How do I interpret covariance in a portfolio context, and why is correlation often more useful?
I understand that covariance measures how two variables move together, but the actual number doesn't seem intuitive. For example, if the covariance between Stock A and Stock B is 0.0038, what does that even mean? Why do we often convert it to correlation instead? A practical portfolio example would really help me understand this.
Covariance captures the directional co-movement between two variables, but its magnitude is hard to interpret because it depends on the units of measurement. Correlation standardizes covariance to a scale of -1 to +1, making it immediately comparable across any pair of assets.
Formula Relationship:
Cov(A, B) = ρ(A, B) × σ_A × σ_B
Or equivalently: ρ(A, B) = Cov(A, B) / (σ_A × σ_B)
Portfolio Example — Meridian Tech and Crestline Energy
Suppose you hold two stocks:
| Meridian Tech (M) | Crestline Energy (C) | |
|---|---|---|
| Expected return | 14.2% | 8.6% |
| Std deviation | 22.0% | 15.0% |
| Covariance (M, C) | 0.0038 | — |
The covariance of 0.0038 tells you the stocks tend to move in the same direction (positive sign), but 0.0038 alone does not tell you whether that linkage is strong or weak.
Converting to correlation:
ρ = 0.0038 / (0.22 × 0.15) = 0.0038 / 0.033 = 0.115
A correlation of 0.115 is quite low — the stocks have only a mild positive relationship. This is valuable information: adding Crestline Energy to a portfolio of Meridian Tech provides meaningful diversification benefit.
Portfolio Variance (equal weights):
σ²_p = w²_M × σ²_M + w²_C × σ²_C + 2 × w_M × w_C × Cov(M, C)
σ²_p = 0.5² × 0.0484 + 0.5² × 0.0225 + 2 × 0.5 × 0.5 × 0.0038
σ²_p = 0.0121 + 0.005625 + 0.0019 = 0.019625
σ_p = 14.0%
Notice the portfolio standard deviation (14.0%) is lower than the weighted average of individual standard deviations (18.5%), illustrating the diversification benefit from the low covariance.
Why Correlation Is More Useful:
- Covariance is unbounded and unit-dependent — comparing Cov(A,B) = 0.0038 with Cov(C,D) = 0.012 is meaningless without knowing the volatilities
- Correlation is bounded [-1, +1] and unit-free, so you can instantly rank asset pairs by co-movement strength
- Both feed identically into portfolio variance; they carry the same information
Exam Tip: For CFA Level I, know how to convert between covariance and correlation, and remember that lower correlation means greater diversification benefit. Questions often present covariance and ask you to assess diversification potential.
Join our CFA Level I community for more quantitative methods discussions.
Master Level I with our CFA Course
107 lessons · 200+ hours· Expert instruction
Related Questions
What exactly is the Capital Market Expectations (CME) framework and why does it matter for asset allocation?
How do business cycle phases affect asset class return expectations?
Can someone explain the Grinold–Kroner model step by step with numbers?
How do you forecast fixed-income returns using the building-blocks approach?
PPP vs Interest Rate Parity for forecasting exchange rates — when do I use which?
Join the Discussion
Ask questions and get expert answers.