Incentives, Not Fuels, Are the Problem

By now, practically everyone in the West has heard the story about forest fires. You know the story I mean: the one about the Forest Service suppressing fires for ninety years, leading to a massive build up of fuels just ready to explode. That's why we've seen so many severe fires in recent years, and why Congress had to give the Forest Service and Department of the Interior nearly $3 billion last year to put out fires and treat hazardous fuels.

Everyone believes this story. I believed it too. But when I tried to find data to back it up, I couldn't. All the data I found indicate that fires today are no worse than would be expected considering recent droughts. This suggests that the billions of dollars Congress is spending on fire are mostly wasted. Worse, if past experience is any indicator, this spending will have serious unintended impacts on forest ecosystems.

To document the fire story, I first looked at the acres burned. Wildland fires burned more acres in 2000 than in any of the previous forty years. But it turns out that the average of the last five years, 4.7 million acres per year, is only slightly more than the 4.6 million acres burned per year in the early 1960s. Since 1960, the five-year average of acres burned has fluctuated between 2.7 and 5.3 million acres depending on the drought cycle.

I became even more suspicious when I looked at firefighter safety. The number of firefighters killed has grown from an average of 8 per year in the 1950s to 17 per year in the 1990s. At first glance, this suggests that fires have grown more dangerous.

Not so. It turns out that the number of firefighters who died from fires actually declined from 6.5 a year in the 1950s to 5.5 a year in the 1990s. But the number who died in aircraft and motor vehicle accidents grew from 1 per year in the 1950s to 6 per year in the 1990s, suggesting that the Forest Service is making a lot more use of such vehicles. Health problems, mainly heart attacks and strokes, caused most of the remaining deathsone-half per year in the 1950s growing to 5 per year in the 1990s. This suggests an aging Forest Service work force, not fuel accumulations.

Then I looked at fire suppression costs. By all accounts, the cost of suppressing fires is growing dramatically. Since 1994 we have had five of the six most expensive fire years in history. But Forest Service economists who looked at this question in 1999 found that virtually all of the increased cost is due to inflation. Despite record fire costs in 1994, and a near-record in 1996, they found "no statistically significant trend" after adjusting for inflation.

Of course, 1999 was another expensive fire year and 2000 proved another record breaker. But there were good reasons for these high costs other than a build up of fuels. First, 1998 through 2000 saw the longest La Niña ever recorded. La Niña is a movement of cool water in the Pacific Ocean that leads to droughts in much of the U.S. Fire managers blamed the expense of fire suppression in 1999 and 2000 on this drought, not accumulated fuels.

The increasing number of homes in the wildland-urban interface is a second reason for higher suppression costs. The West has been the nation's fastest growing region for at least two decades, and FEMA estimates that 38 percent of the homes built in the West are in the wildland-urban interface. The Forest Service has responded by spending extraordinary amounts of money to protect homes.

While the Forest Service is clearly spending more money to protect homes in the wildland-urban interface, I suspect that this problem, too, is exaggerated in much of the West. During the 1990s, the vast majority of homes burned by wildland fire were in California, and most of those were built near chaparral forests in central and southern California.

While chaparral fires are a problem, they don't justify spending billions of dollars to protect every rural home in the West. Wildland fire is responsible for less than 0.2 percent of all structures damaged or destroyed by fire each year. The insurance industry says that so few homes in the wildland-urban interface are lost to fire that it can't generate enough data to justify discounts to people who fireproof their homes with non-flammable roofs and landscaping.

A third reason for increased firefighting costs is incentives. In 1908, Congress gave the Forest Service a blank check to put out fires. This led many firefighters to joke, "The Forest Service fights fires by dumping money on them." But in 1975 the Office of Management and Budget (OMB) started pressuring the Forest Service to control its fire costs.

Congress helped by repealing the blank-check law in 1978. In the 1980s, instead of giving the Forest Service a blank check, it gave the agency a fixed amount for fire suppression each yearusually $125 millionand expected the agency to pay for the deficits from bad fire years out of the surpluses from good years.

That worked until the back-to-back "siege of 1987" and "Yellowstone (and Alaska) fires of 1988." The Forest Service spent $722 million fighting fires in those two years, dipping into Knutson-Vandenberg (K-V) reforestation dollars to cover the deficit. Congress responded by upping the annual fire suppression appropriation to $375 million, but the Forest Service begged and pleaded until Congress finally paid back the K-V fund out of tax dollars in 1990.

This returned the Forest Service the blank check way of thinking. Technically, Congress still gives the Forest Service a fixed amount of money for fire fighting. But if costs exceed that amount, the president can let the Forest Service spend more out of an emergency contingency fund. Of course, the president rarely says "no" to the Forest Service, which has drawn on this contingency fund every year since 1993.

When its check-writing hand was tied by Congress and the OMB, the Forest Service spent an average of $163 million per year on fire suppression in the 1980s. With those constraints removed in the 1990s, per acre suppression costs grew by 35 percent even after adjusting for inflation. Most of the increase seems due to the change of incentives rather than to weather, the wildland-urban interface, or built-up fuels.

If data on acres burned, lives lost, and dollars spent don't support the accumulated fuels theory, then what does? There are no actual measurements of fuel accumulations. While top officials blame recent fires on fuels, all the on-the-ground reports I've read focus on the weather. I am not saying there is no build up of fuels, just that the build up isn't as important as the popular story has led us to believeand that incentives are much more important than any other factor, even the weather. That means Congress is wrongly focused on treating fuels when it should focus on fixing incentives.

After the Cerro Grande fire burned 200 Las Alamos homes to the ground in 2000, Congress began a firestorm of spending on anything to do with fire: presuppression, suppression, hazardous fuel treatments, research, and assistance to states and local communities. The agencies dignify this spending by calling it the National Fire Plan, but it is really just an even more extravagant case of dumping money on fires.

Federal land fire budgets had already exploded from a few hundred million a year in the early 1990s to $1.5 billion in 2000. But the National Fire Plan nearly doubled them again to $2.9 billion in 2001. The Park Service, Bureau of Land Management, and other Interior agencies got some of this money, but the Forest Service spent two-thirds of it.

Budgets for treating hazardous fuels have grown from about $20 million a year in the early 1990s to $400 million a year today. The Forest Service and other agencies say this money is needed to treat the 70 million acres of federal land at a high fire risk due to past fire suppression and accumulated fuels. Because the fire risks are so high, they often don't dare use prescribed fires and instead must thin or otherwise mechanically treat these acres, which is more expensive.

The result is that they are only treating 2.0 to 2.5 million acres a year, which means it will take at least 30 years to treat all condition class 3 acresby which time many of the 140 million acres currently rated moderate fire risk will have grown into a high-risk condition.

The Forest Service says it focuses on high-priority acres such as lands in the wildland-urban interface. But spending hundreds of dollars an acre treating federal lands near private homes will not guarantee the safety of those homes in a really severe fire year. On the other hand, their safety can be guaranteed by fireproofing the homes themselves with non-flammable roofs and landscaping.

Forest Service fire researcher Jack Cohen has shown that homes and other structures will burn only if they have flammable roofs or if there is burnable vegetation or other fuels within 140 feet of the structures. This means we need to treat just one or two acres around each home, not thousands of acres of federal lands that may be near the homes. Instead of treating its own lands, the federal government could share the costs of treating private homes and then let fires on the federal lands burn.

Instead of letting fires burn, however, Congress has nearly quadrupled presuppression costs. Presuppression or preparedness refers to the cost of having firefighters, aircraft, and other equipment and supplies on standby for when fires break out. In the early 1990s, the Forest Service presuppression budget averaged $165 million a year. Today it is $620 million a year. Adding Interior agencies brings the total to more than $900 million a year.

Since 1980, the Forest Service tried to convince Congress that spending more money on presuppression would reduce total fire costs. Having more crews on standby, the theory goes, allows the Forest Service to suppress more fires at a lower cost. The Forest Service backed this up with a computer model called the National Fire Management Analysis System (NFMAS), which supposedly calculated the most efficient level (or MEL) of presuppression funds.

Congress never really bought this theory, however, and typically gave the Forest Service only 70 to 80 percent of MEL for presuppression. After the 2000 fire season, however, Congress is giving the Forest Service 100 percent of MEL. By an amazing coincidence, the Forest Service's calculations of MEL have doubled from about $300 million a year in the mid-1990s to more than $600 million a year today.

The increase in presuppression funding hasn't kept the promise of reducing firefighting costs: Though 2001 was a relatively mild fire year, with less than the average number of acres burned, the agencies spent $151 an acre on suppression, compared with just $44 an acre in 1997 (adjusted to 2001 dollars), when about the same number of acres burned but the Forest Service was only funded at about 80 percent of MEL.

What about the idea of letting some fires burn? If suppressing fires has been so bad, wouldn't it be better to not suppress some of them? And if so, why do we need to spend so much on presuppression and suppression?

Under a 1995 fire policy approved by the secretaries of Agriculture and the Interior, federal land managers can let fires burnthey now call it "wildland fire for resource use"only if they have written an approved fire management plan; that plan allows fire for resource use; the fire meets conditions specified in the plan; and the fire was started by natural causes. As of March 2002, the Forest Service had completed fire management plans for just half of its forests. Few, if any, of those plans allow fires to burn outside of large wilderness areas, which excludes more than 80 percent of the National Forest System. Since 88 percent of fires are human-caused, this means that managers can allow no more than about 1 percent of all fires to burn.

Because of the other restrictions, the actual number is even smaller. In recent years, federal managers have allowed only about 0.3 percent of fires to burn, and those fires have covered only about 2 percent of total burned acres.

So it is not surprising that the Forest Service's NFMAS model presumes that all fires will be suppressed. In fact, although the model is supposed to calculate the value of resource losses as well as of presuppression and suppression costs, it is strangely insensitive to resource losses. Even if the resources being protected from fires were worth nothing, the model would propose about the same most efficient level of presuppression costs.

Lately, it seems that the federal land resources being protect from fires are worth nothing. Timber was the resource whose value the Forest Service traditionally plugged into the NFMAS model, but national forest timber sales are down by more than 80 percent from pre-1991 levels. Most other resources aren't really harmed by fire, and many benefit. It is true that really bad fires can fry soils and lead to erosion, but the Forest Service hasn't ever proven that it can stop those fires.

If resource values don't change NFMAS results, the model is little more than a scam to get more money from Congress. Yet the Forest Service used NFMAS to convince Congress to fund 50 percent more firefighters, hundreds of new fire engines, dozens of aircraft, and scores of new firefighting facilities. And what do we get for it? More fire suppression, which is supposed to be the opposite of what the forests need.

Is there any scientific support for the fuels theory? Recent Forest Service fire documents and papers in scientific journals repeatedly point to one publication in support of the fuels theory: a report concluding that, because of historic fire suppression, "vegetation has accumulated, creating high levels of fuels for catastrophic wildfires and transforming much of the region into a tinderbox." Who wrote this report? The General Accounting Office.

I like the people at the General Accounting Office, but they're accountants, not fire ecologists. Even the Forest Service thought that the report was misleading, saying that changing firefighting strategies, not increased fuels, were responsible for recent increases in acres burned.

The GAO blamed the deaths of 14 firefighters in the 1994 South Canyon fire on excessive fuels. But, as the Forest Service pointed out, "The South Canyon Fire was in oak/brush type vegetation and is not representative of the short fire interval, long needle pine vegetation of the interior western forests. Excessive fuel loading on the South Canyon Fire was not a causal factor."

This brings up the question: How many forests are of the short-interval, low-intensity fire type? These are the forests most susceptible to fuel build-ups and the forests best treated by reducing fuels. According to a recent paper by Forest Service fire researchers, only about a third of the nation's wildlands are in this category, mostly located in the Deep South. Two-thirds of the lands in the West are ecologically adapted to stand-replacement fires, and another 15 to 20 percent are adapted to mixed-severity fires.

This means spending billions on fuel treatments will not stop western wildfires any more than spending billions on presuppression. Fuel treatments are important in the South, where most forests are on a short-interval, low-intensity fire cycle, but in the West they are as much a chimera as NFMAS and MEL.

The ecological danger is that the Forest Service will spend its bounty from Congress on practices that will prove just as harmful in the long run as fire suppression. As it has already done in the South, it may try to convert western forests adapted to severe fires to forest types adapted to low-severity fires. Or it may impose fires at fixed intervals of, say, twenty-five years on forests adapted to varying intervals of, say, five to fifty years. Either way will reduce forest diversity.

The fuel treatment myth is just the latest in the Forest Service's century-long history of misleading Congress to gain bigger budgets and more power. As Ashley Schiff documented in his 1962 book, Fire and Water, to convince Congress to purchase national forests east of the Mississippi, fund western forests, and give the agency more authority over private lands, the Forest Service suppressed and ignored its own research, demonized researchers in other agencies, and knowingly lied about fires and flooding.

After 1950, the Forest Service bragged for decades that it was one of the few government agencies to earn a profit. But its "profits" were mostly based on Enron- and WorldCom-like accounting tricks. Like Enron, it counted receipts that weren't really there, such as the cost of roads built by timber purchasers. Like WorldCom, it treated operating costs as capital costs. More daringly than either Enron or WorldCom, it amortized certain capital costs over hundreds of years and in some cases over eternity (effectively zeroing them out). After correcting for these accounting tricks, the Forest Service did earn a profitin 1969 and four or five years in the early 1950s. Otherwise, it has cost taxpayers hundreds of millions to billions each year.

From 1950 through 1990, the Forest Service supported its timber program on a web of misleading statements and data. It said that Douglas-fir and other forest types "needed" clearcutting to regenerate (they didn't). It said it could sustain 12 billion board feet or more of timber cutting forever (it couldn't). It said that timber cutting benefited most recreation and wildlife (only those kinds that weren't in short supply). Though national forest timber sales have now declined by 82 percent, Congress is still giving the Forest Service nearly as much money as ever for timber management.

Compared with fire, timber is small potatoes. Even in timber's heyday, timber never contributed more than about $900 million a year to the Forest Service budget. Fire added more than $1.9 billion to the Forest Service's 2001 budget and it will probably be more in 2002. Mainly because of fire, the Forest Service's budget jumped nearly 50 percent in 2001.

I like the Forest Service and I respect the people who work for it. But one of the most important jobs of a bureaucracy is to filter information, and the bureaucracy can always be counted on to filter the information in a way that benefits the bureaucracy. Everyone in the bureaucracy can be perfectly honest, yet the bureaucracy itself can't be trusted to tell the truth when it is rewarded for spending more money.

Fire is a perfect example. For years, the Forest Service told Congress that it could scientifically manage the national forests if Congress gave it millions, and then hundreds of millions, for fire suppression. Now the Forest Service admits that fire suppression was a complete mistakeso Congress rewards it for that mistake by giving it billions.

The waste of funds is bad enough. But basing a centralized bureaucracy on a myth about hazardous fuels is likely to do as much damage to ecosystems as basing one on a myth of fire suppression.

If the fuel story is only a myth, why do so many groups believe it? The answer is that it fits their agendas. The Forest Service, of course, uses it to get a bigger budget. But commodity interests also use it to cast blame on environmentalists for stopping timber sales that could have removed some of the fuels. Environmentalists use it to cast blame on timber cutters who leave fuels behind. Both sides are a little bit right but mainly wrong: The real problem is not the timber program or lack of one but the overall incentives facing federal land managers.

I can think of two ways to fix the incentives. One can save money but won't solve the problem of a centralized bureaucracy. The other focuses on decentralization.

The centralized solution requires that Congress stop throwing money at the federal land agencies. Congress could help owners of private land in the wildland-urban interface fireproof their structures. Otherwise, it should tell the federal agencies, "Let the fires burn." In a few years, most forests would have fires at rates not too different from the rates they experienced a hundred or so years ago. This will save billions of dollars, but letting all fires burn might not be the best ecological answer either.

The decentralized solution is for Congress to turn forests into self-funding units. The managers of each unit would have to decide, based on available revenues, how much fire protection they can afford. Some forests might let most fires burn. Others would put some fires out and would rely on insurance for the worst fire years. Congress could create appropriate safeguards for non-market resourcessafeguards that don't exist todaysuch as dedicating a share of forest revenues to non-market purposes.

Because it will fit solutions to the land rather than to whatever myth happens to hold Congress in its grip at the moment, decentralization will protect the environment better than any centralized process. Some timber will be cut, but with no more subsidies it will be less than in the past. Recreation fees will go up, but the people paying those fees will know they are giving managers incentives to protect the things they value.

Environmentalists and commodity interests should stop bickering and work together to end a century of mismanagement by misincentives. Only then will the national forests be truly managed for the people and not just for the bureaucracy that is supposed to protect them.

Randal O'Toole ( is senior economist with the Thoreau Institute ( and author of Reforming the Forest Service.

The Thoreau Institute | Fire Research