At the "town hall" presidential debate last week, moderator Tom Brokaw asked if, in the interest of coming up with alternative forms of energy, we should "find a Manhattan-liek Project ... or 100,000 garages across America, the kind of industry and innovation that developed Silicon Valley." John McCain, while supporting government-funded research, endorsed the latter approach, saying that we should "obviously" leave product development to the private sector.

Barack Obama appears to have the opposite view. During the debate, he asserted that "when JFK said we're going to the Moon in 10 years, nobody was sure how to do it, but we understood that, if the American people make a decision to do something, it gets done." And earlier this year, he told a CNBC interviewer explicitly that he had in mind as president "a Manhattan Project to embark upon that new energy future that we need." In last week's debate, he said that "we're going to have to make an investment, the same way the computer was originally invented by a bunch of government scientists" trying to meet military needs. This is a very distorted view. As it happens, most of the computing technology we use today was developed by private enterprise or by academic researchers with a great deal of independence.

The Manhattan Project, though, was unquestionably a triumph of government planning and engineering. A vast program of research, development, and production was brought to completion in three and a half years, culminating in the successful deployment of a radically new weapon, the atomic bomb, which brought the Second World War to a rapid end. But those who uphold the Manhattan Project as an exemplar of government success often forget that this achievement required methods we would not care to employ today.

The Manhattan Project required billions of dollars, much of America's top scientific talent, and a quarter of American electrical production. These resources were entrusted to the sole direction of General Leslie Groves, the project's director. Groves was subject to virtually no political oversight. His office issued no environmental impact statements. He decided, without any room for legal or other challenge, which technical risks to run, and which health risks to impose on the American public. Radiation was known at the time to be dangerous, and yet the Manhattan Project dispersed nuclear fallout across large areas of the West at the sole decree of General Groves.

Those demanding a new Manhattan Project seldom mean draft scientists into secret duty, spend a lot of money without public oversight, and damn the radiation leaks and other safety consequences. What they usually have in mind is an open spigot of tens or hundreds of billions of dollars to be spent on science and engineering and on subsidies for preferred technologies.

Such a program would have little in common with the Manhattan Project--and would have very poor prospects for success. Speakers who invoke the Manhattan Project or the moon landings as examples of project management imply that all difficulties are engineering difficulties and that these can be solved by a combination of lavish funding and government will. But if this were true, then Soviet and Chinese economic planning would have produced triumph, instead of disaster. Economies cannot be readily engineered. We would have difficulty even articulating our goals in detail, much less forming a coherent plan to achieve them.

A modern economy uses energy in many ways. Increased energy efficiency will require incremental improvements across the board, rather than breakthroughs in one particular technology. We cannot reliably predict how easy it will be to eke out gains in transportation, versus electrical transmission, versus manufacturing, versus residential illumination. Far better to set up incentives for efficiency and get out of the way.

The same holds true in microcosm in many areas. Consider transportation. There are many competing goals: fuel efficiency, safety, performance, low production costs, and more. There are many potential solutions: hydrogen cars, electric cars, hybrid cars, trains, buses, and so forth. We are much better off, as a society, allowing individuals, businesses, and local governments to explore various options, rather than trying to pick one approach, and enforce it by decree. Government planning decisions produced the freeway-oriented development of the last 30 years so often decried by the left. There is no reason to believe that future government policies will be any wiser or more durable.

The computer industry, in which I've been working for the last few years, demonstrates that without government mandates or supervision, the private sector is able to make significant efficiency gains. For the last several years, increased energy efficiency, has become a major goal of the industry. This has been driven not simply by a desire to be climate-conscious, but because it makes good economic sense. Energy is a major cost for companies such as Google and IBM that have hundreds of thousands of computers running all the time. These efficiencies came not from pouring money into the search for one technical breakthrough, but from incremental improvements in many different systems and components.

The Manhattan Project was oriented to a narrow goal: build an atomic bomb, as fast as possible, at any cost. This is a terrible model for economic planning, or even for environmentally conscious or consumer-oriented engineering projects. We don't want fuel efficient cars "at any cost"--a great many costs, both economic and noneconomic, must be factored in and weighted. Correctly weighting the importance of various goals is difficult and is what makes engineering management an art. Large all-encompassing development projects often founder where a swarm of smaller efforts might succeed. The latter describes the computer industry, one of the most successful parts of the American economy. To develop solutions for large-scale national problems, we should try to emulate the computer industry--those 100,000 garages--rather than our wartime command economy.

Research and development are worthy investments. There may even be a role for government subsidy to encourage solution of national problems. But an increase in science funding, spread across dozens if not hundreds of research groups, is a far cry from "a new Manhattan Project." It will inevitably devolve into a vehicle by which Congress can funnel money to favored home-state institutions, rather than a tool for putting money where it will do the most good. When Congress sought to promote space research two decades ago, the result was the Space Grant program. This program spread funding across all 50 states, and across such institutions as the University of Delaware and Brevard Community College in Florida. Such a program may have virtues, but efficiency is not among them. Any "new Manhattan Project" oriented to funding research risks likewise becoming an all-purpose tool for funneling federal dollars to favored academic constituencies.

Tennessee senator Lamar Alexander has a favorite anecdote about his predecessor Kenneth McKellar's support for the original Manhattan Project. Asked to hide a $2 billion appropriation, Senator McKellar responded, "Mr. President, I have just one question: Where in Tennessee do you want me to hide it?" Many of those backing "new Manhattan Projects" doubtless are likewise motivated by the opportunity to spend billions of dollars without being asked any hard questions. Succumbing to this urge will not help overcome national challenges.

Putting aside crass self-interest, most calls for a "new Manhattan Project" are rooted in the perennial fantasy that centralized government planning cures all social ills. This is the philosophy that leads to appointing enough "energy czars" and "drug czars" to fill out a dynasty--and most of these czars have been about as successful as the Romanovs or their successors.

If fuel efficiency is a national priority, by all means, let us tax fuel. If research for its own sake is a goal, we can offer tax credits for corporate grants to researchers. But the phrase "new Manhattan Project" obscures more than it reveals and misleads more than it inspires. The phrase provides cover for all-but-limitless spending of an unusually misguided sort, while offering the public a specious reassurance that nothing more need be done.