The Right Data Makes All the Difference

Key Concepts

• AI-Powered Fundamental Analysis: New Constructs leverages AI, specifically Google’s Gemini 3 through their Finsites agent, to scale and enhance fundamental analysis, not replace it. Data quality is paramount, with their proprietary data validated by academic research demonstrating 8% annual alpha.
• Core Earnings as a Superior Metric: New Constructs utilizes a proprietary “Core Earnings” metric, considered more accurate than reported earnings, and indices based on it have outperformed the S&P 500.
• Expectations Investing: A focus on identifying discrepancies between market expectations and a company’s underlying economic reality, utilizing reverse DCF models and analyzing growth rates implied by current valuations.
• Economic Reality vs. Accounting Manipulation: Prioritizing understanding a business’s cash flow and capital allocation over solely relying on accounting data, recognizing the potential for accounting distortions.
• Valuation Discrepancies: Identifying companies where market expectations are significantly misaligned with their economic fundamentals, presenting potential investment opportunities.

AI and Data Foundation (Part 1)

The discussion began with an exploration of New Constructs’ approach to integrating AI into investment analysis. David Trainer emphasized that AI is a tool dependent on the quality of its input data – “garbage in, garbage out.” New Constructs focuses on building a pristine dataset derived from SEC filings, particularly detailed information found in footnotes, and applying AI to analyze it consistently and at scale. This data quality has been independently validated by research published in the Journal of Financial Economics, demonstrating 8% annual alpha generation from their core earnings data.

A key development is the partnership with Google Cloud, resulting in the Finsites AI agent, powered by Google’s Gemini 3 AI model and New Constructs’ data and proprietary signals. Finsites allows users to query for investment insights (e.g., “Show me the best stocks in the tech sector”) based on New Constructs’ proven ratings. New Constructs currently covers over 3,400 stocks and a total of 10,000 securities. The firm’s methodology is demonstrated by the existence of a Core Earnings Leaders ETF, which has reportedly outperformed the S&P 500 over the past year. Trainer highlighted a Harvard Business School case study recognizing New Constructs’ disruptive potential.

Shifting Focus to Economic Analysis (Part 2)

The conversation shifted to the core of David Trainer’s investment philosophy: an economics-driven approach that transcends traditional accounting-based analysis. He stressed the importance of understanding a business’s underlying cash flow and capital allocation, viewing accounting as merely input data. This allows for a more robust analysis, less susceptible to changes in accounting rules or XBRL standards. The goal is to analyze the economic reality rather than accounting manipulations.

Trainer demonstrated this approach with a live screener example, identifying a healthcare provider with a free cash flow yield exceeding 10%. Despite a recent earnings miss, a deeper dive revealed strong economic indicators: a 120% return on invested capital (ROIC), strong free cash flow, and a low price-to-economic book value of 0.44. The market’s valuation implied a 60% permanent decline in profits, presenting a potential opportunity if that expectation proves incorrect.

Valuation Techniques and Metrics (Part 2)

Trainer explained the use of a reverse discounted cash flow (DCF) model to determine the years of profit growth required to justify a stock’s current price. In the healthcare provider example, less than one year of growth was needed, indicating a significant disconnect between market expectations and potential economic reality. Conversely, an “unattractive” company was presented with a negative economic book value, requiring over 100 years of profit growth to justify its price.

Analysis included tracking shares outstanding (bullish when decreasing, bearish when increasing), ROIC, and breaking down ROIC using DuPont analysis (profit margin and capital turns). The example company showed declining margins but increasing capital turns, potentially due to asset sales. Free cash flow yield was analyzed using both one-year and two-year averages. A key visualization is the economic book value (no-growth value) versus price chart, highlighting the disparity between intrinsic value and market perception.

Case Studies and Contrasting Examples (Part 2)

Kroger was presented as a “long idea” – a company making money with strong market share and a price-to-economic book value of 1.0, implying zero growth expectation. DoorDash was used as a contrasting example, demonstrating the extreme expectations embedded in its valuation, requiring massive improvements in gross order volume to justify its price. Claritive, with a market cap of $744 million versus $4.6 billion in debt, was also briefly mentioned.

The Importance of Independent Diligence (Part 2)

Trainer advocates for “Expectations Investing,” detailed in the book by Alfred Rappaport and Michael Mobes, focusing on identifying discrepancies between market expectations and fundamental realities. He stressed the importance of independent diligence, cautioning against blindly trusting news or social media, as these sources often have a vested interest in selling.

Conclusion

New Constructs presents a compelling approach to investment analysis by strategically combining the power of AI with a rigorous, economics-focused fundamental framework. Their emphasis on data quality, proprietary metrics like Core Earnings, and a focus on market expectations offers a unique perspective for identifying undervalued opportunities and navigating the complexities of the modern market. The core takeaway is that while AI is a powerful tool, it is most effective when applied to high-quality data and guided by a deep understanding of a business’s underlying economics.