Is a public-sector job really as good as a job created in the private sector? I've been wondering about this a lot lately, in part because I just finished a book about the period of the first great American experiment in public job creation, the New Deal. Critics have written that I failed to appreciate the value of New Deal emergency jobs. But the quality of government-paid jobs is also relevant because of the Democratic presidential candidates' interest in that 1930s experiment.

To hear the candidates talk, a repeat of 1930s-scale government job creation is dangerously overdue. John Edwards has proposed that government take the lead in creating types of jobs--"green collar" and "stepping stone"--to serve the two goals of protecting the environment and giving lower earners new skills. Dennis Kucinich is calling for a new green version of FDR's Works Progress Administration.

A structural disaster--the collapse of the levies in New Orleans or the bridge in Minnesota last summer--adds a sense of moral urgency to the debate. Hillary Clinton is warning that "We're trying to build our children's future with our grandparents' infrastructure." Republican Mike Huckabee's talk about domestic infrastructure investment as crucial to our "economic viability" sounds similar.

Academics are backing the politicians up. Bruce Katz of the Brookings Institution recently suggested that intelligent planning is the key to success: "smart policies and investments on infrastructure can foster productive growth in our economy, sustainable growth."

Given this Edifice Complex, the actual quality of New Deal spending, job creation and growth are worth a second look. The record is less impressive than the rhetoric implies.

The New Deal government indeed spent a lot. Nowadays Congress considers a 1% increase in the budget tantamount to treason, or nirvana, or both. President Roosevelt had no time for paltry 1% changes. He nearly doubled the federal budget in his first term. The WPA that Mr. Kucinich mentions spent several billion all by itself. The idea, as the New York Times put it, was for Washington to do work that could "not be undertaken by private industry." A second multibillion-dollar project, the Public Works Administration (PWA), was headed by Harold Ickes, the father of the Clinton adviser. PWA schools, swimming pools or town halls went up in nearly every county of the U.S.

The New Deal also created a lot of jobs--millions. And the New Deal did cause significant business activity. Industrial production--factory activity, basically--came back to 1929 levels around the time of Roosevelt's re-election. All of these outcomes are taken as evidence of public spending's success.

But what really stands out when you step back from the picture is not how much the public works achieved. It is how little. Notwithstanding the largest peacetime appropriation in the history of the world, the New Deal recovery remained incomplete. From 1934 on--the period when the spending ramped up--monetary troubles were subsiding, and could no longer be blamed alone for the Depression. The story of the mid-1930s is the story of a heroic economy struggling to recuperate but failing to do so because lawmakers' preoccupation with public works rather got in the way of allowing productive businesses to expand and pull the rest forward.

What was wrong with those public works jobs? Many created enduring edifices--New York's Triborough Bridge, for example, the Mountain Theater of Mount Tamalpais State Park outside San Francisco, the Texas Post Office murals, which were funded by Henry Morgenthau's Treasury. But the public jobs did their work inefficiently. That was because the jobs were scripted to serve political ends, not economic ones.

One of the saddest accounts of the public-works job culture I came across involved a model government farm in Casa Grande, Ariz. The men were poor--close to "Grapes of Wrath" poor--but sophisticated. They knew that the government wanted them to share jobs. But they saw that the only way for the farm to get profits was to increase output and to stop milking by hand. Five dairy crew men approached the manager to propose purchasing milking machines to increase output. They even documented their plea with a shorthand memo:

"Milking machine would save two men's labor at five dollars per day . . . Beginning in September would save three men's wages or $7.50 on account of new heifers coming in."

The men were willing to strike if they didn't get the machines, though they feared they might lose their precious places on the farm if they did strike. Their fears proved justified. "You're fired," the workers later recalled the manager replying when he saw their careful plan. The government man was horrified at the idea of killing the jobs he was supposed to create. "You're jeopardizing a loan of the U.S. government, and it's my job to protect that loan. You're through, everyone of you, get out."

A related problem was that the New Deal's emergency jobs were short term, lasting months, not years, so people could not settle into them. This led to further disruption. In the very best years of Roosevelt's first two terms, unemployment still stood above 9%. Nine percent is better than horrendous, but it hardly is a figure that induces hope.

One could interject that such arguments do not take into account the context--the paucity of other jobs, the dust storms, the deflations, the homelessness, the incomprehensible real privation of the period. But in the later part of the 1930s, the same model infrastructure projects did their part to prolong that privation. The private sector, desperate, was incredibly productive--those who did have a job worked hard, just as our grandparents told us. But the government was taking all the air in the room. Utilities are a prime example. In the 1920s electricity was a miracle industry. There was every expectation that growth in utilities might pull the country through hard times in the future.

And the industry might have indeed done that, if the government had not supplanted it. Roosevelt believed in public utilities, not private companies. He created his own highly ambitious infrastructure project--the Tennessee Valley Authority. The TVA commandeered the utility business in the South, notwithstanding the vehement protests of the private utilities that served that area.

Washington sucked up much of the available capital by selling bonds and collecting taxes to pay for the TVA or municipal power plants in towns. In order to justify their own claim that public utilities were necessary, New Dealers also undermined private utilities directly, through laws--not only the TVA law but also the infamous Public Utilities Holding Company Act, which legislated many companies out of existence. Other industries saw their work curtailed or pre-empted by government as well.

What about that oft-cited rising industrial production figure? The boom in industrial production of the 1930s did signal growth, but not necessarily growth of a higher quality than that, say, of a Soviet factory running three shifts. Another datum that we hear about less than industrial production was actually more important: net private investment, the number that captures how many capital goods companies were buying relative to what they already had. At many points during the New Deal, net private investment was not merely low, but negative. Companies were using more capital goods than they were investing in.

All this tells us that while some companies were gunning their engines for the moment--the industrial production--they had little hope for productivity gains in the years ahead. Business no longer believed in business. Five years into the New Deal, companies across the country were mounting what Roosevelt himself described as a "capital strike."

People became accustomed to a sort of calculus of frustration. The closer the country got to the prosperity of 1929, the more impossible reaching such prosperity seemed. The 1930s came to be known as the always recovering but never recovered decade. The Dow itself confirmed this pessimistic assessment by stubbornly remaining below 1929 levels through World War II and into the 1950s.

The relevant points for today are simple. The famous "multiplier effect" of public spending may exist. U.S. cities do indeed need new highways, new buildings and new roads, maybe even from government. But these needs should be weighed against damage that comes when officials create projects and jobs for political reasons.

An emergency such as a Great Depression, a Sept. 11, a Katrina, can serve as a catalyst for an infrastructure project and for job creation too. But the dire moral quality of that emergency does not guarantee that the project undertaken in its name will be more efficient than your standard earmark.

In other words, candidates may want to be careful as they climb onto FDR's shoulders. The New Deal edifice may look solid, but it doesn't form a good basis for the American future.