This story was produced by ProPublica Illinois. Learn more.

With questions swirling around his record as Cook County assessor, Joseph Berrios will take the unprecedented step Tuesday of defending his office before the Cook County Board after a Chicago Tribune investigation exposed widespread inequities in the county’s property tax system.

_This story is part of a collaboration between ProPublica Illinois and the Chicago Tribune, continuing the reporting ProPublica Illinois reporter Jason Grotto began on property tax inequity as a member of the Tribune staff._

Published in June, “The Tax Divide” showed the assessor’s office overvalued low-priced homes while undervaluing high-priced ones. These disparities in assessments — known as regressivity — led to inequities in property tax bills, giving the wealthy unsanctioned tax breaks while penalizing low-income residents.

In the weeks since, mounting pressure has put the assessor on the defensive. The county’s independent inspector general opened an investigation; lawmakers at the state and local levels proposed legislation to limit the assessor’s ability to raise campaign contributions from tax attorneys; and a bill has been introduced in the General Assembly that would require greater transparency.

Among the questions Berrios is likely to face at Tuesday’s hearing: Why did his office tout and then quietly abandon a new residential valuation model funded by the MacArthur Foundation designed to reduce regressivity?

After the Tribune reported that the assessor’s office hadn’t implemented the new model, despite issuing a July 2015 news release saying it had, Berrios and his staff criticized the model as well as the experts who led the effort to create it.

Both of those experts — Robert Weissbourd, president of the consulting firm RW Ventures, and University of Chicago public policy professor Christopher Berry — are scheduled to testify at the hearing.

Among other criticisms, the assessor’s office said it had told Weissbourd and Berry before issuing the news release that their model was flawed. Deputy Assessor Tom Shaer produced three emails after the series ran that he said proved the two had “the temerity” to mislead the Tribune into believing otherwise.

Since then, Weissbourd has produced dozens of additional emails showing that the issues raised in the messages Shaer cited were part of a multiyear effort to test and implement the new model and had been resolved before the office produced the news release.

County commissioners also may address the question at the heart of the controversy over the models: how exactly the assessor values the county’s 1.5 million residential properties. The office has refused to disclose crucial details about its methods.

One of Berrios’ recent attempts to defend his record has already evaporated. Late last month, the assessor’s office posted on its website a news release that cited a study by a prominent research organization as proof that the county’s assessments are fair and accurate. Berrios invoked the same study during a rare television interview, also in late June.

However, a close examination of the study shows that its focus was on tax rates, exemptions and other policy issues in all 50 states; it was not an evaluation of how accurately the assessor’s office values property.

Adam Langley, a lead author of the study by the Massachusetts-based Lincoln Institute of Land Policy, confirmed that county officials mischaracterized the report.

“Our study is not about assessment accuracy,” Langley said. “That’s a separate issue.”

The institute issued a statement to the assessor’s office as well as county commissioners last week in an effort to set the record straight.

Two days later, the assessor’s office removed the news release from its site “to do further work regarding it,” Shaer said.

Tuesday’s hearing will see Berrios, who also is chairman of the Cook County Democratic Party, facing off against many commissioners who rely on the party for political support.

Berrios received more than half of his campaign contributions from the property tax appeal industry, the Tribune has reported. Appeals increased dramatically under Berrios, but an analysis carried out in cooperation with the University of Chicago’s Berry found the process makes the assessment system even less fair.

Commissioner Jesus “Chuy” Garcia, who has introduced an ordinance to block the assessor from receiving contributions from tax attorneys and other businesses that file appeals, was among the first commissioners to request that Berrios testify. Berrios will appear at a Finance Committee hearing with the full board in attendance.

“At the heart of all this are the working people who are already hurting from increases in property taxes,” Garcia said in an interview. “We can’t tolerate a system that may be penalizing them. It’s just not fair.”

A Changing Story

In seeking to defend its assessment practices, Berrios’ office sent four pages of bullet points to county commissioners contending, among other assertions, that the assessment model funded by the MacArthur Foundation produced inaccurate results. (The MacArthur Foundation has been a significant funder of ProPublica since it began publishing in 2008.)

“Simply put, this so-called ‘improvement’ is unfit for exclusive, primary or even somewhat primary use,” the document stated.

And in television and radio interviews after “The Tax Divide” was published, Shaer said the office had been optimistic about the MacArthur-funded model at the time but later discovered it produced erroneous results.

Until the Tribune launched its investigation, however, Berrios had offered nothing but public praise for the new model.

A July 2015 news release headlined “Cook County Assessor Berrios Implements New State of the Art Residential Assessment Model” described it as a “breakthrough” that improved accuracy by 50 percent and lowered regressivity by 25 percent.

As recently as April 2016 the assessor’s office was encouraging reporters to look at the model as evidence its assessments were fair and accurate. Officials even gave Weissbourd and Berry permission to provide reporters with a slide presentation showing how the new model improved assessments.

Around the same time, Berrios’ office also provided reporters with studies that concluded the new model outperformed the old one on every measure used by assessment experts around the world to evaluate fairness and accuracy.

When the Tribune analyzed millions of property tax records from the 2015 tax year, however, the results suggested the assessor had not actually implemented the new model.

Presented with those results in August 2016, officials then stated they were using the new model in conjunction with the old one, though they would not disclose how they did so. That was news to Weissbourd, Berry and MacArthur Foundation President Julia Stasch, all of whom had been quoted in the news release.

By then, the Tribune had used open records laws to obtain the new model along with computer records showing the results the office obtained from it. An examination of the computer code showed the office had made numerous errors when it ran the new model.

When those errors were pointed out during a September 2016 interview, the assessor’s office changed its story again, saying the new model had not been implemented because it had “issues” and produced inaccurate results.

The office also said Weissbourd and Berry had been told about the problems, which the two said was untrue.

After publication of the Tribune series, Shaer produced three emails that he said proved the assessor’s office had informed Weissbourd and Berry that the new model was flawed before issuing the news release. One of the emails wasn’t about the model at all; the others, sent in 2012 and in 2013, related to ongoing discussions about how to implement the new model.

The set of emails provided by Weissbourd show that the issues raised in those messages were resolved. There is no evidence in the emails that the assessor’s office alerted Weissbourd or Berry about issues with the new model before issuing the news release.

Two weeks before the 2015 announcement, the office’s head of residential valuation, Al Sarro, wrote to Weissbourd seeking feedback and requesting quotes for the news release announcing the launch of the model.

“Please take a look and add your thoughts,” Sarrow wrote in an email, which was provided by Weissbourd. “And also can you reach out to Chris (Berry) and Julia (Stasch) for any quotes they would like to add?”

When asked to comment on this set of emails, Shaer said the assessor’s office stands by its earlier statements that the model’s developers had been informed it was flawed.

Criticisms Refuted

When Berrios held a June news conference to address the findings of “The Tax Divide,” he brought along a researcher from a for-profit government services company that holds $65 million in county contracts, including a $30 million contract to set up a new property tax computer system for the county.

Richard Borst, of Tyler Technologies, criticized the MacArthur-funded model as well the Tribune’s analysis.

At the news conference, Borst handed out two pages of talking points that said, among other arguments, that the Tribune did not disclose enough technical information to make its analysis credible. But when a Tribune reporter pointed out that a 100-plus-page study had been posted online, Borst acknowledged he hadn’t read it.

Borst also repeated assertions by the assessor’s office that the analysis had included invalid sales involving the liquidation of estates of the deceased.

In fact, the analysis screened out those sales and other transactions that are not considered valid; the study was vetted by three industry experts, including two who helped write standards for the International Association of Assessing Officers.

Borst also took aim at the MacArthur-funded model, saying Berry and Weissbourd’s team had made “rookie mistakes” that caused the model to be “fundamentally flawed.” For example, he said the new model “disregards location” and makes no use of geographic information systems, or GIS, to value property.

However, the new model does include location information. As for GIS, Berry said he and his team of statisticians had wanted to use GIS programs to further improve the accuracy and fairness of assessments, but the assessor’s office told them it didn’t have the necessary technical ability.

“It’s shocking that Mr. Borst is defending the assessor’s regressive assessments, especially without even having looked at their models,” Berry said. “Rather than get distracted by this nonsense, the assessor should get on with the business of fixing their unfair system.”

Borst’s firm declined to make him available to be interviewed for this story.

Asked at the news conference if he had examined the model that the assessor’s office actually uses, Borst said he hadn’t.

“That’s not for me to say,” he said. “That’s not my job.”