In Australia we had the Victorian bushfires. It wasn’t exclusive to News obviously. But our coverage was unique.  We sold an extra half a million newspapers in the week following Black Saturday. Our website traffic more than doubled.  We took an entirely different approach with this story.  We used resources from every newsroom in the country.  Online staff in Brisbane helped the  masthead team in Melbourne moderate the tidal wave of public contributions.  Our editors across the country sent writers and photographers to work with Herald Sun staff on the ground. This gave us more manpower than any other media outlet.

Sounds like the ’system’ working — invest in quality and you get readers. Right? He immediately goes on:

But, what drove readership and web traffic was the content.Readers embraced the opportunity to sign online condolence books and write tributes to victims. We set up online forums so readers could search for news of those lost and those rescued. They told us miraculous stories of those who cheated the flames and heartbreaking accounts of those who didn’t.

So what drove interest was not reporting but user-generated content and social networking! That’s not the traditional model that is the new one. By accident, News provided an outlet for that, realised it drove readership and didn’t realise that that was the point. But wait, he goes on:

Alongside traditional reporting from the scene, we had incredible eyewitness accounts from readers, including amazing pics and video.

News is running a Flickr/YouTube/Facebook combination!

Then he moves on to rail against bloggers. His argument is that they have poor quality and spend their time ripping off news outlets (what does that say about the quality of the news outlets?) and attract readers from those news outlets — those readers who simultaneously used to value the high quality of traditional news outlets but now prefer what the bloggers provide and so “used to know better” but now “no longer no better.” And, by the way, if you are reading this you are one of those transformed readers who are destroying quality and you should get a grip and transform back. (I am sure he will point to the contrast between Sam Wylie and Alan Wood on banks today as a clear example of a quality differential).

Anyhow, he gets better towards the end of the speech with a call to quality to attract readership. The problem here, as I am certain any News Ltd CEO knows, is that quality does not necessarily attract readership. What attracts readership is having something to talk about with peers. Newspaper sites can facilitate that and appear to be successful when they do. The question is: when will their managers realise that.

Most people believe the election results in Iran were rigged. They base this on a couple of arguments. The most obvious is that Ahmadinejad did unreasonably well, including especially in areas where you would expect him to poll poorly - for instance in the home seat of his main opponent Mousavi. This argument would sway most people but it is not scientific. Ahmadinejad could still say that he ran a great campaign and that the people decided he was a safer pair of hands.

Another more statistical argument is that the variability in Ahmadinejad’s vote is too small. By too small, I think they mean less than one would expect from the regional variation that one normally sees. So this does intersect with the first argument and Ahmadinejad’s high vote in some unlikely seats. But it is a distinct statistical view in that we focus on variation across seats rather than overall mean level. Certainly, if one could actually show that the variability of Ahmadinejad’s vote was less than binomial variation this would be pretty damning and suggestive that someone had just made the figures up.

Well, it turns out that the variability of the vote is about 100 times higher than binomial – there is plenty of regional variation overall. I suppose we could compare the variability with a different benchmark – such as the variability of the vote for winners of previous elections – but the argument starts to lose force as there are other explanation of why variability might decline.

So, if we are interested in revealing whether the election count data has been concoted, let’s focus more on the process of human beings making figures up. Humans are pretty bad at making figures up. Human generated data typically looks too good to be true and follows  theory too well. It is well known, for instance, that Mendel’s famous pea data was probably concocted, perhaps by his assistants. Forged signatures can often be recognised by experts as being too consistent. Real signatures are not perfect and vary from day to day.

Humans are especially terrible at generating random numbers. And for a large voting count, for instance 325911 which was Ahmadinejad’s count in the region of Ardabil, the last few digits should be essentially random. On the other hand, if someone were making the numbers up and not concentrating too hard on the unimportant final digits, you might expect to see some tell-tale signs of non-randomness in the those final digits.

This idea is due to Alexandra Scacco and Bernd Baber who have suggested that there is indeed such evidence in the data. They claim that human generated random numbers tend to have too many 7’s and not enough 5’s. And looking at pairs of digits, they claim that human generated digits will have too many adjacent sequences such as 23 and 76.

The data for the 2009 Iranian presidential election are HERE and a graphic of the marginal distribution of the last digit is below. There are indeed too many 7s and not enough 5s. The overall goodness-of-fit of a uniform distribution has a P-value around 8% but this may underestimate the evidence. If we concentrate on the a priori hypothesis of too many 7s and not enough 5s then the chi-square statistic is much, much stronger.  

But is the hypothesis of too many 7s and not enough 5s really a priori? I could not find any evidence on the web for the assertion of not enough 5s but there is some prior reason to look too many 7’s. If we just look at the excess of 7s the P-value is around 0.4%. The excess of 20/116=17.2% sevens over the expected 10% is very suspicious indeed.

One might obtain even stronger results if we concentrate only on those electorates where Ahmadinejad’s vote was likely to be poor. One assumes that the fraudsters would not alter the counts in the electorates where he won. So the random digits in these true counts might dilute the non-randomness in the fraudulent counts.

I also had a look at the last pairs of digits hoping to find something even stronger but I could not recover the results quoted in the Washington post article. Moreover, I could not find any evidence for their claims about adjacent sequences, though I have heard that two digit primes tend to get preferentially chosen. Anyway, I challenge readers to look at the digits and find some really damning evidence of non-randomness (which we would have to correct for the effort you put into the search!)

If I were going to fudge some numbers, I would start multiplying the real counts by some scaling factor, or leave the last digits alone, or even use a random number generator on my mobile phone! I guess criminals and fraudsters are not always very smart.

When the perfect order prevails, the world is like a home shared by all. Virtuous and worthy men are elected to public office, and capable men hold gainful employment in society.  Harmony and trust are strictly observed. All men love and respect their own parents and children, as well as the parents and children of others. Thre is caring for the old, their are jobs for the adults, there are growth for the children. There are support for the widows, the widowers, and for those who are alone and disabled. Every man and woman has an appropriate role… A devotion to public duties left no room for idleness…these are the characteristics of a commonwealth state…

There you go, disability pensions and full employment are clearly there in Confucius, predating Keynes by 2500 years!

The analysis is based on FT ranking data and done at the level of the school. The authors do make an attempt to account for he problem of endogeneity of GMAT scores. They conclude that

Controlling for the average quality of their students and some other schools’ characteristics, average salaries are significantly greater for those schools whose core MBA courses contain a higher proportion of management courses as opposed to courses based on economics or technical disciplines.

However, there is a major problem with the findings it seems to me. They measure the success of the course by average salaries (3 years after graduation). There is no measure of management ability. It seems to me that the authors have fallen into the very mindset that they decry. Measuring manager performance by the salary that the graduate manages to negotiate for themselves is precisely how a (rather poor) economist student might think!

I am not at all surprised that those students who major in management subjects (such as human resource management, negotiations and leadership) are better able to find a high paying job than those who major in Finance and Econometrics, poor tongue-tied geeks that they are.

*via Nick gruen at Clubtroppo

Today Ross Gittins commits this fallacy in arguing for stamp duty on home transactions. The traditional economic wisdom is that such taxes which are non-trivial sum of money in many states, reduce buyers’ willingness to pay for property and so reduce the number of transactions at any given point of time. How much? That is an empirical matter and from what I can gather then effects are real.

The trouble with this advice, however, is that like all neoclassical analysis, it’s based on an erroneous model of human behaviour that assumes the choices we make are always carefully calculated to maximise our material wellbeing.

For the past 20 or 30 years, behavioural economists have been pointing out to conventional economists all the flaws in their assumption that people are always rational, but this seems to have had zero impact on the happy analysis of the tax economists.

There is so much to deal with here it is not funny. First of all, zero influence? Starting with Sunstein and Thaler onwards it seems that tax is one of the major areas where behavioural economics has mattered. Think about this example for one. Second, Gittins argues that because people are spending so much money they don’t care about the tax. I don’t know where the evidence for this is but consider the sales strategy for new properties ‘off the plan to avoid stamp duty.’ That works because people care. Third, just because a land tax is politically unacceptable does not mean stamp duty is good and efficient even if it is a nice little earner for state governments.

Fourth, just because stamp duty might matter does not mean that anyone thinks it is the most important thing for every circumstance. So this statement really misses the mark.

Similarly, only an economist would be stupid enough to imagine conveyancing duty is a significant factor in explaining an executive’s reluctance to uproot their spouse and school-age children and move them to another capital city.

And, finally, in a similar fashion, no one is talking about encouraging mobility the only issue is not discouraging it. Ross Gittins seems to believe that it is not possible to move back to your social network but only away from them. Taxes may well discourage the former while preventing the latter. Which is socially beneficial depends on your perspective.

When it comes down to it, there is no evidence that behavioural economics as a clear and formal theory translates into housing markets through this route. Gittins cites no study or theory that suggests this to be the case only a bunch of unrelated stuff and ranting out economists. Indeed, it is precisely the blind application of general theory that behavioural economists are so against and so Gittins commits these same sins in spades.

Click HERE to open the tool in another window at a much sexier version of the original graphic (for the year 2007). You need Flash 7 and it may take 30 seconds to load but it is worth the wait.

A quick explanation. Each point is a country and upper right means high GDP and high life expectancy. GDP is on a log-scale partly because the distribution is so skew, but also because the relationship is almost linear that way. Colours are different continents – with Africa in blue. Size of the blob is population size – probably not of primary interest here. But the really cool thing here is the last dimension – time. Move the slider back to 1800 and hit the play button to see the data displayed in sequence for the past 200 years.

My original interest in these data was the AID epidemic in Africa. Focusing on the blue swarm of African nations, you will see that there was virtually no improvement in life expectancy until after WW2. Over the past 30 years however, you will also see the blue swarm stagnate to the bottom left. Then, during the past decade about 10 middle income countries just drop through the floor. Pretty stark it is.

You can follow individual countries just by clicking on them (or on their name in the list at the left). Below is the trajectory for Iraq. What do you think happened in 1979?

Other countries whose trajectory has a story to tell – Rwanda but ther civil war was devastating it pre-genocide (1994), Chile which did well after Pinochet (1974) but was doing extremely well in terms of life expectancy before 1974. Check out Russia…it hardly looks like the wall coming down was a resounding success.

Let’s get parochial. Apart from minor glitches after the great depression and WW2, Australia has enjoyed an uninteresting march towards wealth and health. Give me a boring trajectory any old time. For those who ever doubted that NZ was the eighth state, look at their trajectory at the same time as ours. Peas in a pod.

China watchers might like to focus on the middle kingdom around the late 1950’s. I was actually impressed that they managed to get data on life expectancy from a nation that does not like its dirty linen aired. But this thought process led me to a more obvious question – what does life expectancy mean for 2007? This has to be a model projection – it is not really data at all. So I checked the documentation and it is defined as “the number of years a newborn child would live if current mortality patterns were to stay the same.” In China’s case then this would mean that the drop we see is an extrapolation of what would have happened if the great leap forward continued indefinitely. Ditto Rwanda. It is certainly not the case that the babies who were born in Rwanda in 1991 will have a mean life of 24 years.

I look forward to the inclusion of 2008 and 2009 data so we can see the ffect of the GFC and how it is differentially felt in different countries. Look out for the Icelandic bubble to bounce faster than Novak Jokivich’s service routine. You can waste hours playing with this site. There is a huge range of economic, environmental, health, education and demographic measured to select from.

If you want to use this tool for your own data.. .. you can’t. But GapMinder suggest using Motion Chart, which is a free gadget in Google Spreadsheet (an online spreadsheet similar to excel).

Cross Posted at ClubTroppo and Fishing in the Bay

On their ShowtheLoo site is the following blog post which I quote in its entirety.

Nigel Hopkinson said he was disputing a charge of $8562.31 for 73GB of excess data use. The apparent usage was recorded by BigPond between midnight and 5.07am on May 10. He received an automatic email at 8.53am from BigPond, advising his account reached 175 per cent of his monthly 60GB usage allowance. He had left his computer and modem on, he said. Other customers that have been affected by excess data usage claimed to have turned their computers and modems off.

“I had no software running that would cause this and I do not believe with my connection speed, that it is even technically possible,” Mr Hopkinson said.

Another BigPond customer, Eian Mathieson, who is on a 25GB-a-month plan, said he began to worry about his account usage after having gone over the limit two months in a row.

Mathieson downloaded a metering tool, Netmeter, to measure his usage patterns. On one day, BigPond metering recorded more than five times the usage Netmeter reported, he said. On May 27, the BigPond meter showed he had used 1768MB, whereas Netmeter showed 342.6MB for that same day. The next day, the BigPond meter displayed 1650MB of usage and Netmeter recorded 358.62MB, Mr Mathieson said.

Well, add me to the long list of customers with this complaint about BigPond. I am on a 60GB per month plan which I went on a year ago because we seemed to be exceeding our 30GB and I didn’t want to worry about it. The problem is that we now seem to be reaching that limit too. Finally, I woke up and thought about it: how?

We download lots of TV and movies from iTunes but adding that up, it was clear it could not exceed 10GB in a month. There is plenty of YouTube viewing but even there it was hard to image it exceeding 2 or 3GB per month. Where was the rest of it given that we are all out of the house during the day? I tried an experiment switching off the WiFi just in case it was the neighbours, but the meter kept on ticking away. Surely that amount of downloads each month would be noticable on a hard drive somewhere but it was no-where to be seen.

I started to monitor usage closely. Since there are several computers network meters were fine but over-stated usage as they included file transfers. But even with this, they were way under what Telstra’s usage meter was reporting. So I rang technical support. We considered viruses but two things ruled that out: (i) there were no PCs in the house, only Macs and (ii) it was all downloads, with hardly any uploads. Actually, the upload rate seemed high to me but there was a vast asymmetry. I suspect that told them that no computer had been hijacked to send out illicit material and also that I wasn’t downloading illegally and getting into trouble on reciprocal uploads. So that month they reset my limit so I didn’t get extra charges.

But what to do? Telstra’s usage meter doesn’t get the time right so even when we switched it all off for a day it still recorded usage (an incredible 200MB!). So I can’t just monitor stuff for a day, send it to them and say, ah ha! I would have to do it for a week or more which would mean limiting to one computer and a big disruption. So I monitor the usage meter because, believe it or not, our usage is correlated with that and so close to the end of the month we lay off the iTunes, iView and YouTube.

This article has given me pause and I may make another call to BigPond support, or is it billing? I recall they had trouble working out which department was to blame.

Among Dr Gruen’s favourite sites already making government information freely available is one in which the NSW Registry of Births, Deaths and Marriages enables intending parents to check the popularity of names in each year from 1900.

Peter Martin approvingly directs us to the site in question as exemplifying the point of Government 2.0. Here it is. What it shows is a time series graph of 1200 popular baby names over the last century. You can hover over bits of it and find out just how many NSW babies had a name in a given year. Seems need until you ask yourself one question: what is this telling me? Near as I can tell, it is telling you that the birth rate has changed over time and that there are clumps of popularity-weighted boys and girls names at different parts of the alphabet — (yes, the plot is in alphabetical order). So perhaps it is a more convenient way of looking at name popularity than this.

Now, to be fair, it is an example of how data might be presented than is different from what is usually the case within government and so that is fine. Indeed, it is based on this. But the point is that it shows us that we can do better and indeed, it is precisely why the government needs to free the information so that others can provide it in a useful form and innovate on it.

So here is the question: is the NSW baby names data set publicly available as the SMH suggests? I made a few inquiries (and will try and confirm them officially and post an update) but it appears not. You can’t just ask for the data-set so that you can host it on your own site in a useable fashion. It all still resides with the NSW government. So rather than being what the goal of this taskforce, it is, in fact, a clear example of the problem. It is not making information freely available at all. That would require a ‘click here to download a spreadsheet of the data’ button. I’ll be looking out for it.

Despite widespread outrage about mortgage rate increases and the introduction of direct charging of ATM fees, a spokeswoman for the NAB told the Herald only two customers a week had used the “listing and switching” service since it was launched in November.

The chief executive of the Australian Bankers Association, David Bell, also told the Herald that uptake of the service had been very low.

The Government said last year it would inject competition into the banking sector by helping customers “vote with their feet”.

The scheme is a very limited one — allowing customers a list of direct debits and some help re-organising those. But when it comes down to it, that still takes time and paperwork. So it isn’t surprising that there is limited use of what doesn’t look like a very valuable service.

Instead, we need to think larger.

A professor at the Melbourne Business School, Joshua Gans, has called for all bank account numbers to be made portable, like mobile phone numbers, to enable people to switch easily. “The problem is this is not a piecemeal policy issue. This requires a lot of major changes to do it. It does require some serious investment and legislation to be passed.”

There are all manner of estimates as to the cost of this but our experience in telecommunications — an industry with much greater challenges than banking — was that the costs were exagerrated by an order of magnitude and implementation occurred in a timely fashion. The problem is that we also need to deal with switching of accounts in debit (e.g., mortgages) in order to make this work. I don’t have time in this post to outline how that could be done but I think there are mechanisms that could be deployed there too.

That said, the measure of success of any system to reduce switching costs is not how many people use it. Indeed, in principle, the existence of the service is all that is required and by facilitating competition — in particular, to retain customers — it will promote the social good. What we, in fact, want is for consumers to have the option of switching easily so that when they complain to banks about terms and conditions banks will adjust those terms and conditions to stop them from leaving.

Instead, the measure of success is an improvement in competition. And that is where the system is currently not effective. With news that the major banks command major market share and now major profit margins post-GFC, surely the pressure on the government to do something meaningful about bank competition should be on the cards. Resting on the laurels of a functional financial system relative to the mess the rest of the world is in, is simply no excuse.