Kyle Thompson, pictured in this 2010 file photo, is past president of the Oread Neighborhood Association and has seen the decline of owner occupants in the area over the 20 years he has lived there. Empty homes in neighborhoods around town could help unravel the mystery the 5,084 additional people the city says should be here, but the U.S. Census Bureau did not find during the 2010 census.

Local real estate agent John Esau can stand in the driveway of a nice single-family home that he has for sale in Lawrence’s Deerfield neighborhood and see three homes that are sitting completely empty. Sure, the emptiness is a sign of an economic downturn that has been vicious to the housing market. But maybe the view from Deerfield, and a host of other fine Lawrence neighborhoods, also is focusing on something else. Maybe it is a clue — a clue to a mystery that has been lurking in Lawrence like a butler with a guilty conscience or Col. Mustard with a candlestick in the library.

Yeah, Lawrence has a mystery on its hands. For the moment, file it under the category of a missing-persons case — 5,084 missing persons.

The U.S. Census Bureau earlier this year released its finding from its once-per-decade count of Lawrence’s population. The Census Bureau found 87,643 people living in Lawrence as of April 1, 2010. As previously reported, that caught Lawrence City Hall officials by surprise. City planners had estimated that there would be at least 92,727 people living in the city.

That’s a difference of more than 5 percent. Or another way to put it is, with one report, the Census Bureau knocked Lawrence back to 2004, which is when city planners had thought the city’s population topped the 87,000 mark. The Census finding officially marked the decade of the 2000s as Lawrence’s slowest decade for growth since the days of the Great Depression.

If the Census Bureau is right, that is. City officials have a strong suspicion that the Census Bureau simply didn’t count all the housing units that exist in Lawrence.

“We’re very certain that is the case,” said Amy Miller, a city-county planner who oversees Census-related data for Lawrence and Douglas County. “We have a very good handle on our building permits and our housing numbers.”

But thinking it and proving it are two different things. And as the Census Bureau releases more detailed data from its count, little pieces of evidence pop up that raise the question of whether the Census Bureau might be right.

One piece is a number that may surprise some Lawrence residents. The Census Bureau found that in 2010, there were 2,532 living units, either a house or an apartment, that were empty.

That’s about 6.8 percent of all of Lawrence’s housing stock, which is actually far below the 11.4 percent national average, but when you spell it out as 2,500 units, it is eye-opening. But not unrealistic, some say.

“That number wouldn’t surprise me at all,” said Esau, who is an agent with Lawrence’s Keller Williams agency. “One of the things that is disturbing is you can drive around neighborhoods and find several houses that are in preforeclosure. They’re empty, but they’re not on the market yet, and they don’t really show up as empty to a lot of people yet.”

So, are we missing 5,084 people? Or were they ever really here to begin with?

Indeed we have a mystery. Come on, Watson, let’s go take a look around.

The Victims

• The Pocketbook: At this scene, there are big-dollar stakes involved. The Census is used once a decade to determine a community’s representation in the U.S. Congress, but it is used much more frequently to allocate federal funds.

Local Census organizers once even calculated that in a single year — 2008 — Douglas County governments, residents or businesses received $42.2 million from programs that rely on a formula that uses the Census population number as a key component for distributing funds. Those programs included Medicaid, highway funding, food stamp programs, school lunch grants, housing rehabilitation programs and various other social service programs.

The Brookings Institution, a nonprofit that studies Census data, has estimated that in Kansas, every person who goes uncounted in the Census equates to a loss of about $1,100 per year in federal funding. Based on the Lawrence difference of 5,084 people, that’s $5.5 million in funding in a single year or $55 million for the decade.

• The Inner City: Nowhere did the city take more of a beating from the Census than in its core neighborhoods. The Census has released data that shows areas of town that either gained or lost population since 2000. The areas don’t exactly correspond to neighborhood lines, but the figures show a trend of neighborhoods near the center of town losing population while areas on the edge of the city grew in numbers. Kansas University’s campus saw the largest decline in residents — about 1,800 people. But there were other areas, too. An area that includes the Pinckney neighborhood lost nearly 400; an area that includes the University Heights and Centennial neighborhoods lost about 350; an area that includes parts of Hillcrest and Old West Lawrence lost 190; and even an area that runs along the edge of KU’s West Campus lost about 150 people.

If the numbers are right, they’re a big deal. Cities across the country worry, and often spend millions, to reduce a trend of inner-city decay.

“If that is the trend, it would be very significant for us,” Miller, the city’s long-range planner, said. “But I don’t think that is the trend. I’m not buying it yet.”

• The Future: The Census numbers will become the foundation for the next 10 years worth of population estimates made by city planners. Population estimates factor into a host of decisions ranging from when and where to build new schools to decisions about new roads and transit options. Unfortunately for the city, this bit of uncertainty regarding population numbers comes at a time when the city will need to make one of its most expensive decisions ever. The city must decide when to build a new sewer treatment plant, which will likely cost $90 million to $100 million to construct. One of the key factors in the timing of that decision will be when the city thinks it is nearing the 100,000 population mark.

The Suspects

• The Calculator: You’ve heard of a crime of passion. But have you ever heard of a crime of math? Perhaps that’s what we have on our hands here. The city uses a detailed formula to make its annual population estimates. In simple terms, it looks at the number of new housing units constructed, subtracts the number of housing units demolished, multiples the housing units by the average household size and also applies the most current Census estimate of the city’s vacancy rate to all the newly constructed units. Then the city takes that number and adds it to the previous year’s population estimate.

But there is something the city doesn’t do — it does not apply the most recent vacancy rate to all the existing living units in the city. In other words, there were 32,761 living units in 2000, and there were 37,502 living units in 2010. So the 4,741 new units had new, updated vacancy rates applied to them. But the 32,761 existing units from 2000 were calculated using a vacancy rate that was a decade old.

Both a Census Bureau official and a Census expert at KU said they would want to study the city’s formula in more detail but had concerns about its appearance.

“The vacancy rate should be applied to all the units,” said Greg Harper, a demographer for the Census Bureau. “It is a vacancy rate for all the units in a community, not just the new units.”

The city, however, believes its formula is based on guidelines from the Census. Planners also have done some rough calculations that suggest the vacancy rate changes over the last decade wouldn’t account for the 5,000-person discrepancy.

The Journal-World wasn’t able to completely replicate the city’s estimation process, but it did run some calculations that suggested the city’s method could produce a population number that is inflated by at least 3 percent. The city and the Census Bureau’s numbers differ by a little less than 6 percent.

• College Students: The U.S. Census is designed to count students who live in Lawrence while they attend Kansas University as Lawrence residents. But Kansas is unique for having its own state census, which redistributes those students back to their original counties. For the first time this year, KU required students to fill out the state census form at the time of their enrollment, said Xan Wedel, an information specialist at KU’s Institute for Policy and Social Research and a past chair of the National State Data Center Steering Committee for the Census Bureau. Wedel said that new wrinkle is worth investigating.

“That may have created some confusion with students,” Wedel said. “They may have thought they already filled out their federal form when they really didn’t.”

The Census Bureau does do door-to-door checks on addresses that they don’t receive a form from, but those checks don’t begin until mid-May when many students have left for the summer.

• The List Maker: The Census is more than an exercise in counting. It also is an exercise in lists. The Census Bureau sends out forms to every known living unit in a community. But first it must have a list of those living units. The Census encourages cities to participate in a program that keeps their lists up to date, and Lawrence did participate, Miller said.

But Miller said she thinks somehow the Census Bureau did not count all the residences on the city’s list. The city believes that there were 38,884 residences in the city in 2010. Yet the Census only counted 37,502. That’s a difference of 1,382 units, which if all of them were occupied by the average household size of 2.28, would get the city closer to its population estimate but would still leave it short by a couple of thousand people.

But Miller said proving that the Census Bureau missed some residences is the most likely way the city will get the bureau to change the official number. Miller currently is double-checking living-unit data for several neighborhoods and plans to submit a report to city commissioners later this month.

“Right now, it is a huge difference we’re talking about,” Miller said. “We definitely need to get this resolved.”

In the meantime, we still have a mystery. Watson, get me my pipe.