My language ate my savings

Johannes Lenhard
March 10, 2013

I always wondered what ‘plan B’ – the theoretical ‘way out’ every  economist refers to when critiqued for still falling back on ‘homo  œconomicus’-based assumptions – really is. Apparently, behavioural  economics does away with pure self-interest, with supply and demand  magically finding their harmonious balance guided by the market(er)’s  invisible hand (while actually already the Smithian  ‘invisible hand’ must somehow have been wrongly translated from  Scotland; his metaphorical limb was arguably one of ‘moral sentiment’,  of compassion and of providence – much more a natural, a godly rather  than a market hand). But how do behavioural economists work? Are they  experimenting in the lab, conducting test series just like Milgram  trying to find out about the relation between individual morality and  obedience – only under the heading ‘what investment decision does he  take under conditions of electroshock’? Do they absurdly combine  psychological (i.e. uncanny) insights with economic data (i.e. myriads  of already manipulated survey answers), detect (i.e. mystically  oversimplify) causal relationships and publish the newfound truths (i.e.  results of regression analysis) in shiny TED videos?

If the latter is true, Keith Chen is surely an outlier  of this species of economists. He is a vanguard, almost a Žižekian  pop-intellectual fulfilling the necessary clichés (receding hair,  slightly squint-eyed, untucked shirt). His theory is daring and  far-reaching. It could not only influence, but revolutionise our way of  thinking about the economy (at least our way of excusing economic  incompetence). Chen found (TED / paper)  a relationship between language and economic behaviour, more precisely  between a language’s capacity to use the present tense to talk about  future events and the individual’s propensity to invest in the future –  to save. A division between present and future animates people to  perceive the future as disjunct from the present resulting in the  neglect of behaviour benefitting your future rather than your present  self. In tongues such as English, Greek or Portuguese the future has to  be ‘formularised’. The English cannot say ‘It rains tomorrow’ – whereas  in German ‘Es regnet Morgen’ is very common (says a German speaker). The  English language requires the speaker to make a distinction between  present and forthcoming events, while German does not.

Image: Screenshot from the TED presentation

In a second step, Chen compared his assumptions on languages with  economic behaviour in particular countries. The results are thrilling.  The savings rate  in countries with a futureless language on average appears  substantially higher than the rate in ‘futured countries’. On average,  Germans save 10% more of the combined GDP than the British do; the  Chinese save about 20% more than Greeks or Indians. This finding mirrors  nicely what recent figures from Scottish Widows (part of Lloyds Banking  Group) suggest  for the English case: currently, around 15 million Brits are not saving  money for future eventualities, while an astonishing 13% of UK citizens  claim to have no savings at all. So from now on, can we excuse this  ‘propensity to spend’ by pointing to a ‘disabled’ English language?

The link between language and savings behaviour even holds when we control  for variables such as education, the household’s economic situation,  family structure or religion. Chen observes the same pattern for  families in a similarly structured life in the same neighbourhood of for  instance Belgium, one speaking Flemish and one speaking French. The  Flemish family knows how to save, the French doesn’t. The link can be  pushed even further. In terms of statistical correlations, the  connection is not limited to purely economic behaviour alone. Chen  argues in his paper that “languages that grammatically associate the  future and the present, foster future-oriented behavior” outside of  economic matters. He find “that speakers of such languages: save more,  retire with more wealth, smoke less, practice safer sex, and are less  obese.”

This analysis gives us a powerful tool. A powerful excuse. Take the  example of the UK. Now that we know what our language does with us, it  is no wonder that it is in Britain – a country where most people  converse in a futured language (besides English mainly Irish (1%) and Polish (1%)) where the Bullingdon Club squanders £50 notes in front of beggars.  It is not surprising that it is at a Cambridge college where the  majority of fellows barely speak a (linguistic rather than mathematical)  futured language fluently (not to mention non-futured languages), that  data about dining expenses  makes us recall excessive feasts. They cannot change their behaviour –  how is it possible to save money when you cannot talk about it? We are  not to blame them. And we are also not to blame the banks  and the bankers for not saving, for not being more cautious about the  future of their institutions and us. We are to blame the linguists for  not finding out about this earlier.

This is utter nonsense. How far one can push scientific insights –  and how far scientists push their own findings – has to be clearly  differentiated. Chen is aware of the shortcomings of his approach. He is  very careful in interpreting his results and shows an explicit openness  for critique. Geoffrey Pullum for instance expressed substantive distrust in the categories of ‘strong vs weak FTR languages’. In his answer, Chen admits that the categorisation of languages according to the so-called EUROTYP project and Östen Dahl  is problematic. Indeed even in English there are many exceptions to the  rule that you cannot express future behaviour in present tenses. In  light of this, he limits his argument to a statistical relationship  between certain features of language and particular economic  behaviours. In no way does he argue on the contrary for a ‘propensity to  squander’ in countries where futured languages are dominant. In no way  can we use his findings to explain (or even worse: excuse and  naturalise) excessive bonuses. His data explicitly presents a strong  relationship between linguistic and economic data for which further  explanations have to be found. In this way, Chen’s paper is an  auspicious example of how supposedly implausible relationships can  become convincing in interdisciplinary research. Hopefully, however,  only The Sun’s headline (if Sun reporters only watched TED videos), will (for now) try to develop absurd speculations based on Chen’s findings


All by
Johannes Lenhard