What is the correct way to think about historical data?

[k290]

With regards to index funds and stock markets in general. Often we see the disclaimer at the bottom of a unit trust fact sheet: "Historical data is not an indication of future returns". Let's go ahead and ignore that.

Here are the four possibilities I've come up with:

• Historical data is useless because it is does not predict of future returns or volatility
• Historical data is useless because it predicts future returns and future volatility with low probability. (Worse than a coin toss)
• Historical data is useful because it predicts future returns and future volatility with high probability. (Better than a coin toss)
• Historical data is useful because it predicts future returns and volatility

[steveo]

• Historical data is useful because it predicts future returns and volatility

This is my take. Learn from the historical data but don't take it as gospel. Always take a little bit of a risk averse approach towards future returns being exactly as per past returns.

[k9]

Historical data tends to display what happened in different circumstances. It tends to be biased. Imagine you are an investor in the 1880s. Would you invest in the US, that emerging country that just came through a civil war and a sovereign default ? No, you wouldn't. You're not a speculator, you're an investor. You would rather invest in titans like Russia or Germany. This would be safe for the next 40 years, right ? Historical data says so.

Would you invest in stocks ? Yes, maybe a little, but why invest much ? There is no such thing as inflation, since cash is gold, no government can inflate his money, so bonds' inflation risk is a non issue, and dealing with stocks is risky. There is no such thing as public financial reports, when you invest in a stock, you can only rely on its past price and insider rumors. And capital appreciation was not a thing on these days. Money came from dividends. You didn't buy a stock just to see its market value grow, but to earn a dividend flow. Investing was nothing more than comparing stocks yield vs safe bonds yield. Oh, and since information wasn't as fast and as safe as today, markets were all but efficient. That was a very different beast.

Well, anyway, if you wanted to invest in stocks, you chose UK, not US, because the yield was much better and UK was the #1 country in the world, not a loosy emerging country.

So, looking at historical data, our 1880s investor would not put a dime in US stocks and buy lots of British bonds and stocks, and a substantive amount of French, German and Russian treasury bonds.

[DaveR]

Which is correct? I vote none of the above.

Historical data is useful: it teaches you history of markets and behaviors. As a predictor, it is useless: markets are random. Deciding usefulness based on ability to predict markets is a limited view.

That doesn't mean that historical data has no use in forecast models, but it's not a simple coin-toss-like probability reduction.

[k9]

Historical data has a virtue : it shows what can fail. If your investing strategy failed in the past, you know it can fail. If it never failed, you can't know for sure it cannot fail.

[innerscorecard]

Fantastic question and very thought-provoking posts so far.

This is *the* question, isn't it? I tend to think about history as a source of patterns and trends. To the degree that some of these patterns and trends might hold true today as well (which is more or less the case for different things), history can inform one's judgment about what can happen in the future. It's better than knowing nothing. But there are examples and counterexamples of anything you might like to find when looking at macro trends in history.

[daverobev]

I guess the real question is, what will the world look like for the next 50-100 years?

We have advanced so much in so many ways over the last 100, though we are multiplying at an alarming rate and trashing the planet at the same time - both of which seem to be necessary to fuel the growth that has brought so many of us the luxury we live in.

Will we head to a guaranteed minimum income? Will corporations grow and become more powerful than governments? Will capitalism as we know it endure?

Scary to try and think too much. But, I guess, the meme about the only thing politicians fearing being mass demonstrations is true, and with the amount of data we have I'd guess proletariat mollification is probably fairly simple (though there have been plenty of riots in surprising places).

In short: fuck knows. But owning a diversified global portfolio is probably the best we can do (short of tin foil hat, 20 years of canned beans).

[Tyler]

Historical data has a virtue : it shows what can fail. If your investing strategy failed in the past, you know it can fail. If it never failed, you can't know for sure it cannot fail.

I think this is closest to the truth.

Historical data is absolutely valuable. You can look at history and see that stocks can struggle for decades at a time and realize that perhaps diversification into other assets is a good idea even while stocks may be doing great today.  You can see how certain assets like gold and bonds respond in very different ways to economic conditions like inflation and deflation, and balance them in your portfolio accordingly.  You can study real-life retirement scenarios and use them to set practical and realistic spending boundaries.  Importantly, you can read about various funds and portfolios in the past that looked great for a while but eventually failed, and study why that happened and how it potentially could have been avoided. 

I agree that historical data is not a good predictive tool, and anyone who tells you they know precisely how the future will unfold is either fooling themselves or lying to you.  But to ignore history and just pretend like there's nothing for you to learn from it is extremely shortsighted in every walk of life - not simply investing. 

[electriceagle]

Well, anyway, if you wanted to invest in stocks, you chose UK, not US, because the yield was much better and UK was the #1 country in the world, not a loosy emerging country.

So, looking at historical data, our 1880s investor would not put a dime in US stocks and buy lots of British bonds and stocks, and a substantive amount of French, German and Russian treasury bonds.

Since travel was dangerous and expensive, you would be likely to invest wherever you were located. Information on financial instruments was not shared easily, so you would probably invest in real estate or a farm. Just sayin'