New Zealand does actually have a department of Statistics. Their URL is: http://www.stats.govt.nz/default.htm
And in general, it is pretty well set up and would appear to do exactly the sorts of things you would want from such a body. But this is where Statistics get their appellation of “damned lies” from. It’s just not so straight forward is it.
In my first ‘Statistics post’ I drew up a fatuous unemployment statistic. I concluded that we only had half a million people actually doing any work in this country. Officially, the numbers are drawn from the Household labour survey. The numbers there are rather significantly different than I got from my little exercise. The differences come from embedded assumptions and contexts. For instance, the household labour survey assumes that the workforce consists of everyone over 15 yrs, while I assumed that it was 20 yrs. I will leave you to make your own judgements about that. Now I know that they start measuring from 15 yrs because they say so in the section at the end of the report document. In fact they have quite a useful and significant explanation of how and why they come up with the numbers they do. For instance, they have a seasonal adjustment they make to the figures in order to smooth out regular but transitory fluctuations to the numbers. they even state what their methodology is: (All seasonally adjusted and trend series are produced using the X-12-ARIMA Version 0.2.10 package developed by the US Bureau of the Census) So that is quite helpful and useful to know.
They also explain that the survey which provides the official unemployment rate is generated each quarter by taking a representative sample poll of the population using 15,000 households and 30,000 individuals (actually true?). Now in general, I don’t have too much of a problem with that, polling if done properly can produce very accurate results from relatively small samples. However, on the other hand, New Zealand is not a very big country, we are approx 4 million people and we are all already tied into the Internal Revenue Department systems anyway. Why isn’t the employment numbers taken straight from that data base and then corrected if necessary. We don’t have to assume anything, between all the government depts we will have all the information we could every need on everyone in the country, and in consequences, extract all the gross data needed for any statistical analysis. Why are we not?
Well actually we may well, but no-one is saying if they are (there are even laws against data matching between depts). Because now we are getting into the realm of politics and thats insidious little effects on data. IRD will have all the data I am talking about regarding people working, but where does all that data go? Presumably it will be funneled to the minister if he requests it and presumably also it is part of the information flow and process of drawing up the annual appropriations and spending Budget. But it is not like we are going to see that data now is it. Ohh no, that is altogether a very tightly held little secret. Certainly it does come out in the end. When the budget is released, that has “all” the “relevant” data-points enumerated, albeit only after everything has been massaged to fit with the programme as well (you know, like company financial reports… Enron, Worldcom, Parmalat?).
How about if we regularly each week had an Official Information release of all the data that has been presented to, or requested by ministers, if we knew as much as they did about what they are working on! Hmmm, I suspect not, at least not if they have any say in the matter. Afterall, what’s the point of being the high priest if you can’t have your little secrets and lord it over everyone else with your superior knowledge and ability. And we all know just how admirable and intelligent MP’s are </Sarc>. When I was involved in politics a good number of years ago, one of the MP’s made the comment that parliament is actually the House of Representatives. He went on to say, in any society you have a certain proportion of people who are liars, cheats and thieves…. and that parliament is no different. All the characteristics of everyday society is manifested equally in parliament – the good and the bad. Actually, don’t get me wrong, there are a good number of MP’s who do a very good job and put a lot of very hard work and dedication in to what is mostly a pretty thankless job. But that’s only a relatively small part of the whole equation.
When a new MP first arrives in parliament, they get inducted into a system and institution that has been designed and evolved over a very long time. You aren’t going to be leaping in and radically re-arranging anything, instead it is going to normalise/socialise you. What is more, if you play the game correctly by the rules, it will introduce you to the rewards of power, or at the very least the “baubles of office”. So then we have the problem of – power corrupts and absolute power corrupts absolutely. Everyone gets rewarded for staying within the norms and accepting the status quo. What the status quo does is perpetuate a system, where it is now in the best interests of everyone to accept certain presuppositions and to manipulate data to achieve your objectives. Political parties have their agenda’s and government depts have theirs as well. In general they all find they can knock along together OK with a a wink and a nod. Remember “Yes Minister”? The system works, at least insofar as serving the system, but not quite so well insofar as serving the truth or the public at large. Manipulation is built into the system, at this point only a revolution is going to change anything there, albeit history would suggest you only really end up swapping one type of failure for a different type of failure. Still, no need to start off pessimistically, yes?
What we are trying to achieve is some clarity, enlightenment and yes even truth. So that means the Statistics have to be out of the hands of interested parties, whether that be political party’s, government depts and senior civil servant, commercial institutions or any other sort of lobby group. It is far too easy to influence data and the results of data analysis if you can frame the questions or otherwise control the context. The Household Labour Survey is a very interesting example. It attempts to answer questions directly relevant to the political economy while at the same time whole relevant data blocks are not included in the analysis. Who gets to decide that X.Y and Z groups are not relevant or to be included in generating a result? When they say that the unemployment is 5%, that is after the whole question has already been skewed by preconditions about what we are measuring and how. I recommend you take the time to read the whole of the Household Labour report and see what sort of distortions you can foresee in what their methodology would produce.
Unfortunately, it is not just methodology that is at issue, another is just raw content. Have a surf around the Statistics NZ website, in theory they have stats on just about everything you could imagine, except they don’t. Look under any number of headings, click the links, and actually there isn’t any data or analysis or summary at all. So, all form and no substance? What is actually going on? When I looked, there was a statement saying that during July 09 there was going to be a revamp of the website and that things would be re-arranged much better. So arguably, there was no point in them putting up pages that only a short while later would have to be taken down and then reformatted. I.T. is not a cheap exercise and they could be forgiven for husbanding their resources and proceeding as efficiently as practical. Assuming of course that the data does actually make it onto the system at some point and doesn’t just somehow get lost in the mix perpetually, which unfortunately is rather more the style of government depts. If it takes too much trouble for anyone to get around to auditing the performance of a system then any number of sins can slip through the cracks. How many ordinary people I wonder have ever looked at the website, surfed around, discovered issues and raised them forcefully enough so that they get addressed?
It isn’t just political interests which effect results, bureaucratic systems and human nature can also weave their subtle magic and result in the information coming out as either lies, or at best, distortions. As it happens, I just read something today that explored this effect, read here for more if you are interested. What it means is you have to be very careful about data and that rubbish in equals rubbish out. It also means you have to be super cautious about who gets to have a hand in the collection, processing and dissemination of data. Basically, I am rubbishing Statistics NZ and the Household Labour Survey, because how can I not? There is too much room for errors to occur, as well as too many incentives to introduce them. Coupled unfortunately with too little incentive to prevent it – “thats somebody else’s problem…” So, I like the idea of Statistics NZ, pity I just can’t believe in it. Which is a problem because I would dearly like to be able to, and the need for them is so important. The basic form is there of what is needed, can we actually turn it into what it needs to be? It needs to be clearly independant, and it needs to be open to examination, audited even, in a way that is transparent and trusted. lastly, it needs to turn out a product that is clear, accurate and actionable because that in turn will cast a light on the very influences, political and economic that most need the attention and the feedback. Statistics should be affecting politics, not the other way around.
BTW, the Statistics NZ website has a neat “population clock” in its Banner