Correlation, causation and chickens
Correlation does not imply causation — an easily understood and easily forgotten principle. Easily understood through absurd examples; my favorite, from Tyler Vigen’s inspired list, is the correlation between U.S. crude oil imports and … per capita consumption of chicken. That’s right. The more you drive, the more chickens you eat — or vice versa... Easily forgotten because we always try to make sense of this crazy world — so it helps to think we can save the planet by eating fewer chickens — or save the chickens by driving fewer miles.
Economics is especially vulnerable to the causation-correlation fallacy. The Economist’s latest leading article, which castigates Germany for its large external current account surplus, is a perfect example:
A country’s current account balance equals the difference between its savings and investments.
This is always true — it is an accounting identity. We are on even stronger ground than with the chickens: this is not an empirical correlation that might suddenly break down; an accounting identity always holds.
The Economist therefore argues that Germany runs a large current account surplus because it saves too much. In other words, the savings-investment balance determines the current account balance.
Global trade has to be balanced: if some countries run a surplus, others must run a deficit. This is always true — it is an accounting identity. (Your exports are someone’s imports).
The Economist therefore goes on to argue that “To offset such (German) surpluses and sustain enough aggregate demand to keep people in work, the rest of the world must borrow and spend with equal abandon.” And it will end in tears: “In some countries, notably Italy, Greece and Spain, persistent deficits eventually led to crises.”
So: by saving too much, Germany forces other countries into bankruptcy. The logic seems unassailable, but…
…hold on a second: For Germany we said that the savings-investment balance determines the current account . But for Greece, Spain, Italy or the U.S., we now say that the current account determines the savings-investment balance. Which one is it? Does the causality run in a different direction depending on which country you are in? Are we saying it does not matter what you produce or how competitive you are, because your economy “must” mirror whatever Germany is doing? And why Germany and not the US? (By the way, The Economist seems to have missed that Italy has run a current account surplus since 2013, reaching a hefty 2.7% of GDP last year).
IMF chief economist Maurice Obstfeld recognizes that “Because the drivers of current account balances are so very complex, no simple approach to identifying excessive imbalances is likely to give the right answer for every country.” No kidding. The IMF’s more sophisticated approach also concludes that Germany’s current account surplus is “excessive” (though it estimates a “normal” current account surplus at a still hefty 5% of GDP, vs the current 8%); and I agree that Germany should invest more in infrastructure. But that is different from arguing that it’s Germany’s fault if other countries spend themselves into bankruptcy and a productivity black hole. (You can find Germany’s view here.)
Like The Economist, the Financial Times and many others say that Germany forces Italy and other fellow Euro members to “spend with abandon”; but they also argue that economies like Italy’s cannot grow faster because Germany forces them to curb spending with draconian Eurozone fiscal rules. So…is Germany forcing them to spend too much or too little?
I understand the temptation to blame Germany. But instead of mixing up causation and correlation, why not just blame the robots? (No, not the chickens!)
Let me know what you think, here or on Twitter. First, though, click here to prove you are not a bad robot.
