Freakonomics: The Quiet Danger of Non-Inflation-Adjusted Stock Returns

Survivorship bias is behind almost every instance of supposed alpha found when running pattern based studies. That and the bid-ask bounce.

I don't think that makes a difference. If I am understanding the mechanics of equity indices correctly...

The return of the index can be replicated even when stocks get delisted (an investor sells the stock that gets delisted.) When a new stock is added to the index, an investor buys this new stock.

It depends what you are using the index for.

If you investing in a mutual fund that is replicating the S&P 500, for example, it doesn't really matter except for transaction costs.

If you are using the S&P 500 as your baseline (e.g. to decide if an investment strategy is worth the effort) it doesn't matter at all.

If you are using the S&P 500 as a metric about the economy... it might matter. If there is frequent turnover in the composition it might indicate that the market is very volatile... (someone with more knowledge can probably tell me if this is true or not).

Your latter point is often ignored as inflation numbers are often taken as undisputed (or at least "what can you do about it") figures. There is tremendous bias in the calculation of inflation, although the BLS has done a great job of fixing this in the last decade or so. Unfortunately historical data will be incomparable because of these frequent revisions. See the Boskin Report for more details.

The article makes a good point: despite the "masked" non-adjusted returns, you don't really have an alternative to equities over the long haul.

From the linked WSJ article: "All of this might be enough to put investors off stocks entirely, until they consider the long-term alternatives. Measured over the 1978-2008 period, rather than over just one decade, stock performance in real-real terms actually is better than that of just about any other major investment class, Mr. Thornburg found: 4.5% a year. Stocks' ability to keep up with inflation over the very long haul may be their best selling point.

In real-real terms, stocks did better over that period than municipal bonds (2.5% a year), long-term government bonds (2% a year) and corporate bonds (0.2% a year). Real-real home prices were unchanged over those 30 years. Both short-term government bonds and commodities suffered losses. (Mr. Thornburg has experience investing in all these areas, although his mortgage affiliate went bust in last year's housing collapse.)"

I don't understand why using gold makes any sense. For people saving for education, I can see why adjusting for tuition makes sense. For people in Europe, I can see why adjusting to the Euro makes sense. But gold isn't something anyone uses on a large scale basis (in the US anyway...). Unless someone is in the jewelery business, I can't see the purpose. In general, you would want to adjust to whatever you are using money for, which is what inflation attempts to do.