An example of misusing data/GDP/aggregates

What a perfect chance to rant after describing thinking about GDP this morning!  The econ news gods love to provide.

Look, I didn’t really agree with the Business New Zealand piece on New Zealand not being a “low wage” economy either – although I’d like people to be a bit clearer with their definition before they call anything low wage – but what the hell is this sort of article by a “data scientist”. [Note: I like data science and data scientists.  This is not an example of that.]

Now I’m back I’m calling this sort of trash “economics reporting” out – if we are going to publish articles where we are giving an objective rundown of the data we have to know the data and know the research at a basic level.

If we want to state our values (eg there are have nots in New Zealand that we want to support more) we don’t need to do this, we should all be allowed to state what we think matters and doesn’t matter.  But abusing the data and misstating trade-offs isn’t doing that, it is straight lace lying mate – irrespective of whether we disagree with the values associated with the person we are disagreeing with.

Reading it half a dozen times I think I’ve worked out what he is attempting to say.  Real GDP in NZ is probably between 15% and 30% lower in New Zealand than Aussie.  After tax wages are about 32% lower according to a measure they aren’t going to cite, because has no shame about what it does with data (Note: It is probably the OECD disposable income measures).  Therefore wages can go up by increasing the share of income going to labour and we don’t need productivity growth.  Ignore the logical incoherence of that conclusion for now, the very premise of it is flawed to hell.  Let me explain.

Now, aggregates don’t correspond to nice clean single person comparisons so even seeing a difference we can’t say much – as lets not forget we need to ask:

  • If GDP is measured differently in New Zealand (there is an argument part of the GDP gap is Aussie putting more things into their measure).
  • Whether we are comparing data that is measured over the same time period .
  • Whether the figures being compared are using a consistent deflator/PPP adjustment.
  • Given that GDP per capita tells us the market value of income all factors of production/spending on all final goods and services we want to be able to take into account the direct provision of goods and services by government when building an income figure for wage and salary earners!!!

Furthermore he is talking about wages vs other sources of income – self-employment income, benefit income, income from interest payments – has there been some other change in other income sources that explain this difference … we can’t just assume it is all corporate income.  Now we can see that corporate income shares have risen, but we can also see this comes from foreign owners of domestic firms – which implies that there are outflows of domestic production (which are owned by foreign owners).  Hence as we are interested in income in terms of the distribution domestic citizen purchasing power we should be looking at GNI or GNDI not GDP.  If we are worried about outflows of corporate profits then we are worried about the balance of payments issue in New Zealand, which raises the question of why domestic spending has exceeded domestic production in New Zealand … something that is often put at the door of New Zealand productivity being below the level Kiwi’s act as if they are achieving.

This is excluding the issue of measurement error … which given the constant revisions to the data, the difficulties of imputation and aggregation (and how to even interpret such a number), and fact that the population figure is always wrong until after a census update is incorporated, implies we need to take differences of a few percent with a massive grain of salt.

Right, even given all this the measure of GDP he is actually relying on is so unclear after saying he wants to focus on a 2016 income measure.  So lets get on the stats.OECD site and find 2016 values for GDP per capita PPP adjusted:  NZ = 34,628, Aussie 45,087.  The Aussie figure is just a touch over 30% higher … so the actual data is right at the top end of the range he gives us.  The actual difference between the wage measure and the GDP measure is small – especially when the gains often mentioned with respect to productivity are for closing this entire gap.

Now there has been actual work trying to understand why there is a gap in terms of income and GDP.  There has been work looking at the labour share, and noting that the labour share hasn’t fallen as strongly in NZ as Aussie.  But none of the actual research involving an understanding of the data gets mentioned in the article.

Honestly this:

It is a myth that productivity growth, as has long been argued, is needed to support higher wages in the private sector.

At this stage of New Zealand’s economic development, wages growth is as much an equity issue as a productivity issue.

Sounds like economics but is just a straight false statement – if the “productivity gap” is 30% this “issue” is much more substantive than a marginal change in the income share. Furthermore, at least higher productivity does imply higher output and thereby higher income – a change in the income share depends on the policy used, a policy that could well reduce average incomes … note such a change could well be equitable/fair but we can’t evaluate that without policy.

A related note is that, we can’t evaluate productivity truly without a policy either indicating that we actually have to do some work understanding why the data looks how it does in order to discuss it and policy … something that the author is understating in his comparison.

And when you say shareholders need to take responsibility remember that saying that NZ has low productivity doesn’t absolve shareholders of responsibility given they are involved in the production process – it just tells us to ask what is going on to cause this to happen.

If you think the distribution of income is unfair or “unequitable” that is fine.  But make that argument and make it using research on the real trade-offs that exist mate.  If you disagree with Business NZ saying that wages are not low that is fine, make that argument. Instead you’ve betrayed a lack of any perspective of what the data means in order to make an argument about productivity and wages that can’t be backed up by what you’ve stated.

People criticise the politics of someone like Trump (justifiably) as being anti-expert and anti-science.  But this type of reporting and the claim of “data science” in this specific context is just as anti-expert and anti-science!

There is actual research and analysis of these complex issues out there, there are actual frameworks for understanding these ideas, but instead of using that and providing a measured analysis of the data the article appears to have started with a conclusion (a dislike of Business NZ and discussions of productivity) and worked backwards to try to make the data fit into it.  I tell you right now, I’m not accepting that from anyone – I don’t care how crazy the world is right now, on this blog that does not fly 😉