Kozicki Shows Central Bankers Only as Good as Forecasts
For Sharon Kozicki, seeing what’s missing in economic models is as important as seeing what’s there.
Kozicki rose at the Bank of Canada in her native Ottawa to become an adviser to the governor after more than a decade at the Federal Reserve in Washington and Kansas City, where she built a global reputation improving the ways central bankers make forecasts to set policy.
It’s a skill increasingly valued as officials try to understand how the economy changed following the worst financial crisis since the Great Depression. Old economic relationships have become less reliable because the world they describe has passed, she says.
“You have to be very mindful of what’s not in your models,” Kozicki, 49, said in an interview at the Bank of Canada’s headquarters. “It doesn’t mean that you throw out your model, but you may need to bring some judgment into that framework to characterize what your model doesn’t capture.”
At the Fed in the 1990s, Kozicki saw that its economic model didn’t adequately explain how people formed expectations about inflation. Her analysis, which was praised by Chairman Ben S. Bernanke, would help the central bank as it developed a new forecasting tool to adjust to an era of slower price increases. Her “intellectual rigor” also impressed Mark Carney when they worked together, said the former Bank of Canada governor.
“She’s a first-rate economist, but she’s also a first-rate thinker,” said Tom Hoenig, who was Kozicki’s boss for a decade at the Kansas City Fed.
Kozicki’s progression in econometrics -- which blends economics and statistics -- came naturally. After excelling in math at high school in Ottawa, she went to the University of Toronto. When she needed an elective, “I was looking at courses that were outside math and science, and economics sounded really interesting.”
Once Kozicki began studying the field, “I loved it,” she said. “I always liked real-world stuff,” and “it seemed to just be something that made me think a lot, that fit in well with some of my own strengths.”
Kozicki received a master’s degree in economics at the University of Toronto, focusing on econometrics. Her professors -- including former Bank of Canada special adviser Angelo Melino and Dale Poirier, a former visiting economist at the International Monetary Fund -- helped steer her to the doctorate program at the University of California at San Diego, where she worked with Robert Engle and Clive Granger, who won the Nobel Prize in economics in 2003.
A key lesson she learned was not to be bound by prevailing ways of thinking about economic relationships. “When you’re on the econometric side, a lot of what you’re doing is building a toolkit,” Kozicki said. If the models don’t adequately explain what’s going on, “maybe you need to actually build another tool.”
The school provided “a great environment” for graduate students, Kozicki said. “People were very serious in what they were doing, but in their interactions with other people, very relaxed.” Aside from econometrics, she studied macroeconomics, which looks at the behavior of an economy as a whole, as well as finance, trade and international issues. Kozicki would draw on all these areas later in her career.
Her thesis -- assessing common features across different data sets -- identified ways economists could see similar elements in economic cycles among different countries and patterns of volatility in financial market data, she said.
Kozicki assumed while growing up that a career as an academic was in her future, given the encouragement she received from her mother and father, a professor of chemical engineering.
“I had never imagined coming out of graduate school I would have ended up as anything other than a professor,” she said. “It just seemed like that’s what you go and get a Ph.D. for.” Working at the Bank of Canada as a summer intern in 1987 and 1988 opened the door to a career in making policy, rather than simply studying it.
When Kozicki received a recruitment pitch in 1992 from the Fed’s Board of Governors in Washington, she was impressed by “the diversity of types of jobs, types of questions that came up, and also the importance of solid research and analysis.”
Kozicki became part of the Fed team building an economic projection model that would reflect changes in the U.S. economy, particularly the impact of financial deregulation and the way the U.S. was increasingly tied to the global economy through trade and investment.
“When I arrived, they had an old-style macro model that they were using to generate forecasts and also to run simulations,” Kozicki said.
She also wrote research papers “at the intersection of macroeconomics and econometrics,” adding a layer of analysis critical to central banking. A number of those were done with Peter Tinsley who, as deputy associate director of the board’s research and statistics division, helped recruit her to the Fed.
She and Tinsley saw that the central bank’s equations had an unrealistic view of how inflation expectations are formed; this led to insights about the role they play in monetary policy. The two collaborated on a series of papers that Bernanke cited in a 2004 speech.
“Ms. Kozicki is the smartest researcher I have had the good fortune to work with, and on several occasions saved me from public exhibitions of my ignorance,” Tinsley, now a professor at University of London’s Birkbeck College, said in an e-mail message.
Kozicki’s skills were augmented by her time at the Kansas City Fed, where she and her husband were offered full-time jobs. She arrived as an economist in 1995 and by 2003 had been promoted to vice president. Her responsibilities included pulling together the staff’s forecast for then-President Hoenig, reaching out to companies in the district for their assessment of the economy and explaining policy to business groups and the public.
She welcomed the breadth of responsibilities. “You’re still trying to get your research published, but you’ve got that policy side of things and then that outreach,” Kozicki said. “I originally liked economics for that ‘real world,’ and this was really ‘real world.’”
Her experience with businesses was invaluable because Kozicki would use the anecdotes she heard to make judgments about the model and its accuracy, Hoenig said by phone from Washington, where he’s now vice chairman of the Federal Deposit Insurance Corp.
“One thing I’ve learned about models is whenever I rely on them 100 percent, I’m usually 100 percent wrong,” Hoenig said. Kozicki “was able to work the econometric models as well as anyone,” he said. “She understands their value and understands their limitations.”
After a decade in Kansas City, Kozicki took the opportunity to move closer to family and return to her hometown of Ottawa, joining the Bank of Canada in 2006 as research director. By 2008, she had been appointed deputy chief of the international department, just as the global economy was tumbling into turmoil following the bankruptcy of Lehman Brothers Holdings Inc.
“Clearly a challenge at the onset of the crisis was getting a sense of the severity of the situation,” Kozicki said. She examined what would happen if policy interest rates were reduced to zero, and how to account in the bank’s models for the ways unconventional monetary policies might be implemented.
Her experience at the Fed also gave her insight into how central bankers in Canada’s most important economic partner were viewing the turmoil.
Kozicki’s work was crucial to a team effort in supporting the bank’s governing council, according to Carney, now Bank of England governor.
“Her profound understanding of macroeconomic forecasting, combined with superior research and analytical skills has enabled her to make a tremendous contribution to the bank during the recession and recovery in Canada,” he said in an e-mail. “Her exemplary research on inflation expectations and unconventional monetary policy, in particular, stand out.”
Kozicki said she tends to “think more about the assumptions underlying the econometrics in a model and whether these assumptions continue to be reasonable in the current environment. I think a lot about whether models are missing important ingredients for understanding and capturing the economic forces that are important today.”
During the crisis, for example, she said equations of the U.S. economy didn’t adequately handle people’s net worth: how the value of their assets, such as houses, was changing relative to their liabilities.
The value “went down a lot, but they still owed all this debt,” Kozicki said. “That splitting out was not in the models. It ended up being really important,” including for understanding U.S. consumer behavior.
The return to the Bank of Canada meant Kozicki was exposed to “all of the sorts of things that the main office of the central bank gets involved in,” such as the international issues. It also allowed her and her husband to pursue their passion for sports.
“The day after we arrived, we went out and bought cross-country skis,” she said. “Kansas City, it’s hot,” and less conducive than Ottawa to train for the triathlons in which she regularly competes. “I like doing things outdoors, and it’s a great city for that” because “everything that nature has is out there.”
Kozicki was promoted to head the bank’s Canadian Economic Analysis department in 2010, and became adviser to Governor Stephen Poloz in August, just one step below deputy governor. While her rise has been quick, she points to those who have helped her on the way and is taking the time to help others.
“I really like to be able to sit down with newer economists to the institution,” she said. “That part of me that thought of being a teacher or professor is still there.”
To contact the reporter on this story: Paul Badertscher in Ottawa at email@example.com.
To contact the editor responsible for this story: Chris Wellisz at firstname.lastname@example.org.