Low-Correlation Concentration
Concentrate in distinct ideas without letting one theme control the portfolio
- Difficulty
- Advanced
- Time to result
- ~ongoing to results
- Steps
- 6
- Confidence
- 96%
A concentrated portfolio does not become safely diversified merely because it holds several company names. The relevant question is whether its largest positions respond to the same underlying economic force. Begin by identifying each holding's principal return drivers, then group holdings with shared dependencies such as artificial-intelligence spending, interest rates, commodity prices, or consumer demand. Examine how those groups behave together, particularly during stress. Keep the portfolio focused enough to maintain high active share, but combine companies whose businesses and catalysts differ materially. Andrew Slimmon illustrates the mechanism with large positions in Nvidia, cement producer CRH, and asset manager Ameriprise: technology, basic materials, and finance exposures that are less likely to move in unison. Correlations must be monitored continuously because apparently distinct positions can converge around one dominant thesis.
Origin
Extracted from Masters in Business, where Andrew Slimmon describes the correlation discipline behind his concentrated equity portfolios.
Core principles
- 01Concentration should reflect conviction rather than a single hidden macro bet.
- 02A small number of holdings can remain diversified when their return drivers differ.
- 03Correlation matters more than superficial differences in company names.
- 04The largest positions should not all succeed or fail for the same reason.
- 05Portfolio risk must be evaluated at the combined-position level.
How to run it
- 1
Define each investment thesis
Write down why each major holding should produce returns and which variables could invalidate that thesis. Focus on economic drivers rather than sector labels alone.
Pro tip Express each thesis as a short cause-and-effect statement.
Watch out Vague theses make hidden overlap difficult to detect.
- 2
Map shared dependencies
Group holdings that rely on the same technology cycle, commodity, rate environment, customer base, regulation, or macro outcome. Treat differently named companies as one risk cluster when their success depends on the same cause.
Pro tip Include suppliers and customers in the same thematic map when appropriate.
Watch out Companies in different official sectors can still be highly correlated.
- 3
Examine correlation
Review historical co-movement and reason through how positions may behave during a market shock. Give extra attention to the portfolio's largest weights.
Pro tip Compare both ordinary periods and stressed periods because correlations often rise in crises.
Watch out Historical correlation can understate a newly emerging shared exposure.
- 4
Break dominant clusters
Trim, replace, or resize holdings when one theme controls too much of the portfolio's outcome. Preserve the strongest idea while adding return drivers that do not depend on it.
Pro tip Diversify the source of expected returns, not simply the ticker count.
Watch out Adding weaker names from the same theme does not reduce thematic concentration.
- 5
Balance conviction and independence
Build a limited set of meaningful positions whose theses are individually strong and collectively distinct. Ensure no single economic development can invalidate most of the portfolio.
Pro tip Ask what each new holding contributes that existing holdings do not.
Watch out Excessive diversification can turn the portfolio back into a closet index.
- 6
Monitor changing relationships
Revisit the dependency map as business models, market narratives, and macro conditions evolve. Adjust before a once-diverse portfolio becomes dominated by a common factor.
Pro tip Repeat the review after major earnings reports or large changes in position weights.
Watch out Correlation risk can grow silently as one winning theme expands within the portfolio.
In the wild
A global concentrated portfolio holds large positions in Nvidia, cement producer CRH, and asset manager Ameriprise. Each is a meaningful conviction, but their principal business drivers differ across technology, basic materials, and finance, making it less likely that all three will rise or fall for the same reason.
→ The portfolio maintains concentration and high active share without depending entirely on one thesis.
A 20-stock portfolio owns a chip designer, semiconductor equipment maker, data-center utility, and AI software vendor as four of its five largest positions. The manager recognizes that all depend heavily on continued AI capital spending, trims two positions, and reallocates to businesses driven by infrastructure replacement and household savings.
→ A reversal in AI spending no longer threatens most of the portfolio's largest positions simultaneously.
Common mistakes
Concentrating only the best recent winners
A manager's top-performing ideas may all reflect the same successful theme. Moving them into one focused fund can magnify a single point of failure.
Counting tickers instead of risks
Several stocks do not provide meaningful diversification when they share the same economic driver. Portfolio construction must examine dependencies and co-movement.
Using concentration without monitoring
Position weights and correlations change over time. A portfolio that began with balanced drivers can become dominated by one appreciating cluster.
Is it for you?
Best for
It is best for active equity managers and investors building focused portfolios from a limited number of high-conviction positions.
Not ideal for
It is not ideal for investors unable to research individual businesses or tolerate the volatility of a concentrated strategy.
From the transcript
“The problem is a lot of times those best ideas are highly correlated.”
“focusing on what is the correlation of the stocks in the portfolio are supremely, supremely important.”
“making sure they don't all zig and zag together”
From the episode
At the Money: Avoid Closet Indexing