With the economy once again teetering on the edge of recession, policy makers will inevitably propose another round of stimulus spending. You can bet on it--just as you can bet that any such spending won't help the economy. From the beginning, the Obama administration has misdiagnosed the problem and implemented policies that are indefensible.

Consider the problem of economic stabilization. In the old days, U.S. recessions were short and recoveries sharp: Between World War II and 1990, the average rate of growth in gross domestic product (GDP) in the five quarters after a recession was 6.8%.

For the stimulus optimist, the temptation of moving some of that blockbuster growth into the recession is certainly overwhelming. But getting the timing right is a daunting--or even impossible--task.

It's hard enough to call the onset of the recession correctly and arrange the spending so that it happens at precisely the right moment. But the successful policy maker must also remove the stimulus at a moment of unusually high growth. If not, the drag from disappearing stimulus could easily push the economy back into recession.

"It was a mistake for the Obama administration to pursue short-term Keynesian stimulus." -- Kevin A. Hassett

Keynesians tend to assume that government spending has a big positive effect on economic growth. Others disagree. But if the impact of increasing government spending is large, then the impact of removing it is also. So policy makers better be sure that the boom is around the corner.

And all these are just short-run considerations. Here's the real dirty secret of Keynesian policies: They are sure to have a negative effect in the fullness of time.

Every stimulus effort has not two but three stages. When the stimulus is imposed, there is some positive short-run increase in GDP. When the stimulus is removed, there is an approximately equal and opposite reduction in GDP. But after that, the stimulus must be paid for with higher taxes or ongoing borrowing--causing a further reduction in GDP. Thus the total impact of the Keynesian policy is negative over its life. This fact is visible even in the fine print of Congressional Budget Office analyses so often cited by stimulus apologists, such as its 2009 finding that the Obama stimulus would reduce output in the long run.

Obama administration officials should have known all this as they set out in 2009. Financial crises inevitably create lengthy periods of slow economic growth, as research by economists Carmen Reinhart and Kenneth Rogoff has shown. The typical duration of the employment downturn after a financial crisis is 4.8 years. Another study by Ms. Reinhart and her husband Vincent Reinhart found that economic growth rates tend to be lower for as much as a decade after financial crises.

Given this lengthy period of slow growth, it was a mistake for the Obama administration to pursue short-term Keynesian stimulus. Such a policy might be wise if the economy were in a typical recession, which can be expected to last a bit less than a year and be followed by a recovery with sharply higher growth. In such a case, adding a percent or two of growth during the recession might be worth having a slower recovery.

But in the lengthy, slow‐slog out of a financial crisis, the stimulus hangover arrives before the recovery has taken off. Temporary stimulus therefore hurts the economy when it is removed and again when it is paid for. The hangover is virtually guaranteed to arrive at a moment when it can push us back into recession.

Worse, aggressive stimulus sets off a kind of Keynesian death spiral in which nervous politicians adopt repeated stimulus packages in order to avert near-term distress, the cumulative effect of which can be ruinous.

Keynesians might dispute this logic, saying that their policy prescription helps avert "catastrophic panic" in the economy. But there is no established connection between higher government spending and the mental health of consumers and investors, and there likely never will be. A policy that reduces long-run output should, if anything, increase the likelihood of panic.

The arguments against recent stimulus actions are so strong that one wonders whether big government ideology is the true source of the Democrats' Keynesian enthusiasm. President Obama's plan all along may have been to expand government spending massively in the name of temporary stimulus, and then fight to never let it decline.

This week's budget deal suggests that the policy course is changing, but going cold turkey on stimulus isn't enough. This isn't the moment for temporary fixes but for significant and ambitious reforms.

Given the scale of the problem, we should pursue two paths. The first is tax reform that broadens the tax base, lowers rates, and reduces taxes on capital income. Over a decade, fundamental tax reform can deliver enduring benefits large enough to offset the residual negative effects of the financial crisis.

Second is entitlement reform. Given the massive imbalances that exist today, consumers likely have little faith that current programs will remain in place throughout their lifetimes. Accordingly, cuts to entitlements that phase in gradually will likely have little impact on consumers' perceived lifetime wealth, as the benefit cuts are already factored into their thinking. If so, government can reduce promised benefits without causing consumption to go down. Such a reduction would restore market confidence and create the budget space to ensure successful tax reform.

Perhaps now that the Keynesian approach has so visibly failed, policy makers can finally see clearly enough to do the right thing.

Kevin Hasset is a senior fellow and Director of Economic Policy Studies at AEI.