NZ ‘slide down anti-corruption ranking’?

Yesterday Hamish Rutherford at Stuff reported NZ’s anti-corruption record slipping: watchdog.

New Zealand is slipping down the ranks of the least corrupt countries, with watchdog Transparency International accusing the Government of “astonishing” complacency.

After topping the Corruption Perceptions Index (CPI) for seven years in a row until 2013, the 2015 survey ranked New Zealand behind Denmark, Finland and Sweden. In 2014 New Zealand was ranked second, behind Denmark.

The survey draws scores from a range of other surveys to give an overall rating of the perceptions about corruption for 167 countries. In 2015 New Zealand scored 88, a marked fall from the 91 it scored in 2014.

Care should be taken reading too much in to minor changes in ratings and rankings on what are a collation of “perceptions of public sector corruption” with margins of error in the data analysis.

There was also an Opinion piece by Rutherford: Slide down anti-corruption ranking should be a wake up call.

While being ranked fourth out of 167 nations in the annual corruption perceptions index might sound impressive in isolation, it is hardly the gold standard, or what it used to be. New Zealand was ranked top of the index year after year until 2013.

Being among the least corrupt nations does not quite have the same ring to it. Nearly as good as the best. There, or thereabouts.

Those who have seen first hand the abuse of the law protecting our right to information might take some cold satisfaction from the slow slide in the rankings, or at least will not be surprised by it.

And Labour leader Andrew Little has picked up on this. NZ loses squeaky clean corruption-free reputation:

A second consecutive drop in a global transparency index busts the notion that New Zealand would remain corruption-free, Opposition Leader Andrew Little says.

“It is an indictment on the Government that we are now lagging behind several other countries.”

“New Zealand’s highly-regarded international reputation for fair dealing has been eroded. It is shameful that the standards of transparency under John Key’s leadership have slipped so far,” Andrew Little says.

So how big has the ‘slide’ been? From first to second and now to fourth out of 167 countries, and from a rating of 91 to 88.

But this is a ‘perception of corruption’ and Graeme Edgeler points out that the drop could be due to:

Have looked at the methodology and data underpinning the Transparency International Corruption Perception Index to see why New Zealand fell.

It is because the “global business executives” who respond to surveys for the IMD World Competitiveness Center answered the question:

“Bribing and corruption: Exist or do not exist”

on a scale of 1-6 ever so slightly different last year than in 2014.

Limiting freedom of information? Search warrants on journalists? Environmental governance? That made no difference whatsoever.

Instead, it was that one question, where probably one or two NZ CEOs changed a 6 to a 5 this year, because who knows why.

Now Rutherford and Little may have raised valid points about concerns about corruption in New Zealand, but this survey hardly signifies a ‘slide’. It is more like a minor tweak.

Edgeler also tweeted:

Last year we had a better rating than in 2012. It’s just others had improved more.

More importantly, however, it’s not a corruption rating, but a perception of corruption rating. If the government was to crack down on corruption, the perception of it would increase and we’d do worse.


I’m not saying it’s a good thing, it’s just one of many stupid things about coverage of the index and the index itself.

My point is that these indexes are generally stupid, and nobody should read anything into changes in them.

Fair point. The survey may be of minor curiosity value but an attempt at measuring perceptions and opinions with a very small sample related to New Zealand there must be a significant margin of error.

Edgeler continued on Twitter:

I started a post. Lost it, and probably can’t be bothered starting over, so I’ll do a short tweet summary instead.

TI doesn’t do it’s own research for this any more. They use data from questions around corruption in other’s surveys. Ten surveys overall, but not all cover all countries (eg the African Development Bank Governance Survey).

Data from seven of the surveys is relevant to New Zealand. In three of those surveys, our result is unchanged. In three of those surveys, our scores increase.

Only in one of those surveys does New Zealand drop.

I tweeted the single relevant question from that survey earlier. 6100 self-selected executives around the world responded to that survey.

If New Zealand had it’s fair share of executives in that 6100, that’s about five people. One of them choosing a 5 on a scale of 1 to 6 on one question is the sole reason we’ve dropped in TI’s ranking.

Sorry, that’s drops in two indexes 🙂

3 unchanged + 3 increases + 2 drops <> 7 surveys so not sure what other number changes.

Bryce Edwards:

The score of 88 is the average of the seven independent assessments figures of 92, 93, 81, 83, 98, 88, 83.

Averaging 7 assessments assumes they have equal weighting. I presume they are not of equal input.


Now compare them to the numbers from the previous year.

These numbers are all in Transparency International’s Corruption Perceptions Index 2015.

Their press release is brief, doesn’t mention New Zealand and lacks details. It just says “the Corruption Perceptions Index is based on expert opinions of public sector corruption”.

Their “Short Methodology”:


In Frequently Asked Questions they say:




So the CPI is an ‘indicator of perceptions of public sector corruption’ and ‘not a verdict on the levels of corruption of entire nations or societies’.

From Technical Methodology:

Each Country’s CPI score is calculated as a simple average of the available rescaled scores.

The CPI score will be reported alongside a standard error and 90% confidence interval which reflects the variance in the value of the source  data that comprises the CPI score.


The seven sources for each of the top five countries were the same.

With tight results plus Standard Deviation and Standard Error on perceptions of public service corruption I think you need to be cautious about reading too much into slight changes in ranking.