Author: Louise Pryor

Categories
LouisePryorNews

Hello again!

This is the first issue of an occasional newsletter. It’ll be a bit like the one I used to send out until a few years ago (which is archived on my website), but will be a bit unlike it too, as my interests have, naturally enough, changed over the years.

After four immensely interesting and fun years at the Board for Actuarial Standards, where we developed a brand new set of principles-based technical actuarial standards from scratch, I’m now freelancing again. I’m specialising in risk, modelling and governance, basically helping people to understand and manage – in the broadest sense – the models they use and the risks associated with them. My website gives more details, as you’d expect. If you’d like to know more, do get in touch.

I’m also blogging again. Sometimes my definition of what’s relevant is loose in the extreme, but I hope you’ll find some of it interesting. Recent topics include the new modelling,  Benford’s law and modelling and what Google’s all about (and by the way, it’s not search or advertising). If you’ve got any comments on these or other posts I’d love to see them — just add them to the blogs. Future issues will highlight what I consider some of my more interesting blog posts, or ones that have given rise to interesting discussions. Do let me know if there are topics that you think should be covered.

 

 

Categories
Actuarial Data

The new modelling

Data is the new modelling. That is, it’s where all the sexy stuff is going to be over the next few years. Over the last few years, in the insurance industry at least, modelling has been where its at. Driven largely by Solvency II, a huge amount of effort has gone into building and, now, validating, hugely complex financial models.

But now, in the insurance industry as well as others, data is coming to the fore. After all, what is a model without data? And, as we all know, Garbage In, Garbage Out is one of the fundamental tenets of computing. The FSA has pointed out that data is a key area for the successful introduction of Solvency II and has produced a scoping tool that will help them assess a firm’s data management processes.

And it’s not only Solvency II. At GIRO last week there was an interesting debate over whether telematics will be at the heart of personal motor insurance in ten years’ time. The thing about telematics is that it produces large quantities of data. With the Test Achats case meaning that gender won’t be able to be used as a rating factor, insurers are going to be looking for other ways of coming up with premiums, and other factors they can take into account. The thing about gender, of course, is that it doesn’t take much data. It’s just a single bit in the database. Other rating factors may have more predictive power, but it’s harder to get at them.

We’re seeing this everywhere, though. As computers continue to get more powerful, and data storage gets ever cheaper (how big is the disk drive on your laptop? — even my phone has 16GB), doing things the rough and ready way with only limited data has fewer and fewer advantages. Big data is becoming mainstream: look at Google, for instance. And why did HP buy Autonomy?

You mark my words, a change is gonna come.

Categories
Data risk management

Fiddling the figures: Benford reveals all

Well, some of it, anyway. There’s been quite a lot of coverage in on the web recently about Benford’s law and the Greek debt crisis.

As I’m sure you remember, Benford’s law says that in lists of numbers from many real life sources of data, the leading digit isn’t uniformly distributed. In fact, around 30% of leading digits are 1, while fewer than 5% are 9. The phenomenon has been known for some time, and is often used to detect possible fraud – if people are cooking the books, they don’t usually get the distributions right.

It’s been in the news because it turns out that the macroeconomic data reported by Greece shows the greatest deviation from Benford’s law among all euro states (hat tip Marginal Revolution).

There was also a possible result that the numbers in published accounts in the financial industry deviated more from Benford’s law now than they used to. But it now appears that the analysis may be faulty.

How else can Benford’s law be used? What about testing the results of stochastic modelling, for example? If the phenomena we are trying to model are ones for which Benford’s law works, then the results of the model should comply too.

Categories
Society

Ada Lovelace day

I missed Ada Lovelace day! It was last week, and is intended to recognise women in maths, science and technology. Ada Lovelace was Byron’s daughter, and worked on Babbage’s Analytic Engine, and has been called the world’s first computer programmer.

Categories
Data

What Google’s about

What is Google about? Is it about search? or advertising? Actually, it’s probably about data.

I’ve been reminded of this by a couple of recent articles I’ve read. The Economist’s Babbage blog has a good piece on the Google Internet bus – a free, mobile cybercafe that operates in India.

“It has covered over 43,000km and passed through 120 towns in 11 states since it hit the road on February 3rd, 2009. Google estimates that 1.6m people have been offered their first online experience as a result.”

And also as a result, Google has a huge amount of data. I was reminded of Google’s appetite for data by a recent review of several books about Google in the London Review of Books. In 2007 Google started up a directory inquiry service in the USA:

“You dialled 1-800-4664-411 and spoke your question to the robot operator, which parsed it and spoke you back the top eight results, while offering to connect your call. It was free, nifty and widely used, especially because – unprecedentedly for a company that had never spent much on marketing – Google chose to promote it on billboards across California and New York State.”

People wondered why Google was doing this – it definitely wasn’t a money making exercise, and indeed only lasted for three years. What Google was doing was collecting a huge amount of phoneme data that it could use in voice recognition technology. It is almost certainly doing the same with its bus in India.

The availability of large amounts of data has changed various aspects of technology in many ways. Off the top of my head, it’s influenced voice recognition, natural language understanding and translation (which are rather different from each other) and of course has had a huge effect on marketing in general. Then there’s the forecasting of epidemics, and various initiatives to make data freely available to all. There is also now much more demand for large amounts of data storage, both physical (disk drives etc) and in software (database technology). It’s a chicken and egg situation with data storage, of course.

 

Categories
Environment

Does locality matter?

I’m a complete glutton for soft fruit, and love it when the shops are full of local (or at least british) berries. But there are some times of the year when there just isn’t much local produce around, and sometimes I really like the idea of some fresh green beans, say, even when they aren’t in season over here. But I always feel rather guilty about buying air-freighted fruit and vegetables – surely the carbon costs of all those air miles can’t be a good thing?

So I was really interested to read this blog entry in the Observer. There’s apparently a good argument that western consumers actually have a moral duty to eat strawberries out of season. There are several reasons – food miles within the UK are by no means negligible, and sometimes are actually more than for imported produce. Also, it’s the people growing fruit and veg in sub-Saharan Africa who will suffer disproportionately from climate change, so is it right to make them suffer in order to avoid climate change? And, anyway, if you buy imported organic produce you’re probably saving more emissions from the production of synthetic fertilisers and pesticides than you’re causing from air freighting. So the argument is that if you buy organic and Fair Trade you are outweighing the food miles.

It’s clearly not a simple matter, and I bet there are good arguments pushing in both directions. And in many ways life would be a lot easier if there were simple answers to questions like this. But from a purely selfish point of view I’m glad that I needn’t feel too guilty about occasionally indulging my taste for foreign fruit and veg.

Categories
Actuarial Uncertainty

Getting rates in a mess

Another good blog post from Understanding Uncertainty: for once not based on a howler from the British press. Instead, it’s based on a howler from the German press – High suicide rate in German forces serving abroad – every fifth soldier takes his own life. They actually meant to say one in five of deaths among soldiers serving abroad.

The post goes on to discuss whether the suicide rate is higher or lower than would be expected, and along the way gives some explanations of the concepts behind exposed to risk (without actually using the term). A great example of how to explain something that can get pretty technical in an uncomplicated way. If only most actuaries could do the same.

Categories
Society

Women’s suffrage – old hat?

You might think that the struggle for women’s suffrage is over, especially with the recent news about Saudi women. But you’d be wrong, according to this interesting blog.

Categories
Uncertainty

Critical thinking

A very good blog post in the Guardian by Jon Butterworth, discussing the faster-than-light result (or possible result). One thing that wasn’t at all clear from the mainstream press coverage was that the whole thing depends on probability distributions. It’s obvious when you think about it (it’s not as if you can label an individual neutrino and send it from one place to another, hundreds of miles away), but it does change the complexion of the whole thing. It also makes it rather more complex, of course, and as usual there is massive scope for disagreement about statistical techniques and whether everything has been allowed for.

I’m looking forward to the next chapters in what I’m sure will turn out to be a long story.

Categories
risk management

Arbitrage or speculation?

A good piece in The Register about the difference between arbitrage and speculation, and how it can all go horribly wrong.