Do Robo-Advisors Beat the Market? No, and They Are Not Trying To
Last updated August 2026
Short answer
No, and the more useful point is that they are not attempting to. A robo-advisor assembles index funds, which are built to match their benchmarks rather than exceed them, then charges a fee on top. The expected result is the market minus costs, and that is the design working rather than failing. Most people asking this are comparing a diversified portfolio holding bonds against the S&P 500, which holds neither bonds nor anything outside the US. Judge one on four other things instead. Walnut is informational and is not an investment adviser.
Someone asking this has usually been told that automated investing is sophisticated, and reasonably expects sophistication to show up as higher returns. It does not, and finding that out from a statement two years in feels like a failure. It is worth understanding before that happens, because the product is fine and the expectation was wrong.
Four reasons it cannot, structurally
1. It buys the market on purpose
The portfolio is assembled from index funds tracking broad markets. An index fund is designed to match its benchmark, not exceed it, so a portfolio of them cannot beat the thing it is made of. This is the design rather than a shortfall in execution.
2. The fee comes off the top
Whatever the funds return, the management fee is deducted from it every year. A portfolio built to match the market and then charged a fee returns slightly less than the market, reliably, in the same way that arithmetic is reliable.
3. It holds bonds, and you are comparing against stocks
The comparison people make is against the S&P 500, which is entirely stocks. A robo portfolio with a bond allocation will lag it in strong years and fall less in bad ones. That is not underperformance, it is a different question being answered.
4. It is diversified beyond the index you are watching
International and small-company holdings drag when large US companies lead and help when they do not. Judging a globally diversified portfolio against the single index that happened to lead is a comparison chosen after the fact.
The third and fourth are the ones causing most of the disappointment, and neither is really about the robo-advisor. They are about which benchmark got picked. A portfolio holding bonds and international stocks compared against an all-stock US index is being measured against something it was deliberately built not to be.
What to judge one on instead
| Question | Why it is the right one |
|---|---|
| Did it match its own benchmark | The right question. A blended benchmark matching the actual allocation, not the S&P 500 |
| What did it cost, all in | Management fee plus the underlying fund expense ratios. Both, added together |
| Did it keep you invested | The largest determinant of your outcome, and the one nothing else on this list measures |
| Did harvesting produce anything real | In a taxable account only, measured against your actual tax bill rather than as a headline |
The third row is the one no review scores and the one that decides outcomes. A portfolio that returned slightly less and that you held through a bad year beats a better portfolio you abandoned, and the difference between those two paths is far larger than anything in the fee comparison.
Get a recommendation for your situation
Walnut is the AI that knows your portfolio: ask anything in plain English, research any fund, and get an honest second opinion. On the broker you already use, read-only, and you approve every trade. Walnut is not a registered investment adviser.
Three things a robo can genuinely add, none of them returns
You stayed invested through a decline
The gap between what funds return and what investors in those funds actually earn is well documented and comes almost entirely from buying and selling at the wrong moments. A managed account that is boring to look at closes part of that gap, and it dwarfs any plausible edge from security selection.
You started at all
Money that sat in checking for three years earned nothing while the decision was postponed. Against that baseline the return difference between good options is a rounding error.
Harvested losses reduced a real tax bill
In a taxable account with gains to offset, this produces an actual number in your favour. It is the one robo feature that can add measurable value beyond market returns, and it is bounded and situational rather than a performance edge.
Only the third produces a number you can point at, and it applies in taxable accounts alone. The first two are worth more and are impossible to measure for any individual, because the comparison is against a version of you who did something different. See how harvesting actually works.
If beating the market is genuinely the goal
Then a robo-advisor is the wrong purchase and it will not become the right one. What that goal requires is active selection, which means either choosing holdings yourself or paying someone whose record you have examined over a long enough period to distinguish skill from luck, and the evidence on how often that succeeds after costs is not encouraging. The realistic version is to keep the core of the money in something broad and cheap and to run any conviction in a deliberately smaller account, where being wrong is affordable.
Related: are robo-advisors worth it, and why robo returns are so hard to compare.
FAQ
Do robo-advisors beat the market?
No, and they are not built to. A robo-advisor assembles index funds, which are designed to match their benchmarks rather than exceed them, and then charges a fee on top. The expected result is the market minus costs, which is the design working correctly rather than failing.
Why is my robo-advisor underperforming the S&P 500?
Almost certainly because it holds bonds and international stocks and the S&P 500 does not. A diversified portfolio lags a concentrated one when that concentration is what is leading, and falls less when it is not. Compare against a blended benchmark matching your actual allocation instead.
What return should I expect from a robo-advisor?
Roughly what a portfolio of that stock and bond mix returns, minus the management fee and the underlying fund costs. There is no separate robo-advisor return: it is determined by the allocation you were assigned, which is why the questionnaire matters more than the provider.
Can any robo-advisor beat the market?
Not by design, and a provider claiming otherwise is describing active management with a different name. What can produce results better than a passive baseline is tax-loss harvesting in a taxable account, which reduces your tax bill rather than raising your pre-tax return.
Are robo-advisors better than doing it myself?
On returns, essentially identical, because you would be buying similar funds. The differences are cost, which favours doing it yourself, and behaviour, which frequently favours the managed account. Which matters more is a fact about you rather than about the products.
How should I judge a robo-advisor then?
Four things: whether it matched a benchmark reflecting its actual allocation, what it cost all in, whether it kept you invested, and whether harvesting produced a real number on your tax return. Only the first two appear in most reviews, and the third is the one that decides outcomes.
Do robo-advisors use AI to pick investments?
Mostly not in any meaningful sense. The allocation comes from long-standing portfolio theory, and the automation is in the plumbing: rebalancing, harvesting, contributions. The word appears in marketing far more often than anything predictive appears in the product.
Is a robo-advisor worth it if it just matches the market?
Matching the market at low cost is a perfectly good outcome and better than most people achieve unaided. The question is whether the fee buys enough of the other things, staying invested, harvesting, aggregation, to justify itself against a target-date fund that does much of it for less.
Related articles
Walnut is informational and is not an investment adviser, and nothing here is investment advice. Past performance does not indicate future results, and outcomes depend on allocation, costs and timing rather than on any provider.