Let your baby play in the mud or the value of experiments

Expert Advice or Who to believe

Adapted from the book Mindware – Richard Nisbett

The Calm Investor: Learning from experiements

We get it all the time. Advice meant to improve our careers, our children, our bodies, our lives. Only, a lot of it is often contradictory. And unlike twenty years ago when there were three national dailies, two television channels and 1 radio station,  there’s an unending stream (literally) of tweets with links to online self-improvement resources.

It’s the same with investing. “Time to buy consumer staples” “Get out of commodities” “Cyclicals are the next mega trend” or even “Sell Everything!“, the warning from the good folks at RBS in Jan 2016. With #Brexit, there’s another rash of pronouncements about the markets and what investors should be doing.

How should you make your critical decisions, about investing or even life in general, for example whether to let your baby be exposed to germs or to deploy an antiseptic environment for her safety?

Outcomes of Natural experiments

“Learn from other people’s mistakes. Life is too short to make them all yourself.” – Sam Levenson.

Handy advice, especially if your decisions will impact the future health of your baby.

A natural experiment allows a comparison between two (or among several) cases that are generally similar but differ in some way that might be related to an outcome variable of interest.

Suppose you knew that certain groups of people have fewer severe allergies as adults…

Group 1 < (has fewer allergies as adults compared to) Group 2

  • Former East Germany < Former West Germany…
  • Russians < Finns
  • farmers < city dwellers
  • attended day care < those who didn’t
  • had pets as infants < those who didn’t
  • had lots of diarrhea as infants < those who didn’t
  • were born vaginally < were born by Ceasarian section

Empirical data suggests these are all true. These are all natural experiments in that they are largely different in one key independent variable that might explain the outcome of the question. Each test therefore tests the hypothesis that early exposure to bacteria confers resistance to allergy, as well as to an abnormal, mistakenly protective overreaction against substances not normally present in the body.

It would be safe to assume that former East Germany and Russia are probably less hygienic than their counterparts. It’s easy to see that children raised on farms, or have pets, or are exposed to other children, are exposed to a wider range of bacteria. Similarly, those born vaginally are exposed to the range of bacteria in the mother’s vagina.

Therefore, it’s a fair bet at this point to say that those exposed to wider array of germs as infants are less likely to be afflicted by allergies later in life.

Natural to Proper Experiments

At this point, you might be fairly convinced of the hypothesis, but you’re hardly likely to let your baby crawl through a sewer to build immunity for a healthy adulthood. After all, isn’t one of the biggest biases we’re all susceptible to is to mistake correlation for causation?

So, a double-blind, randomized control sort, with babies assigned to an experimental high-bacteria exposure condition versus a control, low-bacteria condition. Both the experimenter and the participants (mothers) should be ignorant of (blind to) the condition their baby is assigned to. In short a proper experiment.

Now, if it turned out that the experimental high-exposure babies had fewer problems with allergies and did not have any other long-term problems compared to the low-exposure set overall, then you could conclude definitively that you should expose your baby to a wide range of bacteria.

Conducting Proper Experiments

Fortunately, the use of animal models preclude having to run such a morally questionable experiment – would you volunteer your baby for such an experiment? Therefore, the same experiment was conducted on young mice, which are phylogenetically close to humans. Instead of a high-exposure environment, scientists created a germ-free environment and left the control group in that while the test group was exposed to normal lab conditions which have a variety of germs.

The result: The germ-free mice developed abnormal levels of killer T cells in parts of the colon and lungs. These T cells attacked even non-threatening substances, with resulting inflammation, allergy and asthma.

Experiments and Investing

Historical analysis of asset class returns over multiple business cycles are outcomes of natural experiments that tell you which asset classes are likely to give you sustainable returns. Which is how you can form the hypothesis that equities are better than fixed deposits over the long run. What’s missing is the proper experiment of actually owning those assets and studying your reaction to market fluctuations and transaction costs associated with any market activity.

Inadvertently, a fundamental piece of basic investment advice encourages proper experiments – diversification. I would qualify that to mean “selective minimalistic diversification“.

By buying nominal amounts of various asset classes, even those you do not like, and holding over at least a couple of business cycles, investors can arrive at their conclusions of what really works. So, while I do not believe gold can ever be a sustainable investment, I have about 1% of my portfolio in a gold ETF that is supposed to mimic the price of physical gold. If over the course of several years if this experimental part of the portfolio had delivered returns superior to the mostly carefully selected set of stocks, I would need to revisit my hostility to the yellow metal. So far, the results haven’t disproved my hypothesis of the value of gold as an investment, Gold is the worst investment in history.

So to be confident about how stuff works, turns out there is no better way than to rely on experiments, natural and proper. And certainly, don’t blindly believe the experts.


What do you think?