Technology

An analytics platform custom-built from the ground up to deliver the insights fund investors need to make better decisions.

Conclusions, not data.

"Past performance is not indicative of future results." This ubiquitous disclaimer highlights a major obstacle for asset allocators: quartile rankings shift, styles drift, and relative underperformance is all too common.

With an ever-growing number of investment options available, including new passive and "smart beta" alternatives, a systematic decision framework is mandatory: the fund selection challenge has become too complex, and the costs of sub-optimal decisions too high, to rely on simple screens and informal judgments.

Decision criteria must be customized to the specific parameters of the situation; there is no one "best" fund in a category for all investors. And they must be ready to withstand intense future scrutiny by those with the benefit of hindsight.

How does your selection process score on validity and reliability?

// Validity

The extent to which the metrics used consistently and meaningfully predict future outcomes, and are therefore valid decision criteria.

// Reliability

The extent to which the process would produce the same results if repeated with the same data.

Challenges We're Focused On

Missing Data
Track records of unequal lengths and strategies with short histories can be framed as a missing data problem.
Comparability
Relative performance evaluation requires an appropriate peer group or index. Novel benchmarking techniques can improve accuracy and comparability, even for unique strategies.
Causal Inference
Observed excess returns relative to an appropriate benchmark must be separated into skill, risk and random variation components.
Importance Weighting
The large number of often-conflicting due diligence data and statistics available require quantitative isolation and weighting of the most predictive factors.
Dynamic Change
Skill, risk taking and the external environment can all vary over time. This requires building dynamic, path-dependent models.
Robust Decisions
Decision science techniques can help identify decisions that are robust to deep uncertainty and most likely to meet multiple objectives.

How We're Solving These Challenges

previous arrownext arrow
Bespoke Simulation Engine

Our proprietary investment strategy simulation engine, EPSE, accommodates sets of complex constraints, enabling us to perform custom benchmarking analyses at a new level of accuracy and precision.

 

Designed by experienced Portfolio Managers, our software can replicate a wide range of realistic strategies, biases and behaviors to uncover detailed insights about the drivers behind performance, as well as accurately model the opportunity set available to the active manager. This enables better estimation of manager skill, even with a limited track record or sparse peer group.

Dynamic Factor Framework

Factor models and style analysis can shed light on the sources of returns, but imputed factor exposures can vary significantly depending on the time period and model specification used.

 

Replicator is our custom-built modeling tool for multi-period attribution analysis with time-varying factors. It also performs decomposition and bottom-up synthetic replication of active strategies based on trading rules programmed by actual Portfolio Managers.

 

Factor exposures and dynamics can be used to assess the stability of a fund's performance (for example, to determine its suitability for use in a portable alpha strategy), as well as its quality.

StateSpace
Time Series Modeling

StateSpace is our collection of time series analytical techniques which powers evaluation of allocation timing decisions as well as the stability of a strategy's return profile. These techniques involve decomposing a time series into trend and random components, as well as modeling path dependence and mean reversion.

 

By estimating regime switching models, StateSpace can determine how a strategy's performance depends on environmental conditions, such as the economic environment. Using mulitple statistical techniques, we can also test whether and how a change in a fund, such as a new Portfoio Manager or a growth in assets under management, has altered its alpha generation capability.

Premortem Analysis
Robust Decision Planning

Our scenario analysis and risk evaluation framework, Premortem, facilitates thorough understanding of decision alternatives which involve multiple objectives and difficult tradeoffs.

 

Designed to address the limitations of traditional case-based scenario analysis, our tool enables clear specification and exploration of all preferences, uncertainties and priorities involved in a choice – including multiple time horizons and differing beliefs. Premortem's transparency garners acceptance from decision makers, while its sound methodology ensures the consistency and robustness of the decisions and adaptation plans it yields.

Classifying ESG Behavior

The complex and multi-dimensional nature of ESG issues, combined with a lack of standardization, oversight and verification have contributed to a "rhetoric-reality gap" between stated and actual approaches to responsible investment and ESG integration. These challenges cannot be overcome by traditional case study and questionnaire-based diligence approaches.

 

Empirically's RealESG capability comprises a set of fully objective, quantitative techniques to evaluate a manager's ESG approach at both the strategy and firm levels. It facilitates improved reliability, comparability and accuracy in assessing and monitoring ESG credentials.

next arrow
Slider

 

Not your father's attribution table.

Select an asset class to see how this technology answers critical diligence questions.

Or, View Our Solutions to learn how organizations like yours are applying Empirically's capabilities.