Your Portfolio's Beta Is Meaningless If R-Squared Says It's Not Even Related to That Benchmark
6 min read
Loading...
6 min read
You check your portfolio's beta on NetWort's Portfolio Health card and see 0.85, next to a badge that reads "Calm." That looks like a clean, reassuring signal: your portfolio moves less than the market. But beta only measures the slope of a relationship between your portfolio and whatever benchmark it was compared against. It says nothing about how strong, or how real, that relationship actually is. A second number, R-squared, answers that question directly. Skip it, and a 0.85 beta computed against a benchmark that barely explains your portfolio's moves can look exactly as convincing as one computed against a benchmark that actually drives most of what happens to your money.
R-squared, written R², measures how much of one thing's movement is statistically explained by another thing's movement over the same period, on a scale from 0 to 100%. In portfolio terms: how much of your portfolio's swings can be accounted for by the benchmark's own swings on those same days.
The simplest way to remember it: R-squared is the correlation coefficient between your portfolio's returns and the benchmark's returns, squared. A correlation of 0.90 gives an R-squared of 0.81, or 81%. A correlation of 0.22 gives an R-squared of about 0.05, or 5%. This definition is consistent across independent finance-education sources, including Foliolytic's investing glossary and the Institute of Business & Finance's risk-metrics primer written for financial advisors.
R² = (correlation between portfolio and benchmark returns)²
Beta tells you the size and direction of a relationship: for every 1% the benchmark moves, how much your portfolio tends to move with it, on average. R-squared tells you whether that relationship is strong enough to trust in the first place.
Multiple finance-education sources converge on similar working bands for how much to trust a beta once you know its R-squared:
Above 70%. A strong, reliable relationship. Beta is meaningful.
Between 40% and 70%. A moderate relationship. Beta gives some signal, but treat it as a clue, not a settled fact.
Below 40%. A weak relationship. The benchmark barely explains what happens to the portfolio, and beta against that specific benchmark is close to statistical noise: a number the math produced, not a stable description of how the two actually move together.
Say you're comparing two hypothetical portfolios against IHSG, the Indonesian benchmark index. Illustrative numbers only.
Portfolio A holds large IDX banking and consumer names, BBCA among them. Its monthly returns correlate with IHSG at 0.90.
Portfolio B is crypto-heavy, built mostly around Bitcoin and a couple of other large-cap coins. Its monthly returns correlate with IHSG at just 0.22, because crypto largely trades on its own cycle, independent of what Indonesian equities do that day.
| Portfolio A | Portfolio B | |
|---|---|---|
| Correlation with IHSG | 0.90 | 0.22 |
| R-squared | 81% | 5% |
| Beta vs IHSG | 0.87 | 0.82 |
Both portfolios show a beta just under 1, which on its own reads as "slightly calmer than the market" for either one. R-squared says that reading only holds up for Portfolio A. IHSG genuinely explains 81% of what happens to Portfolio A month to month, so its 0.87 beta is a real, usable description of how it behaves relative to Indonesian stocks. IHSG explains only about 5% of what happens to Portfolio B. Its 0.82 beta is arithmetically real, correlation and covariance can always be computed, but it describes almost nothing about how Portfolio B will actually behave, because IHSG was barely involved in the first place. What actually drives Portfolio B's swings is crypto-specific: exchange flows, funding rates, the wider crypto cycle, not the Jakarta Composite.
NetWort's asset detail pages show beta for a single stock with a plain-language label instead of a raw number: "Calm," "Follows Market" or "Amplified," depending on the value. The Portfolio Health card on your dashboard shows a portfolio-level beta the same way. Neither view currently surfaces R-squared alongside it, so nothing on the page tells you whether a given beta is a real signal or the kind of statistical noise Portfolio B produced above.
There's a second detail worth knowing if you hold a mostly Indonesian portfolio. NetWort's beta calculation auto-selects its benchmark by what your portfolio actually holds (backend/app/calculations/beta.py): IHSG for a portfolio that's mostly IDX-listed, the S&P 500 for one that's mostly US-listed, and similar rules for London and Hong Kong. That's the right design, since a beta measured against the wrong market is not a small error. But the tooltip text on the Portfolio Health card currently describes the number as measured "vs the S&P 500" regardless of which benchmark your own portfolio actually used, so an Indonesia-heavy portfolio's beta can be computed correctly against IHSG while its tooltip still names the wrong index. Worth knowing if you're reading that number closely, and one more reason not to take a beta figure, from NetWort or anywhere else, at face value without checking what it was actually measured against.
NetWort doesn't currently show R-squared as its own number. Until it does, treat any beta figure, from NetWort or anywhere else, with more skepticism the less obviously related the asset and the benchmark are. A beta computed for a large IDX bank stock against IHSG deserves real trust. A beta computed for gold, Bitcoin, or a single volatile small-cap stock against a broad index deserves a second look, because a low but real correlation can still produce a beta number that reads exactly as clean and confident as a high one.
Check your portfolio's beta in the Portfolio Health section of your dashboard, and treat it as a starting point rather than the final word, especially for anything that isn't a plain, domestic stock. If the relationship to the benchmark feels shaky, your portfolio's own volatility describes its actual risk without needing any benchmark relationship to be real in the first place. For the current IHSG level and the wider Indonesian macro picture behind any benchmark comparison, see the macro context page.