Cato Policy Analysis No. 165 November 18, 1991

Policy Analysis

The District Of Columbia:
America's Worst Government?

by William A. Niskanen

William A. Niskanen is chairman of the Cato Institute.

Executive Summary

Washington the Capital is a symbol of democracy and America. Washington the city is a symbol of almost everything that sincere and thoughtful men know is wrong with democracy and America. Washington the Capital is the hope of world freedom. Washington the city is overcrowded, badly housed, expensive, crime-ridden, intolerant, with inadequate transportation, schools, and health facilities. It staggers under a dilapidated and hopeless governmental organization, and its problems are rapidly getting worse.

--Alden Stevens

Although written in 1941, those words are disturbingly descriptive of the Washington of 1991. Washington the capital is still the marble mecca of democracy. And in some ways it has improved over the past half century. Washington the city is now more cosmopolitan, more tolerant, and air-conditioned. Commercial development has greatly increased the property tax base. The District of Columbia, which provides the services of both a state and a city government, now has substantial home rule. The government of the nation's capital, however, is probably the worst in the nation--maybe not as bad as the city government of Detroit but clearly worse than the government of any state. District expenditures per resident, excluding the cost of special federal services in Washington, are about twice the national average of state and local expenditures. Despite the high level of District spending, the performance of students in the public schools is lower than in any state, and the murder rate, infant mortality rate, and several other measures of social pathology are higher than in any state. A local satirical revue recently described the District Building (Washington's city hall) as "the nation's first work-free drug zone." Such are the record and the reputation of the District of Columbia that Mayor Dixon recently inherited.

Sharon Pratt Dixon brings to her new position a record of personal integrity, good management, and an overwhelming election victory. That will not be enough. One should remember that Mayor Barry first came to office with a reputation for high energy and political entrepreneurship. Honesty, good management, and politics-as-usual will not be sufficient to resolve the major problems of the District of Columbia. This analysis summarizes the major dimensions of the cost of District government, social conditions in Washington, and the radical changes that may be necessary to improve those conditions.

Comparative Government Spending and Social Conditions

The magnitude and characteristics of the problems of the District government are best illustrated by comparing government expenditures and various measures of social con- ditions in Washington with those elsewhere in the nation. Most of the comparative data are for 1987 or 1988, the lat- est years for which they are available, and most of the data are from the Statistical Abstract of the United States, a government publication that is available in almost every library. Any concerned resident could have made the same comparisons, but the public has paid very little attention to the lessons to be gleaned from these data.

General Economic and Demographic Conditions

General economic and demographic conditions in Washington, with one exception, compare favorably with those in the rest of the nation; the one exception is the unusually high poverty rate (Table 1).

Table 1
General Economic and Demographic Conditions, 1988
  Nation District Percentage Difference(a)
Personal income per resident $16,489 $21,389 +30
Average salary $21,871 $30,254 +38
Employment rate 62.3% 66.5% +7
Client groups (percentage of residents)
Poor (1979) 12.4 18.6 +50
Unemployed 2.7 2.8 +1
Age 17 and under 26.0 22.3 -14
Age 65 and over 12.4 12.4 0

Source: Statistical Abstract of the United States, 1990, pp. 23, 399, 408, 437, 458.

(a)Calculated from unrounded totals.

The problems of Washington are apparent. What may be less apparent is that most of the general economic and demographic conditions are better than the national average. Average personal income per resident is 30 percent higher than the national average, and higher than in all but two states. The average salary of those who work in Washington is nearly 40 percent higher than the national average, and higher than in any state. The percentage of Washington residents who are employed is also unusually high.

The most striking condition in Washington is the combination of a high average income and an unusually high poverty rate. The only states with higher poverty rates are those on the lower Mississippi, where average incomes are the lowest in the nation. The variance of (pretax and pretransfer) income in Washington is higher than in any state and is the primary explanation of the unusually high political demand for transfers and social services. The other client groups for government services, however, are either about the same or a lower percentage than elsewhere. The unemployed are a lower percentage of the labor force but about the same percentage of the resident population. The percentage of Washington residents age 17 or younger is lower than in any state. The percentage age 65 and over is the same as the national average. And Washington's high population density (not shown) should make it possible to economize on police and highway services.

In summary, Washington's general economic and demographic conditions, even the high poverty rate, are not a sufficient explanation of the combination of unusually high District expenditures and unusually poor social conditions.

District Expenditures, Employment, and Earnings

Expenditures. Expenditures by the District are much higher than are needed to provide good government services. District expenditures per resident are about twice the national average of state and local governments; they are higher than in any state except Alaska (Table 2). Moreover, direct expenditures by the District do not include the cost of those government services that some of us value most highly--specifically, the District's contribution to Metro and the federal expenditures to maintain and police the parks, the museums, and the zoo--costs that residents of other major cities bear directly.

Employment. Employment by the District government is much higher than average. Average earnings, however, may be insufficient to attract a quality workforce.

Government employment in the District is about 1 for each 11 residents, nearly 80 percent higher than the average state and local government employment rate, and higher than the rate in any state (Table 3). Again, the District total does not include the many federal workers who provide local government services in Washington. The District government is hugely overstaffed, by far more than the 2,000 positions that candidate Dixon pledged to cut or the 6,000 positions that the Rivlin Commission (a special commission of private citizens that reviewed the District budget) recommended be cut. A level of District government employment equal to the average state and local government employment per resident, even though few states are models of efficiency, would permit a reduction of 25,000 positions.

Table 2
State and Local Expenditures per Resident, 1987
  Nation ($) District ($) Percentage Difference
Direct general expenditures 2,685 5,163 +92
Capital outlay 404 886 +119
Total 3,089 6,049 +96

Source: Statistical Abstract of the United States, 1990, p. 281.

Table 3
State and Local Government Employment and Earnings, 1987
  Nation District Percentage Difference
Employees per 10,000 residents 496 888 +79
Average monthly earnings $2,090 $2,706 +29

Source: Statistical Abstract of the United States, 1990, p. 303.

Earnings. Average earnings of District employees are nearly 30 percent higher than the national average of state and local employees, higher than in any state except Alaska. District employee earnings, however, are not significantly out of line with other earnings in Washington. Nationwide, earnings of state and local government employees are about 15 percent higher than average earnings; the earnings of District government employees are about 7 percent higher than average earnings in Washington. Indeed, there may be a case for increasing the average earnings of District employees if necessary to recruit and maintain a higher quality workforce.

Public Education

District expenditures for primary and secondary schools are reasonable, but student performance is lower than in any state (Table 4). Expenditures for the University of the District of Columbia are extraordinarily high and have little apparent return.

Expenditures for education are the largest share of the District budget, as is the case in the rest of the nation. The education share of the current budget, however, is unusually low for two reasons: the relatively low public school enrollment in the District and the very high expenditures for other services.

Table 4

Public Education, 1987-90

(table omitted)

Source: Statistical Abstract of the United States, 1990, pp. 139, 144, 154, 159, 302.

Costs per student in the District's primary and secondary schools are nearly 30 percent higher than the national average. That may seem reasonable, given the high salaries necessary to recruit government employees in Washington. One is chastened, however, to recognize that the average cost per student in District schools is about 50 percent higher than the cost at Archbishop Carroll High School and about three times that at the Catholic elementary schools in the Washington area.

The more serious problem is the very low performance of District students; by several measures it is lower than in any state. Only 55 percent of District ninth graders in 1985, for example, graduated from District high schools in 1989--a percentage that is much lower than the national average. (That may understate the relative graduation rate in the District, however, because it does not include those students who graduate later from other school systems.) The average SAT score of District students hoping to enter college is 20 percent lower than the national average, and lower than in any state in which students take that exam. (The average score probably overstates the average performance of District students, because only the best 20 percent of District students take the exam.)

District support of public higher education is limited to the University of the District of Columbia, where public expenditures per student are more than twice the national average. That university should now be recognized as a failure. It has been plagued by scandal, incompetence, and pervasive mediocrity since its inception and now risks losing its accreditation--a bad investment for both the District and the students. The recent offer of a teaching position in criminal justice to the former mayor was only its latest offense.

Courts, Police, and Corrections

District expenditures per resident for courts, police, and corrections are more than four times the national average. Despite those very high expenditures, the reported crime rates in the District are higher than in any state and most major cities (Table 5).

Table 5
Courts, Police, and Corrections, 1987 and 1988
  Nation District Percentage Difference(a)
Costs and employment (1987)      
Share of current budget 7.9% 17.4% +121
Expenditure per resident $211 $900 +327
Employment per 10,000 residents      
Police 27.2 74.1 +172
Corrections 16.5 59.6 +261
Reported crimes per 100,000 residents (1988)      
All violent crimes 637 1,922 +202
Murder 8.4 59.5 +608
Property crimes 5,027 7,993 +59

Source: Statistical Abstract of the United States, 1990, pp. 171, 181.

(a)Calculated from unrounded totals.

Expenditures for courts, police, and corrections are the second largest component of the District budget, a share that is more than twice the national average. Expenditures per resident are more than four times the national average, much higher than in any state. Most of the higher expenditures reflect a much larger number of public safety employees. The numbers of police and corrections employees per resident are each several times the national average. And a roughly equal number of federal and Metro police, not included in the District data, also serve in Washington. The total number of police in Washington is nearly six times the national average per resident.

Crime rates in the nation's capital, however, are a national scandal; they are higher for each type of crime than in any state. The rate of violent crime is over three times the national average, even though it is lower than in 11 other major cities. The murder rate in 1988 was over seven times the national average, the highest in any major city, and it increased in 1989 and 1990. The rate of property crime is nearly 60 percent higher than the national average, although it is about average for the major cities. The combination of very high expenditures for public safety and unusually high crime rates is the single most perplexing problem facing the District government.

A Special Analysis of Police and Corrections

Many readers will correctly observe that it is not strictly appropriate to compare government expenditures and social conditions in Washington with the average for the states. Washington is a major city, and major cities face special conditions that may increase expenditures or cause unusual social problems. The available data, unfortunately, do not permit a direct comparison of Washington with other major cities. The costs of those District services usually financed from a state budget are not separately identified. Many city governments do not finance education and public welfare. And data on most economic, demographic, and social conditions are not available by city.

Fortunately, statistical analysis of the data available for each state and the District provides a way to sort out the issue. This paper summarizes that type of analysis for only police and corrections services. Such analyses are complex and never fully definitive, but, at least conceptually, they could also be applied to other state and local government services.

The primary questions of interest are the following: How much do District expenditures, police employment, crime rates, and the like (the data presented in Table 5) differ from those of the rest of the nation, given the conditions that explain the systematic variation of those "dependent" variables among the states? This analysis uses a sample of the 48 continental states plus the District. The full set of statistical results is described in the Appendix.

Most of the variation in expenditures per resident for police and corrections and in the number of police and corrections employees per 10,000 residents appears to be explained by five conditions: federal revenues per resident, personal income per resident, the violent crime rate, the property crime rate, and population density. In turn, most of the variation in crime rates appears to be explained by three conditions: the percentage of the population that resides in metropolitan areas, the percentage of the population that is poor, and population density. (One of the most disturbing results of this analysis is that increases in police services, within the observed range, do not appear to reduce crime rates.) Most of those conditions suggest that both District expenditures for police services and crime rates in Washington should be expected to be higher than the national average. The more disturbing conclusion is that each of those variables is higher than in other, "normal" large cities that have the same conditions (Table 6).

Table 6
Comparison of a "Normal" Large City and the District
  "Normal" City District Percentage Difference
Costs and employment (1987)      
Expenditures per resident $601 $898 +49
Employment per 10,000 residents      
Police 53.7 74.1 +38
Corrections 40.1 59.6 +49
Reported crimes per 100,000 residents (1988)      
All violent crimes 876 1,922 +119
Murder 12.5 59.5 +376
Property crimes 4,336 7,993 +84

The top line of Table 6 indicates that the District spends about 49 percent more per resident for police and corrections than would be expected in a jurisdiction with the same federal revenues, personal income per resident, crime rates, and population density. For comparison, the fourth line indicates that the rate of violent crime is about 119 percent higher than would be expected in another jurisdiction that is also 100 percent metropolitan and has the same poverty rate and population density. A comparison of Table 6 with Table 5 indicates that some of the differences between the District and national average conditions are explained by the economic and demographic conditions specific to the District. Even under those conditions, however, expenditures per resident and crime rates in the District are still substantially higher than the expected levels, and the difference is statistically significant in all cases. District officials should not be held responsible for the differences summarized in Table 5. Elimination of the differences summarized in Table 6, however, would be a reasonable goal.

Public Welfare

District expenditures per resident for public welfare are much higher than the national average, even though the average benefits per recipient are not unusual (Table 7). That is the result of a large recipient population and an unusually large welfare staff.

Expenditures for welfare are the third largest component of the District budget, a much higher share than the average of state and local budgets. Expenditures per resident are over 2.7 times the national average, and public welfare employment per resident is nearly twice the national average.

The finding that surprised me is that average benefits per recipient are not unusual. Average welfare benefits are slightly lower than the national average, despite the high cost of living in Washington. Average unemployment benefits are 30 percent higher than the national average, but that percentage difference is not quite as high as the percentage difference in average salaries.

The more perplexing issue is why the recipient populations are so large. The population of welfare recipients, for example, is 72 percent greater than the national average, compared with a poverty rate (in 1979) that was 50 percent above the average. And the population of the insured unemployed is 44 percent above the national average, although the number of unemployed per resident is about the same as the national average. Those differences may be due either to a different composition of the recipient populations or to more relaxed eligibility criteria. The data, by themselves, do not indicate that Washington is unusually generous to either welfare recipients or the unemployed.

Table 7
Public Welfare, 1987 and 1988
  Nation District Percentage Difference(a)
Costs and employment (1987)      
Share of current budget 12.3% 17.4% +42
Expenditure per resident $329 $899 +173
Employment per 10,000 residents 17.1 32.2 +89
Recipients and benefits (1988)      
Recipients per 100 residents 6.1 10.5 +72
Average monthly benefits $374 $347 -7
Insured unemployed per 100 residents 0.8 1.2 +44
Average weekly benefits $145 $188 +30

Source: Statistical Abstract of the United States, 1990, pp. 281, 302, 362, 367, 368, 399.

(a)Calculated from unrounded totals.

There is more reason to question the large size of the District's welfare staff.

Public Medical Care

District expenditures per resident for public medical care are more than twice the national average, but average health conditions are much worse (Table 8).

Expenditures for public health and hospitals are the fourth largest component of the District budget, a slightly higher share than the average for state and local budgets. Expenditures per resident are more than twice the national average, and the number of District medical care employees per resident is nearly 2.5 times the national average. There are many able professionals in the District medical care system, some of whom work beyond the call of duty, but anyone dependent on that system would testify that it was substantially overstaffed with low-quality support employees.

Table 8
Public Medical Care, 1987
  Nation District Percentage Difference
Costs and employment      
     Share of current budget 8.7% 9.4% +8
     Expenditure per resident $234 $486 +149
     Employment per 10,000 residents 53.6 133.7 +149
Vital statistics      
     Infant deaths per 1,000 births 10.1 19.3 +91
     Total deaths per 1,000 residents 8.7 11.9 +37
     Expected life at birth (1980) 73.9 69.2 -6

Source: Statistical Abstract of the United States, 1990, pp. 73, 76, 78, 281, 302

Across the nation, the average health status of the population generally increases with average income, the share of employment that is low risk, and the share of the population age 18 through 64. On that basis, the average health status of Washingtonians should be among the highest in the nation, whatever the level of public expenditures for medical care. Unfortunately, that is not the case. The number of infant deaths per 1,000 live births is nearly twice the national average, higher than in any state. Total deaths per 1,000 residents are nearly 40 percent higher than the national average, and higher than in any state, despite the small share of the population in the most vulnerable age groups. And, as of the 1980 census, the expected life of District residents was nearly five years shorter than the national average, and shorter than in any state. The combi- nation of very high public expenditures for medical care and unusually poor health is probably the second most challeng- ing problem facing the District government.

Highways and Other Services

Highway construction and maintenance are the one service for which District expenditures are relatively low (Table 9). District expenditures per resident for all other services not previously addressed, however, are unusually high.

Expenditure per resident for highways is slightly below the national average, but even that level of expenditure is surprisingly high. District road mileage per resident is one-ninth the national average; the weather is generally clement; and the District's record of repairing road damage is notoriously bad.

District expenditure per resident for the wide range of other government services not previously discussed is more than twice the national average. Most of the services that I use and value as a District resident are in that group, and the quality of those services seems adequate, but I have no explanation of why the expenditure for them is so high. The two activities that the District seems to pursue with any zeal are enforcement of parking regulations and assess- ment of residential real estate.

Table 9
Highways and Other Services, 1987
  Nation District Percentage Difference
     Share of current budget 8.0% 3.6% -55
     Expenditure per resident $214 $187 -13
All other services      
     Share of current budget 28.5% 34.5% +21
     Expenditure per resident $765 $1,786 +133

Source: Statistical Abstract of the United States, 1990, p. 281.

After so much data mongering, the mind wanders. . . . Journalists will have to complete the record with their sto- ries of local saints and sinners.

A Radical Agenda for the District

Mayor Dixon promised Washington a District government that is clean, competent, tough-minded, and caring. She deserves our best wishes. There is reason for hope, with the cooperation of the District Council, that her promise can be realized. And that would change the District's record of scandal and corruption. The mayor's approach, however, will not be sufficient to change the District's record of very high expenditures and very poor government services.

This section summarizes a "radical" agenda for the District--suggestions for changes in the way the District is organized and operates that offer some hope of alleviating the most serious problems in Washington. These changes may not be "realistic," in that they stem from a view of the problems that differs fundamentally from the conventional perspective. Given the magnitude of the District's problems, however, these changes deserve serious consideration. In the end, what proves to be realistic will depend on perceptions of what works.

Self-Deception and Shibboleths

One of the major barriers to acknowledging and addressing the problems of Washington is the pattern of self-deception and shibboleths that characterizes local political discussion. The necessary first step toward substantial reform is to recognize the self-deception and to reject the shibboleths.

The major self-deception, as is often the case, is about money--the illusion that a substantial increase in District revenues either is possible or would resolve the major problems. The favorite illusion is that someone else will pay the bills, either through an increased federal grant or a commuter tax. Federal revenues are now nearly 40 percent of total District revenues, excluding the direct federal cost of several local services in Washington. Neither Congress nor the Bush administration has expressed much willingness to make a further increase in the federal grant. Nonresidents already pay a substantial part of District revenues as commuter and tourist customers and nonresident owners of District corporations and property. The District does not have the authority to levy a commuter tax, and Congress has expressed no willingness to authorize such a tax.

District officials should also recognize the limits on local tax revenues. Despite the large share of District revenues that comes from the federal government, local taxes consume a larger share of personal income than they do in all but two states. More important, District individual, corporate, and sales tax rates are substantially higher than in Maryland and Virginia. The population of Washington has declined 20 percent in the past 20 years, and higher local tax rates would increase the loss of those residents who contribute most to Washington and thus might not increase long-term tax revenues.

The perception that higher revenues would alleviate the major problems of Washington is also an illusion. The District spends nearly twice the national per resident average of state and local governments without providing a single high-quality government service. In the absence of major changes in the organization and operation of the District, there is no reason or evidence that more money would make much difference.

The major local shibboleths involve the political status of the District, specifically the commitment to full home rule and statehood. Those objectives have no relevance to any major problem in Washington. The largest employer and property owner in any jurisdiction has substantial political influence. The federal government generally interferes too much with state and local governments, but there is no case that the relative federal influence on the District is too great. As long as federal revenues are nearly 40 percent of total District revenues, the federal government will and should exercise substantial influence on the activities of the District.

The case for District statehood is even less relevant; statehood would provide several opportunities for District politicians to advance their careers but would provide no significant benefits for District residents. There is a good general case that every citizen should have representation in Congress; that could be achieved easily by reincorporating the District into Maryland. The fact that statehood advocates do not endorse that option suggests that they have some other more parochial objective.

General Organization, Employment, and Revenue Reforms

Many District officials have aspirations that the District become a state. They should take that idea seriously. The major organizational change that should be considered is to restructure the District as a state, rather than as a unitary state-city. There are very few economies of scale in government services--certainly not in primary and secondary education, most police services, street maintenance, and most other local services. Individual communities in Washington could be granted the authority to form separate cities, to select the local services that each will provide, and to set their own property tax rates. The District could retain responsibility for higher education, courts and corrections, public health and welfare, highways, and whatever local services the new cities do not provide and would retain all taxes other than the property tax. The District also could provide grants to the poorer cities to increase the revenues available for local services. The structure of cities in the District and the division of services between each city and the District should not be implemented by a master plan but should be permitted to evolve on the basis of some common rules. The primary benefits of that type of structure are that it permits increased diversity in the level, character, and quality of local services and induces greater efficiency. As the capital of the world's major federal nation, the District should take the idea of federalism seriously.

In many cases, Washington would be better served by a District that was a financer of public services, rather than both the financer and the provider. There is a good case for government financing of a limited range of services; there is no good case for the same government's organizing and supplying those services. Many services now provided by the District could be better supplied under contract with private organizations or other governments or through vouchers for services to specific individuals. Moreover, the authority to contract for services, in whole or in part, would increase the incentive for District agencies to improve their performance.

The District is now hugely overstaffed--a condition that reflects, in part, a perspective that District employment is an instrument of political favoritism, job creation, and welfare policy. A rejection of that perspective is requisite to important changes in employment practices. The important changes would be to increase the salary and qualifications of higher level managers, to substantially reduce the number of middle managers, and to tighten the quality standards for District employees. Another valuable change would be to eliminate all residency requirements for non- political positions, greatly increasing the pool from which District employees are selected. For those few workers subject to emergency recall, there is a basis for setting a time limit from residence to work station, but that need not preclude residence in a nearby suburb. The District will never provide high-quality services until District employment is considered a contract to supply specific services, rather than a sinecure.

The District should consider two changes in its revenue base but not try to increase total revenues. Over time, the base for the property tax could be gradually shifted from buildings to land by attributing future increases in assessments to the general price of land in the site area rather than to the value of the building on a specific site. That would increase the efficiency of land use and eliminate the current disincentive to improve buildings. If the property tax base were shifted to new cities in the District, of course, those cities should have this option.

The second change would be to collect congestion fees on all vehicles entering or leaving the District during peak hours. Simple technology is now available to do that with out tollbooths. Expected congestion could be reduced to any desired level and the use of Metro increased. The revenues from those fees should be divided equally between the District and the jurisdiction in which the vehicle is registered, thereby reducing the opposition of adjacent governments to the fees.

The best use the District could make of additional revenues from the recommended changes would be to reduce those tax rates that are higher than in adjacent jurisdictions. The two proposed revenue measures would substantially increase the incentives to live, shop, and invest in Washington and would reduce the dependence of the District on federal employment and revenues.


Washington has a relatively small population of families who care about education but cannot afford a private school. That is both an effect and a cause of a poor public school system. The parents and students who choose not to participate in the District system are those on whom better schools are most dependent. That situation has led to a progressive decline in the performance of the District schools with no prospect for recovery.

Only a radical change in the organization of local schools has any potential for improving their record. There are ample reasons to maintain a high level of District spending per student; both average incomes and average salaries in Washington are unusually high. The District, however, should phase out of the business of organizing and managing the local schools.

The most promising reform has two essential characteristics: parental choice and school-based management. Full parental choice would permit every Washington family to send their children to any public or private school in the area, subject only to a District-specified core curriculum and a nondiscrimination standard. An amount of money equal to a share of the average District spending per student at each grade level, maybe in the second prior year, would be transferred to any private school selected.

Full school-based management would permit each school to set its own tuition, teacher credentials and salaries, and all other standards affecting the curriculum, pupil- teacher ratios, textbooks, and the like. The reform should probably be phased in, one year at a time beginning with kindergarten, both to permit a gradual adjustment of the public schools and the expansion of other schools and to avoid a windfall to those families who now send their children to private schools. Those who are most concerned about education should recognize that this reform would increase local political support for school financing, both because current resident families would have greater options and because more families who value education would choose to live in Washington. Over time, this reform would improve education, change the composition of Washington's population, and probably reduce crime. The District must now make a choice between managing a poor school system and financing better education.

The University of the District of Columbia should be recognized as a failure and be subject to the public equivalent of a Chapter 11 bankruptcy and reorganization. There are a dozen good colleges and universities in the Washington area and no apparent reason to maintain an unusually expensive and low-quality District university. A scholarship to any resident high school graduate in an amount less than the average District spending per student at UDC would be more than sufficient to pay tuition at the other universities. One wonders whose interests are served by maintaining the pretense of higher education at UDC, surely not those of the students or the Washington community.

Courts, Police, and Corrections

The District and the federal government spend an extraordinary amount for courts, police, and corrections in Washington without providing much public safety. The rate of violent crime in Washington is several times the national average and many times that in other nations. The sad fact is that America's capital is one of the most dangerous cities in the world. The high crime rate may be due primarily to cultural attitudes and social conditions, but that is not a very helpful observation in the short run. The immediate challenge is to make much more effective use of the large number of police already employed in Washington.

Many readers of these comments will recognize that the magnitude of this problem overwhelms my understanding of the feasible measures that might be most effective. For that reason, I offer the following suggestion as a concerned and moderately informed resident of Washington, not as an expert on police and crime.

The first step toward improved police effectiveness is to change the priorities for the use of police. The District, like most governments, has passed more laws than can be feasibly enforced; for that reason, all police forces must set priorities among the many laws they are charged to enforce. The highest priority use of police should be to protect people against crimes against which it is difficult to protect or insure oneself. The highest priority should be to reduce the extraordinarily high rate of violent crimes, even at the cost of less enforcement of laws against property crimes or other crimes, such as drug use and prostitution, that are offensive but not directly threatening.

The second step is to change police tactics to reduce crime rates, even at the cost of lower arrest rates. For Washington, that requires a higher visible presence of police in areas of high violent crime and the substitution of civilian personnel for most police in desk jobs. That tactic has been most effective in Charleston, South Carolina; was recently adopted in New York City; and was strongly endorsed by the police experts on the Rivlin Commission.

The District also spends an extraordinary amount on corrections. The major opportunities for economy in this area appear to include speeding up court procedures, reducing the use of incarceration for nonviolent crimes, and contracting with nearby governments or private prisons.

Welfare, Medical Care, and Other Services

The District's unusually high spending for public welfare and medical care, along with many of Washington's social problems, is primarily due to one condition: around 60 percent of recent births in Washington are to unmarried women, nearly 2.5 times the national average. Moreover, most of those births are to adult women; only 17 percent are to teenage mothers, some of whom are married. Nevertheless, births to teenagers are about 35 percent above the national average. Unmarried women and their children are now the dominant family type of a substantial part of the Washington population. And the government, in effect, has become the marriage partner of last resort. Many of Washington's depressing social problems--from high infant mortality to low school performance to high welfare caseloads and a high crime rate--derive primarily from that condition. It is unclear, however, whether that condition would be responsive to changes in social policy.

District welfare payments per recipient and the standards for welfare eligibility are not unusual; Washington is not a welfare magnet. The problem is that unusual conditions demand unusual policy responses, and no such responses have been tried. The most difficult policy challenge is to discipline irresponsible behavior by adults without penalizing innocent children. None of the alternatives is easy, but none should be precluded for that reason. The federal welfare legislation of 1988 already requires that an unmarried teenage mother (with special exceptions) live with her mother to be eligible for welfare, but teenage mothers are a small part of the District's welfare population.

One option is much more rigorous child support enforcement, recognizing that fathers may be difficult to locate or may not have much apparent income. A second option is to eliminate the increase in welfare payments per mother with any future increase in the number of dependent children, an option that is less desirable because it would penalize dependent children if it were not effective in limiting their number. A District government that is not willing to consider such options must accept responsibility for a large underclass that is costly to others and dangerous primarily to its own members. The high cost and poor performance of other District services such as medical care, housing, and social services derive from the same cause and are not further addressed.


The government of the District of Columbia may or may not be America's worst. That issue is provocative but not important. The sad fact is that the District government does not bring credit to the nation's capital, does not use its huge resources effectively, and continues to face a distressing array of social problems. Mayor Dixon may be effective in reducing the corruption and the more conspicuous waste that mushroomed during the administration of her predecessor, and she deserves our support. The major problems of the District will remain unresolved, however, unless the new mayor and District Council acknowledge them and are willing to consider radical changes that offer some promise of a government that is worthy of the nation's capital.

Statistical Appendix

This appendix presents the results from six linear regressions that are the basis for the special analysis of police and corrections summarized earlier.

Definition of Variables

PCE = expenditures per resident for police and corrections (1987).

POL = police per 10,000 residents (1987).

COR = corrections employees per 10,000 residents (1987).

VCR = violent crimes per 100,000 residents (1988).

MDR = murders per 100,000 residents (1988).

PCR = property crimes per 100,000 residents (1988).

CON = constant.

DCD = 1 for D.C., 0 for all states.

FRR = federal revenues per resident (1987).

PIR = personal income per resident (1987).

MET = percentage of residents in metropolitan areas (1987).

POV = percentage of residents below poverty line (1979).

LDN = population per square mile (1988), natural logarithm.

All data are from the Statistical Abstract of the United States, 1990.

Notes on Interpreting the Regressions

For each regression, the top number in each row presents the effect of a one-unit increase of the specific independent variable on the dependent variable. For example, a $1 increase in federal revenues per resident appears to increase state and local spending per resident for police and corrections by about 11 cents. Similarly, a 1 percentage point increase in the share of a state's population that lives in metropolitan areas appears to increase the violent crime rate by about 11 crimes per 100,000 residents. The numbers in parentheses are the standard errors of each coefficient. R2 is the share of the variance of each dependent variable that is explained by the combination of the independent variables. S.E.R. is the standard error of the regression, measured in the units of the dependent variable.

The interpretation of the expenditure and employment regression is rather straightforward. The demand for police services is a positive function of federal revenues, personal income, and crime rates and a negative function of population density. The crime rate regressions are more disturbing. The level of crime appears to be primarily dependent on the poverty rate, the degree of metropolitanization, and population density and does not appear to decline in response to a larger number of police. Indeed, adding the POL variable (police per 10,000 residents) to the crime rate regressions yields a positive coefficient at high levels of statistical significance. Although other studies have found a similar effect, I cannot accept the conclusion that additional police increase crime. For that reason, I have reported only those crime rate regressions that do not include the police expenditure and employment variables.

The results of the regressions are interesting and suggestive but not definitive. Much more analysis would be necessary for a thorough understanding of these relations. For this paper, the primary purpose of the analysis was to estimate the DCD coefficients that measure how much police expenditures, police and corrections employment, and crime rates differ in the District, given the conditions that explain those variables across the nation. For that limited purpose, this limited analysis should be sufficient.

1991 The Cato Institute
Please send comments to webmaster