Archive for August, 2010

When is paying off debt not really savings?

August 31, 2010

A few days ago Ezra Klein was posting about tax cuts and the usual claim that poor people spend them while rich people don’t. He cited some studies that seemed to show that poorer people (not poor, mind you, nobody surveys them) were more likely to say that they were going to use tax-cut money to pay off debt rather than spend.

It struck me at the time that this was questionable, because for poorer household debt is one of the things that constrain consumption.  If you’re paying 20-plus percent on your outstanding balance, knocking that down by $500 means another $10 a month you can spend on food and gas. Or even more bluntly, if you’re pretty much at the limit of what you can carry on your cards, paying down $500 on your debt means not only immediately improved cash flow but also $500 that you can spend during the coming month or couple of years as your debt slides right back up to the limit. Either way, it’s not really savings as typically thought of, it’s debt management in the service of fairly immediate increased consumption.

In a macro sense, this would normally count as savings, because the money is available to lend out again, but the availability of money to lend at most financial institutions really isn’t particularly correlated with how much people are paying them back. So if it’s savings in that sense, it’s technical only.

What we would like to ask people about tax cuts (for purposes of stimulus) is not really whether they’re going to spend the money immediately versus paying off debt or putting it in savings/investment, it’s how it’s going to change their non-debt-service consumption. To statisticians, the interest you pay on loans is consumption just like food or gas or televisions — you’re paying for the use of someone else’s money. But for individuals, not so much.

Advertisements

Those who do not understand statistics….

August 26, 2010

BBC News – Walk ‘could track down criminals’, researchers claim

With almost 2,000 walking sequences recorded in a database, Mr Matovski claims a 95% success rate.

Gait can also be measured at a distance – an advantage over other forms of biometric identification.

However researchers found that “extreme changes” in clothing can affect recognition levels.

Obviously this stuff will eventually get better. But first: 95% accuracy is terrible for identification purposes. If everyone in the world were a criminal, it would be fine, but when the vast majority of people walking down the street aren’t the criminal you’re looking for, most of your hists are going to be false positives.

Second: apparently changing your gait isn’t so hard. So you’ll just of people putting rocks in the their shoes, wearing knee braces, wearing heavy stiff coats when they go to rob a bank…

Good news on GPS tracking

August 10, 2010

US appeals court bashes warrantless GPS tracking • The Register

Attorneys from the American Civil Liberties Union and the Electronic Frontier Foundation filed a friend-of-the-court brief in the case arguing that absent a warrant, the planting of the device was an illegal search under the US Constitution’s Fourth Amendment. The appeals court on Friday firmly rejected federal prosecutors’ arguments that the suspect had no reasonable expectation of privacy because the vehicle’s whereabouts could have been easily tracked using human surveillance.

“It is one thing for a passerby to observe or even to follow someone during a single journey as he goes to the market or returns home from work,” Judge Douglas H. Ginsburg wrote. “It is another thing entirely for that stranger to pick up the scent again the next day and the day after that, week in and week out, dogging his prey until he has identified all the places, people, amusements, and chores that make up that person‘s hitherto private routine.

Would be surprised if it’s overturned, but really the warrantless aspect has made it way too easy to just use it on everybody.

Wait till they get face recognition working on CCTV…

Is it geology or is it CGI

August 10, 2010

Giant Crystal Cave in the Mexican Desert | Amusing Planet

The giant obelisks are formed from groundwater saturated in calcium sulphate which filtered through the cave system millions of years ago. These water warmed by an intrusion of magma about a mile below, began filtering through the cave system millions of years ago. When, about 600,000 years ago, the magma began to cool, the minerals started to precipitate out of the water, and over the centuries the tiny crystals they formed grew and grew until 1985, when miners unwittingly drained the cave as they lowered the water table with mine pumps.

Some of the most amazing cave pictures

Not neat to be a neet?

August 1, 2010

BBC NEWS | UK | Education | Neets ‘should not get benefits’

MPs are suggesting adopting a system use in Holland reduce the number of 16 to 25-year-olds not in education, employment or training, known as Neets.

They said the Dutch equivalent of jobseeker’s allowance was dependent on being in work, education or training.

At the end of 2009, nearly 15% of 16 to 24-year-olds were classed as Neets.

And 9.3% of the youngsters classified as Neet at the end of 2009 were aged 16 to 18.

And the government looks set to miss its target to reduce that figure to an average of 7.6% for a whole of 2010.

The false positives just keep coming

August 1, 2010

Medical Daily: Reading terrorists minds about imminent attack

“Since 9/11 preventing terrorism is a priority,” Rosenfeld said. “Sometimes you catch suspicious people entering a building. You suspect that they’re terrorists, and you have some leads from the chatter. You’ve heard they’re going to attack one city or another in one fashion or another on one date or another. Our hope is that our new complex protocol – different from the first P300 technology developed in the 1980s – will one day confirm such chatter in the real world.”

I’m sure that there’s always some secret that only the terrorists and the surveillance geeks know, that could be used as an effective probe for this technique. Because otherwise you’re going to test positive for every malcontent who’s thought about carrying out a terrorist attack in some particular way on a particular date in a particular place, every security officer or first responder who’s worried about defending against some particular terrorist attack, every news junkie who’s considered that a particular place on a particular date might be a terrorist target… And there are roughly a million times as many of those people as actual terrorists, so unless you’re 99.9999% accurate, you’re diverting resources from something that might actually help.

Schneier has made this point over and over again: tagging the people who are evildoers, on a day to day basis, is not nearly as important as not mistakenly tagging the people who aren’t.