MDT: A day in the life

Just what is it that quant managers do? We asked Dan Mahr, Head of Federated Hermes’ MDT Advisers, exactly that question, and his response was enlightening. One thing we learned: it’s not all algorithms and machine learning — the human interface is important too. 

Q: What does a working day at MDT look like?

At MDT Advisers, portfolio construction is a daily, data-driven process powered by proprietary machine learning tools.

Each night, the team downloads data from vendors, recalculates company characteristics, and runs every stock in the domestic equity universe through our regression tree model to generate fresh forecasts. These forecasts feed into our proprietary portfolio optimizer, which rebalances each strategy and produces a trade list for review.

The day begins with a trade review process, where human oversight plays a critical role. While MDT’s models operate mechanically, the team seeks to ensure that the data is accurate and that no material news has emerged that the model may not yet reflect. This step is not about subjective overrides, but about validating inputs and understanding the model’s behavior.

Beyond daily operations, most of our time is spent on research and model enhancement. Idea generation is central to MDT’s process, and we believe the firm’s commitment to building tools in-house — from back-testing engines to risk models — gives us exceptional flexibility and control.

Q: How does the portfolio optimizer work in practice?

Our proprietary optimizer is a key component of MDT’s portfolio construction process. It integrates multiple dimensions, including:

• Alpha forecasts from our regression tree model
• Hard risk constraints applied consistently across portfolios
• Statistical risk models to predict volatility and tracking error
• Trading cost models, including market impact and liquidity considerations

The risk constraints — which are consistent across all our portfolios — along with a statistical risk model that makes predictions for volatility and tracking error, help guide us to more consistent outcomes.

The goal of this optimization process is to ensure that our portfolios are aligned with the model’s current alpha forecasts, while carefully managing the balance between risk and return, and avoiding excessive turnover.

This disciplined blend of automation and oversight helps us to maintain precision, adaptability and transparency in our daily investment operations.

Q: How has MDT’s process adapted to changes in the market in recent years?

We’ve witnessed notable shifts in market structure over the past few years — the ever-increasing market share of passive investing, the rise of zero commission and social media driven retail trading, and the increased role of high frequency and algorithmic traders — all of which have influenced how quantitative strategies operate.

Rather than attempting to pinpoint a single cause for this change, we emphasize the importance of maintaining active strategies designed to capitalize on inefficiencies — regardless of their origin.

This means we don't need to know what's in the driver’s seat for market behavior. The important thing is having strategies that are active and able to take advantage of inefficiencies when they present themselves.

Despite the challenges and uncertainties, day to day, we remain energized by the dynamic nature of markets. The constant influx of new data, unexpected macroeconomic developments and evolving investor behavior make quantitative investing a continuously engaging field.

There's always new information out there. There are always curveballs coming from a macro perspective, risks that you’d never seen before that all of a sudden manifest themselves. From my perspective, it's a great place to be and a really exciting place to be applying my technical background.