Here's a fun little tool my brother seems to be involved in Little Big Cloud. Each url on the domain is a tag cloud - if you go there you're prompted to provide more tags and they automatically link to each other. I think the concept is great. Add typed links and you have a semantic web on one page. I think it is a little annoying it keeps asking me for keywords everytime I go to the same page (I'd rather see a seperate url to add, also from a search engine perspective) and some filtering for double keywords from the same user would make it less spamable, but other than that I think it is a nice little hack.
Saturday, June 16, 2007
We're in the
So in theory you could jus go to the form and try all numbers between 100 and 999 and see which one would give you the biggest discount. I tried a few and it seems that there was something to this theory. Sometimes it said invalid code, but often it said something like, code not valid for this date or code not valid for this region. Hmm, what could try 900 codes without getting bored?
So I fired up my python editor and wrote a small script. This turned out to be slightly more complicated than thought the site needed a cookie and stored in the cookie some of the route information. Weird. Also it, seemed to generate the form on the fly based on I don't know what. But I had my little firefox developer toolbar, converted the post form to get, looked at the cookies I had and pasted it all in my Python script and after 20 minutes or so it actually worked. I added a sleep statement to the script to not send out all queries at once (I'm trying to get a discount here, not to annoy the server) and let the script run for another 20 minutes.
Result: It is not just V185 that gives you 10% off. V311 also gives you 10% off! Oh well, at least now I know for sure.
Thursday, June 7, 2007
Unemployment insurance is a tricky thing. You want to give people something they can fall back on when they get fired, but at the same time you don't want to make freeloading an attractive option. In the Netherlands we have (had?) an employment office that was supposed to take care of the whole process, from finding out what you were good at, finding you a matching job and paying out your monthly unemployment money. It was only the last thing that they really succeeded in. Governments in general are better at taxing and spending than finding jobs for the needy. So can we use the free market to solve this problem?
It seems tricky since Social security is meant to take the sharp edges of the market, but there are some options. For example, the temp offices in the Netherlands turned out to be much better at finding jobs than the employment office ever was. They make money on it of course, while if you want to be cynical, the employment office becomes more important the more unemployed people there are. So what if we would let these temp offices grow out to become a replacement of the employment office, including them offering unemployment insurance? It would nicely align the interests. These employment companies would get a cut of the salary of the people they insure (as unemployment insurance premium) just like the temp offices do (as a service fee). If the people lose their jobs, they would have a strong incentive to find them a new job (because they no longer get the premium and now have to pay out unemployment money instead), so they would take an active interest. Also, since the money they make is a percentage of the income of the people they insure, they have an incentive to find the best paying jobs for their' unemployed. They won't just ship them off to McDonalds just to get them off the books and if they do, because right now there is nothing else available, they will keep their eyes open should something better come along.
It's a nice scheme, I think, but it suffers from some of the problems that market based health insurance has too. One is that most people don't think that much about bad things and left to their own devices might just go without insurance (which happens a lot in the US) or get very crappy one to save money. The other thing is that the whole solidarity concept is lost. For health insurance this means that some people have to pay way higher premiums than others to the extent that some people won't be able to get insurance at all. Employment offices might charge very high premiums or refuse to insure the vulnerable too, like the 55 year old coal miners. In short we can make social security more efficient by making it less social.
I think both of these issues can be solved if we take the individual out of the equation. It would go something like this. Every quarter, the government would draw lottery-style 1% of all people that need insurance. Their previous insurance will be canceled. This group will be further divided in plots of say 25.000 people. Employment companies can now bid on these plots. The government will say, if any of these people lose their jobs, we want you to pay them 80% of their last salary for 1 year and after that 80% of the minimum wage. What percentage of their income would you want to make that happen? All employment companies put in bids of the form, I take 5 plots for a 2.7% premium, and the lowest bidding companies will get the deals. The actual premiums paid will of course by averaged out, so for the individuals it doesn't matter whether they change plot or not.
The bidding companies don't really know whom they will get, but it is a representative sample of the population, so they can draw up their computer models and work out what they can afford. They'll lose a little money on the 55 year old miner, but make some back on the 25 year old lawyer exactly what solidarity is about. Since there are multiple providers, prices should be kept near the market minimum. They can also concentrate on the longer term, since they'll have the people in the plot for the foreseeable future, with an attrition rate of 1% per quarter (or whatever the government sets), so they might consider paying for education in some cases if that would lead to better chances on the labor market for their people.
To some it might seem odd that as an individual you just get a letter from the state saying your unemployment insurance is being handled by company X, but that is actually not so different from the statist approach where the government decides which of its bureaucrats will handle your case, only a little better. What happens if any of those companies goes bankrupt? Nothing much really, the plots that they are currently holding will just be resold. Premiums could go up a little if the bankrupt company was underbidding or they could go down if it was overbidding and just didn't know how to run its business.
In interesting option for the employment companies could be to sell their obligations on the capital markets. A more or less steady income stream set against the risk of a general rise in unemployment is something that can be nicely repackaged in the current financial system and would reduce overall risk to the unemployment companies (i.e. they wouldn't all suddenly go bankrupt if the unemployment rate would jump, since they would in effect have reinsured the risk on the capital markets).
An open question is how to handle change. What if the government decides that 80% of the last income is too much and wants to reduce this to 75%? Obviously this is good news for the employment companies, but how good the news is in money is harder to say. A solution could be that the government just pockets the money and from then on starts selling plots at 75%. Within not too wide margins this would probably work out quite ok (it would create a market with different rates, just like the government bond market has right now.
A propos health insurance, you could probably make something like this work there too. Health insurance companies would bid on plots of people to be insured, either for a percentage of their income or a fixed fee. They would then have to provide a certain level of health care, no matter who they happened to have drawn. We could have that way a market driven health insurance system with government set minimum level of service, while at the same time having an equal premium for everybody.
One more area where this principle could be applied in a slightly different way is in government bonds. On an abstract level, the government borrows money now to invest into something and promises to set aside a slice of future tax income to pay for that money. How big a slice is not always that clear at the beginning of course and has gotten many a government into trouble later on. What if the government would just make this explicit? We need 10 billion now to build a new airport and we want to pay off the debt in 25 years. What slice of tax income would you need to give us that 10 billion? The capital markets would have an answer.
Maybe this is the way to solve the looming pension crisis. Lots of countries have pay-as-you-go pension schemes that combined with an aging population lead to hugely under funded pension plans. In other words the slice of tax receipt that needs to spend on these pensions will go up in the future, but nobody knows by how much; more over in effect it is the future generations that will have to pay for the current mess. Instead we could just go to the capital markets and say, if we want to keep this sort of pension scheme running until 2100 or so, what percentage of future income would you require to handle the pay-outs? We might get quite a scare when we see that number. On the other hand if that number is real, better be scared now and adjust the parameters.