Elroy Dimson

Why not 1918 or 1917 or 1960?

It was a war on.

People couldn't trust prices.

So the series began after the wartime turmoil was out of the way.

That was easier to do.

The data was more reliable.

You take a more extreme case.

There was a Barclays publication of a similar nature.

looking at Germany, which waited until after the turmoil of the Second World War and the succeeding events were out of the way.

And so if you look at Germany, and you include the recovery period after Germany recovered, but leave out what happened in the war, you again have a misleading number.

And so we discovered that almost everywhere, almost every country,

was one where if you use the standard index over the standard period, that was easier to do.

You didn't have so much in the way of data collection, but performance was very overstated compared to what happened if you imposed a common start date on all of the different markets that you looked at.

I would say that that is the most important of the biases that we eliminated, but there are others as well.

If you take markets which are important at the date you're compiling data, you're more likely to incorporate ones that have done well and more likely to leave out ones which had started tiny and got smaller still.

There is a bias there in terms of choosing markets, which is similar to the choices you have to make within a stock market looking at individual stocks.

There are relatively few markets that simply didn't succeed, leaving out a couple of major geopolitical casualties, the demise of the Russian stock markets,

was important.

The acquisition by the state of not only Russian but Chinese resources when China moved towards a communist framework.

But mostly, there aren't markets which start up, do okay, and then just completely disappear.