Investment Podcast #36 - Rethinking Genuine Diversification

In this episode of the Ensemble Investment Podcast, host James Whelan and guest Eric Marais, Investment Specialist and Head of Clients at Orbis Investments, delve into the critical challenge of genuine diversification in today's highly concentrated markets. Despite decades of emphasis on diversification, many portfolios appear diversified on paper but behave like concentrated bets due to widespread ownership of the same high-growth, U.S.-centric, technology-heavy stocks—particularly those tied to AI and semiconductors. Marais argues that passive benchmarks and even many active funds have become dangerously homogenous, leaving portfolios vulnerable to systemic risks when market leadership shifts. The discussion highlights how valuation discipline, style diversification (especially value), geographic breadth (including overlooked markets like Korea and Japan), and active currency management are essential tools for building portfolios that can withstand multiple future scenarios. The episode concludes with a strong emphasis on preparation over prediction, advocating for a 'blank sheet of paper' approach to portfolio construction that prioritizes unique, undervalued opportunities across sectors and geographies. Key takeaways include: 1) True diversification requires more than just asset allocation—it demands diversification across styles, geographies, and valuations; 2) The dominance of growth stocks, especially in the U.S. and tech sector, creates systemic risk and reduces portfolio resilience; 3) Value investing remains relevant and offers compelling opportunities, particularly in overlooked markets and sectors like biopharma and mid-cap U.S. stocks; 4) Active currency positioning can mitigate risks from an overvalued U.S. dollar; 5) AI’s impact is broader than just the 'big tech' names—it’s also creating efficiency gains in logistics and other operational businesses. The episode underscores that the most effective diversification is intentional, disciplined, and rooted in deep fundamental analysis.