This is an exceptional time—a time in which many of the normal rules of the Dismal Science don't apply because, as Paul Krugman puts it, "depression economics" is in the driver's seat. The normal benefits and costs of government borrow-and-spend policies are overturned for now—and for as long as the crisis of high unemployment lasts.
Yet I find that many people do not understand why arguments that make perfect sense in normal times do not apply today. Let's run through the arithmetic—first in normal times, and then in a financial crisis like this one.
In normal times, a boost to government purchases:
• Produces a limited increase in production and employment;
• Creates a substantial increase in national debt;
• And requires that this new debt be financed at a sizeable interest rate.
Consequently, only government spending initiatives that promise a high value for the dollar are worth undertaking. Consider a $100 billion boost to government purchases. In a normal year, the Federal Reserve will worry about inflation, and raise interest rates somewhat to offset the inflationary impact of the fiscal boost. The multiplier effect of the purchases will therefore be something like 0.4—we will spend $100 billion on government purchases and gain perhaps $40 billion in extra production and associated employment out of it. Those who earn that extra $40 billion will pay taxes—perhaps $16 billion. So the net impact on government debt is this: by spending an extra $100 billion we will have added some $84 billion to the national debt.
That debt must then be amortized. At a five-percent-per-year long-run real rate of interest on government bonds, amortizing that debt will cost Americans $4.2 billion a year.
But wait—there is more. The Federal Reserve's compensatory increase in interest rates will also reduce investment. Because of the $100 billion in government purchases, perhaps $60 billion of private investment that would have been made won't be made—it will be crowded out. As a result of the lower capital stock, some $6 billion a year of income that would have been earned won't be.
Thus the net cost will be a reduction in Americans' disposable incomes of $10.2 billion per year.
That is not an attractive bargain: to purchase $40 billion of extra production now at the expense of a reduction in incomes of $10.2 billion in every year in the future. That is a usurious real interest rate of 25.5 percent. No sane economist would recommend a policy with such costs and benefits.
But how different everything looks in those rare times—like now—when depression economics applies!
First, more government spending will not lead the Federal Reserve to raise interest rates to fight inflation; the Federal Reserve has pushed interest rates to the floor right now and wishes it could push them lower still—perhaps another 5 percent into negative terrain. Given that, the multiplier on government spending is not 0.4, but more like 1.5. In other words, we do not get $40 billion of additional production and employment for $100 billion in government spending: We get $150 billion.
How? The federal government spends money to buy something it would not otherwise have purchased, thereby taking that product out of business inventories. Typically, businesses would respond by saying, "Our goods are flying off the shelves; we should raise our prices." As businesses raise their prices, (most of) what the federal government bought would be offset by a decline in private purchases because private buyers, confronting increased prices, decide to hold off on purchases.
But, again, now is not normal. Now, businesses respond to the federal government's purchases by saying, "Demand for our goods is greater than we thought; we don't need to fire as many people." The workers who they don't fire retain incomes they would not otherwise have had, and so they are a bit freer with their spending than they would be otherwise. Add up the extra spending by the federal government, and the extra spending by people who would have been fired but for the spending by the federal government, apply the appropriate offsets, and you wind up with a multiplier that most forecasters—by which I mean people who actually make their livings selling forecasts to businesses—think is probably around 1.5. Thus we get $1.50 of economy-wide spending for each $1 of stuff bought by the federal government.
Second, that boost to production creates a substantial reflow in taxes that makes the spending program lunch not free, but cheaper; $150 billion of added production leads to $60 billion of additional tax revenue, leading to only $40 billion in increased debt.
Third, depression economics means not only that there is no offsetting Federal Reserve interest rate increase, but also that the government can borrow at uniquely favorable terms: 2 percent per year in real terms for the next 30 years. Amortizing the $40 billion of additional government debt requires only $800 million a year in additional interest payments and taxes.
Fourth, the absence of interest rate increases means that there is no crowding-out of private investment. Private-sector incomes down the road are unchanged—or increased.
So the net cost of gaining $150 billion in increased production and incomes this year is $800 million a year going forward. It's not a free lunch—they take away my union card as an economist if I start claiming that things are free lunches—but it is a very cheap lunch: like getting a 2 lb. lobster with all the trimmings for $1.95.
Clearly, a second round of stimulus right now would be a very good deal for the American economy. But how big should the fiscal boost be?
Given the arithmetic, it is a no-brainer that we should be doing an extra $100 billion in stimulus. In fact, it is a no-brainer that we should be doing several $100-billion tranches.
How many? Well, the arithmetic holds until:
• Further tranches of additional deficit spending stoke inflation and lead the Federal Reserve to offset their impact on production and employment;
• Further tranches lead to a deterioration in the long-term corporate bond market and crowd-out private investment;
• Further tranches undermine confidence in U.S. long-term public finances and must be financed at higher Treasury interest rates.
Last December, Lawrence Summers and the rest of Obama's incoming National Economic Council feared that increasing the size of the stimulus program from $800 billion to $1.2 trillion would bring the above factors into play. It is now reasonably clear that they were overly pessimistic about the effects of additional short-term government spending, in large part because they were overly optimistic about the state of the economy.
So how much should the government spend? I would favor starting with $100 billion next month, and continuing with an additional $100 billion every month thereafter. Meantime, we should keep an eye on the bond market and inflation forecasts. As long as the terms on which the U.S. government can borrow are exceptionally advantageous, and the unemployment rate remains exceptionally high, the benefits of government spending will continue to be exceptionally large.