Robo-Advisers for Beginners: Four Jobs They Do Well, One They Can't

The week the autopilot gets tested#

The Federal Reserve announces its decision tomorrow afternoon, and for once the argument is not about direction. A quarter-point rise is widely expected, which would be the first increase since 2023. The two-day meeting began this morning, on the dates in the Fed's own calendar. Oil has climbed about 20% this month.

If you opened your first investment account any time in the past three years, you have never watched a rate rise land on it. Which makes this a fair week to ask whether a piece of software ought to be handling it for you.

Plenty of people are asking. The Financial Conduct Authority reckons roughly 7 million UK adults hold £10,000 or more in cash without having considered investing any of it, and fewer than one in ten take regulated financial advice at all. Automated investing was built to sit in that gap.

What a robo-adviser actually is#

Strip away the branding and a robo-adviser (spelled "robo-advisor" in the United States) is three pieces of plumbing bolted together.

First, a questionnaire. You answer perhaps a dozen questions about your age, your goal, when you need the money and how you feel about losing some of it. Second, a model portfolio. The software maps your answers onto one of a fixed set of allocations, usually built from exchange-traded funds, which are baskets of shares or bonds that trade like a single share and charge very little. A "60/40" portfolio means 60% shares and 40% bonds. Third, a set of standing instructions. The software buys the funds, and then keeps buying and selling small amounts to hold your mix steady as prices move, a process called rebalancing.

The legal position is less exotic than the word "robo" suggests. In the United States these firms register as investment advisers under the Advisers Act and owe clients a fiduciary duty, exactly as a human does. In the UK they are FCA-authorised and sit under the same Consumer Duty.

Scale has followed. Vanguard Digital Advisor was reported at roughly $312 billion in mid-2024 and Wealthfront at about $93 billion in March 2026, on one compilation of company disclosures that mixes reporting bases.

The four jobs an algorithm does better than most beginners#

The first is having an allocation at all. A beginner who buys three funds they read about in January and two more in June owns something nobody designed. A model portfolio is at least deliberate.

The second is rebalancing. Vanguard's own research puts its value at 26 to 86 basis points a year, a basis point being one hundredth of a percentage point. It is unglamorous work that almost nobody does by hand.

The third is automating contributions, which removes the monthly decision about whether this is a good moment to buy. No study cleanly isolates that effect, so treat it as plausible rather than measured.

The fourth is tax-loss harvesting, and it is the one with real numbers attached. The technique means selling a holding that has fallen, booking the loss against your tax bill, and buying something similar so you stay invested. Chaudhuri, Burnham and Lo tested it over US data from 1926 to 2018 and found an average annual benefit of 108 basis points before costs, falling to 82 once the American wash-sale rule binds, with trading costs taking roughly another 13. The average also hides a wide spread: about 213 basis points across the volatile 1926 to 1949 stretch, about 51 in the calmer years from 1949 to 1972. Wealthfront tells prospective clients the feature "can typically cover our annual fee more than 6x over" for Classic portfolio holders, which is the firm's own marketing claim about its own product and should be read that way.

British readers should note that most of this does not apply to them. Money inside an ISA or a pension has no capital gains tax to offset, so there is nothing to harvest.

There is also evidence that automation helps most when things are bad. Liu, Yang and Wen analysed daily portfolio and transaction data from an online platform and found robo-adviser users outperformed matched human investors by 12.67% across the COVID-19 crash window, in work published online in Production and Operations Management in June 2023. The software cut risky holdings; the humans mostly froze. In normal conditions the two groups performed about the same. That is one platform and one crisis, and 12.67% is the gap over the episode rather than an annual figure, so the result travels less far than the headline suggests.

What it costs, including the bits that never appear on the fee page#

Three costs hide outside that table.

Fund charges sit on top of the advisory fee and come out inside the funds, so you never see them leave your account. Vanguard's headline advisory fee is 0.20% for its all-index option, which revenue credits cut to somewhere around $15 to $20 per $10,000 depending on what you hold. The fund expense ratios are charged separately.

Cash is the second. In June 2022 the SEC fined three Schwab investment adviser subsidiaries $187 million, comprising about $52 million in disgorgement and interest and a $135 million penalty, over the cash held in Schwab Intelligent Portfolios between March 2015 and November 2018. The firm's own internal analyses showed the allocations "would cause clients to make less money even while taking on the same amount of risk", and Schwab earned the spread by sweeping that cash to an affiliate bank. Enforcement director Gurbir Grewal put it bluntly: the cash level "was decided by how much money the company wanted to make". Schwab advertised no advisory fee. The revenue came out of the cash instead.

The third is dispersion. Condor Capital's Robo Report tracks roughly thirty 60/40 portfolios. In the first quarter of 2026, net-of-fee returns ran from +2.56% at Empower down to -0.87% at both E*Trade Core SRI and Robinhood, a spread of more than three percentage points in one quarter among products that describe themselves in almost identical language. The label tells you how something is sold, not what is inside it.

Where the questionnaire stops asking#

The SEC ran an examination sweep of firms providing electronic investment advice and published its findings in a risk alert on 9 November 2021. Nearly all the advisers examined received deficiency letters. Examiners found firms that "relied on just a few data points to formulate investment advice", and most failed to check periodically whether a client's circumstances had changed. More than half had problems with their advertising, including hypothetical performance applied retroactively without proper risk disclosure.

Those are compliance findings, not proof that any particular portfolio was wrong. They do point at the structural limit. A dozen questions cannot see a credit card balance at 24%, a business with lumpy cash flow, a spouse's pension, an inheritance that is coming, or a marriage that is ending. Each of those changes the right answer more than the choice between a 60/40 and a 70/30 allocation does.

Regulators have noticed the same hole from the other side. The FCA's targeted support regime went live on 6 April 2026, creating a middle rung between generic guidance and full advice, where firms can make suggestions to groups of consumers who share characteristics. The FCA estimates at least 18 million people could receive it over the next decade. The regulator's view is that neither a questionnaire nor a £2,000 advice bill serves most households.

The one thing they still cannot do#

Every mechanical advantage described above assumes you leave the account alone, and nothing in the software makes you. The liquidation button sits on the same screen as the balance, in your pocket, available at eleven at night after a bad day of headlines. No algorithm stands between your thumb and that button, and none was built to.

This is the largest single component of what human advisers are paid for. Vanguard's framework assigns behavioural coaching a value of 150 basis points a year, the biggest item in a total of roughly 3%, and consistent across every region it studied. Vanguard is careful to say the benefit is "very irregular" rather than an annual entitlement, and it arrives in lumps during exactly the sort of week the Fed is about to deliver.

How large is the gap it fills? That is disputed, and the argument is live. Morningstar's Mind the Gap study, published on 6 August 2026, found the average dollar in US funds earned 8.7% a year over the decade to 31 December 2025 against 9.9% for the funds themselves, a shortfall of 1.2 percentage points from badly timed buying and selling. Municipal bond fund investors captured only 49% of their funds' returns; US equity fund investors captured 97%.

Then in May 2026 four academics took the same data apart in the Financial Analysts Journal and concluded that poor timing costs investors about 0.10% a year, roughly twelve times less than Morningstar's estimate, arguing the gap calculation attributes to timing what other factors explain.

So the size of the effect is unsettled. The mechanism is easier to see. The Carlson School study found robo users did better because the software cut risk while their human counterparts sat still, and that advantage lasts only as long as the software still holds the account.

The fee ladder in the table above is partly a ladder of who talks you down. Betterment charges 0.25% to run the portfolio and 0.65% to add a person you can ring. That 40 basis point difference is what one large firm reckons it is worth to have someone answer the phone on the day you decide to sell everything. Whether it is worth it to you depends on how you behaved the last time prices fell.

Key takeaways#

1. Robo-advisers do four mechanical jobs competently: setting an allocation, rebalancing, automating contributions and, in taxable accounts, harvesting losses. The measured benefits are real but modest and vary enormously by period.
2. Cost is not the same as the advertised fee. Fund charges sit on top, and cash allocations cost Schwab $187 million in SEC penalties when they were not properly disclosed.
3. These products are not interchangeable. Net returns on comparable 60/40 portfolios varied by more than three percentage points in the first quarter of 2026 alone.
4. The questionnaire is the binding constraint. The SEC found firms advising on "just a few data points" and rarely rechecking whether a client's life had changed.
5. No robo-adviser can hold you to your own plan. Vanguard values that function at 150 basis points a year, and rival academic estimates of the cost of bad timing differ by a factor of twelve, so treat any precise figure with caution.

Frequently asked questions#

Are robo-advisers worth it for beginners?

For someone with a lump of cash, no investments and no appetite for research, a 0.25% fee buys a diversified portfolio and removes several decisions. Whether it beats buying a single global index fund yourself depends on whether you would actually do the rebalancing. Nothing here is a recommendation about your own money.

What is the difference between a robo-adviser and a financial adviser?

A robo-adviser manages an investment portfolio according to rules. A financial adviser also considers debt, tax, insurance, property, pensions and family circumstances, and is available to talk you out of a decision. The fee difference, commonly 0.25% against 0.5% to 1%, mostly pays for that second list.

Do robo-advisers beat the market?

They are not designed to. Most hold index funds and aim to match a benchmark after costs. The Robo Report's data shows some trailing their own benchmarks by one to three percentage points a year over multi-year periods, largely on fees and allocation choices.

Is my money safe with a robo-adviser?

Custody protections apply in the normal way, through SIPC in the US and FSCS in the UK, though these cover firm failure rather than investment losses. The SEC has separately criticised robo firms for misrepresenting what SIPC actually protects, so read the disclosure rather than the marketing.

What happens to a robo portfolio if markets fall this week?

The software will hold the allocation and rebalance towards it, which means buying more of whatever has fallen. That is the design. Whether you let it happen is the open question.

Does tax-loss harvesting work in the UK?

Not inside an ISA or a pension, where there is no capital gains tax to offset. It only matters in a general investment account. UK share-matching rules also work differently from the American wash-sale rule: a repurchase within 30 days is matched against the sale, so the loss does not arise in the first place.

Should I use the hybrid tier with a human adviser?

It costs roughly 40 basis points more at Betterment and requires $500,000 at Vanguard's Select service. The case for paying it rests on whether you expect to want a conversation during a bad month.

Glossary#

Robo-adviser An FCA-authorised or SEC-registered firm that builds and manages an investment portfolio through software, usually from a short questionnaire.

Exchange-traded fund (ETF) A basket of shares or bonds that trades on an exchange like a single share, typically tracking an index at low cost.

Rebalancing Buying and selling to return a portfolio to its target mix after prices have drifted.

Basis point One hundredth of a percentage point. 150 basis points is 1.5%.

Tax-loss harvesting Selling a holding at a loss to offset taxable gains, then buying something similar to stay invested.

Wash-sale rule A US tax rule that disallows a loss if you repurchase a substantially identical asset within 30 days. The UK has no direct equivalent: its 30-day rule instead matches the sale against the repurchase, so the loss is never realised at all.

Fiduciary duty A legal obligation to act in the client's best interest, which applies to registered investment advisers whether human or automated.

Targeted support A UK regulatory category live since 6 April 2026, allowing firms to make suggestions to groups of consumers with shared characteristics without providing full personal advice.

References#

1. Financial Conduct Authority, PS25/22: Supporting consumers' pensions and investment decisions: rules for targeted support.
2. Financial Conduct Authority, "Millions of people set to get extra help with investments and pensions decisions", 11 December 2025.
3. US Securities and Exchange Commission, Division of Examinations, Observations from Examinations of Advisers that Provide Electronic Investment Advice, 9 November 2021.
4. US Securities and Exchange Commission, Schwab Subsidiaries Misled Robo-Adviser Clients about Absence of Hidden Fees, 13 June 2022.
5. S. E. Chaudhuri, T. C. Burnham and A. W. Lo, An Empirical Evaluation of Tax-Loss-Harvesting Alpha, Financial Analysts Journal, Q3 2020.
6. J. A. Fulkerson, B. D. Jordan, T. B. Riley and Q. Yan, Bad Timing Does Not Cost Investors 15% of Their Funds' Returns, Financial Analysts Journal, vol. 82 no. 3, May 2026.
7. Morningstar, Mind the Gap 2026, published 6 August 2026.
8. Carlson School of Management, University of Minnesota, Robo-Advisors Outperform Human Investors Amid Market Crash, reporting research published in Production and Operations Management, June 2023.
9. Vanguard, Putting a value on your value: quantifying Vanguard Adviser's Alpha.
10. Condor Capital Wealth Management, The Robo Report, Total Portfolio Returns, data to 31 March 2026.
11. Betterment, Pricing.
12. Wealthfront, Pricing.
13. Vanguard, Digital Advisor and Personal Advisor.
14. Board of Governors of the Federal Reserve System, FOMC meeting calendar, 2026.
15. Kiplinger, September Fed meeting: live updates and commentary.
16. Investing in the Web, The largest robo-advisors by AUM in 2026.