The Bangladeshi government is going to revise the GDP statistics and the new numbers will show that the economy is larger than we thought. Therefore, the government has been doing better than we thought – funny how that works, isn’t it? The numbers telling us how well the government has done are being revised up and improved by government.
However, while being a cynic is the correct manner of dealing with any and every government thing aren’t as bad as that. We do in fact need to revise our economic statistics from time to time and all countries do so. The point being that economies change in their structure, not just their size, so we need to change our structure of data collection to deal with this.
What’s actually happening here is that we’re going to change the “base year” from 2006/7 to 2015/6. Don’t use this description as an answer in an economics exam but roughly what we’re doing here is using a base year to define that structure of the economy.
We’ve got garments over here, leather there, agriculture, banking, and so on. Then, until it’s time to change the base year again – every decade or so – we’ll not worry about changes in the structure at all, we’ll just watch changes in the numbers we collect from the structure we’ve identified.
For year to year changes, this is probably good enough but it gets further and further out of line the longer we keep using that same base year.
WhatsApp is very much more productive, obviously. But there is no fee to use it nor does it carry advertising. So, in GDP, it doesn’t really exist
For example, when we created that model of the economy, that structure, for the 2005 base year – e-commerce wasn’t really a thing in Bangladesh. Now it is, but we’re not recording it as being anything. So, we probably should change the model to incorporate those changes in the basic structure.
So, yes, GDP will rise, making the government look good, just because of this change in the way we count it. But it’s still something we should be doing.
At which point, what every serious econometrician does (econometricians are the economists who actually add up the numbers in the economy, the ones with lots and lots of digits on their calculators) is to rebase the time series.
We go back over history and reproduce all of our old GDP numbers using this new model and structure that we’ve just designed. The Bangladesh Bureau of Statistics are serious econometricians so this is what they’re going to do – we’ll then have GDP numbers all the way back to the start of the Bangladeshi economy in the early 1970s by both the new 2015 model and the older 2005 one (and 1995 and so on).
The importance of this is that we want to be able to compare over time but we cannot do that if we’ve a change in our model. So, every time we change the model we do the work again to make sure we can compare over time.
There is another way to do this of course. Sir John Cowperthwaite essentially ran the Hong Kong economy by doing very little indeed for some decades. He refused to allow anyone to even attempt to record GDP numbers on the grounds that “some damn fool will only try to do something with them.”
Not an attitude much in evidence these days but it did seem to work for Hong Kong.
So, a good idea for Bangladesh then, an economy growing at 6% a year for decades – doubling in size every 11 or 12 years – is going to see substantial structural change and that’s what we try to capture by doing this rebasing.
However, we’ve all got another problem entirely with our GDP numbers. This is a global problem for, as Google’s economist, Hal Varian says, GDP doesn’t deal well with free.
We know very well there’s a technological revolution going on. The smartphone is the fastest adopted technology ever in the history of our species. What we can all do on the internet advances by leaps and bounds. But when we look at the economic statistics to see the effect we can’t find it at all.
This is really just a flaw in our very idea of GDP.
For we only record transactions at market prices and if something’s free we’re not recording it at all. In more detail, we record Google’s or Facebook’s contribution to the economy as either everything they spend money on as a company – primarily wages and profits – or as all of their sales, which is pretty much exclusively advertising.
The actual value to us all of the existence of those two is very much higher than that but we’re just not recording it.
Even more extreme is WhatsApp. A major telephone company might employ 150,000 people to provide telecoms for hundreds of millions of people. WhatsApp uses 200 engineers (I checked with Facebook as to how many do this) to provide telecoms of a kind to some 1 billion people.
WhatsApp is very much more productive, obviously. But there is no fee to use it nor does it carry advertising. So, in GDP, it doesn’t really exist.
The wages of those 200 engineers are in GDP, buried within Facebook’s accounts, as a cost, but the value of the output isn’t recorded anywhere. The effect of WhatsApp on the real world is an increase in wealth for those billion.
The effect in our economic statistics is in fact a fall in productivity – we’re recording the labour that goes into it but not the benefit coming out. This is madness, of course, but no one really has any idea about how to solve the problem.
That Bangladesh is to revise its GDP statistics is a good idea, this is something which should be regularly done. All of us out here would also love to hear any bright ideas about how we revise these economic statistics more generally.
For the basic restrictions of the idea are such that we know very well that the numbers we’re measuring are becoming ever more detached from what we hope to measure, which is how rich are we all getting?
Tim Worstall is a Senior Fellow at the Adam Smith Institute in London