Correlation risk
The risk that several positions move together because their prices are correlated, so losses arrive at the same time. Two trades in closely linked markets can behave like one larger position.

- Level
- Intermediate
- Reading
- 2 min
- Sources
- 4Original authorInternational bodyRegulatorBook
- Updated
On this page· 6 sections
Two positions can sit in different markets and still lose money on the same afternoon. If their prices tend to move together, one shift in the market reaches both, and the account carries more risk than either trade shows on its own.
That hidden overlap is correlation risk. It grows when several positions share a driver, such as the same currency on one side of every pair.
How it works
The correlation coefficient r measures how strong a straight-line relationship between two series is, and it always lies between −1 and +1[4]. At +1 the correlation is perfectly positive, at −1 perfectly negative, and at 0 there is likely no linear relationship[4]. A strong correlation says nothing about cause[4].
Markowitz built portfolio theory on measuring risk for the portfolio as a whole. A portfolio’s variance is the variance of a weighted sum, so it contains a covariance term for every pair of holdings[1].
Bank capital rules add risks up the same way. Within a group of risk factors, the Basel market risk standard combines weighted positions under a square root that holds a correlation term for each pair[2]. For two positions of size a and b with correlation ρ, that is √(a² + b² + 2ρab). At ρ = 1 the result is a + b, the full sum. At ρ = 0 it is smaller, and at a negative ρ smaller still.
The same standard does not trust correlations to stay put. Under its sensitivities-based method, banks repeat the calculation with correlations multiplied by 1.25 (capped at 100%) and with lower ones, to address the risk that correlations rise or fall in financial stress, and they hold capital for the largest of the three results[2].
Example
Say an account holds two positions, each with a typical daily swing, measured as a standard deviation, of $200.
- Unrelated prices, ρ = 0: √(200² + 200²) ≈ $283 for the pair.
- ρ = 0.8: √(40,000 + 40,000 + 64,000) = √144,000 ≈ $379.
- ρ = 1: $400, the same as one position twice the size.
- Opposite directions on two instruments with ρ = 0.8 act like a correlation of −0.8: √(80,000 − 64,000) ≈ $126.
At 0.8 the two trades carry about 95% of the swing of one doubled trade.
Watch out
Correlation is measured on the past. The Basel stress scenarios exist because it can change when markets are under strain[2].
Different names can hide one exposure. Investor.gov makes the point for funds: someone holding several may think they are diversified, and it suggests checking that their top holdings differ[3]. Several currency pairs that share one currency raise the same question.
A correlation near zero is not proof that two prices are unrelated. Data with a curved pattern can show a correlation of 0[4].
At TRONEXO
On the simulated Challenge accounts and TRONEXO Accounts, hedging is prohibited, and that includes correlation hedging: offsetting positions on different but reliably correlated instruments. See the FAQ.
Sources
- Original authorThe Nobel Prize (nobelprize.org)Foundations of Portfolio Theory (Harry M. Markowitz, Nobel Lecture, 7 December 1990)Checked November 9, 2026
- International bodyBank for International Settlements (Basel Committee on Banking Supervision)Minimum capital requirements for market risk (BCBS d457, January 2019)Checked November 9, 2026
- RegulatorU.S. SEC (Investor.gov)Asset Allocation and DiversificationChecked November 9, 2026
- BookOpenStax (Rice University)12.3 The Regression Equation (Introductory Statistics, Illowsky & Dean)Checked November 9, 2026
