The peak, single-year impact on actual GDP under SA.2028 occurs in 2025, with a loss of $76 billion (1996 dollars), or about 0.4 percent of GDP. The largest percentage change in actual GDP, 0.5 percent, occurs in 2011, where the estimated loss in actual GDP that year is $57 billion.

This cost is not trivial. But the Iraq war is costing some $10 billion per month (2006 dollars). In that context, the program is not all that expensive after all -- and worlds away from Novak's estimate.

The larger point is this: Experience from other efforts to reduce pollution suggests that costs inevitably turn out to be lower than projected in advance.

As David Suzuki wrote:

A study a few years back by the Economic Policy Institute of Washington DC found that in almost every single case they looked at, the costs of complying with environmental regulations were far lower than industry - and even governments - claimed they would be. For example, electric utilities in the U.S. claimed that it would cost $4-5 billion per year to meet the 1990 Clean Air Act. But by 1996, utilities were actually saving $150 million per year.

CFCs are another example. When a phase-out of these substances that damage the ozone layer was announced, many industries claimed that alternatives did not exist or were too expensive. In 1993, car manufacturers said the CFC regulation would increase the price of all new cars by up to $1,200. Just four years later, the industry admitted that costs were already down to as little as $40.

What industry seems to "forget" is that once regulations come into force, market pressure leads to development of new, innovative ways to reduce pollution.

Suzuki notes other reasons costs tend to be lower than estimated:

First, much of the touted costs are for capital equipment that is usually much more efficient and cleaner than old, dirty equipment. These costs more appropriately should be considered capital investments, which end up reducing overall operating costs. Second, technologies change and improve, and once adopted on a mass basis, these technologies benefit from economies of scale that result in lower costs. Finally, in complying with regulations, industries are forced to rethink standard business practices that may have been wasteful or unproductive.

Despite all the evidence that costs of regulation are overestimated, it's unlikely that more accurate estimates will ever be embraced by those opposed to regulation, like Novak. For them, high cost estimates are a strategy, part of the politics of fear: "if we implement this regulation, it will destroy our way of life."

Effective, but wrong.