Tech Trends

Posted on by David Wright

I’m beginning to think that there are only hardware trends in tech (faster, cheaper) and social trends in software.

My favourite question to ask people in tech is: “what about your job could not be done 15 years ago?”. Most of their answers have to do with standards (common protocols emerging) and organizational will towards technological solutions (“my boss 15 years ago didn’t understand the benefits and the employees didn’t have the skills to implement”).

This says to me that innovation is not cross-pollinating between software industries. The fact that google and 4square exist doesn’t have any direct implications for insurance company efficiency.

To understand where the next breakthrough will come, take today’s technology and add higher median computer familiarity within an organizational system.

The days have arrived where back offices of companies are populated by middle-aged men and women who do nothing but administer computer systems all day. That’s a pretty interesting development.

So, what’s next?

The Power of Data Mining

Posted on by David Wright

Reading about Emerging Adolescence in the NYT, I was particularly annoyed at this:

People can vote at 18, but in some states they don’t age out of foster care until 21. They can join the military at 18, but they can’t drink until 21. They can drive at 16, but they can’t rent a car until 25 without some hefty surcharges.

Rental companies’ decisions are based on insurance data and so are based on behaviour. The rest of these are political artifacts, long ago dressed up in the ‘logic’ of that age but really resting on ugly political deal-making, warped by public choice failures and surviving thanks to status quo bias.

But our author redeemed himself eventually:

The scientists found the children’s brains were not fully mature until at least 25. “In retrospect I wouldn’t call it shocking, but it was at the time,” Jay Giedd, the director of the study, told me. “The only people who got this right were the car-rental companies.”

This is why the auto industry has the capacity to be the most competitive insurance market of all – standardization of coverage and vast amounts of high quality data.

A Little Blast of Seratonin

Posted on by David Wright

I read something on the Internets today that tickled my bias. Always get the warm & fuzzies when that happens.

Here are the pertinent two quotes (on why Yahoo failed):

In the software business, you can’t afford not to have a hacker-centric culture.

and…

So which companies need to have a hacker-centric culture?… The answer is: any company that needs to have good software.

So what if every company needs to have good software?

The Future (no less!)

Posted on by David Wright

David Leonhardt has kicked off quite a discussion with an excellent piece in the NYT. Here’s the key bit:

In 2008, only 13.2% of the labor force was unemployed at some point. That compares to 18.1% in 1980, and 22% in 1982.

Real wages, which normally fall during recessions, have risen in this one. Even nominal wages are up.

Arnold Kling’s view (?) -this quote is out of context, but I think still gets the message across.

I am inclined to view what is happening today as the death of the pre-Internet economy.

Arnold goes on to say that this trend may change education and health care, which suffer from Baumol’s cost disease, of course.

I think he’s skipping a step, though, and is perhaps too influenced by Robin Hanson, the ultimate long-range thinker. I think that Arnold’s out-of-context quote is more normative than positive, more what-should-be than what-is.

The what-is is, to me, that, at the margin, the non-technical are having much more trouble getting a job.

I know that any time I have any influence in a job hiring situation, I push hard for someone who has math/science/programming skills or, at the very least, inclination.

As administrative roles are replaced by capital and ‘engineers’ are hired to run them, the unskilled labour either gets pushed, wrongly, into sales and fails or has to tool up with some technical skills.

Two effects: one, get those skills; and, two: the ways in which those skills can/will be deployed are exponentially increasing.

Inflation and Insurane

Posted on by David Wright

I like this presentation by Bob Hartwig at the III (haven’t finished it yet).

I find his treatment of real vs nominal variables frustrating, though. The CPI measures what people spend on stuff and is ultimately driven by wages, imho.

Wages might drive many insurance prices (rating basis for insurance costs for many companies is their payroll). But, realistically, insurance rates are not set by some theoretical link between the driving variable (wages or whatever) but by claims experience. Any changes from ‘inflation’ are competed away if the claims don’t show up. What drives claims? Well, of course they’re driven by changes in some components in the CPI, among many other things (legislative changes, exogenous shocks, etc). But this year’s claims cost is often driven by the CPI level from when the claim occurred, which may be years in the past. So there’s a lag. Tough to predict.

Bob’s an economist, so he lives and breathes ‘economic’ variables and probably doesn’t often dig into his preconceived notion of what they measure. Inflation is different depending on what you’re doing. I imagine the correlation between changes in the driving variable for an insurance policy (wages or whatever) and the actual claims cost changes is much smaller than one would think. I’d bet real money it’s less than 0.5 in the short/medium term (1-5 years).

Looking for Reliable Data

Posted on by David Wright

Mandel on offshore R&D spending.

He makes the point that in fact that there is, in fact, a dearth of reliable data on things that matter and an embarrassment of data on things that don’t* (see: the Twitter firehose).

I think this thought. My pet view is that analytical capabilities are becoming more and more commoditized and that real ‘value added’ in the future will depend, to a greater and greater degree, on building data sets.

*Obviously, by “matter” I mean things of non-promotional economic value. The Twitter data could “matter” a helluva lot to advertisers.

You Digest Data With Your Gut

Posted on by David Wright

Listening to this week’s econtalk with Ed Leamer in which Russ and Ed discuss (among many other things – recommended) this article (weird, I first found this article. WTF?) about Tiger Woods:

Mr. Woods is such a dominating golfer that his presence in a tournament can make everyone else play significantly worse. Because his competitors expect him to win, they end up losing; success becomes a self-fulfilling prophecy.

The article then progresses to the ridiculous:

Ms. Brown argues that the superstar effect is not just relevant on the golf course. Instead, she suggests that the presence of superstars can be “de-motivating” in a wide variety of competitions, from the sales office to the law firm.

Ed and Russ appropriately cut this conclusion up, but I want to focus on their response to the former point. When presented with a counter-intuitive assertion, these two economics professors start talking about their private experiences and those of their friends to validate it. That’s not science, is it?

But that’s the problem that Ed Leamer talks about in his book and on a previous Roberts podcast. In social science, data can cast a new light on a situation, but if the only narrative you have is your own you’re not going to change your mind or learn anything. The likelihood is, of course, the weekend warrior experience is totally irrelevant when considering the psychological context of playing in a tournament with Woods and so, without a narrative (ie, a plausible setting and mechanism of causation), you have no empathy and so no ability to understand the conclusion.

Narrative without data is bullshit. Data without narrative is incomprehensible.