Key Takeaways
- • Intraday data does not improve the accuracy of daily value-at-risk (VAR) calculations, according to research by Ralf Wohrle, portfolio manager at Metzler Asset Management, a Frankfurt-based hedge fund.
- • The findings contradict assumptions among data providers that more sophisticated risk managers benefit from larger datasets and increased computing power when modeling daily loss forecasts.
- • VAR model testing reveals that expanded intraday datasets fail to enhance predictive accuracy for daily risk assessments, challenging the industry's push toward bigger data infrastructure.
Summary
Ralf Wohrle, portfolio manager at Metzler Asset Management in Frankfurt, has published research challenging a widely held assumption in risk management: that intraday data improves the accuracy of daily value-at-risk calculations. His analysis tested multiple VAR models and found that the incorporation of intraday data does nothing to enhance the precision of loss forecasts used in daily risk modeling.
The research directly contradicts the growing appetite among data providers for intraday datasets. These providers have assumed that as computing power has increased and risk managers have become more sophisticated, larger and more granular datasets would translate into better risk models. However, Wohlre's findings suggest this assumption does not hold in practice for daily VAR applications.
The implications are significant for institutional investors and hedge funds investing capital in infrastructure designed to capture and process intraday market data. If intraday information provides no marginal benefit to daily risk forecasting accuracy, the cost-benefit analysis for such systems comes into question. The research indicates that practitioners may be able to achieve equivalent risk measurement results using less computationally intensive approaches, potentially reducing operational complexity and infrastructure expenses without sacrificing model accuracy.
The research directly contradicts the growing appetite among data providers for intraday datasets. These providers have assumed that as computing power has increased and risk managers have become more sophisticated, larger and more granular datasets would translate into better risk models. However, Wohlre's findings suggest this assumption does not hold in practice for daily VAR applications.
The implications are significant for institutional investors and hedge funds investing capital in infrastructure designed to capture and process intraday market data. If intraday information provides no marginal benefit to daily risk forecasting accuracy, the cost-benefit analysis for such systems comes into question. The research indicates that practitioners may be able to achieve equivalent risk measurement results using less computationally intensive approaches, potentially reducing operational complexity and infrastructure expenses without sacrificing model accuracy.
Go deeper than the headline
Research every firm in this story, track the ones that matter and get the daily briefing — free for 7 days.