Chicago has long been a city divided by race and class, a metropolis with starkly different crime rates, economic realities and educational opportunities depending on where you live. But there’s another division in Chicago and Cook County, one that for years has gone unexamined even as it pits rich against poor.
An unprecedented analysis by the Tribune reveals that for years the county’s property tax system created an unequal burden on residents, handing huge financial breaks to homeowners who are well-off while punishing those who have the least, particularly people living in minority communities.
The problem lies with the fundamentally flawed way the county assessor’s office values property.
The valuations are a crucial factor when it comes to determining property tax bills, a burden that for many determines whether they can afford to stay in their homes. Done well, these estimates should be fair, transparent and stand up to scrutiny.
But that’s not how it works in Cook County, where Assessor Joseph Berrios has resisted reforms and ignored industry standards while his office churned out inaccurate values. The result is a staggering pattern of inequality.
From North Lawndale and Little Village to Calumet City and Melrose Park, residents in working-class neighborhoods were more likely to receive property tax bills that assumed their homes were worth more than their true market value, the Tribune found.
Meanwhile, many living in the county’s wealthier and mostly white communities — including Winnetka, Glencoe, Lakeview and the Gold Coast — caught a break because property taxes weren’t based on the full value of their homes.
As a result, people living in poorer areas tended to pay more in taxes as a percentage of their home’s value than residents in more affluent communities. Known as the effective tax rate, the percentage should be roughly the same for everyone living in a single taxing district.
But the Tribune’s analysis shows the rates became skewed in favor of wealthier residents.
The assessor’s office says it does not check its own work for fairness and accuracy, as is standard practice for assessors around the world.
So the Tribune stepped in, compiling and analyzing more than 100 million property tax records from the years 2003 through 2015, along with thousands of pages of documents, then vetting the findings with top experts in the field. The process took more than a year.
The conclusion: Residential assessments have been so far off the mark for so many years that the credibility of the entire property tax system is in doubt.
Berrios, who took office in December 2010, declined to be interviewed for this series. But through an interview in September with top aide Thomas Jaconetty, and then in written statements in 2017, the office defended its assessments as fair and accurate and said it strongly disagreed with the Tribune’s findings.
“The (Cook County assessor’s office) believes the valuation and uniformity opinions formed by the Chicago Tribune are not sufficiently credible,” the office said in a statement.
The latest powerful politician to hold his office, Berrios is chairman of the Cook County Democratic Party and controls three active campaign funds where he’s raised more than $5 million since 2009, more than half of which came from property tax attorneys and the businesses associated with them. Among his strongest allies: Illinois House Speaker Michael Madigan and Ald. Edward Burke, both property tax lawyers who benefit from the county’s broken system.
It isn’t as if Berrios is unaware of the problems with his assessments. In July 2015, he boasted in a news release that his office had adopted new, state-of-the-art computer models designed to improve accuracy and address persistent inequalities.
But the Tribune’s examination of the computer programs officials say they used to value residential properties in 2015 shows Berrios continued to use the old, faulty model.
“My understanding at the time of the press release was that they were going to replace the old model with the new one,” said Robert Weissbourd, who led the team that created the new model and was quoted in the release. “The new model was technically, by every measure, more accurate, fair and transparent.
“If they are not taking advantage of it, that would be deeply troubling. There’s millions of dollars at stake for the county’s poorest residents.”
In Chicago’s West Garfield Park neighborhood, Gail Bates knew there was something wrong with the county’s valuation practices as soon as she opened her first mailing from the assessor.
A caseworker for the Illinois Department of Children and Family Services, Bates purchased her two-flat in 2009 for $119,000. According to state law and county ordinance, her property tax bill should be based on her home’s market value.
Yet that same year, the assessor valued the property at $210,500, and the office continued to overestimate its value for years afterward.
Bates didn’t appeal in 2009 or in 2012, the year of the next triennial reassessment. Finally, in 2015 she filed an appeal and won a 10 percent reduction.
“Why does it have to be so unfair?” asked Bates. “We are the people who can least afford it.”
In an interview last fall, Jaconetty, the county’s deputy assessor for valuation and appeals, dismissed the need for updated assessment models, arguing that homeowners would be relieved to know that the assessor is not solely relying on the “purity of mathematics.”
“Would they be as concerned about their assessments being based purely on math and driven by equations? Or would they feel better knowing there was a human being involved?” he asked.
That philosophy helps explain why the county appears to rely so heavily on “hand checks,” a process in which property values are adjusted manually on a case-by-case basis.
Because the office doesn’t keep statistics on hand checks, it’s impossible to tell exactly how many are done each year. When the Tribune filed an open-records request for documents detailing how the practice works, Berrios refused. The newspaper filed suit, seeking to obtain the records.
The assessor’s office also does not use computer-based mapping programs that are a common feature in modern assessment systems and relies on an outdated mainframe computer from the early 1990s as it calculates estimated values for 1.8 million properties in the county. Officials say the county is updating its computer system, but the process is expected to take years.
Jaconetty stressed that residents who believe their property is overvalued have the option to appeal — and are encouraged to do so.
But when the Tribune partnered with the University of Chicago’s Harris School of Public Policy to study appeals filed by homeowners, the paper found that the process makes an already unequal system even less fair. That’s because owners of high-priced homes are far more likely to appeal.
In defending the office, Jaconetty also noted that under Berrios the county has sent out property tax bills on time, which hadn’t been the case for about 35 years. Because of this, officials say, local governments saved millions in borrowing costs for loans used to tide them over until property taxes come in.
But that timeliness came as accuracy suffered.
Since 2009, the Tribune found, Cook County’s assessments have been so inaccurate they violated standards set by the International Association of Assessing Officers, a professional organization that develops guidelines used around the world.
Valuing property in the wake of the 2008 housing market collapse was no simple task. But even in a time of unprecedented upheaval, the office says it did not run the studies that assessors carry out to measure the accuracy and fairness of their valuations.
“That’s insane,” said Peter Davis, an assessing expert who helped write the standards for the International Association of Assessing Officers. “There’s no justification for not doing the studies. That’s how you find problems.”
Widowed and living alone in a four-bedroom home near Melrose Park, Barbara Garner decided in 2010 to downsize.
She swapped her 2,000-square-foot, two-story house — one of the biggest in the area — for a nearby single-story home built just after World War II. Her new place is smaller than 800 square feet.
So Garner was shocked when she learned her new tax bill would be nearly the same amount she’d been paying for her old home, more than $4,000 a year.
“I blew a gasket,” she said. “I moved here to save money but instead I was paying the same amount in taxes.”
Garner’s tax bill was higher than she expected because the assessor overshot the price she paid for the house by a factor of two, the Tribune found. The county had valued the home at $164,640, just months before she bought it for $75,000.
Up in Chicago’s Lakeview neighborhood, meanwhile, the assessor estimated the value of a 2,600-square-foot house at $787,000 in 2012. That is just over half the price a couple paid for it the same year. Five years later, the valuation is still 30 percent short of the $1.4 million sale price, even though home values in the neighborhood have soared.
Accurately estimating a property’s market value is crucial because that is the first number that goes into calculating property taxes. If the estimate is faulty, the fairness of the entire bill is thrown into doubt.
To determine if assessments are fair and accurate, assessors usually conduct sales ratio studies, which compare actual sale prices to their estimates. Cook County says it doesn’t do the studies, relying instead on research conducted by the Illinois Department of Revenue. But that research comes out years after property taxes are calculated and doesn’t provide granular details on individual neighborhoods.
The Tribune checked the assessor’s accuracy over 13 years, including at the neighborhood level, by obtaining data on more than 1.4 million sales since 2003 and matching them to assessment records. Three assessment experts, including a former executive director of the International Association of Assessing Officers, vetted the Tribune’s analysis.
In its written statement, the assessor’s office contended that any privately conducted sales ratio study is invalid because it would never be admitted into a court of law. In fact, such studies have been accepted as evidence in court cases around the country for decades.
No mass appraisal ever produces exact results, so assessing experts use statistical tests to ensure that any errors fall within reasonable limits.
The most common measure of accuracy is called the coefficient of dispersion, or COD. Assessment experts say the highest acceptable score is 15, which basically means the average error rate was 15 percent.
Prior to 2009, Cook County had scores of around 15, the Tribune’s analysis found. Since then, however, the scores have been as high as 31 in some townships, meaning assessments were deeply unfair.
In Chicago, scores for the city’s eight townships ranged from 16 to 31 for the 2012 reassessment. The scores for the 2010 reassessment in the north suburbs and the 2011 reassessment in the south and west suburbs also exceeded the standard of 15, sometimes by a wide margin.
“If you get CODs over 20, or certainly 30, it’s professionally unacceptable,” said Richard Almy, a former executive director of the International Association of Assessing Officers. “It calls into question the credibility of the whole tax system.”
A high score means the system is riddled with errors, but it doesn’t say whether the errors are skewed in a way that favors certain groups. That’s where two other tests — price-related differential and price-related bias — come in.
The tests can help determine whether the property tax system suffers from regressivity: the overvaluing of low-priced homes and undervaluing of high-priced ones.
The Tribune found regressivity problems exploded in Cook County beginning in 2009, when regressivity scores exceeded industry standards for most of Chicago’s townships. In 2012, the scores got even worse.
Comparing assessed values to actual sales also reveals inequities in the system. In the township covering Chicago’s Northwest Side, for example, the 2012 median assessment level was 10 percent higher than what is called for in county ordinances. The township that includes Lincoln Park, the Gold Coast and downtown, meanwhile, was nearly 20 percent lower than it should have been.
Areas of the city where gentrification had taken hold saw values come in low, while homes just outside those neighborhoods were more likely to be overvalued. Small apartment buildings in the trendy eastern parts of Humboldt Park and Logan Square were likely to catch a break while many on the west sides of the neighborhoods got hammered.
Bungalows in the far south suburbs of Chicago Heights, Lynwood and Ford Heights were far more likely to be overvalued while luxury homes in Wilmette and Winnetka were undervalued, some by as much as half.
The 2015 reassessment for Chicago was slightly better than in 2012, as regressivity for the entire city fell to acceptable levels. But the scores still failed to meet professional standards in many townships, the Tribune found, and the county as a whole continued to suffer from severe regressivity.
Moreover, if condominiums are excluded from the analysis, error rates and regressivity violated standards across the board in 2015. The assessor’s office historically has been better at valuing condos, the Tribune found. One former official said in a court deposition that the value of a condo was estimated by tracking sales in the same building — a different method than the flawed computer models used for single-family homes and small apartment buildings.
The pattern of regressivity in Cook County meant that the tax burden — the fixed amount of money that the county is going to collect — was being distributed inequitably as well.
By comparing sale prices to property tax bills, the Tribune and the University of Chicago calculated the effective tax rate for Chicago and found it differed widely by neighborhood between 2009 and 2015, even though the rate should have been roughly the same for everyone.
For example, the average effective tax rate in North Lawndale and Little Village was around 2 percent – about two times higher than in wealthier areas like the Gold Coast and Lincoln Park.
The fix that wasn’t
The 2009 reassessment for the city of Chicago was so problematic that it was clear something needed to be done.
The assessor’s office had overvalued some lower-priced homes by as much as 150 percent, the Tribune’s analysis shows. And complaints were pouring in, especially from neighborhoods where property values had collapsed in the aftermath of the housing market crash.
James Houlihan, the assessor at the time, turned to outsiders for some help.
The MacArthur Foundation agreed to fund a grant for another nonprofit group, LISC, to examine how foreclosures had affected housing prices. RW Ventures, a consulting firm, was tapped by the nonprofit to oversee the project, and Weissbourd, president of the firm, drafted U. of C. public policy professor Christopher Berry to lead the technical aspects.
Berry and RW Ventures’ chief statistician, Michael He, discovered high rates of regressivity in the system and delivered the findings to Houlihan, along with a proposal to attempt to fix the problem.
“I was struck by the level of regressivity,” said Berry. “It wasn’t something I came in expecting to see. We thought maybe we could do something about it.”
The following year, in 2010, MacArthur provided another grant to LISC so it could hire RW Ventures and Berry to develop a new computer model for the county that would accurately value homes and cut down on regressivity.
Regressivity is an inherent problem in mass-appraisal models because they rely on averages. The models take certain characteristics of homes that sold and figure out average values for each one — the price per bathroom, per square foot and so on. Those values are then used to estimate market prices for everyone.
But there’s a lot of information the models don’t capture, such as remodeling, and that can affect the market value. So, one house with two bedrooms and two baths might sell for $300,000 and another with the same characteristics might go for $100,000. The mass appraisal models would average the two, and the resulting value of $200,000 would be too low for one house and too high for the other.
The housing market crash exacerbated the regressivity problem because the models rely on the previous five years of sales. Sales prior to the crash had been included in the models, driving up the assessor’s estimates for many properties, especially in poorer areas where values had collapsed.
After months of testing, Berry and the RW Ventures team settled on a novel approach: Divide properties used in the models into segments based on sale price — high, medium or low — and compute values separately. That way, two houses with similar characteristics but wildly different sale prices wouldn’t skew the estimates for all homes.
“We really didn’t know when we first got into it how successful we would be,” said Berry. “That’s why we tried so many different kinds of models.”
By 2010, Berry’s team was done with refinements and confident in what it had developed. Studies conducted by the team found that its new model outperformed the assessor’s old model on every measure.
One benefit was that the new model greatly decreased the number of “hand checks” that would be required to fix obvious mistakes. Because no model is perfect, assessors use hand checks to address outliers, values that are clearly wrong. Experts in assessments say hand checks should be kept to a minimum and any changes should be thoroughly documented.
Before the new model could be implemented, however, Houlihan announced his retirement. Berrios became the new assessor in December 2010.
For more than two decades, Berrios had been a member of the Cook County Board of Review, an elected three-member panel that fields assessment appeals. Weissbourd said that after Berrios assumed office, Weissbourd urged him to deploy the new model.
Instead, Berrios waited.
Nearly a year later, in late 2011, MacArthur released another round of grant money to develop a manual and training program for the assessor’s office. But not until July 2013 — 2½ years after the new model had been built — did Berrios allow Weissbourd’s team to begin showing staff how to use it.
Another two years passed before Berrios issued a news release in July 2015 stating that he had “implemented a new state-of-the-art residential assessment modeling technique that assesses the value of homes in different price ranges to improve accuracy.”
Among other things, the release stated that the new models improved accuracy by 50 percent and cut down on regressivity by 25 percent. “The technique was developed to address ‘regressivity,’ which results in higher-priced homes being under-assessed,” the news release said.
“We created new assessment models ensuring the fairest and most accurate assessment possible,” Berrios was quoted as saying. “Good government is fair and its information clear.”
MacArthur Foundation President Julia Stasch provided a quote for Berrios’ news release that praised him for his “leadership and persistence in demanding a more sophisticated assessment model.”
But the new model was never fully implemented. Stasch and the public never knew.
In the dark
The creators of the MacArthur-funded computer program also were left in the dark.
After analyzing the 2015 reassessment for city properties, the Tribune found that the values weren’t as accurate or fair as would be expected if Berrios had used the MacArthur-funded model. So the newspaper contacted Weissbourd and Berry in the spring of 2016.
Both said they were under the impression that Berrios was using their work.
“We spent weeks training staff,” Berry said. “It just seemed unimaginable that they wouldn’t have implemented the model.”
Making requests under the Freedom of Information Act, the Tribune obtained the models that the assessor’s office said it used for the 2015 reassessment. A thorough review of the computer code found few differences between those models and the ones used previously.
So the newspaper asked the assessor’s office specifically for the new model, as well as any results the office had produced from it.
The Tribune found that just 2 percent of those results matched the market values that the office sent to taxpayers.
Informed of the Tribune’s findings, Stasch of the MacArthur Foundation said in a statement that she had expected that the new model would replace the old one.
“Increasing accuracy and equity of property tax assessments would benefit hundreds of thousands of local homeowners,” she said, “particularly in low- and middle-income neighborhoods.”
In the September 2016 interview, Jaconetty told the Tribune the office used both models but would not say how it reconciled the two. He also could not identify any other jurisdiction in the country that uses two models or cite any studies conducted by the office that justify the practice.
“This is not a situation where two models work better than one,” said Weissbourd. “It’s not a matter of opinion. If you have a thermometer that’s accurate to 5 degrees and someone comes along with one that’s accurate to 1 degree, there’s no reason to use both thermometers.”
Pressed further by the Tribune in that interview, Jaconetty provided a different explanation. He acknowledged that the office did not rely much on the new model — it has “issues,” he said — and instead used the old model as a baseline.
But only a small fraction of the estimated values sent to taxpayers in 2015 matched the results from the old model, raising further questions about how the office arrived at its valuations that year.
Jaconetty also tried to discredit Weissbourd and Berry, saying they have a “vested interest” in promoting their model.
In fact, the assessor’s office owns the new model, and since turning it over seven years ago Weissbourd and Berry say they haven’t done any work in the assessment field. (Berry, a tenured professor at U. of C.’s Harris School of Public Policy, teamed with the Tribune late last year to examine regressivity and the impact of appeals on assessments as part of a graduate-level course.)
“If they are not using our model and are continuing to use the old model, then they missed an opportunity to reduce regressivity, which would have helped people who own the lowest-priced homes in the county,” said Berry. “That would be inexcusable.
“It’s not like we’re saying ours is perfect. There is a lot more work to be done. But our model was a major improvement.”
This month, the assessor’s office contended in a written statement that it had told Weissbourd and Berry as early as 2010 that “sole and exclusive use” of the new model was “grossly unreliable” and “creates distortions in uniformity.”
Weissbourd said that assertion was “a one hundred percent fabrication and really disturbing.”
“There was never any indication they were unhappy with the new model — or not using it — until the Tribune informed us of their comments, years later,” he said.
Said Berry: “Our model substantially improved uniformity, and every other industry-standard metric, and we gave them the analysis to show it.”
The assessor’s 2015 news release about the new model did not mention any problems, instead touting improvements in accuracy and fairness. “The assessed values produced by the model are more reflective of what the homes’ real market values are or what they would sell for,” it stated.
Berry also said he pointed out to the assessor’s office that the new model’s more accurate valuations would have boosted the market value of some homes dramatically — presenting a potential political issue for Berrios.
“I suggested that they may want to phase in the new model, since it would significantly increase assessments for high-priced homes,” Berry said. “But no one ever followed up with me about it.”
How the county assessor calculated its 2015 valuations remains unclear, even after the Tribune examined the two models available to the assessor’s office. Very few of the dollar figures sent to taxpayers that year align with either model.
Were adjustments made through hand checks or some other way? What methodology was used?
The assessor’s office would not say.
In December, Cook County Associate Judge Neil H. Cohen ordered the assessor’s office to provide the Tribune with documents related to its methods for valuing property, including hand checks. The newspaper had filed suit under the Freedom of Information Act after the assessor’s office denied its request, contending that the public wasn’t allowed to see how its process works.
“That’s not what FOIA was meant for, to hide things from the citizens,” Cohen said in court. “FOIA was meant to disclose and have total transparency of its elected officials who work for the people.”
Berrios is appealing the ruling — at taxpayer expense. So the documents remain secret.
‘I just can’t keep up’
Whether assessments are fair isn’t just a technical matter. The issue directly affects the financial futures of people whose homes are their biggest investments and whose taxes are a source of constant anxiety.
In July 2010, Kenyetta Braxton-Williams scraped together enough money to make a down payment on a modest $82,000 yellow brick home in Calumet City.
Braxton-Williams, who earns $12 an hour working at a homeless shelter her grandmother founded on the South Side, said her grandmother was adamant that she buy a home and get settled.
“The word she used was stable,” said Braxton-Williams. “She just wanted the best for me and my daughter.”
At first, Braxton-Williams said her payments — including mortgage and taxes — were $600 per month, doable if the family watched expenses.
Then she received her tax bill notice in 2011, stating that the assessor’s office had valued her home at $147,550. That was almost double what she had paid for it.
“I love my house, but I know it’s not worth that much. And they know it’s not worth that much,” she said. “In the six years I have lived here my payments have gone from $600 a month, to $800, to $1,200 and now they are $982. I just can’t keep up.”
In a written statement, the assessor’s office said its assessment was fair and argued that the 2010 sale was “plainly not an arms-length, fair-cash-value transaction” because the home was purchased from the trust of a deceased person. However, the assessor’s own data show that the office treated it as standard transaction, even labeling it a “pure market” sale.
In 2012, Braxton-Williams decided to appeal her assessment and by extension her $4,776 annual tax bill.
County officials agreed to lower the market value on her home to $93,630, still nearly 15 percent higher than the amount she paid for it.
Since then, officials have increased the estimated market value of the home to $98,560.
“When I bought this house I had no idea what I was getting into,” she said. “My credit is shot. I’m $17,000 in arrears, and now they keep threatening to take my home. I love my home, and it’s not like I have anyplace else to go.”
Braxton-Williams said her struggle to pay her tax bills and mortgage has given her a new sense of empathy in her job as intake coordinator at the homeless shelter, A Little Bit of Heaven.
“There have been many times,” she said, “when I thought I might be in that same situation.”
John Chase and David Kidwell contributed.