Big Data: The Next Frontier in Risk Management

Panel Composition & Objective:

• The Milken Institute 2018 Global Conference panel, moderated by Stacey Worden, features C-level executives from Trend Micro, Sigma Ratings, GeoQuant, and Starling Trust Services.
• The collective goal is to lower the cost of risk management by quantifying previously unmeasurable risks, thereby enabling more robust and enthusiastic risk identification.

Trend Micro (Ed Cabrera):

• Data Volume: Daily threat data has grown from less than 1 terabyte in 2008 to over 6 terabytes, 15 billion log lines, and 300 million sensors today.
• Threat Evolution: Cyber risks have shifted from financial/reputation risks to operational risks due to the speed of automation and social engineering (e.g., Business Email Compromise).
• Mitigation Strategy: Launched a "writing analysis DNA" program to fingerprint C-suite communication styles, distinguishing malicious impersonation from legitimate correspondence.
• Future Outlook: Predicts exponential risk growth driven by IoT adoption and expects a convergence of security, HR, and cultural data analysis within privacy-centric frameworks.

Sigma Ratings (Stuart Jones):

• Market Focus: First AI-powered non-credit rating agency targeting management, governance, and financial crime risk in emerging and frontier markets where data is scarce.
• Data Strategy: Overcomes public data gaps by combining public domain information with "opt-in" proprietary control data from companies themselves.
• Business Model: Avoids "blackmail" perception by framing the opt-in as a mechanism for companies to improve their public "business integrity" rating and transparency.
• Forward-Looking: Anticipates business integrity and governance ratings becoming as critical as traditional credit ratings within the next five to ten years.

GeoQuant (Mark Rosenberg):

• Product Function: A "seismic monitor" for politics that quantifies political instability, social polarization, and policy risk at daily or multiple-daily frequencies.
• Differentiation: Replaces subjective expert-based analysis with objective, transparent models derived from non-traditional indicators (e.g., social polarization along ethnic/religious lines).
• Competitive Advantage: Provides high-frequency, comparable metrics that allow for direct integration into investment portfolios, unlike static reports like the CIA World Factbook.
• Bias Management: Acknowledges inherent bias in social sciences but aims to minimize it through objective baselines and continuous human expert validation.

Starling Trust Services (Steven Scott):

• Core Metric: Measures "conduct risk" by analyzing corporate communications metadata and HR data to detect behavioral shifts signaling misconduct or toxic culture.
• Industry Context: Addresses a $1 trillion+ global spend over the last decade on failed governance and compliance efforts, driven by $320 billion in regulatory fines for misconduct.
• Behavioral Thesis: Operates on the premise that behavior is a social contagion derived from peer trust rather than rational actor economics; women are identified as more adept at building trust dynamics.
• Forward-Looking: Predicts that human expert opinion will become "completely obsolete" in 5–10 years, replaced by a hybrid machine-human model for risk assessment.

Regulatory & Ethical Landscape:

• Privacy vs. Security: Regulatory frameworks (e.g., GDPR) have not yet caught up with new data models, creating a "feel in the dark" environment regarding privacy invasions for risk mitigation.
• China Comparison: Panelists note China's competitive advantage in big data due to lack of privacy laws, enabling aggressive social credit scoring, whereas US/EU systems face significant normative and legal barriers to similar government data access.
• Bias & Transparency: All panelists emphasize "augmenting" rather than replacing human analysts to ensure accountability and mitigate algorithmic bias.

Industry Trends & Investment:

• Market Consolidation: The RegTech/RiskTech sector is currently fragmented and over-hyped in areas like transaction monitoring but is moving toward consolidation into industry utilities with network effects.
• Venture Capital Gap: Investors often lack domain expertise, leading to a "catch-up" phase where only firms with deep, specific domain knowledge will survive as industry standards.
• Efficiency Drivers: Financial institutions currently face "mind-boggling" manual workloads (e.g., 3,000 staff monitoring trade transactions) due to blanket high-risk classifications in 60% of the world, driving demand for automated granularity.