If you have enough to manage your own Index comprised of individual stocks with no one stock too heavily weighted (risk), it can work out better than funds.
Take for instance a low cost Vanguard S&P500 tracker. Looking at their yearly performance compared to the 'benchmark'
Vanguards yearly benchmark was lower than the S&P Index total return by respectively
0.84% 0.73%
Their fund actual performance however was above benchmark, reducing the difference of S&P Index Total Return minus Vanguards funds total return to respectively
0.488% 0.46%
So looking at the fund factsheet the impression is that its low cost (0.07% expense ratio) and actually outperformed the benchmark by around 0.3%/year (after costs). But in practice it lagged the S&P Total Return by around 0.5%/year. A whopping 0.8%/year difference!
And Vanguard is one of the best/lowest cost choices.
How did such 'error' (that makes the fund look good) occur? Simply because they benchmark to a net total return version of the S&P index, but in some cases they don't actually incur those taxes/costs. By security lending just prior to x-dividend to a more dividend tax efficient party, and then receiving those shares back again after having gone x-dividend, together with a payment for having lent the stock (that compares to the gross dividend), they benefit and relatively outperform the net of taxes benchmark.
That 0.5% actual cost might make a big difference to a investor. A casual glance might be taken as the fund nigh on matched the S&P total return after low 0.07% costs (expense ratio) and as stocks provide a 4.7% real return (or whatever) I can base my spending/lifestyle on a 5% rate. But if in practice you lag that by 0.8%/year then over time that compounds out to a poorer than expected result. And that's for one of the best choices. For others the impact can be much more significant.