Yves here. This post from MacroBusiness provides a good point of departure, and I’ll provide some comments further down.
By Sell on News, a global macro equities analyst. Cross posted from MacroBusiness
A little known fact about John Maynard Keynes, detailed in Jane Gleeson-White’s book “Double Entry” is that he was responsible for the development of national economic statistics and that he expected them to be aggregated only on a temporary basis.
It was being done for the war effort, and would, he reasoned, not be necessary afterwards. This certainly puts “Keynesianism” in a different perspective, and poses the intriguing question: where would we be without economic statistics?
The Economist recently had a leader “Don’t Lie to Me Argentina” in which it accused Argentina of some kind of unforgivable treachery for politicising its economic statistics. As if economic statistics aren’t political in their very nature (a heavy bias towards capital and against labour, for instance).
So in contrast to H&H [a fellow MacroBusiness blogger], who enthuses that without economic data we are “naked, bereft of meaning” I wish to present a very different perspective. I wish to briefly examine what it would mean not to have economic statistics. Here are a few implications, I submit:
1. We would have to stop being lazy in the way we construct meaning and do the work of creating meaning ourselves.
The worship of economic statistics encourages a certain passivity of mind because it presents us with a picture ready made that we can then seek to interpret. Trouble is, that picture is heavily biased. Imagine, for instance, if it included unpaid housework as was proposed in Keynes’ time? Economics just presents transactions and makes little distinction between good transactions and bad ones.
A natural disaster, for instance, is generally thought to be bad, but in statistical terms it is not because typically the reconstruction creates a lot of economic activity (witness the Japanese growth figures post Fukushima). What happens is that transactions are not seen as a reflection of reality; rather reality has to be fitted into the transactions. “We all must change our behaviour because GDP is not growing fast enough, or productivity is not improving enough”.
2. We would embrace a broader sense of meaning, one that did not involve just what can be measured.
Most economic growth statistics measure the exchange of consumer goods, because it is easy. Much harder to measure assets, because they are not continually transacted — that was why the asset bubbles in America were ignored for so long, because they are hard to measure – and harder again to measure long term infrastructure investment. It is impossible to measure culture, yet culture is essential to well being. Indeed, well being is not really measured, and when they have tried to use broader measures it is generally found that life has improved little despite the economic growth.
3. We would not have a financial/economics sector purporting to understand what they do not understand.
For example, the “inter-relationships” between various economic indicators (such as Friedmanites v Keynesianism). This is for the most part an intellectual fraud. There are the obvious conclusions – you can’t spend more than you own, for instance – that derive from housekeeping (that being the etymology of economics). But anything beyond that is either unknowable or a circular argument (for instance Friedman’s maxim inflation is always and everywhere a monetary phenomenon is a tautology dressed up as insight).
4. We would have a greater sense of how we can impel affairs as thinking creatures with free will, rather than being pushed about by the “economic system”.
There is a reason why economists are so poor at anticipating the future. Economic statistics are always retrospective and tell us little about what people are going to do – and it is what people DO that shapes the future. Of course the past will shape what people do, but it does not determine it. Money is a social construct and transactions are social arrangements. They are subject to individual and collective will, not the logic of a mathematical system.
5. We would not have the giant casino that is amusingly referred to as the global capital markets.
The use of algorithms, which poses deep dangers to the system, is only possible because of the blizzard of data and statistics. That is exactly what is being codified and manipulated. Intriguingly, one of the geeks made an interesting comment, saying that a 2-3% variation, which makes all the difference in GDP statistics between acceptable growth and recession is all but inevitable in his geek world. he could not understand why it is considered so significant. But that is our world, dominated by economic statistics.
So there are five reasons why we would probably be a lot better without all that data. Far from being naked, bereft of meaning, I would suggest we could put on some decent clothes and find some more substantial, dare I say it, real meaning.
Yves here. While I would come up with a different list of implications, I’ve long been struck with the fetishization of numbers, particularly when all numbers are not created equal. Transaction data, say the price at which a particular stock traded at for a specific order, is a hard number. But the overwhelming majority of economic statistics are “soft” in that they are constructions, and statisticians care a great deal about consistency over time. That is useful up to a point. The procedures that produce consistency mean that if the measurement fails to capture or underweights an issue that is becoming increasingly prominent, it will have obtained consistency at the expense of representativeness (the notorious non-farm payrolls “birth-death adjustment” is a classic of this type). And the caliber of official statistics had decayed due to reduced staffing thanks to budget cuts.
I wrote about some of the problematic behaviors that result from the propensity to give quantitative information, no matter how poor it may actually be, great weight, in an article, “Management’s Great Addiction.” One the symptoms, and it takes place way too often, is when people rely unduly on a single or very few metrics to analyze a complex phenomenon (think of the use of VAR to measure firm wide risk!). If you must measure to get a grip on a complicated situation, use more rather fewer views in (provided you don’t go for more just to have more but can find pretty reliable measures of different, important aspects of a complex situation).
And perhaps the most important habit to adopt is to be aware of the limits of your knowledge.