What it is

Correlation measures the degree to which two holdings move together. It runs from +1 (they move in perfect lockstep) through 0 (their moves are unrelated) to -1 (they move in exact opposition). Position overlap is what happens when several names in a portfolio carry high positive correlation: on paper they are distinct tickers, but in practice they rise and fall as a group, so the portfolio behaves as if it holds one large position rather than several small ones.

This matters because almost every risk control - position sizing, sector caps, the comfort of "I'm in five different stocks" - silently assumes the positions are at least somewhat independent. When correlation is high, that assumption breaks, and the portfolio's true risk is far larger than the count of positions suggests. Correlation is the hidden variable that turns a diversified-looking book into a single concentrated bet.

How it works

Returns are correlated when the same underlying force drives them. Two regional bank stocks share exposure to interest rates, credit quality, and banking-sector sentiment; two semiconductor makers share the chip cycle; two oil producers share the crude price. When that shared driver moves, all the related names move together - and the diversification benefit you were counting on simply is not there.

The danger is that overlap is usually hidden. It does not show up in the ticker list and it does not show up in a naive position count. It hides in three common forms:

• Sector overlap: five "different" stocks that all sit in the same industry.
• Factor overlap: several names that are all high-growth, all high-beta, or all rate-sensitive, even across different sectors.
• Macro overlap: positions that all depend on the same single outcome - a soft landing, a falling dollar, a commodity boom.

The second trap is that correlation is not constant. In calm markets, holdings that look independent can drift apart and lull you into adding more of each. In a stress event, correlations across risk assets snap toward +1 and everything falls together - exactly when you needed the diversification most. So a portfolio that looks diversified in quiet times can reveal itself as a single bet on the worst possible day.

How to use it

The practical job is to size across correlated names rather than treating each in isolation. A workable checklist:

1. Group before you size. Before opening a position, ask which existing holdings share its primary driver - sector, factor, or macro theme. Mentally (or literally) put it in a bucket with its correlated cousins.
2. Budget risk per theme, not just per position. If your rule is "risk at most 1% of equity per trade," add a second rule: "risk at most 2-3% of equity per correlated group." Two highly correlated names sized at 1% each are really a single 2% bet, and your theme budget should treat them that way.
3. Treat correlation as size, not just count. Two names at correlation 0.8 are closer to one and a bit independent positions than to two. The higher the correlation, the more you should shrink the combined size to keep the group's risk inside your limit.
4. Stress-test for the bad day. Ask: "If this whole theme dropped 15% tomorrow, what would my account do?" If the answer is uncomfortable, you are overlapped, regardless of how many tickers you hold.

Worked example

Suppose you risk 1% of a $50,000 account per position, so $500 of risk each. You like three regional banks and open all three, feeling "diversified across three names." But their correlation is roughly 0.85 - they trade almost as one. A bad bank-sector headline hits all three stops at once, and you lose about $1,500, or 3% of the account, in a single event. You did not take three independent 1% risks; you took one 3% risk on the banking sector. Had you grouped them first and capped the banking theme at 1.5% total, you would have sized each at roughly $250 of risk and kept the real exposure where you intended it.

Strengths & limits

Thinking in correlated groups is powerful because it aligns your risk controls with how the portfolio actually behaves under stress, not how it looks on a quiet day. It catches the concentration that position counts and sector labels miss, and it stops you from quietly turning five small trades into one oversized bet.

The limits are worth respecting. Correlation is a backward-looking, unstable number - it is estimated from past data and changes with the regime, so treat it as a rough guide, not a precise dial. It also says nothing about direction or magnitude; two assets can be uncorrelated yet both crash for unrelated reasons. And forcing artificial "diversification" by adding low-correlation positions you do not understand simply swaps concentration risk for ignorance risk. Use correlation to avoid accidental overlap, not as an excuse to own things you cannot explain.

Key takeaway: Correlated positions behave as one bet, so hidden concentration hides in sector, factor, and macro overlap; size across correlated names by budgeting risk per theme - and remember correlations spike toward +1 exactly when you need diversification most.