these funds would complain if the Census didn’t go their way. There is a built in bias amongst those in the business of serving the under-class to want things to appear as bad as possible for any given city.

In addition, the Census governs the structure of legislative districts. Obviously, politicians hoping to preserve or expand their power want the Census to go their way.

One has to take the words of Census naysayer’s with a grain of salt.

After telling a friend involved in Trenton funding about my intention to write an article de-bunking the Mayor’s position, she sent me a quick list of articles supporting the “under-count” theory.

Her list included positions from:

  • A non-profit organization that benefits from more poor people
  • A division of UCLA that works with urban community groups
  • A TV show that is inviting state employees to complain about loss of Federal funding to New Mexico
  • A City of San Francisco funded study meant to convince investors that Census is wrong
  • A New York based poverty advocacy group that doesn’t cite a study, they just make the claim the count was wrong. They are advocating more congressional power.

A list of articles is at the end of this paper[i]

All of these groups have something to gain from adjusting the Census upwards for their areas. The Mayor has something to gain as well. Loss of Federal Funding will expose the dramatic weakness in Trenton’s budget. Federal and State funding prop up Trenton’s ability to function as a city. Non-local funding makes up a whopping 86% of our municipal and school budget.

The Decennial Census and the ACS are great tools

Economists live for the census. How else can they evaluate the experiences of one city with those of another? We have to have consistently measured data.

Much of what I have written over the past six years was informed by an analysis of ACS and census data.

In “A Vision and Plan for Trenton”, I was able to compare changes in per capita income in Jersey City, Newark and Trenton. In “Go Trenton! Beat Clifton!”, I could compare the relative success of Hamilton, Trenton and Clifton. In “The Case against Affordable Housing” I put Trenton’s growth in context with Mercer County, and the entire state.

Without unbiased data from the census bureau, it would be impossible for critics to fairly evaluate the efficacy of local revitalization policy in a city like Trenton.

[i] Articles complaining about census undercounting







My recent Letter to Editor on the Census, Trenton Times August 2, 2008 The Truth is Out There. My letter is at the end.

Share and Enjoy:
  • Print
  • Digg
  • StumbleUpon
  • del.icio.us
  • Facebook
  • Yahoo! Buzz
  • Twitter
  • Google Bookmarks

2 Responses to “Embrace bad data, don’t shoot the messenger”

  • Hard to say whether this is Palmer’s mis-understanding or the Times’s, but the ACS is a sampled product, not a census. Indeed, sampling is just what the critics of the decennial census have always wanted as a supplement. Is he really saying he doesn’t trust the Census bureau’s sampling? Or just that he’s focused on the 2010 census? Very hard to tell from the article, and I’d give him the benefit of the doubt in this case.

  • Maybe I should have made that distinction as well. I felt that the distinction between sampling and census was a bit narrow for the reader in this context.

    That said, of course you are correct. ACS is a sampling that extrapolates Census (i.e. counted) data.

    I might have given the Mayor benefit of the doubt, except this isn’t his first time complaining about surveys (he derides Morgan Quinto every year) or his first time playing fast and loose with statistics (he boasted about average housing prices that were Trenton MSA not Trenton city). The Mayor is welcome to defend his comment in this space.

    The modus operandi seems to be to complain about numbers in order to throw the public off the trail of poor performance. My job as a critic is to pay attention and correct the errors.

Leave a Reply