The surplus of ’99
On March 11, Dan Rostenkowski issued a challenge: Adopt my plan to fix the budget deficit or come up with a better one.’ Challenge accepted.
TRUE OR FALSE: To balance the budget in the 1990s we must either increase taxes or cut spending from current levels.
The right answer is “false.” If you responded true,” however, I won’t take off too many points. There is probably no more widespread myth about the U.S. budget system than the supposed need to raise taxes or cut spending.
Let’s consider how the myth got started. We do have a budget deficit of some $160 billion. For the Congress to balance the budget in a single year, it would have to enact some $160 billion in combined tax increases and spending cuts. Since most of our budget is uncontrollable,” in that it is devoted to defense, Social Security, Medicare, and interest on the debt, cutting $160 billion is impossible. So the pundits say we need a major tax increase as well.
The fallacy in this analysis is that no responsible economist would suggest that the United States try to reduce its budget by $160 billion in a single year. The result would almost inevitably be a very sharp recession, or even a depression, as the government desperately sucked spending power out of the economy. Most sensible people favor narrowing the gap by $30 to $40 billion per year, ridding ourselves of the deficit over four or five years. That requires neither raising taxes nor reducing spending below current real levels, for the current tax code will automatically provide $75 to $80 billion more in income-tax revenue each year for so long as economic growth continues at its current rate. With an automatic $300-billion revenue increase already scheduled for the next four years, we could increase spending by $140 billion (roughly enough to keep up with inflation) and still bring the deficit down to zero in just four years.
The Congress does not need to raise taxes because taxes are going up anyway. There is no voodoo involved. Both real growth in the economy and continued inflation effectively increase average tax rates and tax revenues over time. In the 1960s Keynesians called this effect a fiscal drag” or a fiscal dividend,” depending on whether they were advocating more stimulus or arguing that such stimulus was largely self-financing. The table on page 35 summarizes this relationship for each of the major types of taxes the government collects, in each case showing the effect on tax revenue of a 1 per cent increase in the economy’s real output as well as 1 per cent inflation. TART WITH the personal income tax. We are all aware of the phenomenon known as bracket creep, by which inflation pushes us into higher tax brackets without our real income increasing. The indexing provisions of the Economic Recovery Tax Act of 1981 (ERTA) curbed this highly destructive practice. With indexing, if both your income and prices rise 1 per cent, your tax payments-not your tax rate-will also rise 1 per cent, and taxes will take the same percentage of your income as before. As a general rule, therefore, 1 per cent inflation will cause personal income-tax revenue to rise only 1 per cent. But because not all parts of the income tax are fully indexed, tax revenues do rise slightly faster than inflation.
Of more importance is what happens when you get a real raise, not just an inflationary increase. Since indexing does not apply to changes in real income, you may be pushed into a higher tax bracket when your income rises. Even if your tax bracket does not change, your tax payments will go up because you will be paying on a larger income. Moreover, your average tax rate will rise because a smaller part of your income will be protected by basic exemptions and deductions.
Consider the case of a four-person family earning $30,000 in 1989. It is allowed an exemption of $2,000 for each person plus a standard deduction of $5,000, for a total of $13,000. Its taxable income is then $17,000. The tax rate on taxable income up to $31,000 is 15 per cent, so this family pays a tax of $2,550, which is 8.5 per cent of its total income, for an average tax rate of 8.5 per cent. Now assume the family’s real income increases by 10 per cent, to $33,000. Its deductions are unchanged, so its taxable income rises to $20,000. The tax on $20,000, at the standard 15 per cent rate, is $3,000. The family’s tax payments rise 17.6 per cent ($450 on top of $2,550) though the family’s income went up only 10 per cent. On average, every 1 per cent rise in the family’s income caused a 1.76 per cent rise in its tax payments. The family’s average tax rate rose from 8.5 per cent to slightly more than 9 per cent. This increase occurred even though the family remained in the same 15 per cent tax bracket. A greater fraction of the family’s income was taxed at the 15 per cent rate, and a smaller percentage at the zero rate. Note that this is not a case of bracket creep, in which inflation increases the rate applied to the same real income: this family is paying a higher average rate because it really is making more money. This is fundamental to any progressive tax code. The family is still better off.
If everyone in the United States got a 10 per cent raise, tax revenues would rise about 15 per cent. Ordinarily this would imply that every 1 per cent of real growth would produce about 1.5 per cent extra revenue. But not all real economic growth can be attributed to salary increases for those currently working; about one-third is due to new workers joining the labor force. New workers add to tax revenues, but they do not produce the same average increase per point of economic growth that occurs when existing workers get higher wages. The actual average effect is thus about 1.3 to 1.4 per cent extra tax revenue per point of economic growth.
CORPORATE income-tax revenues also rise much faster than the economy, as long as the economy is expanding. Corporations pay a flat tax rate, and so are not subject to bracket creep. But corporate taxes are increased both by inflation and by real growth because of the way that the tax system defines corporate profits.
Imagine the production process as a giant pipeline: raw materials, labor, and capital enter one end, and finished products come out the other. Corporate profits are the difference in value between what goes into the pipeline and what comes out, or the difference between costs and final sales.
When the economy slows down, fewer finished products are purchased and items build up in the pipeline, a process known as inventory accumulation. Corporations thus incur the costs of producing the inventory without having any final sales, and corporate profits fall.
When, on the other hand, the economy speeds up, existing items in the pipeline, already largely paid for, are ready to move. Final sales increase faster than input costs, and corporate profits rise quickly.
This “pipeline effect” is most dramatic at turning points in the economy. When the economy turns up or down, corporate profits can easily change as much as 3 to 5 per cent for every I per cent change in the economy. As business expansion continues, the ratio drops to about 1.4 points of corporate profit for every point of economic growth. Eventually, rising demand for products can be met only by new investment: the pipeline itself must be expanded in order to accommodate the greater flow of products. This is where the tax system plays a part. Suppose a business invests in expanding the capacity of its pipeline by 10 per cent. We might expect sales, profits, and costs, including capital costs, to rise together in proportion and the capital costs of a 10 per cent larger pipeline also to be 10 per cent higher. But the tax system does not allow corporations to deduct the capital costs of the bigger pipeline immediately; these added costs must be depreciated over many years. Because the costs of the expansion cannot be fully counted, profits appear to rise much more quickly than economic activity expands. Thus corporate profits, and therefore corporate taxes, increase more than point for point with economic activity during periods of real economic growth.
Inflation has a similar effect. Again, consider the goods moving through the pipeline. During periods of inflation, the value of the inventory in the pipeline rises with the general price level. The corporation does not really gain, because the cost of replacing the materials at the beginning of the pipeline also rises. But the tax system views the rising value of inventory as a source of corporate profit. This is particularly true for corporations that use the first-in-first-out (FIFO) method of accounting, under which the cost of products is based on the original cost of materials.
The tax system also fails to take into account the effect of inflation on the capital costs of the pipeline itself. When inflation is high, the depreciation allowances for the cost of the pipeline are far lower than the cost of actually replacing the pipeline at inflated prices. The result, since corporations cannot fully deduct the true capital costs of production from their profits, is an artificial rise in both profits and taxes. All these factors combine to increase corporate tax payments about 1.4 per cent for every 1 per cent increase in the price level. (That 1.4 is a conservative estimate; for the expansions of 1965-9 and 1982-5, the actual figures were 1.6 and almost 1.9.)
Social Security tax payments tend to rise point for point with both real economic activity and inflation, because they are indexed to the price level and have a proportional rather than progressive rate structure. Unless social-insurance tax rates change, these revenues will continue to grow almost exactly as fast as the economy.
Excise taxes, such as those on gasoline, alcohol, and cigarettes, included in “Other” in the table above, are generally levied on a “cents per unit” basis, and so are totally unresponsive to inflation. Sales of goods taxed in this way tend to increase more slowly than the expansion rate of the economy. As a result, revenue from these sources does not rise as fast as overall economic activity. Altogether, however, tax revenues will grow significantly faster than the economy, and this rapid expansion will balance the federal budget if we are at all sensible about spending. The only question is how fast we can expect the economy to grow.
Long-term economic projections are chancy, but let us consider two possibilities. The long-term growth of the U.S. economy since the end of World War II, averaging out recessions and expansions, has been 3.2 per cent per year. Though there is no reason to expect that we will do better in the future, there is also no reason to expect that we will do worse. We might reasonably project an average growth of 3.2 per cent over the next twenty years. If we limit our base to more recent years, say from 1981-the peak of the last business cycle-through 1986, which includes the deepest recession we have experienced since the Great Depression, we get an average growth in the economy of only 2.7 per cent per year. Projecting that figure forward would produce what we might call the bearish long-term forecast. The table on this page shows what would happen to revenues with these rates of growth over the next twenty years, assuming inflation continues at the average rate of the past several years, or about 4 per cent. In both cases, revenues grow quickly over time, and an increase of just 0.5 per cent in average real growth adds $530 billion per year by the twentieth year. This is an important indication of just how sensitive the U.S. budget is to continued economic growth.
To reduce the deficit we do not need to cut spending. But we must show some self-control. After all, the Congress could spend every penny collected in revenue and then some. To illustrate the long-term budget needs of the country, the spending” column in this table employs a five-year flexible freeze” such as President Bush championed during his campaign. The flexible freeze limits spending growth to the level of inflation for five years, then allows real program growth of 2 per cent per year. Thus, program spending rises at a 4.0 per cent annual rate through 1994 and at a 6.1 per cent annual rate thereafter. We assume the interest rate on the national debt, which cannot be frozen, will be about 7.6 per cent annually (that was the average rate of interest paid on outstanding publicly held debt during 1989, according to the 1990 Federal Budget).
This table shows that a tax increase is completely unnecessary. If spending is limited to the growth of inflation until 1994, we will have either a balanced budget (with slow growth) or a $40-billion surplus (with average growth). After 1994, a large surplus arises and begins to grow even though spending is allowed to grow faster than inflation. By 2004 we will be showing annual surpluses of roughly $800 billion and will have paid off the existing national debt.
Of course, this will not happen. Given such enormous revenues, the Congress will try to increase spending. Assuming the freeze holds until 1994, both spending and taxes will be 19.5 per cent of GNP in that year. Over the next 15 years, taxes would automatically grow to between 22.3 and 22.9 per cent of GNP. The Congress could drastically increase the size of government and the government’s share of the economy if it appropriated all of the extra tax revenue for that purpose.
Instead, the country should insist on a series of tax cuts in the latter part of the 1990s. The $300-billion surplus forecast for 1999 would easily finance a 15 per cent reduction in tax rates across the board. If the entire surplus were applied to reducing the income-tax rates in the current code, the bottom rate would be reduced from 15 per cent to 10 per cent and the top rate reduced from 28 per cent to 20 per cent. Form 2000, the personal-income-tax code I have proposed [cf. “On the Right,” April 161, is designed to produce roughly the same amount of revenue as the current code. So if in the early 1990s we adopted Form 2000 with its single rate of 19 per cent, we would be able to reduce that rate even further in the late 1990s. Assuming we control spending for the next five years, the country will be able to afford substantial tax relief by the end of the century.
ALL OF THIS may seem magical given the headlines about the government’s current fiscal crisis. Yet these calculations do not assume any behavioral changes on the part of the public, or any sharp drop in interest rates, or any unusual rate of economic growth. The only magic involved is the magic of normal economic growth compounded year after year, coupled with permanent restraint on the growth of government spending.
Present-day pundits, stuck in their static analysis of a single year’s budget, never calculate the effect of revenue growth on the prospects for the U.S. budget. Hence the persistent calls for tax increases. But tax increases would tend to slow the rate of economic growth and reduce future revenue growth. Much better to keep taxes at current rates-or lower them-and assure continued economic growth not only for the Treasury but for all Americans.