Beating the Market by 4% Sounds Like Skill. Jensen's Alpha Asks Whether You Just Took More Risk to Get There
7 min read
Loading...
7 min read
IHSG was down roughly 23 to 24% year to date as of early September 2026, according to Mandiri Sekuritas and Kontan's own market coverage. If your portfolio only fell 20% this year, you technically beat the market by 4 percentage points. That sounds like skill. Jensen's Alpha is the number that checks whether it actually was, or whether you just carried more risk than the index and got lucky it didn't hurt more.
Jensen's Alpha compares your portfolio's actual return to the return a standard pricing model says it should have earned, given how much market risk it carried. That model is the Capital Asset Pricing Model (CAPM), and the risk measure it uses is beta, how much your portfolio tends to move for every 1% the market moves. A beta of 1.5 means your portfolio typically swings 50% more than the market, in both directions.
α = Portfolio return − [Risk-free rate + Beta × (Market return − Risk-free rate)]
The part in brackets is CAPM's predicted return: the risk-free rate you'd earn doing nothing, plus your beta times however much extra the market returned above that risk-free rate. Whatever your portfolio actually returned above or below that predicted number is Jensen's Alpha. Positive means you beat what your own risk level should have produced. Negative means you underperformed it, even if your raw return looked fine.
Michael C. Jensen, then at the University of Chicago, introduced this measure in "The Performance of Mutual Funds in the Period 1945-1964," published in the Journal of Finance in 1968. He used it to test whether mutual fund managers could beat the market on a risk-adjusted basis, not just a raw one. His finding, that the average fund's Jensen's Alpha was close to zero or slightly negative even before fees, became one of the founding pieces of evidence behind index investing: most active managers were not adding value once the extra risk they took was priced in.
Picture two hypothetical portfolios that both returned -20% this year, while IHSG returned -24%. Both "beat the market" by exactly 4 percentage points. The risk-free rate anchor is Bank Indonesia's BI Rate, held at 5.75% at its most recent policy meeting in August 2026.
Portfolio A is defensive, built mostly around large, stable names. Its beta against IHSG is 0.6, meaning it should typically move only 60% as much as the index.
Portfolio B is aggressive, concentrated in smaller, more volatile positions. Its beta against IHSG is 1.5, meaning it should typically move 50% more than the index, in either direction.
| Portfolio A (defensive) | Portfolio B (aggressive) | |
|---|---|---|
| Actual return this year | -20% | -20% |
| IHSG's return | -24% | -24% |
| Raw outperformance | +4 points | +4 points |
| Beta | 0.6 | 1.5 |
| Risk-free rate (BI Rate) | 5.75% | 5.75% |
| CAPM expected return | -12.1% | -38.9% |
| Jensen's Alpha | -7.9% | +18.9% |
Both portfolios put up the exact same headline: beat IHSG by 4 points. Jensen's Alpha tells a completely different story about each one. Portfolio A's low beta means it should have been shielded from most of the market's fall, CAPM predicted it would only drop about 12.1%. It actually fell 20%, worse than a defensive portfolio should have, so its Jensen's Alpha is negative. Portfolio B's high beta means it should have fallen even harder than the market, CAPM predicted roughly -38.9%. It only fell 20%, far better than that much risk should have produced, so its Jensen's Alpha is strongly positive.
There is no fixed pass mark, but the sign and size carry a consistent meaning.
Positive alpha. The portfolio earned more than its beta alone would predict. This is the closest single number gets to evidence of real, risk-adjusted skill, or luck that has not yet run out.
Alpha near zero. The portfolio performed almost exactly as its own risk level predicted. Neither impressive nor disappointing once risk is accounted for.
Negative alpha. The portfolio earned less than its beta predicted, even if the raw return number looked acceptable or the portfolio still beat the index in absolute terms, the way Portfolio A did above.
A single year's alpha, positive or negative, is weak evidence on its own. The same way the information ratio separates a fund that beat its benchmark once from one that did it consistently, a Jensen's Alpha worth trusting is one that holds up across several years, not one good or bad stretch.
Jensen's Alpha is only as reliable as the beta it is built on, and beta is only reliable when R-squared says the benchmark actually explains your portfolio's movement. Feed a low-R-squared beta into the alpha formula and the entire result inherits that noise: a crypto-heavy portfolio's beta against IHSG, for instance, describes almost nothing real, so any alpha calculated from it is not trustworthy either.
CAPM itself is also a single-factor model. It explains return using market risk alone, and says nothing about size, value or momentum tilts that other, more advanced models (like the Fama-French three-factor model) separate out. A portfolio can show a positive Jensen's Alpha against a broad index simply because it leans toward small-cap or value stocks, categories that have historically outperformed for reasons that have nothing to do with the manager's own skill.
You need three inputs: your own return over a period, the risk-free rate for that period, and your portfolio's beta against a benchmark that actually fits it.
NetWort's Portfolio Health card already shows two of these three: your real return (calculated from your actual transaction history, not a tidied assumption) and your beta against the benchmark NetWort auto-selects for your holdings. It does not yet combine them into a Jensen's Alpha figure for you. Until it does, you can compute it yourself: take your IRR from Portfolio Health, your beta from the same card, the current BI Rate as your risk-free rate, and IHSG's return over the same period as your market return, all from the macro context page, and plug them into the formula above.
Do that before assuming a year of beating the index means your process is working. A defensive portfolio, like BBCA and similar large-cap holdings tend to produce, that merely loses less than the market in a down year is not automatically outperforming on a risk-adjusted basis. It might just be doing exactly what its low beta predicted, or worse.