As a theoretical exercise, I can see a possible strategy involving individual stocks. Let's toss out 95% of people out of the picture before we even start. Most people underperform the market whether in stocks, actively managed funds or index funds because they're chasing performance. (Looking at the fund inflows/outflows at Vanguard, it happens even with index funds.) For the purposes of our discussion, our comparison is against a buy-hold asset allocation plan.
Bogle's research shows index funds are at the 50% percentile for performance before expenses. If index funds capture the whole of market activity, it should be right smack dab in the middle. It's after expenses that index funds then jump to the 85% percentile. Then after taxes, it jumps to the 90-95% percentile. So we have the first part of our answer -- expenses and taxes. The next part of our answer comes from the Efficient Market Hypothesis. In a nutshell, you can get higher long-term returns in exchange for enduring higher risk. With these two points in mind, let's dwelve into "stock picking". How could we construct a strategy to out-perform an index?
First, we completely ignore all pretense of the ability to pick individual stocks. Random dartboards have beaten wall street analysts and the monkey is beating Cramer as we speak ( http://cramerwatch.org
). But this also means randomly picking enough stocks out of an index will reproduce index performance as you increase the number of stocks chosen. So to uncouple tracking to the index and introduce more volatility, we decrease the number of stock chosen. Over a long enough period, the performance of 20 random companies from the S&P500 would have the same return as the overall S&P500 but with a higher volatility. The benefit of the higher volatility will be rebalancing bonus - you get higher highs to sell off profits and lower lows to buy new shares.
Second, instead of randomly picking 20-40 stocks from each category, we try to subdivide the asset classes to eliminate those stocks in the transition zone. For example, the S&P500 can be divided into 250 growth stocks and 250 value stocks. Except about 50-100 in both categories are right on the line in between the two where any share price changes can change their status. Instead, we want to pick from those stocks deep in growth for growth classes and deep in value for value classes. Total randomness is probably best -- it probably won't perform any better or worse but at least you won't have wasted time on futile research. What this does is accentuate the asset class differences to (1) fully capture the returns of each class and (2) increase rebalancing bonus.
Now here comes the caveats. Because we're only picking the extremes of each class, our turnover will be far higher than an index fund. Where a full index would turnover only the small percentage in the buffer zone between large/small/growth/value, a strategy that picks the extremes would easily turnover 50% of a portfolio or more. Even LTCG at 15%, high turnover would be a big tax drag. This means you absolutely need to attempt this strategy in a tax-deferred account.
So we need to find the right account to keep expenses low. If we target 20 stocks per large class and 40 per small class (more stocks needed for riskier small caps), we are looking at roughly 120 stocks and another 10-15 ETFs for those classes too hard to capture with individual stocks (international). At 50% turnover, that's a minimum of 65 trades a year. There are free trade offers but whether these trade offers will remain free forever is unknown. So let's assume $4/trade over an investment lifetime -- that translates to a minimum of $104K needed to match an index fund at 0.25% ER.
Hence, in order to make use of this strategy, you need rather big retirement balances in order to split off a comfortable percentage for a higher risk sub-portfolio. I can see two possibilities. (1) You are self-employed making more than $180K a year which lets you contribute the max $45K a year into a SEP-IRA. (2) You have about 20 years of significant contributions and growth in your 401K and you've recently switched jobs allowing you to transfer to an IRA brokerage account.
Finally you need discipline and nerves of steel as you will see bigger drops than the benchmark indexes. Whatever the computer says to pick during rebalancing, do it because if you can't stick to the original parameters of the strategy, it will be a sure loser. Of course, it may be a loser anyways because that's the definition of risk premium. The higher risk and volatility is guaranteed -- higher risks is only a chance.