I am not sure I explained myself well so here is what others said:
"Keynesian theory sounds good, and it would be nice if it made sense, but it has a rather glaring logical fallacy. It overlooks the fact that, in the real world, government can’t inject money into the economy without first taking money out of the economy. Put more bluntly, Keynesianism only looks at one-half of the equation. It conveniently ignores the fact that any money that the government puts in the economy’s right pocket is money that is first removed from the economy’s left pocket. As such, there is no increase in what Keynesians refer to as aggregate demand. The bottom line is that Keynesianism doesn’t boost national income, it merely redistributes it.
The real-world evidence also confirms that Keynesianism is a failure. Indeed, it was a failure even before Keynes published The General Theory in the mid-1930s. In his four years, Herbert Hoover was a poster-boy for big government. He increased taxes dramatically, including a boost in the top tax rate from 25 percent to 63 percent. He imposed harsh protectionist policies. He significantly increased intervention in private markets. Most important, at least from a Keynesian perspective, he boosted government spending by 47 percent in just four years. And he certainly had no problem financing that spending with debt. He entered office in 1929 when there was a surplus and he left office in 1933 with a deficit equaling 4.5 percent of GDP. Needless to say, Hoover’s big-government Keynesian experiment was not very successful since growth went down and unemployment went up.
Unfortunately, other than being a bit more reasonable on trade, Roosevelt followed the same approach. The top tax rate was boosted to 79 percent and government intervention became more pervasive. Government spending, of course, skyrocketed — rising by 106 percent between 1933 and 1940. This big-government approach didn’t work for Roosevelt any better than it did for Hoover. Unemployment remained very high throughout the 1930s and overall output did not get back to the 1929 level until World War II.
Other Keynesian episodes generated similarly dismal results, though fortunately never as bad as the Great Depression. Gerald Ford did a Keynesian stimulus focused on tax rebates in the mid-1970s. The economy did not improve. But why would it? After all, borrowing money from one group and redistributing it to another group does nothing to increase economic output. Tax cuts only boost the economy if they reduce the tax penalty on work, saving, and investment — i.e., lower tax rates, not gimmicks.
More recently, George W. Bush gave out so-called rebate checks in 2001 and 2008, yet there was no positive effect in either case. And Bush certainly was a big spender, yet that didn’t work either. Not that this should be a big surprise. Surveys of the academic literature reveal that even left-wing international bureaucracies are producing research showing that bigger government hurts economic performance by misallocating national resources.
Japan’s experience also shows the foolishness of Keynesianism. Throughout the 1990s, Japanese politicians tried to use so-called stimulus packages to jump-start a stagnant economy. But the only thing that went up was Japan’s national debt, which more than doubled during the decade and now is far above even Italy when measured as a share of GDP. The economy, not surprisingly, remained stagnant.
If Keynesian spending doesn’t make sense from a theoretical perspective, and also fails every time it is tried in the real world, why do politicians keep trying the same approach? Your guess is as good as mine, but the answer probably has something to do with the fact that politicians love to spend other people’s money, and Keynesianism is a convenient rationale."
http://sites.google.com/site/redhotright/articles/myth--government-spending--stimulates--the-economy
