It’s a rare talent to double assets under management in five years. Yet that is what Mark Steed, CIO of the Arizona Public Safety Personnel Retirement System, has done by harnessing data analytical prowess: The fund’s $19.6 billion in AUM is up 100% from what it was when he took over as CIO in 2018.

Steed’s system is 67.2% funded as of its most recent financial report, up from 45% when he assumed the helm five years ago. A separate ranking for employees hired since 2017 is 108% funded.

After starting out at J.P. Morgan Chase and Beneficial Financial Group, Steed joined PSPRS in 2007 as a portfolio analyst and worked his way up through the ranks, managing every asset class, to deputy CIO in 2015. Three years later, he got the top investment job after CIO Ryan Parham retired.

One of Steed’s initial moves was to consolidate PSPRS’ asset categories into three from 10. The switch both made its structure simpler and easier to ensure diversification. The goal was to increase “optimizing returns while keeping risk in check,” Steed told CIO at the time. “To do this, we wanted to be as honest and careful as possible to categorize between uncorrelated types of returns. From these three distinct categories, the idea is to have the asset classes and investments compete for capital.”

A more recent prescient maneuver was to beef up the fund’s cash stash. After the big market run-up of 2021, Steed and his staff figured that rising inflation and rates and a slowing economy were going to hobble returns in 2022. They pushed the cash account up to near 15% of assets, from slightly less than 2%.

“Holding more cash made sense, given the lower opportunity cost with higher rates,” Steed explains, adding that the “dry powder” gives PSPRS the flexibility to buy enticing assets when they appear. “It’s easy to overlook the power of simple solutions.”

The PSPRS handles 200-plus municipal DB plans, and many initially were reluctant to make meaningful contributions to Steed’s fund. Steed realized he needed to improve his organization’s investment return record to entice the local governments to pony up more. “Fortunately, the strategy was successful, enticing cities and towns to contribute an additional $5 billion to our plan,” he recounts.

Undergirding all this success was the building up of PSPRS’ internal data analytics team. That involved moving to fully digital systems for front-, middle- and back-office functions, thus increasing efficiency by removing manual work and human errors. In addition, Steed used machine learning and natural language processing to systematically analyze data from such documents as quarterly reports. “This provides deeper insights from vast amounts of unstructured data,” he explains.

All that has applicability for peers, in Steed’s view: “We hope our work developing this idea moves the industry toward more evidence-based portfolio construction grounded in tangible plan objectives.”

—Larry Light