Why predictions make asses out of all of us

Feb 11, 2013 18:46 IST

#Crystal Ball   #HowThisWorks   #Sensex   #US Dollar  

By R Jagannathan

When the US economy was downgraded in 2011, it seemed as if it was only a matter of time before the world started bailing out of the dollar and into emerging market currencies. The dollar era was over - or that's what I thought. I still think so. But nothing of the kind has happened so far.

When the Lehman crisis stuck, I happily predicted that the Sensex will hit 60,000 in around five years' time. I still think so. At 19-20k right now, it needs to triple in the next four years. But I am beginning to look more and more foolish 18 months down the line after I made the statement.


If Chidambaram gets his fiscal deficit right, it may be because inflation has pushed money GDP higher than ever. Getty Images

A few months ago, this writer also wrote that the fiscal deficit this year was likely to hit 6 percent. Now, the finance minister says it can be contained within 5.3 percent. Wrong again?

Having egg on my face has not cured me of the temptation to look into the crystal ball, but I will readily admit that one needs luck to get things right more often than not. To get things right, many things have to happen together-which sometimes don't happen.

The reasons why predictions work or don't work are many, but primarily three.

First, predictions often relate to getting relative changes right. You come first in class if someone scores less, and not necessarily because you scored more. You can score low and still come first. You can predict that the slowest deer will be eaten by the fastest lion, but if another deer goes lame, even the slowest deer will get away.

So if I make a prediction that the US dollar will lose out, it can go wrong for several reasons: the other currencies and economies may be doing worse even though the US may not be doing any better.

If Chidambaram gets his fiscal deficit right, it may be because inflation has pushed money GDP higher than ever, or because he has shifted all critical expenses from this year to the next, or many more things.

Second, you can be wrong most of the time and still be right in the end. This is what happened with the 2008 financial crisis. Many people - from Nouriel Roubini to Raghuram Rajan - said that the sheer growth in sub-prime lending will lead to a bust. But they were wrong most of the time - in 2005, 2006, 2007, and most of 2008. They became correct only several years down the line, when Lehman went belly-up.

This is what Nassim Nicholas Taleb calls the turkey problem. A turkey is killed on Thanksgiving, but for most of its life it lives a good life and is well fed. It does not know till the last day that its life is going to end. For 99 percent of its life, it has been right. The wise old turkey which warned the young turkey to try and escape to avoid turning up on someone's dinner table would thus have been wrong most of the time.

The same holds for the US dollar. It will continue to remain the world's top currency as long as the world thinks the other currencies are worse. But spiralling American government debts will one day have to debase the currency. We can't say if that will happen in 2013, 2014 or 2017 or much later.

The world is anyway shifting away from the dollar - into gold and other assets. The day an overwhelming majority of investors starts thinking so, the dollar could face a rout. But that day is not right now.

The reality is that long-range forecasts based on fundamentals are often off the mark because they are distorted by short-term liquidity surges or other unexpected events. In 2008, the US looked like it had messed up. In 2011 and 2012, it looked like it was Europe that got it wrong. In 2015, if China gets into an economic or political crisis, both US and Europe will look even better.

Third, one can be right for the wrong reasons. In the 19th century, one could have confidently made the assertion that 99 percent of the population will never die of cancer. This does not necessarily mean cancers were unknown or that people were healthier then (though, one could say the survivors were certainly looking healthier). But in those days most people did not live beyond 40. It means they didn't live long enough to develop serious cancers. They may have died for being bitten by a rabid dog, when no anti-rabies vaccine was in sight, or even starved to death in a severe drought.

In 2009-10, inflation in India suddenly touched zero. Was this a triumph of monetary policy? Chances are, the sheer drop in oil prices post-Lehman did the trick. Once the trick ended, the inflation rate went back to being what it was. This can happen again in 2014 if oil prices crash.

Fourth, the markets do not rise or fall in a linear fashion. Which is why it is always risky to guess long-term market direction. For example, between 2008 and now, the Sensex has essentially given zero returns - over five years. Adjust for inflation, and the Sensex would have given seriously negative returns. This does not mean Sensex 60,000 is a pipedream.

Money flows in or out very quickly these days. Given the right circumstances - some question-marks about the US and European economies, a smart burst of reforms in India, and a few other favourable circumstances - and it can move up very quickly. In 2003-2008, the Sensex went up six-fold in five years.

So when will the Sensex hit 60K? I certainly don't have the answer. But I do believe that when the stars are right, the Sensex will not take five years to move from 19-20K to 60K.