Niall Ferguson, writing in the NYT (Dec. 7, 2003):

GUNS or butter: this is the choice historians conventionally say that governments face. Either they can build up their military capabilities to wield power abroad, or they can aim to increase their citizens' living standards.

In "The Rise and Fall of the Great Powers," the Yale historian Paul M. Kennedy developed this zero-sum model into a sophisticated theory of how empires work. In essence, you need wealth to be able to fight your rivals, but if you devote too much money to war, your wealth tends to stagnate. That's because (according to the theory) investment in the arms industry is less conducive to long-term economic growth than investment in sectors that ultimately satisfy some kind of consumer demand.

A simpler version of this idea suggests a trade-off between military spending and personal consumption. "Guns" are paid for by raising taxes, and this leaves people with less money to spend on "butter."

The Bush administration is currently engaged in an audacious — some would say reckless — experiment to disprove this theory. To judge by his actions, President Bush's response to the question "Guns or butter?" is: "Thanks, I'll take both." This, in short, is the guns and butter presidency.

It's generally a safe assumption that, in politics as in life, you can't have it both ways. But there are exceptions — provided you get the timing right. Today's economic circumstances mean that, in the short run, the administration can actually afford to spend billions simultaneously on conquest and on consumption.

In the long run, this double or nothing strategy has dangers — but, as Keynes remarked, in the long run we are all dead. All Mr. Bush needs to stand a good chance of re-election is 12 more months of guns and butter. In short, President Bush's second term depends on his being President Both.

Many a government has been impaled on the horns of the guns and butter issue. In the runup to Thanksgiving, however, two measures symbolized the Bush administration's conviction that it can grab those horns and take a ride. The first was approval of a $401 billion military appropriations package for next year, the biggest ever. The second was Congress's approval of a Medicare overhaul that increases the spiraling costs of the system by adding a drug prescription benefit...

To critics of the White House, the rapid shift of the federal budget from surplus to deficit is a sign of profligacy — part of what they would call the Enronization of public finance. It is true that there are real constraints on how much the administration can have of both guns and butter. Yet these constraints may prove to be weaker (or, to be precise, further away in time) than Mr. Bush's critics anticipate.

First, recall that the United States has broken the guns or butter rule before. Under President Ronald Reagan, substantial increases in military spending coincided with comparable increases, relative to gross domestic product, in personal consumption — that proportion of G.D.P. that the public, as opposed to the government, spends.

From 1979 to 1986, military spending leaped from 4.6 percent of G.D.P. to 6.2 percent, while personal consumption rose from 62 percent to 65 percent. Nor was this unprecedented. From 1965 to 1968, Lyndon B. Johnson managed to combine the guns needed to fight the Vietnam War with the butter of the Great Society — not to mention the ballooning consumer society.

The crucial point, of course, is that in the short term at least, fiscal policy is not a zero-sum game: a government can easily increase military spending without reducing consumer demand if it finances the higher spending by borrowing rather than taxation (and provided taxpayers do not view borrowing as future taxation and reduce consumption in anticipation).

The downside is that such debt-financed fiscal policies led to inflation in the past . In the late 1960's and in the late 1980's, deficits were partly financed by printing dollars, which ultimately led to higher prices.

The good news for Mr. Bush is that this is unlikely to happen now.