You are using an outdated browser.
Please upgrade your browser
and improve your visit to our site.
Skip Navigation

Let's Start Over

A Budget Cure-All

At least credit Richard Darman for creative accounting in managing to squeeze the new White House budget into the spending categories and constraints of the 1990 budget agreement. But his efforts were doomed from the start; the agreement is obsolete, and everyone knows it. It was cobbled together before the cold war ended and before the recession started. The question is: What's to replace it?

The real scandal of the budget that President Bush sent to the Hill was not so much the expected deficit in 1993 ($351.9 billion), but the paltry sum that the White House expected to invest in the future ($133.1 billion). This isn't new, of course. For years federal borrowing has exceeded federal investment. To solve the problem, Democrats have talked about creating a "capital" budget or an "investment" budget, designed to make clear how much the nation is investing in the future. Bill Clinton has called for an investment budget. (I'm helping him in the campaign, but he shouldn't be burdened by the following musings.) Yet no one has ever provided a detailed blueprint. How might it look? A sane budget would be one whose spending categories and constraints, and whose sources of revenue, are logically connected to the generation of Americans that benefits from them. In particular: a budget that reveals how much we're borrowing from the future compared with how much we're investing in the future.

Begin with the basics. Ideally, any federal budget should be divided into three categories corresponding to different time-dimensions of government responsibility: 1. Past obligations to a previous generation of workers, to guarantee them a minimally decent standard of living. Included here would be most spending for Social Security and Medicare. 2. Present obligations to ourselves, to ensure that all Americans enjoy a safe, secure, and minimally decent standard of living. This category would include spending on welfare, criminal justice, and defense. 3. And future obligations to coming generations, to ensure they have the capacity to generate a good standard of living. This means spending on education, training, civilian research and development, and infrastructure.

Although the American population obviously doesn't divide neatly into separate generations, it's still useful to think about the needs of Americans ten or fifteen years from now as separate from their needs today and from the needs of people no longer able to support themselves. For example, revenues to support the retirement of the past generation of workers should come from workers' (and employers') contributions. People can debate about how progressive this system should be -- what payments should be required of workers with differing incomes and what benefits should go to retirees of different wealth. And they can debate about whether (and to what extent) the system should be financed on a pay-as-you-go basis, with current workers picking up the tab for retirees. But the system should finance itself -- and be separated from the rest of the federal budget. No dipping into the other categories. Likewise, neither of the other two categories should be able to raid it for funds.

The second budget category -- to maintain the current safety, security, minimal living standard of Americans -- is logically financed through taxes on current incomes (including investment income). Here again, we can debate how progressive this tax should be, but the principle is the same: revenues to pay for the current living expenses of Americans should come from the current incomes of Americans. There's no reason why one generation should be able to borrow from future generations to finance its own standard of living. Nor should they tap into their own retirement savings (by, say, using surpluses in the Social Security trust fund to offset current budget deficits).

Thus, so long as marginal tax rates don't change from year to year, spending in this second category cannot grow faster than improvements in the incomes of Americans. This seems logical. How can we expect to consume collectively, at a faster rate than we produce? (One small wrinkle: to enable the government to provide extra relief during troughs when incomes stagnate or dip, and to cool the economy during recoveries when incomes surge, the yearly spending cap might be based on a rolling average of national income gains during preceding years.)

The third and last category -- investments in the future capacities of Americans to produce wealth -- is appropriately financed through borrowing that must be repaid in the future, when the nation reaps the benefits of these investments. Such investments are subject to a different constraint than is spending in the other two categories. No self-respecting business executive fails to borrow money if the expected return is much higher than the borrowing costs. Similarly, so long as the likely returns to the nation's future productivity are high enough, borrowing in order to finance such investments makes sense regardless of the amount. In fact it would be irresponsible not to do so. Borrowing and investing shouldn't stop until the cost of additional borrowing is no longer worth it -- given the expected return on such investments relative to what private investors could get on their borrowings.

So much for theory. This sort of budget is logical, and the public can readily understand it. But how would it work in practice? There are at least four big obstacles.

What's an investment? The first and most obvious problem is to decide what programs fit into which category. Since no one wants to have his favorite government program vulnerable to the spending constraints of either the first or the second categories, there's likely to be a stampede into Category Three. After all, the definition of an "investment" is hardly airtight. And when politics tangles with epistemology, politics almost always wins.

But this problem may not be insurmountable. Last Sunday afternoon, with nothing better to do now that football season is behind us, I took out a copy of the White House's new budget proposal and ticked off the items that seemed to me to be justly included within each of the three categories. Category One was fairly straightforward: Social Security and Medicare, including some additional retirement programs like veterans' benefits. But how to distinguish between Categories Two and Three -- spending for today and investments for tomorrow? It wasn't nearly as difficult as I imagined. I simply asked myself whether there was any reason to believe that the program would pay off in the form of improved national productivity a decade from now.

Who's to say that my assumptions would be widely shared? Luckily, I had a test. Officials in the White House's Office of Management and Budget undertake a similar exercise each year (presumably with more thought and care than I gave to mine) as part of the budget presentation. The analysis is buried in the back of the budget document. I found their "investment" calculations for FY 1993 and compared them with the ones I'd done.

The two sets of calculations were remarkably alike. For example, I put $13.1 billion of 1993's science, space, and technology programs into my Category Three; OMB dubs $10 billion of these programs "investments." I put $26.5 billion of 1993's transportation spending into the category; OMB puts $24.8 billion. I placed $31 billion of education, training, and social services into it; OMB's number is $37.6 billion. For a few programs, OMB is more generous: I didn't put a dime of federal spending on income security, justice, veteran's benefits, housing credits, or general government revenues into my Category Three; OMB tosses over $9 billion worth of these programs into it. There was one big difference. I nominated $20.3 billion worth of programs related to the health and nutrition of children; OMB includes just $10.1 billion of these. Would my number have been closer to theirs if I had more information about these programs? Would theirs be closer to mine if I had presented my arguments to them? Maybe.

The two sets of calculations totaled up about the same. Under my calculation, in 1993 the White House is planning to invest $124.2 billion in the nation's future -- or about 8.2 percent of what it plans to spend that year. OMB's total is $133.1 billion -- or 8.8 percent. Either way, a paltry portion of the federal budget.

My point is that there may not be that much disagreement over which programs fall into which budget category. Politicians will try to shoehorn their favorites into the investment category, of course. But the danger can be minimized (although not eliminated) by allowing into Category Three only programs that have been so designated by the White House with the concurrence of majorities of the House and Senate Budget Committees.

How to measure the returns on public investment? It's hard enough to predict the benefits of private investments, which at least show up on the bottom line. It's nearly impossible to predict how this or that program will affect the nation's capacity to generate wealth in the future. There are simply too many factors to take into account -- and no clear-cut way to measure improvements in national productivity anyway. For years academics have been engaged in lively debates (lively, that is, for academics) over the value of public investments -- and even livelier ones about the outcome of investments in education. Of course, some estimates are broadly agreed upon, such as those showing that children in Head Start are far more likely than other poor children to complete their educations and be self-supporting and law-abiding in later life. It's just that, by their very nature, attempts to measure the future productivity gains of public investments are imprecise and subject to debate.

And yet, these controversies notwithstanding, there's a remarkable degree of agreement on the rough magnitudes of the returns on such public investments. Almost every economist and policy wonk who has studied the subject has concluded that the United States has woefully underinvested in roads, bridges, sewer systems, worker training, child nutrition, child immunization, civilian research and development, and a host of other areas where the private sector can't be expected to do it all. Almost everyone recognizes that other nations, notably Japan and Germany, are investing far more, as a percentage of their national products, than we are.

Moreover, no one disputes the fact that federal spending on such investments has been declining as a percentage of GNP. Infrastructure spending dropped from 1.14 percent of GNP in 1980 to 0.75 in 1990. Spending on education dropped from 0.51 percent to 0.37 percent. Non-defense research and development, from 0.42 percent to 0.31 percent. So even though exact measures are impossible, it is possible to set broad goals for public investment over, say, the next decade.

How to draw down the debt? By 1994, according to Congressional Budget Office projections, the accumulated federal debt will total over two-thirds of the year's gross national product. So even if we raised taxes to cover current expenditures (Category Two), any additional borrowing to finance investments in future productivity (Category Three) would cause yearly interest charges to grow almost as large as those on my Visa bill.

That's why we also need a rule to draw down the accumulated debt as the national economy begins to grow faster -- reaping the benefits of these new investments. Again, it's only logical that Americans in the future, who will enjoy a higher standard of living because of investments made today, should pay not only the interest on the debt but also bear some responsibility for reducing it. One possibility: an income tax surcharge that's triggered in any year when the economy grows faster than, say, 4 percent -- the proceeds of which would be applied solely to debt reduction.

How to get from here to there? Even if we could agree on these categories and principles, we'd still have to face today's mess. Right now, Categories One and Three are being used to bail out Category Two: that is, the Social Security surplus that's been created to help finance baby-boomer retirees is being used instead to offset the budget deficit, and most of the remaining deficit is being used to finance present living standards rather than to invest in the future. So we need a plan that over some period of years will raise tax revenues and reduce Category Two consumption until they become equal. Simultaneously, we have to reduce borrowing and increase Category Three investments until they're equal.

Can we do it? Seems awfully difficult. But remember that as public investment increases, we should expect the entire economy to grow faster. That means tax revenues will grow as well. I'm not suggesting a painless transition -- only that one of the real objectives is to foster economic growth. And to the extent that we invest for the future, it will be that much easier to fulfill our current needs as they arise. But why should the public buy it? Why won't we (and our representatives in Washington) continue to prefer borrowing from the future to finance current consumption, without investing in the future? These are rhetorical questions, of course, to which most Washington-watchers have every right to give a cynical answer. I can offer only one small reason for hope. Most adult Americans do have children, and are at least a bit concerned about how their children will live. At best, a budget like the one I'm suggesting may handicap our efforts to deny that we're treating them far worse, collectively, than we would ever dream of treating them individually.