Where Figures Fail; R. Higgs

Employed with caution, they can provide valuable information. How Much Has Government Grown? One widely used measure is gov ernment spending (federal, state, and local) for final goods and services relative to the gross national prod uct (GNP). In the early twentieth Robert Higgs is Professor of Economics at the Uni versity of Washington. He is popular as a lecturer on economic and monetary affairs. His writings include numerous articles as well as books on The Transfor mation of the American Economy, 1865-1914, and Competition and Coercion. century, this share stood at 6-7 per cent and fluctuated little from year to year. During World War I, the federal government's procurement and mobilization efforts drove up government's share to over 21 per cent of GNP. After the war it de scended as rapidly as it had as cended, stabilizing during the 1920s at a level only slightly higher than that of the prewar era.

With the onset of the Great Depression, government's share in creased from about 8 percent in the late twenties to a plateau of 14-15 percent during the New Deal era. Notably, the initial rise in 1930-32 occurred because GNP fell precipi tously while government spending for currently produced goods and services remained roughly the same. After 1933, government spending increased but GNP rose at about the same percentage rate; hence gov ernment's share stabilized again. The level of 1932-40 was roughly twice that of the pre-Depression era, which testifies both to the vitality of the New Deal spending programs and to the absence of complete economic 151 152 THE FREEMAN March recovery prior to the defense build up. The massive mobilization of the early forties produced by far the greatest government share ever at tained. At the peak, in 1943-44, government spending for currently produced goods and services com manded over 46 percent of GNP. Of this, defense-related purchases ac counted for nine-tenths. Rapid de mobilization dropped government's share during the period 1946-50 to the range of 11-15 percent, slightly below the prewar level.

The outbreak of the Korean War led to another steep increase of gov ernment's share, from about 13 per cent in 1950 to 22.5 percent in 1953. Although some retrenchment fol lowed cessation of the fighting in Korea, government's share has re mained at a high level, about one fifth of GNP, ever since. Notably, this postwar plateau of the past three decades holds government's share of GNP at approximately three times the level s~stained before World War I. By this measure, then, govern ment during the twentieth century has become three times as impor tant in relation to the economy. Looking at all government expen ditures, not just those for currently produced goods and services, rela tive to GNP reveals a different tem poral pattern. This broader mea sure, which includes the various governmental transfer payments, shows that government did occupy a somewhat higher level after World War I: 10-12 percent in the postwar period as opposed to 6-7 percent be fore the war. It shows also the famil iar leap in the early thirties and a plateau at about 18-21 percent in the late thirties; the huge increase during World War II followed by a complete reversal by 1948; a sharp rise between 1950 and 1953; and, unlike the index discussed above, an upward trend from the mid-1950s to the early 1980s that lifts the gov ernment's budget outlays from about 26 percent to 36 percent of GNP.

This upward tend'ency during the 1960s and 1970s occurred entirely because government transfer pay ments, mainly within the social se curity system, increased faster than GNP. By this measure, government has grown about five to six times larger relative to the economy dur ing the twentieth century. By in cluding government transfer pay ments as well as purchases of currently produced goods and ser vices, one obtains a measure better calculated to reveal the emergence of the modern welfare state. Not surprisingly, this emergence ap pears fairly sustained over the past seventy years, especially during the post-1956 era. Another commonly employed in dex of the size of government is its employment share. The temporal pattern revealed here differs from 1983 WHERE FIGURES FAIL 153 that shown by either of the expen diture indexes. Government em ployees increased slightly faster than the labor force even before World War I, reaching a share of almost 5 percent on the eve of the war. This share jumped to over 6 percent dur ing the war, fell back slightly in 1920-21, then drifted slowly up ward during the twenties, reaching 6.5 percent in 1930.

How MeasureEmployment? The history of government em ployment during the 1930s raises unusual, indeed unique, complica tions. To tell this tale, one must de cide what to do about the "emer gency workers." These people worked on programs administered by such "emergency" work-relief agencies as the Civilian Conservation Corps, the National Youth Administration, the Federal Emergency Relief Adminis tration (under which a state relief agency operated in each of the states), the Civil Works Administration, and the Works Progress Administration. At the time they were not consid ered "regular" government employ ees. Subsequently, economic statis ticians counted them as unemployed members of the labor force, a proce dure that has created confusion and controversy among economists in their description and analysis of the labor market during the Great Depression. If one follows the conventional practice, counting the emergency workers as unemployed, then the government's share of the civilian labor force appears to have re mained almost constant during the 1930s, falling slightly between 1931 and 1933 before rising slowly to 7.2 percent in 1939. This measure indi cates that government's share jumped much higher between 1939 and 1944, when it reached an un precedented 11.1 percent. Thus, World War II seems to have stimu lated a huge expansion of govern ment's direct importance in the ci vilian labor market.

If, however, one treats the emer gency workers as government em ployees-and it is hard to see why they should not be so treated-the course of history looks completely different. Now the jump in govern ment's employment share is seen to have occurred between 1930 and 1936, with especially large in creases in 1933 and 1934. At the peak in 1936, government workers of all kinds constituted 14 percent of the civilian labor force, more than twice their share in 1930. After 1938, as the ranks of the emergency workers thinned, government's share dimin ished; and by 1943, when only a handful of these peculiar workers remained, government employed only about 11 percent of all civilian workers, its share having dropped three full percentage points from the earlier peak. According to this more 154 THE FREEMAN March defensible index, government's en hanced role in the labor market grew out of the Great Depression. Not un til 1966, after two decades of steady postwar growth, did government's employment share exceed the level it had reached in 1936.

Immediately after World War II, government's employment share fell almost to 9 percent before starting a longterm march upward. Since the late 1960s it has stabilized in the neighborhood of 15 percent-only slightly above the peak attained in the 1930s. Notably, during this postwar period, the lion's share of the increased government employ ment has taken place at the state and local levels. Between 1947 and 1981, federal employment rose by less than a million workers, while state and local government employ ees added almost 10 million to their ranks. Comparing the most recent share to that at the turn of the cen tury, one finds that government's relative weight in the labor market has become almost four times greater. The Essence of Big Government We could continue to examine quantitative measures of the growth of government (e.g., tax revenue, funds borrowed, loans made) but lit tle would be gained by doing so.

Though each such index throws some light on the question at issue, each in a fundamental sense does not tell us what we really want to know. The basic difficulty arises because the quantitative measures of the size of government do not correspond closely-sometimes not at all-with the underlying essence of govern ment, which is coercive power. Government can grow bigger, much bigger in terms of its expen diture or employment share, and still not become Big Government. What distinguishes that capitalized levia than is the scope of its effective au thority over economic decision-mak ing. Under certain easily imagined circumstances, government could be very big yet rigidly limited. It might be necessary, for example, for gov ernment to spend and employ at high rates merely to deter external ag gression, maintain domestic order, and enforce private property rights. Such a government would severely confine the scope of its activities, yet its performance within its own lim ited sphere would require a large share of the economy's resources. In dividual citizens would remain free to decide for themselves all the basic economic questions about what and how to produce, determining in this ongoing market process the distri bution of income and wealth. Unde niably, government's command over resources by means of taxation and expenditure would diminish the ag gregate of options available to citi zens. But within the constraints set by their after-tax incomes, citizens 1983 WHERE FIGURES FAIL 155 would remain free to determine among themselves the allocation of the economy's resources. The heavy tax burden would indicate nothing more than the true cost of preserv ing an orderly and free society.

Differences Not Measured All quantitative indexes of the size of government share a common de fect: their changes may indicate ei ther changes in the scope of effective governmental authority or merely alterations in the level at which government operates within a con stant sphere of authority. On the one hand, for example, government may increase its expenditure and em ployment to extend its regulatory powers over previously unregulated dimensions of private economic de cision-making. On the other hand, it may increase its expenditure and employment to enlarge or improve the judicial system in order to pro vide quicker and more accurate en forcement of existing private prop erty rights. The two cases differ completely in their implications for the nature and workings of the po litical economy, yet the standard quantitative measures cannot dis tinguish them. Further, quantitative indexes may register little or no change even when the substance of governmental power changes enormously. For example, it takes the same resources to oper ate the Supreme Court no matter whether that high tribunal's deci sions give the owners of private property much or little protection from the intrusions of government and other citizens. Many regulatory agencies operate on tiny budgets, yet they exert far-reaching effects on the allocation of resources.

American government in the twentieth century has been loath to nationalize industry outright; regu lation of privately owned industry has been the preferred means of governmental control over resource use. In its most important economic effects, this continued toleration of nominal private ownership may matter little. But the accounts look different. When the owners of industrial properties spend billions of dollars at the behest of the Environmental Protection Agency or the Occupa tional Safety and Health Adminis tration, these expenditures are recorded as "private." The usual in terpretation, that private expendi ture signifies voluntary choice, is wholly unwarranted under such conditions. And the commonly cited quantitative indexes of the size of government completely fail to cap ture this critical feature of our mod ern Big Government. Governmental expenditure and employment are derivative manifes tations of the power of government.

Before it can spend or employ, gov ernment must obtain the authority 156 THE FREEMAN to promote a specified public pur pose. If such authority cannot be ob tained (a legislative question) and sustained (a judicial question), then the matter is settled: no authority, no program. If expanded authority is obtained and sustained, govern ment may exercise its newly ac quired power at various levels. Ex ercising this authority at higher levels of resource cost does not imply a wider scope of governmental authority. During the past three decades, for example, Social Security outlays for old-age pensions have mushroomed by billions upon billions of dollars. Yet the upward trend of these ex penditures does not signify any new accretions of Big Government dur ing that period. Government has possessed the authority to make these payments ever since the Social Securi ty Act was passed by the Con gress and upheld by the Supreme Court in the 1930s. The events of the thirties created new potential for governmental activity of this kind; subsequent events have determined only the degree to which that poten tial would be utilized. Like a limited government, Big Government may operate over a wide range of societal resource absorption.

In Sum Quantitative measures of the growth of government are certainly convenient; the newspapers report them daily, and everyday discus sions make constant reference to them. Used with caution, they can convey valuable information. Yet we must not forget that these indexes describe only derivative manifes tions and not the essence of Big Government. That essence is coercive power. Its extent cannot be accurately mea sured by indexes of government ex penditure, taxation, or employment. The coercive power of government has its source in the actions of leg islatures, regulatory agencies, and courts. These governmental bodies can, and often do, extend their sway without expending a single addi tional dollar or hiring a single addi tional employee. Of course, addi tional expenditure, taxation, and government employment frequently follow sooner or later. By the time they do, however, the citizen who opposes the growth of government discovers that the horse has long since escaped the barn: entrenched bureaucrats and program benefici aries make it virtually impossible to undo what has been done.

Limiting the scope of modern Big Government requires more than containing its expenditure, taxa tion, and employment. It is essential that the creation and extension of governmental authority be re strained at its sources. Here as else where, an ounce of prevention works more effectively than a pound of cure.