Understanding Differences in Regional Poverty Rates by Elizabeth T

November 15,1994 eCONOMIC GOMMeNTORY Federal Reserve Bank of Cleveland

Understanding Differences in Regional Poverty Rates by Elizabeth T. Powers and Max Dupuy

Ahe U.S. poverty rate is among the Poverty is the product of a multitude of most widely used indicators of the factors, including market conditions, The official U.S. poverty rate, often nation's economic success—and of how demographic characteristics, and fiscal cited as a measure of our nation's that success is shared. However, the policy. Its accurate measurement is com- economic strength, hides a huge varia- national rate masks tremendous variation plicated by interregional differences in tion across regions. Here, the authors in the regional numbers. In 1992, for the cost of living and the quality of life, examine the reasons for this disparity, example, the official poverty rate ranged among other factors. focusing on differences in demographic, from a low of 8.3 percent in New Eng- economic, policy, and cost-of-living land to a high of almost 16 percent in the A common feature of all these determi- factors across the nine U.S. census East South Central states. nants is their great variation across divisions. Results show that although regions. The burdens of recession are all of these elements have some influ- If we accept the poverty rate as a legiti- generally distributed unevenly over the ence on the poverty gap, economic mate yardstick of national achievement, country, as illustrated by the disparate factors are key. then policymakers have several reasons experiences of the West and Midwest to be concerned about this disparity. during the last downturn. Demographic First, interregional equity is a legitimate types that are overrepresented among the In this Economic Commentary, we policy concern in and of itself, analogous poor, such as female-headed households, employ a simple method (see box) to to concerns about how income is distrib- are more prominent in some areas than attribute poverty differences across the uted among individuals. In fact, inter- in others. State fiscal policies, which are nine standard census regions to demo- regional poverty differentials may be presumably uncoordinated, are an obvi- graphic, economic, policy, and cost-of- another dimension along which the well- ous source of interregional poverty vari- living factors.1 Our results show that being of the nation is judged—a premise ation. Federal fiscal policy also plays a economic factors are the primary source that casts doubt on the desirability of role through both intended and unin- of the disparity, with weak economies having means-tested cash transfers (com- tended interregional transfers. Finally, accounting for a large share of the above- monly termed "welfare") determined by the official poverty level is not adjusted average poverty rates seen across the the states. Second, even if equalization for regional cost-of-living differences, United States. Surprisingly, government of poverty across regions is not a policy which may contribute to large and per- fiscal policy, although poverty-reducing goal, understanding the marked and persistent disparities in measured rates. on an absolute basis in every region, is sistent differences in the regional statistics Recent legislative interest has made it less effective in areas with weak econ- can shed light on the factors contributing particularly important to understand how omies. Our findings also show that to poverty. And finally, it is important to such adjustments might change percep- demographic patterns play an important determine whether interregional dispari- tions of the poverty picture, since the role. Interestingly, a crude cost-of-living ties truly reflect meaningful differences flow of federal dollars to the states would correction would result in a very different in well-being. be affected by changes in the measured poverty picture, implying a major change distribution of poverty. in the flow of funds from the federal government to state coffers. DECOMPOSING REGIONAL POVERTY DIFFERENCES account for a significant portion of the Our method for decomposing differences in regional poverty rates can best be illus- total value of transfers, particularly for 4 trated by working through a comparison of the East North Central (ENC) region and low-income families. Consequently, we the United States. First, families are sorted into three classifications: female-headed count as income the value of some nor- 5 households with children, two-parent families with children, and other family units. mally excluded in-kind benefits. Data from the March Current Population Survey are used to compute the fraction of each type of household in poverty in both the ENC states and the United States as a This more complete measure does not whole. To compare the difference in poverty rates attributable to demographic charac- change the basic picture of interregional teristics peculiar to the ENC, we compute the total poverty rate using U.S. family- poverty. Table 1 presents the official specific rates, but apply the ENC proportions of the three types. Comparing this rate family poverty rate for the nine census with the actual U.S. poverty rate reveals whether poverty in the ENC is above or below regions in column 1 and our measure in average due to the special demographic characteristics of the region. Dividing this column 2. Official rates range from a difference by the U.S. poverty rate yields the entry of 3.25 percent for the ENC low of 8.3 percent of families in New demographic factor in table 2. Policy effects are derived similarly. England to more than 15 percent in the South Central divisions. As has histori- cally been the case, the South has the TABLE 1 FAMILY POVERTY RATES AND highest poverty rates by far. The North- RANKINGS, 1992 east has the lowest overall rate, thanks to Official After Taxes, Transfers, Adjusted for the influence of New England. Poverty Region Rate and Some In-Kind Benefits Cost of Living rates in the West have been rising Northeast steadily and now surpass those of the New England 8.30 (9) 7.73(9) 9.16(8) Midwest. The difference between the Mid-Atlantic 10.80 (5) 10.29 (5) 11.70(3) rates in columns 1 and 2 is due entirely Midwest to tax and in-kind transfer policy. Note East North Central 10.27 (7) 9.68 (7) 9.58(6) that all of the adjusted rates are lower, West North Central 10.12(8) 9.09 (8) 7.92 (9) indicating that the effect of in-kind transfers dominates, but that the poverty South ranking of the divisions remains South Atlantic 12.18(3) 11.36(3) 10.08 (5) unchanged. Unless otherwise stated, the East South Central 15.81(1) 14.87(1) 12.56(2) term "poverty rate" will refer to this tax- West South Central 15.36(2) 14.20 (2) 11.43(4) and transfer-adjusted measure. West Mountain 10.36 (6) 9.85 (6) 9.31 (7) • Decomposing Poverty Pacific 11.50(4) 11.02(4) 12.85(1) Differences: An Overview United States 11.68 10.97 10.72 The first step toward understanding NOTE: Regional rankings are in parentheses. regional poverty differentials is to SOURCE: Authors' calculations based on data from the March 1993 Current Population Survey. decompose them into their demographic, policy, and economic components. We compare regional poverty to the national • Measuring Poverty rate is determined by computing the rate for income reported in 1992, the 6 In the late 1950s, the federal government fraction of all families whose pre-tax latest year for which data are available. developed poverty thresholds based on cash income falls below their family- One advantage of using 1992 data is that cash income. At that time, data revealed appropriate levels.2 it was a fairly typical year—neither a that the average American family devoted peak nor trough of the business cycle (at about one-third of its budget to food. Since the development of the original least not nationally), and free from any Thus, a subsistence food budget was poverty statistic, there have been two major fiscal policy changes that would multiplied by three to yield a subsistence important policy changes that might dramatically affect the level or distribu- income, or poverty line. The poverty line affect the measurement of poverty. First, tion of poverty rates. Table 2 compares varies by family size, composition, and the tax burden on low-income families each division's family poverty rate to the ageof householder. Larger families are has increased since the late 1950s, when national average and, implicitly, to each assigned higher poverty thresholds, while it was about zero.3 Thus, we calculate our other, based on regional characteristics. older families face lower ones. With poverty measure based on after-tax and (The divisions are listed from least to minor modifications, these guidelines cash transfer income. Second, "in-kind" greatest poverty for convenience.) The have been in use since 1961 and are transfer programs (primarily Medicare, first column is the percent deviation of updated yearly using the Commerce Medicaid, food stamps, and housing the region's rate from the nation's. A Department's Consumer Price Index for subsidies) that did not exist in 1961 now negative number indicates that the all items. The official family poverty regional rate lies below the national TABLE 2 PERCENT DEVIATION OF POVERTY RATE FROM NATIONAL AVERAGE BY FACTORS, 1992a Demographic Government Economic Region Total Conditions Policy Conditions'1 New England -29.50 -6.50 -9.76 -13.25 West North Central -17.14 -4.47 -5.86 -6.81 East North Central -11.75 3.25 -6.41 -8.59 Mountain » -10.21 -1.85 1.14 -9.50 Mid-Atlantic -6.20 -1.69 0.22 -4.74 Pacific 0.44 0.51 1.45 -1.52 South Atlantic 3.60 0.24 1.24 2.12 West South Central 29.44 0.27 6.89 22.28' East South Central 35.61 4.55 3.72 27.34 a. National average = 10.97. b. Residual after demographic factors and government taxes and transfers are taken into account. SOURCE: Authors' calculations based on data from the March 1993 Current Population Survey.

average, while a positive number indi- To get an overview of the primary Finally, the West and East South Central cates the opposite. For example, New causes of regional poverty differences, poverty rates are 129 percent and 136 England has a family poverty rate of 7.7 it is convenient to group the nine census percent of the national rate, respectively. percent, about 30 percent below the divisions into those with below-average, Demographics can account for only a national average of 10.97. average, and above-average poverty. We modest portion of the disparity in the designate the four regions with poverty East South Central division, which has These divisional deviations from the rates less than 90 percent of the national the highest concentration of single-parent U.S. average can be attributed to several average as the below-average group— families in the nation. Government pol- factors. The demographic factor conveys New England and the West North Cen- icy also has somewhat below-average the role of the relative regional distribu- tral, East North Central, and Mountain effectiveness at reducing poverty in both tions of family types in poverty. For states. Demographic compositions divisions. The most significant influence example, the West North Central divi- favorable to lower poverty are fairly on the huge poverty gap between the sion's poverty rate is 4.5 percent below important in New England and the West South Central states and the rest of the the national average due to the region's North Central area. Unfavorable demo- country appears to be a less robust econ- more favorable demographic conditions graphics are overcome by other factors omy: More than 75 percent of the area's (primarily an above-average concentra- in the East North Central division, and deviation from the national rate is tion of two-parent families), accounting have little effect in the Mountain states. accounted for by economic conditions. for about one-quarter of the total difference. The "government policy" factor What all four divisions do have in • A Closer Look at measures the net impact of all federal, common is above-average economic Government Policy state, and local taxes and transfers on performance. Perhaps surprisingly, the While we have seen that government poverty. Government policy appears to contribution of government policy to policy plays an important role in areas of do a better-than-averagejob of poverty below-average poverty in New England extreme poverty, its overall impact can relief in New England, for example, and the North Central states is nearly as mask crucial differences between state reducing its rate nearly 10 percent below large as that of relative economic success. and federal policy. The detailed compo- the national level and accounting for a The salutary effects of government policy nents of the "government policy" cate- full one-third of the difference between are particularly important in the East gory from table 2 are presented in table the region's rate and the U.S. average. North Central division: If policy had its 3. Both federal and state policies exert Finally, "economic conditions" captures typical impact, poverty rates there would large and often offsetting influences on both transitory (business-cycle-related) be little better than the national average. poverty rates.7 However, because federal and long-run regional differences in eco- and state taxes do not have great differ- nomic status. The effects are not surpris- The poverty status of the Mid-Atlantic, ential impacts on poverty over regions, ing: Relatively strong economies have Pacific, and South Atlantic divisions we do not describe their effects in detail. below-average poverty, while less-than- follows the national average quite closely. average economic performance boosts While these areas appear to be similar to States have primary responsibility for the poverty rate in the two poorest the nation in demographic, policy, and cash welfare payments, so state transfer southern divisions from 22 to nearly 30 economic factors, we shall see that the policy is an obvious source of inter- percent above the national average. Pacific region is actually much different regional poverty differences.8 Consistent from the national norm along several with the known facts about welfare, dimensions. TABLE 3 PERCENT DEVIATION OF POVERTY RATE FROM NATIONAL AVERAGE BY PUBLIC POLICY FACTORS, 1992a Total State Total Total Federal Federal Federal and Local State and State and Region Government Government Transfers Taxes Government Local Transfers Local Taxes New England -9.76 0.90 -1.05 1.95 -10.66 -11.11 0.45 West North Central -5.86 -9.45 -8.59 -0.86 3.59 3.90 -0.31 East North Central -6.41 -6.59 -6.75 0.17 0.18 -0.38 0.56 Mountain 1.14 -5.64 -5.10 -0.54 6.78 7.78 -1.00 Mid-Atlantic 0.22 1.90 1.34 0.56 -1.68 -2.14 0.46 Pacific 1.45 10.80 10.26 0.54 -934 -10.24 0,90 South Atlantic 1.24 -2.01 -1.41 -0.60 3.25 3.45 -0.20 West South Central 6.89 237 2.42 -0.04 4.52 5.52 -1.00 East South Central 3.72 1.93 2.44 -0.52 1.79 2.08 -0.29 a. National averages 10.97. SOURCE: Authors' calculations based on data from the March 1993 Current Population Survey.

table 3 illustrates that the poorest states federal welfare program—food stamps • Adjusting for also have relatively less effective transfer —helps to equalize the generosity of Cost-of-Living Differences programs. Part of this can be explained state welfare programs, thus reducing Poverty guidelines have long been criti- by the wider gap between pre-transfer interregional inequality, the vast major- cized for ignoring regional differences income and the poverty line in these ity of federal transfers are not welfare in the cost of living. If, for a given regions.9 That is, poorer regions have but rather non-means-tested cash benefits amount of income, a family can attain relatively more individuals whose pre- (primarily Social Security) that remove more or better food and clothing, a larger transfer income is well below the poverty many elderly from the nation's poverty apartment, etc., in Iowa than California, line, whereas relatively better-off regions rolls. Since Social Security benefits are then perhaps having a single poverty line have more people whose pre-transfer based on lifetime work experience and for both states makes little sense. Based income is near the line. Hence, even if average wages, higher benefits flow both on recent cost-of-living estimates, the all nine divisions offered identical wel- to areas with solid labor market histories California poverty line would presum- fare benefits, the poverty-reducing (as table 3 makes apparent for the two ably have to be set as much as 20 percent effect of state transfer programs would midwestem divisions, where poverty higher to enable the state's lowest- be smaller in the South. This effect is is reduced more than 5 percent below income families to consume the same reinforced by state policy, since southern the national average by federal transfer quantity and quality of items as does an welfare benefits are known to be well policy) and to retirement havens, particu- officially poor family in Iowa. below the national norm. Evidence from larly in the South Atlantic and Mountain table 3 also indicates major differences regions. The Earned Income Tax Credit Recently, policymakers have expressed in state transfer policy, as can be seen has become a significant transfer program interest in implementing state-specific from the wide variation in the effective- to the poor and near-poor, but because it poverty lines. Because federal aid for ness of state transfers for areas with targets workers, it does not have a neutral the Head Start program, community similar poverty rates (for«xample, the poverty-relieving effect across regions development block grants, home energy Pacific and South Atlantic divisions). either. assistance, and remedial education for poor children is distributed to states Table 3 also shows that the effectiveness The Pacific region is a case where large according to their poverty rates, the of federal transfer policy varies a great policy effects are masked at the aggre- potentially large changes to relative deal across census regions. Federal gate level. Federal transfers have the rates from a cost-of-living adjustment transfer payments to individuals are weakest poverty-reducing effect there, are of more than just academic interest. made for the most part without regard to but this is almost completely offset by residence. The question, then, is what state and local transfers. However, the To explore the poverty implications of accounts for this tremendous variation? outcome of these opposing forces is not regional price indexing, we use an index a neutral government policy. The Pacific of state prices to adjust the official pov- 10 As before, there is the measurement states have a well-below-average poverty erty line in each state. The resulting issue based on the poverty "gap." How- rate for single-parent families (due to rates are presented in the third column of ever, it is also clear from comparing relatively more generous state transfer table 1. Because the poverty line is low- regions with similar poverty rates that a policies), but the plight of the elderly ered in states with below-average costs wide disparity exists in the effectiveness and others counting on federal dollars of federal transfers. There are several appears to be relatively worse. reasons for this. Although the primary TABLE 4 PERCENT DEVIATION OF POVERTY RATE FROM NATIONAL AVERAGE BY FACTORS, ADJUSTED FOR THE RELATIVE COST OF LIVING, 1992a Demographic Government Adjusted for Economic Region Total Conditions Policy Cost of Living Conditions1* West North Central -26.20 -4.61 -5.99 -10.02 -5.57 New England -14.58 -6.70 -9.99 20.26 -18.16 Mountain -13.23 -1.91 1.17 -3.15 -9.33 East North Central -10.71 3.35 -6.56 1.08 -8.58 South Atlantic -6.02 0.25 1.27 -9.37 1.83 West South Central 6.60 0.27 7.05 -18.08 17.35 Mid-Atlantic 9.10 -1.74 0.23 16.59 -5.98 East South Central 17.09 4.69 3.80 -14.10 22.70 Pacific 19.82 0.52 1.49 19.38 -1.57 a. National averages 10.97. b. Residual after demographic factors and government taxes and transfers are taken into account. SOURCE: Authors' calculations based on data from the March 1993 Current Population Survey.

and raised in states with above-average This different picture of poverty should and are probably not an appropriate costs, poverty rates decline in the Mid- be interpreted cautiously, however. Liv- basis for policy. west and South and rise dramatically ing costs also reflect the presence of for the high-cost coastal areas. These amenities that people value, such as • Conclusion changes are large enough to alter the warm weather and proximity to recrea- Cost-of-living adjustments aside, dis- poverty status of some divisions signifi- tional resources. As is true for any other parities in economic conditions across cantly. The Pacific region, which con- desirable goods, high demand bids up the states are the most important determi- tains the high-cost states of California, price of access to these items. Although nant of regional poverty variation in the Alaska, and Hawaii, zooms from a posi- it is difficult to put a price tag on ameni- United States. Some of these economic tion of average poverty to the poorest of ties, they make some contribution to differences reflect transitory economic all divisions. By contrast, the low-cost individuals' well-being—including the shocks whose effect on income might West South Central division, formerly low-income population—implying that appropriately be smoothed by govern- among the poorest areas, boasts a pov- it is probably inappropriate to adjust the ment policy. More persistent differences erty rate little different from the national poverty lines of any two regions by the in economic performance—such as the average after cost adjustment. The New full, conventional cost-of-living differ- long-standing gap between the South and England and East South Central divi- ential. The true interregional distribution other areas—are well-known puzzles. sions are so firmly ensconced at the of poverty may be intermediate between Whether these permanent differences poles of poverty that they remain near the extremes of columns two and three should engender permanent redistribu- the extremes even after huge cost-of- in table 1. tory schemes depends on one's beliefs living adjustments. about why these gaps exist and whether It is tempting to conclude that an appro- alternative policies would provide better Table 4 recreates the calculations from priate state-by-state cost-of-living incentives. table 2, but includes the cost of living as adjustment would result in a superior an additional factor. Because poorer measure of poverty. However, this is not Another significant finding is that cur- areas tend to have lower costs and richer necessarily the case, since the cost of rent government transfer policies tend areas higher costs, overall interregional living within a state often varies by more to have below-average effectiveness in inequality is reduced. With the excep- than that between states. This is particu- poorer regions, enhancing measured dif- tion of the Mountain and West North larly true when comparing rural and ferences in regional poverty rates. Social Central states, areas with below-average urban areas. For example, rural family Security, the most effective poverty- costs also demonstrate below-average poverty in a high-cost state might be reduction program by far, obviously is economic performance. Thus, extremely overstated, while urban poverty in an not geared toward interregional income high poverty rates no longer appear to overall low-cost state might be under- smoothing, but rather is linked to recipi- be a "southern problem," because stated. These biases could be as great or ents' past labor market performance. It although economic forces continue to greater than those that arise from apply- seems unlikely that much sentiment drag down the South Central regions, ing the same poverty line cutoff to the could be aroused among policymakers cost-of-living adjustments substantially entire country. The point is not that these or the public for reducing the unequal mitigate these forces. adjustments could not be made, but that regional impact of federal transfers on crudely deflated poverty measures, such measured poverty through changes in as those presented here and elsewhere, should be viewed as highly preliminary non-welfare programs. One alternative and Pacific (Washington, Oregon, California, brought up to the poverty line, which reduces would be to federalize the welfare sys- Alaska, Hawaii). the effectiveness of transfers in areas with low pre-transfer poverty. tem, since the poverty-reducing effec- 2. For example, in 1992, the poverty thresh- tiveness of state transfer policy varies old for a two-parent nonelderly family with 10. None of the available price data are ideal greatly from region to region. This two children was $14,228. for such an adjustment. We use an index con- structed by the American Federation of would not be a complete solution, how- 3. After peaking in the mid-1980s, the tax Teachers from data collected by the American ever, because most of the differences effect on poverty has declined, however. seen in regional poverty rates are not Chamber of Commerce Researchers Associa- 4. These programs deliver benefits in the attributable to state welfare programs. tion. Since costs are assumed to be those of a form of vouchers for goods and services middle-manager's family, the index is some- rather than cash. what inaccurate for poor families. • Footnotes 5. We exclude Medicaid and Medicare, since 1. The census divisions are as follows: New the determination of their equivalent values England (Maine, New Hampshire, Vermont, in cash remains controversial. Massachusetts, Rhode Island, Connecticut), Elizabeth T. Powers is an economist and Max Mid-Atlantic (New York, New Jersey, Penn- 6. This article was written before revised Dupuy is a research assistant at the Federal sylvania), East North Central (Ohio, Indiana, population weights for the March 1993 Cur- Reserve Bank of Cleveland. Illinois, Michigan, Wisconsin), West North rent Population Survey became available. The views staled herein are those of the authors and not necessarily those of the Fed- Central (Minnesota, Iowa, Missouri, North 7. Local transfers and taxes are included in eral Reserve Bank of Cleveland or of the Dakota, South Dakota, Nebraska, Kansas), the state categories. South Atlantic (Delaware, Maryland, Wash- Board of Governors of the Federal Resen'e 8. Unemployment insurance and workers' ington, D.C., Virginia, West Virginia, North System. compensation payments are also included in Carolina, South Carolina, Georgia, Florida), state transfers, but because they have a mini- East South Central (Kentucky, Tennessee, mal impact on poverty rates, we do not dis- Alabama, Mississippi), West South Central cuss them separately. (Arkansas, Louisiana, Oklahoma, Texas), Mountain (Montana, Idaho, Wyoming, Col- 9. This bias is somewhat mitigated, how- orado, New Mexico, Arizona, Utah, Nevada), ever, since areas of low poverty have relatively fewer people who can potentially be

Federal Reserve Bank of Cleveland BULK RATE Research Department U.S. Postage Paid P.O. Box 6387 Cleveland, OH Cleveland, OH 44101 Permit No. 385

Address Correction Requested: Please send corrected mailing label to the above address.

Material may be reprinted provided that the source is credited. Please send copies of reprinted materials to the editor.