The global economy is in a deep, synchronized recession, and governments around the world are moving mountains to stop job losses and wealth destruction. Monetary authorities are pumping massive liquidity into credit markets and working with finance ministries to prop up, sustain, and nationalize major financial institutions. Nearly every government in Asia, Europe, and North America is pursuing some vigorous form of Keynesian fiscal stimulus policy, defined generally as debt-financed consumer-oriented tax cuts and substantial increases in government spending to push up aggregate demand in the hope that economic output, jobs, and incomes follow. President Bush signed a $152 billion stimulus bill in 2008 and President Obama signed a $787 billion stimulus bill early in 2009, and the ranks of unemployed continue to swell. Despite the paucity of results, some policymakers are suggesting the need for a third round of debt-financed spending.

The U.S. recession that began at the end of 2007 is different from that in Japan, which differs from those in Russia and Germany. This observation is important to understanding the recession triggers. Perhaps without exception, every country in recession today contributed to its own economic weakness in some material way, either through the actions of its citizens, institutions, or public policies.

Weakness first appeared in the United States in the housing sector, spread to the financial sector leading to a credit crunch that sapped the rest of the economy. Export-dependent countries like Germany and Japan suffered disproportionately from a collapse in international trade. Europe eventually succumbed in 2008 to the financial distress that swept the United States. From Iceland to Italy, financial institutions engaged in irresponsible, high-risk, highly leveraged lending similar to the lending in the U.S. mortgage market. Nor are the shocks over--the United States is facing new troubles from commercial real estate while Europe is badly exposed to dubious lending to emerging markets, especially in Eastern Europe.

Ongoing research will eventually determine what went wrong in the credit markets and in the regulatory architecture, and policymakers will attempt to respond. The immediate task, however, is to restore economic growth. In addition to very innovative, aggressive monetary policy responses, policymakers have pursued massive doses of Keynesian fiscal stimulus. The U.S. government alone may borrow up to $2 trillion in 2009 to finance its fiscal policy stimulus policy, equivalent to nearly 15 percent of gross domestic product (GDP).

Fiscal policy as an umbrella term refers to policies involving government revenue, spending, and debt issuance. In macroeconomics, fiscal policy may simply refer to whether the budget is balanced, in surplus, or in deficit. In public finance, the meaning of fiscal policy is more textured, involving the composition of government spending, distinguishing between direct consumption, research, infrastructure investment, and so on, and the kinds of tax systems imposed to collect revenues, such as property taxes, individual and corporate income taxes, and value-added taxes.

A stimulative fiscal policy in the newly revived Keynesian tradition increases the budget deficit from one year to the next to raise aggregate demand through either increased government spending or reductions in tax levels with the expectation that increases in output and income will follow. The federal government has twice applied Keynesian stimulus during the current recession with no evidence of improvement. There is good reason to believe that this brand of fiscal policy will not help the economy recover this time, or in the future.

A History of Ineffectiveness

The theory behind Keynesian stimulus is simple enough. The economy is underperforming; for whatever reason total demand from the private sector--consumption, investment, and the international sector--plus government demand is inadequate to allow the economy to operate at full employment. The proposed solution is to increase public-sector demand and let output rise to meet the higher level of demand. Expressed in these terms, the efficacy of fiscal stimulus would hardly seem debatable.

That something must be seriously amiss with Keynesian theory is apparent in the simple observation that if fiscal policy were so readily effective at raising output and lowering unemployment, countries with persistently underperforming economies would have been doing it for years. Some have tried, but their economies continued to underperform stubbornly, nonetheless, while their government debt burdens continued to rise.

The 1960s and 1970s were the golden age of Keynesianism. Policymakers embraced persistent budget deficits combined with accommodative monetary policy to fine-tune the economy and increase employment. This approach failed. As Christina Romer, Chairman of President Obama's Council of Economic Advisers, noted in a paper published prior to her government service, "The economic ideas of the 1960s and 1970s that led to expansionary policy also led to inflation and real instability."[1]

Europeans shared in the dream of fine-tuning the economy while justifying additional spending, with similarly lackluster results. As James Callahan, the former Labour Prime Minister in Great Britain, said in 1967, "We used to think that you could spend your way out of recession and increase employment by cutting taxes and boosting government spending. I tell you in all candor that the option no longer exists, and that insofar as it ever did exist it only worked on each occasion since the war by injecting a bigger dose of inflation into the economy followed by a higher level of unemployment as the next step."[2]

Japan of the 1990s is the modern poster child for Keynesian stimulus, having embarked on massive government infrastructure projects producing wonderful new roads, bridges, waterworks, and airports.[3] Net government debt rose as a share of the economy from 15 percent in 1990 to 60 percent in 2000.[4] Japan was left with beautiful infrastructure, a mountain of debt, and the now-resumed lost decade.

Recent experience in the United States with Keynesian policy is no less discouraging. The United States ran a budget deficit in 2008 of $459 billion, or 3.2 percent of GDP, up from a deficit of 1.2 percent of GDP in 2007. This increase of 2 percent of GDP represented a powerful dose of Keynesian stimulus and yet the recession accelerated markedly. Again, an explicit policy of Keynesian stimulus failed.

According to the Congressional Budget Office (CBO), the U.S. government is expected to run a deficit of $1.8 trillion in 2009, or 13 percent of GDP.[5] This would amount to a stunning $1.4 trillion of Keynesian stimulus--nearly 10 percent of GDP. Despite this massive jolt of deficit spending, the CBO and others project the real economy to decline in 2009. The numbers tell the story in black and white. Either these forecasters believe the economy would have contracted by 11 percent or more in 2008 but for the stimulus, or they believe massive Keynesian stimulus will be as ineffective in 2009 as more modest stimulus was in 2008.

For 2010, the CBO projects a deficit of $1.4 trillion under President Obama's budget, a decline of $393 billion, or 2.7 percent of GDP. Under the Keynesian theory, the deficit needs to rise slightly to have a neutral effect on the economy in the short run. A drop in the deficit of 2.7 percent of GDP under this theory is then massively contractionary. Keynesians should be in panic about the economy's immediate future. Most forecasters, including the CBO, appear calmly to ignore this phantom contractionary pressure in their own economic forecast. Apparently, forecasters outside the political realm do not believe in Keynesian theory, either.

Why Keynesian Stimulus Fails--The Second Half of the Story

Simple observation has its place, but how does the Keynesian stimulus approach break down in theory? Keynesian stimulus theory ignores the second half of the story: Deficit spending must still be financed, and financing carries budgetary consequences and economic costs. Proponents generally acknowledge the long-term budgetary costs, but ignore the offsetting near-term consequences that render Keynesian stimulus useless.

In a closed economy, government borrowing reduces the pool of saving available for private spending, either investment or consumption. Government lacks a wand to create real purchasing power out of thin air (with the fleeting exception of monetary expansions, discussed below). Government spending or deficit-increasing tax cuts increase demand as advertised; and government borrowing reduces demand by the same amount, for no net change.

The dynamics in an open economy are slightly more complicated, but the final outcome for output is unchanged. An open economy permits a government to finance its deficits by importing savings from abroad as the United States has done for years, rather than by tapping domestic sources. However, an increase in deficit spending met by an increase in net imports of foreign savings must, in turn, be matched by an increase in net imports of goods and services to preserve the balance of payments. Thus, the increase in domestic demand due to deficit spending is fully offset by a reduction in demand arising from an increase in net exports. Once again, Keynesian stimulus has no effect.

What if the extra government borrowing soaks up "idle savings" in an underperforming economy, proponents may ask. In troubled economic times those who can save more often do so, directing their savings toward safe investments like Treasury Bonds and bank deposits. However, these cautious savers almost never withdraw their savings from the financial system entirely by stuffing cash into mattresses. Aside from the occasional mattress stuffer, even savings held in the safest of instruments are not idle but remain part of the financial system, working to find their most productive uses through the available channels. Borrowing to finance Keynesian stimulus, then, remains a subtraction from the funds available to the private sector.

Suppose widespread fear spurred savers to engage in rampant mattress stuffing, withdrawing purchasing power from the economy and creating large amounts of truly idle savings. This has happened before, and could be happening now to some extent. Surely, Keynesian stimulus works in such cases. Highly unlikely. Nothing about a flood of government bonds engulfing capital markets to finance a surge in wasteful government spending is likely to convince nervous mattress stuffers that their concerns are misplaced. Idle savings, then, remain idle, making deficit spending a competitor for an even smaller pool of available private savings. Worse, mattress stuffers are likely to increase their mattress-based, economically idle saving in the face of a surge of profligate, irresponsible government spending. Keynesian "stimulus" would then be an economic depressant.

Printing Money to Make Fiscal Stimulus Work

In the last theoretical refuge for Keynesian stimulus, suppose the monetary authority broke its commitment to independence and opted explicitly to buy up the Treasury's debt issuance under a Keynesian fiscal policy. Government cannot create real purchasing power by whim, diktat, or debt, but the monetary authority can create the illusion of purchasing power through a policy of monetizing debt and increasing cash liquidity in the economy. Combining an obliging monetary policy with increased deficit spending may create the illusion that fiscal policy is effective, but as Prime Minister Callaghan commented, it is temporary and only an illusion.

In most countries the monetary authority's independence is a foundational policy principle. The monetary authority may buy significant amounts of Treasury notes and bills in pursuit of its own expansionary monetary policy as the Federal Reserve has done for many months in extraordinary quantities. But these actions would be driven by monetary policy considerations of maintaining strong long-run growth consistent with low and stable inflation. The monetary authority would take these actions whether or not the fiscal authorities embarked on a stimulative policy. The central bank's policy goal is the same as the Treasury's in this instance--to resuscitate the economy--but the central bank is ultimately pursuing its policies independent of Treasury policy.

However, the monetary authority may opt to subordinate its policy rules and objectives to fiscal policy, repeating the 1970s experiments in the United States. The previous outcome of loose monetary policy was economic stagnation coupled with high and rising inflation, what came to be known as "stagflation." The policy failed to produce sustained economic growth because market participants learn quickly. Businesses, investors, workers, and savers recognized the shift toward an inflationary monetary policy, interpreted it correctly, and reflected higher inflation in their pricing and expectations. In so doing, they nullified the potential stimulative effects of the policy and the Federal Reserve was forced to adopt a contractionary counter-inflationary policy resulting in the deep recession of 1981- 1982. Even a compliant central bank cannot make Keynesian policy effective unless the central bank can consistently and persistently fool the markets.

Casual empiricism suggests that Keynesian stimulus policy does not work, and the theory behind the policy fails on inspection. What does empirical research indicate?

Empirical Insights on Keynesian Effectiveness

One approach to testing the efficacy of debt-based fiscal stimulus is to turn to the data and learn what stories it tells. Unfortunately, few have attempted this task in recent years. This may be due to a focus on the emergence, development, and parameterization of a new consensus model in macroeconomics, the so-called New Keynesian model.[6] Also, until recently most of the developed world (other than Japan) had been relatively immune to significant business cycle swings, thus dampening the demand for research on countercyclical fiscal policies in industrial nations. Part of the reason may also be the strong consensus, before recent events, that Keynesian stimulus was ineffective and that studies reporting statistically insignificant results confirming the consensus view are rarely published.

Perhaps Robert J. Barro's analysis of fiscal stimulus efficacy is the most well known and controversial. Barro argues the clearest evidence of fiscal policy effects is likely to be found when spending ramps up rapidly during wars.[7] Examining U.S. fiscal policy in the periods surrounding World War II, the Korean War, and the Vietnam War, Barro's analysis suggests a fiscal multiplier of 0.8, meaning even at its most effective, the increase in output was a fraction of the increase in government spending.

Barro further theorizes the wartime multiplier is likely to be much greater than the peacetime multiplier, with a peacetime multiplier likely near zero so every extra dollar of government spending actually replaces a dollar of private spending leaving output unaffected. Paul Krugman among others have criticized Barro's results, noting that the wars themselves and the often attendant wage and price controls would have diminished the effectiveness of fiscal policy.[8] However, none of his critics has as of yet provided an empirical analysis challenging Barro's results.

Using a purely statistical approach, Andrew Mountford and Harald Uhlig find an unexpected increase in government spending, beyond what would occur through automatic stabilizers, "weakly stimulates the economy": a 1 percent increase in spending increases output by about 1.3 percent after one year.[9] Proponents of extra increased spending as Keynesian stimulus may take comfort in this result, but then they must also acknowledge that the authors find "a deficit-financed tax cut is the best fiscal policy to stimulate the economy."

An alternative approach to ferreting out fiscal multipliers is to use macroeconomic models to simulate policy effects. Modern macroeconomic models are abstract, mathematical representations of essential elements of the economy as described by theory. They may be simple or complex, derived from underlying principles or constructed from suggested broad relationships. Their great advantages are that they force a degree of specificity on the part of the model builder and offer as a reward the ability to examine economic interactions consistently and in great detail.

The downside to all economic models is that they often provide uncertain illumination for policymakers because the models ultimately only report what their builders have designed into them. Economic models are inherently abstract representations of an economic phenomenon, dependent on the state of economic theory and the quality and availability of data. Given these limitations it can be difficult to discern whether an interesting result reflects the model or the economy the model is intended to represent.

Christina Romer and Jared Bernstein, Chief Economist of the Office of the Vice President, provide a recent example of the model simulation approach.[10] They averaged the output from policy simulations using two quantitative macroeconomic models--one in use at the Federal Reserve Board, and one from an unnamed private forecasting firm. Romer and Bernstein found that an increase in government spending of 1 percent of GDP increases output by 1.6 percent.

In contrast, John Cogan and his colleagues[11] used a state-of-the-art macroeconomic model constructed by Frank Smets and Rafael Wouters.[12] The Smets-Wouters model embodies the "new Keynesian" approach to macroeconomic analysis. Among the differences from older models, such as those used by Romer and Bernstein, Smets-Wouters includes forward-looking, or rational, expectations. Cogan found the impact in the first year of a Keynesian stimulus to be "very small" and that the multipliers are less than one as consumption and investment are crowded out.

As the above discussion on monetary policy suggests, the policy of the central bank can have a powerful influence on the economy and thus on the apparent effectiveness of fiscal policy. Eggertsson used a model similar to Smets-Wouters to examine these questions.[13] Her analysis explored the consequences of increased government spending in the two cases in which monetary policy is and is not explicitly coordinated with fiscal policy. Uncoordinated policies need not mean that monetary and fiscal policies have divergent goals. For example, both monetary policy and fiscal policy may react to economic weakness, a threat of deflation, or off-target inflation. As defined by Eggertsson, the lack of coordination in policies means that in reacting to macroeconomic conditions the monetary authority's actions may be coincidental to fiscal policy, but not specifically intended to support fiscal policy. On the other hand, if the monetary authority sets aside its usual guidelines to subordinate monetary policy to fiscal policy goals, it is considered to be coordinated with fiscal policy.

Eggertsson found fiscal policy very effective if monetary policy is explicitly supportive, producing a fiscal policy multiplier of 3.76. However, if monetary policy remains independent, as economists and financial markets generally assume, the multiplier becomes exactly zero and fiscal policy is completely ineffective. This latter result is fully consistent with the theoretical discussion above and is generally consistent with Cogan et al., who also explicitly assumed that the monetary authority remains fully independent of fiscal policy.

Stepping back, Eggertsson's monetary policy focus, while understandable coming from a member of the New York Federal Reserve staff, is perhaps not on point as a test of Keynesian stimulus. Eggertsson's results derive from a stylized model intended to explore a specific question of economic policy and thus depend critically on the effects of fiscal policy on inflationary expectations. These are important issues but do not address the underlying rationale for Keynesian fiscal stimulus of increasing aggregate demand. The real message of Eggertsson for current policy is to underscore the point made above that despite their sophistication modeling exercises sometimes address the modeler's interests more than they do the economic processes relevant to policymakers. This is not a criticism of Eggertsson or any user of such economic models, but rather a caution to those who might interpret and apply their results.

Keynesian Stimulus Theory Comes and--Thank Goodness--Goes

Recent experience with Keynesian, deficit-based fiscal policy to provide short-term stimulus to the economy very much agrees with the theory described above and essentially agrees with the corpus of modern empirical research: Keynesian stimulus does not work. As the economy was experiencing a mild recession in 2008, the budget deficit jumped by 2 percent of GDP, partly as a result of the $152 billion stimulus signed into law by President Bush. The economy was presented a sizable dose of Keynesian stimulus and remained in the doldrums with employment dropping by 590,000 workers from the third quarter of 2007 to the third quarter of 2008. Then the recession worsened significantly.

Consequently, President Obama inherited a budget deficit of almost $1.7 trillion, representing a massive dose of Keynesian stimulus equal to 8.7 percent of GDP, almost three times larger than the largest such peacetime increase. He then advocated even more, and signed a $787 billion package of spending hikes and ineffective tax reductions, pushing the projected 2009 deficit to above $1.8 trillion. Despite the massive initial and the additional deficit spending, the unemployment rate rose by almost two full percentage points in the first half of 2009, and Administration officials are cautioning that unemployment will rise above 10 percent.

The Keynesian stimulus theory fails for the simple reason that it is only half a theory. It correctly describes how deficit spending can raise the level of demand in part of the economy, and ignores how government borrowing to finance deficit spending automatically reduces demand elsewhere. Exculpatory allusions to idle saving simply do not wash in a modern economy supported by a modern financial system. Deficit spending does not create real purchasing power and so it cannot increase total demand in the economy. Deficit spending can only shift the pattern of demand toward government-centric preferences.

Empirical research rarely provides a simple, single answer to a policy question, and examinations of Keynesian stimulus are no exception. Yet the available results consistently indicate that, using a modern macroeconomic model and treating monetary policy carefully, Keynesian stimulus's short-term effects lie somewhere in the narrow range between slim and none. Keynesian stimulus produces debt, not jobs.

Bad policy ideas rarely go away forever. Circumstances change, memories fade, political fashions come and go. The current global experiments with Keynesian fiscal stimulus will fail as they have failed before. Unfortunately, the price of learning this lesson yet again is an unnecessarily prolonged recession, a weaker recovery, and millions more lost jobs--and, of course, the massive increases in public debt.

J. D. Foster, Ph.D., is Norman B. Ture Senior Fellow in the Economics of Fiscal Policy in the Thomas A. Roe Institute for Economic Policy Studies at The Heritage Foundation.