He credits that math and science background for giving him the skill set he’d need to eventually revolutionize the field of quantitative market analysis.
Kolanovic’s first Wall Street position came in 2003 as a derivatives strategist for Merrill Lynch. It was there that he started looking at flows around option hedging — what types of flows there were, and what type of market impact they were having.
He eventually became interested in passive indexes as well. His main focus in that area was monthly or quarterly rebalances, which he says can produce predictable flows that can ultimately impact the direction of the market.
Kolanovic’s work in the space eventually led him to take a close look at risk premia strategies — which target absolute returns through long-short positions across a variety of factors and asset classes — in 2010 and 2011. The logic was the same: there are predictable flows associated with them, and you can estimate how much market impact they’ll have.
All of these inroads culminated in the series of incredibly accurate market calls Kolanovic has made over the past few years. In fact, his forecasts were so spot-on that his research started to move the market.
It was the ultimate sign of respect from investors, who voraciously devoured Kolanovic’s reports and traded accordingly.
At this point, given his immense track record, it can be helpful for market participants to fully understand how Kolanovic is analyzing various market forces and making directional recommendations. That’s where we come in.
Business Insider conducted an exclusive interview with Kolanovic, where he shed some light on his overall process for quant analysis. His comments are provided in full below:
Looking at options
“In the options complex, a lot of information is available on exchanges. So you can calculate the convexity content of options. It’s not 100% transparent, so we do need to take some views on supply and demand, buyers and sellers.
“You can make estimates, then give yourself some room for error. Once you have the estimate, if the market moves, you know roughly what the flows will be.”
“For systematic strategies, it’s more patchwork. For instance, for trend-following strategies you can look at various industry publications and indices. It’s not perfectly clean-cut as to how these funds are classified, and what types of strategies they pursue.
“You can also look at some broad indices of these types of systemic managers, and then you can analyze daily regressions versus the S&P or other asset classes. Once you know the asset pool, and market beta, you can estimate exposures and dollar flows. You can see how quickly betas to equities or bonds change.
“It’s a combination of model output performance, fund performance, and asset-price data. If you look at the options complex, it’s relatively easy to generalize. Trend-following strategies are relatively easy to generalize. And volatility-controlled type strategies also have many commonalities. Some strategies follow short-term signals, some intermediate, some look at realized volatility, some implied — it’s going to be a gamut of signals.”
Making key assumptions
“One needs to make assumptions. We can run a survey of clients to see what types of signals they use. We can look at performance and try to back out some average signal. One thing is sure when volatility is higher: you will see de-risking from most of the the strategies.
“You end up with a number of analyses, each with a different level of certainty. Then you try to reconcile them, understand where differences come from, and estimate another uncertainty band. Then you arrive at some conclusions. It’s largely science, math, and statistics, but there’s also an element of intuition and art.”
The role of liquidity
“Once you get estimates of flows, you need to assess the liquidity situation. If liquidity is depleted, it could have an outsized impact. That’s another element of the equation that you need to account for.
“And on top of that, there’s going to be fundamental data and event risk. Very positive headlines can neutralize some of the outflows. You also have to monitor news-driven flows coming from fundamentals or the macro side.”