I don't say this merely because the FTSE 100 is only a subset of the total UK market: it represents four-fifths of the All-Share index. Nor is it because the index is weighted by market capitalisation and so gives 35 times as much weight to HSBC as to Intu Properties.
Instead, there's another reason: inflation. If we adjust the FTSE 100 index for this (using the consumer prices index) it is still around 26 per cent lower than it was at its peak in December 1999 (I say around because we don't yet have inflation data for March, so I'm estimating it). The index is also lower than it was in 2007. To put this another way, the index would have to be around 9800 now to be at a record high in real terms.
In saying this, I'm not just playing with numbers. Wealth matters because of what it buys us. A unit of the FTSE 100 buys us one-quarter less of goods and services today than it did in 1999. In fact, this understates the FTSE's loss of purchasing power because the CPI excludes housing. The FTSE 100 today buys you only 38 per cent as much of a house as it would have done in 1999.
We've suffered 18 years of real losses on the FTSE 100. Such a long losing streak isn't unprecedented. The Bank of England has data going back to 1700. These show us that it was only in the late 1990s that prices in real terms regained their 1909 peak. And even that long gap wasn't unprecedented: shares didn't return to their 1720 peak until 1850.
You might be surprised by this. Shares are supposed to be (or to have been) a great long-term investment. So how can they spend decades under water?
Simple. It's because the FTSE 100 index is misleading in another way. It excludes dividends. And over the long term these matter enormously. If we look at the FTSE 100's total return index in real terms, it is today 34 per cent above the level it reached in December 1999. And it has been mostly above that level since January 2013.
Most of the long-term returns on the FTSE 100 come from dividends, not prices. This is true even for those lucky enough to buy at low points. If you'd bought at the index's post-bubble real-terms trough in January 2003, you'd have made a total (pre-tax) return by now of 6.7 per cent a year. Only 2.8 percentage points of this came from price appreciation. The rest - over half - came from dividends.
This will probably continue to be the case. Let's make some simple assumptions. Let's assume shares are fairly priced, so the dividend yield won't systematically shift from its current 3.6 per cent. And let's assume dividends grow in line with real GDP growth in the western world, which we can call around 2.5 per cent a year. Then we should see prices rise by 2.5 per cent a year on average but a yield of 3.6 per cent, implying that dividends will make up over half of total returns.
Of course, you can quibble endlessly with these assumptions. But it is only if shares are very under-priced now, and/or if the world economy grows very quickly, that dividends won't make up the bulk of long term returns.
Yes, price changes can be the bulk of short-term returns if we get lucky. But over longer periods it's dividends that matter more.
The point here is not that investors should favour stocks that pay lots of dividends: that's a different issue. Instead, it's simply that over the long-term dividends matter enormously, as does inflation. Simple stock market indices ignore both of these. For a longer-term investor they are therefore very misleading.
