How to compare ETFs and funds

Two instruments aiming at the same target, global equities say, can start at different points, in different currencies, with different costs. Looking only at last year's return almost always leads to the wrong conclusion. Comparing ETFs and funds properly means putting them on the same starting line, checking them across more than one time window, and understanding how much they move together before you pick one.

ETFs and funds: what actually differs

An ETF trades on an exchange all day, like a stock: you can buy it at 10am and sell it at 3pm, at a price forming continuously on the market. A fund is priced once a day: whoever buys or redeems shares gets the NAV calculated at day's end, without knowing the exact execution price in advance.

The second difference is cost. It isn't an absolute rule, but it's a strong pattern: passively managed ETFs and index funds have structurally lower TER, because they track an index instead of paying a team to pick stocks. Actively managed funds, the ones trying to beat the market, cost more on average whether or not they succeed. A TER of 0.20% a year, typical of many index ETFs, and a TER of 1.5% for an ordinary active fund describe two different cost structures, not two quality tiers of the same product.

Third difference, discussed less often but real: ETFs publish their holdings almost daily, funds usually only quarterly. You know what you own more often with the first.

Then there's how an ETF replicates the index. Physical replication actually buys the securities that make up the basket, or a statistically representative sample if the basket is too wide to replicate in full. Synthetic replication, more common on niche indices or markets that are hard to buy directly, uses a derivative to deliver the same return without holding the underlying securities. Neither is automatically better: synthetic adds counterparty risk, physical can cost more in tracking deviation on illiquid markets.

None of these differences alone tells you what to pick. An active fund with a higher TER can be justified if it beats the index net of costs, but that's a claim to check case by case against the data, not to assume.

Why the raw price isn't enough: normalized performance

Comparing two price charts on different scales, a €45 ETF and a €210 fund, fools the eye even when the numbers are correct. Normalized performance fixes this by rebasing every instrument to the same starting point, zero, and showing cumulative growth from there. Whoever returned more over the whole period is obvious at a glance, regardless of the absolute starting price.

The period you pick changes the answer. An ETF and a fund compared over the last three years can produce one winner, and the same comparison over the last ten a different one: neither number is wrong, they're just describing different windows. That's why a serious comparison looks at more than one period, not just the one that happens to favor the instrument you already prefer.

Rolling returns: don't judge on a single period

Rolling returns break the history into many moving windows of the same length, one year, three, five, and calculate the annualized return for each, shifting the window one step at a time across the whole series. The result isn't a single number but a line showing how annualized performance changed over time.

The practical difference: a fund can post the best 10-year CAGR thanks to two exceptional years, while lagging the rest of the time. Rolling returns expose exactly this scenario, a strong long-term return actually driven by a handful of lucky windows, something the final number alone hides.

Head-to-head: win rate and pairwise comparison

With two or more instruments loaded over the same period, the head-to-head comparison answers a precise question: how many times, across all comparable rolling windows, did one instrument beat the other? A fund with a 60% win rate was best in six windows out of ten.

With three or more instruments in the comparison, the overall ranking isn't enough to answer "who beats whom": that's what the pairwise comparison is for, a matrix pitting every instrument against every other one, one pair at a time. A fund can have a modest overall win rate and still systematically beat the exact competitor you care about. The matrix shows that; the aggregate ranking doesn't.

Correlation: how much they move together

Before replacing a fund with an ETF, or keeping both in the portfolio, it's worth checking their correlation. A coefficient close to +1 means the two instruments rise and fall together almost all the time: owning both adds little real diversification, even if on paper they look like two different choices. The further the coefficient sits from +1, the more the second instrument brings something genuinely different to the portfolio, not just a duplicate under a different name.

Risk-adjusted metrics: how much return for how much risk

Sharpe, Sortino and Calmar answer a question CAGR alone doesn't ask: how much return you got for the risk you took on. Two funds with the same return can have had very different paths, one nearly flat, the other with wide swings or a deep drop partway through. The three metrics isolate exactly that difference.

Sharpe divides return in excess of the risk-free rate by total volatility; Sortino does the same but only counts downward swings, so it rewards an instrument that rises often and falls rarely; Calmar weighs return against maximum drawdown. For Sharpe and Sortino, a value between 1 and 2 counts as good, above 2 excellent; for Calmar the bar is lower, between 0.5 and 1 good, above 1 excellent.

When you're comparing two or more instruments, check whether the three metrics tell the same story. The highest single number isn't enough on its own. An instrument with a high Sharpe and a low Calmar delivered good average returns but took a heavy drawdown somewhere along the way, something total volatility won't show you but Calmar will. For the detail on each metric, and when one matters more than another, see the guide Sharpe, Sortino and Calmar ratios explained.

A numbers example

Two global equity instruments, same reference index on paper. ETF A has a TER of 0.15% and physically replicates the index. Fund B, actively managed, has a TER of 1.2% and tries to pick the best stocks within the same universe.

On five-year rolling returns, Fund B wins in 7 windows out of 10, a 70% win rate: in the periods where active selection worked, it genuinely outperformed the index. On cumulative return over the full period, though, once you account for the higher TER paid every year, the edge narrows to almost nothing. Correlation between the two stays high regardless, 0.92: even when the active fund wins, most of its movement is still explained by the same market the ETF tracks too.

This is a constructed example meant to show how the three readings, rolling returns, cumulative performance and correlation, fit together, not a comparison of two real instruments. The actual result depends on the data you load.

In BacktestFolio

ETF/Fund Analysis lines up this guide's five readings on the same instruments over the same period: Normalized Performance (Return) for the cumulative comparison from a shared starting point, Rolling Returns Analysis to see how annualized return shifts window by window, Head-to-head (Win rate & pairwise) for the direct comparison, the correlation matrix to understand how much real diversification they add to each other, and risk-adjusted metrics, Sharpe, Sortino, Calmar, for return relative to the risk taken.

You can try it free in the Analysis section, loading two or more instruments by ticker or ISIN.