As Hurricane Dean roars through the Gulf of Mexico, it reminds everyone how dangerous the weather can be. But it should also remind everyone how poorly the National Hurricane Center has been at predicting storms.

Despite dire predictions from the National Hurricane Center, no hurricanes hit the U.S. last year. This year they predicted as many as 10 hurricanes. Fortunately, Dean will most likely miss us.

All this raises a question: Is the government’s free weather prediction service so bad that it is worth paying for private companies to predict the weather?

With all the blame still going around about Hurricane Katrina’s devastation, one fact has missed getting much attention: private weather forecasting companies predicted the threat to New Orleans well before the National Weather Service. In fact, AccuWeather issued a forecast that the hurricane would hit New Orleans 12 hours earlier than the government service.

This is hardly something new. Private companies with a lot at stake would often rather pay for private forecasts than rely on the “free” forecasts from the government. Hugh Connett, the president of Bridgeline, a gas pipeline company in Louisiana, claims that the government’s hurricane forecasts are too imprecise. He says that private companies such as AccuWeather do it better, because they give more accurate predictions and provide hour-by-hour forecasts of a storm’s path.

His position is not ideological – Connett’s firm monitors the past accuracy of hurricane forecasters to make sure paying extra for the private service is worth it.

It is not just for hurricanes that private forecasting comes out on top. A new study by Forecast Watch, a company that keeps track of past forecasts, found that from Oct. 1, 2006, through June 30, 2007, the government’s National Weather Service did very poorly in predicting the probability of rain or snow. Comparing the National Weather Service to The Weather Channel, CustomWeather, and DTN Meteorlogix, Forecast Watch found that the government’s next-day forecast had a 21 percent greater error rate between predicted probability of precipitation and the rate that precipitation actually occurred.

In looking at predicting snow fall from December 2006 through February 2007, the National Weather Service’s average error was 24 percent greater.

“All private forecasting companies did much better than the National Weather Service,” the report concludes.

The government doesn’t do any better with forecasting temperature. For the largest 50 cities in the U.S. over the last year, ForecastAdvisor.com ranks the National Weather Service’s overall predictions for high and low temperatures as well as precipitation as dead last among the six weather forecasting services they examined.

It has only been in the last several years that comparisons between government and private weather companies have been possible, as the National Weather Service has made its data more readily available. But none of this should be very surprising. Incentives matter. If the private companies don’t do a good job, they go out of business. Government agencies never even shrink.

Unfortunately, these aren’t the only problems with the National Weather Service. The service has been accused in the past of withholding government aircraft reconnaissance of hurricanes for up to 11 hours before releasing the information to private companies or the public.

A more serious problem is that the National Weather Service gives away its services for free. In 2004, the National Weather Service began shelving a 1991 policy that had barred the weather agency from offering services that private industry could provide. The Weather Service also now offers much of its data on the internet for free. With an $882 million budget for 2007, the National Weather Service is certainly not free and it can afford to give away a lot of services that its private competitors can’t.

There was a good reason for this 1991 ban on competition. During the 1980s, private meteorology services saw a chance to make money by providing television stations with specialized forecasts that the National Weather Service hadn’t been offering. But soon after the private companies began providing this service, the National Weather Service started giving stations the same specialized forecasts for free, driving the private forecasting companies out of the business of providing these forecasts. That is a sure way to discourage future innovation. Private forecasters are much more established now, but the government giving away free services still hurts them.

In the case of anything other than government, people would instantly call this behavior what it is: predation. But, ironically, government enterprises represent a much bigger predatory threat than private companies. It is rare for private firms to destroy competitors by selling their products below the cost of production because doing so is very expensive. And new competitors can always enter the market in the future. But it is easy for the government to sell below cost. The Weather Service increases its clout and turf by giving services away for free while passing the bill on to taxpayers.

The new data comparing the National Weather Service and its private competitors should give us pause.

Scaling back the National Weather Service isn’t just about accurately predicting tomorrow’s high temperature. People’s lives and livelihoods are at stake in getting this right. With a budget of almost a billion dollars, this may be one place where cutting back on government will save both money and lives.

John Lott is the author of Freedomnomics, upon which this article is based, and a senior research scientist at the University of Maryland