Final Paper Social Issue

profileprincesingh1
california_poverty_measure.pdf

The California PoverTy Measure

1

C h r i s t o p h e r W i m e r , m a r y b e t h m at t i n g ly, m at t l e v i n ,

C a r o l i n e D a n i e l s o n , a n D s a r a h b o h n with research support from Tina Tran, Lucas Manfield,

Shannon McConville, and Bonnie Bui

In collaboration with the Public Policy Institute of California

T his research brief presents initial results from the newly- released California Poverty Measure (CPM). The CPM, which is jointly produced by the Public Policy Institute

of California (PPIC) and the Stanford Center on Poverty and Inequality, is our best estimate of economic disadvantage across and within California. It improves on the official poverty measure (OPM) and the Supplemental Poverty Measure (SPM) in ways that will be discussed in some detail below. The CPM can be used to provide county-level estimates of poverty, to explore how cur- rent policy is affecting poverty rates, and to examine the potential impact of certain proposed changes in policy. Because California will be facing key decisions in the future about how to address poverty, we need to be able to assess how proposed changes in its safety net will affect Californians. The CPM is a partial but important step in that direction.

We address five questions in this brief: (1) How much poverty is there in California and how do estimates of poverty vary across the main competing measures of poverty? (2) Does poverty vary much across California counties? (3) How do patterns of poverty vary by demographic characteristics? (4) By how much do social safety net programs reduce poverty rates? And (5) Which demo- graphic groups benefit the most from safety net programs?

This brief provides only some of the key results coming out of the CPM and focuses particularly on demographic and county- level variability in poverty. For more information about the CPM, including a detailed discussion of the impacts of the safety net, the depth of poverty, and other key findings, see our companion publication and technical appendices (available at www.ppic.org/ main/publication.asp?i=1070).1

A Portrait of Poverty within California Counties and Demographic Groups

The Stanford Center on Poverty and Inequality

Key findings

• The CPM indicates that 22.0% of Californians were living in poverty in 2011. This is 1.5 per- centage points lower than the 2009-2011 sup- plemental Poverty Measure (sPM) for California (23.5%), but 5.8 percentage points higher than the official poverty measure (OPM) for California (16.2%). The differences arise principally be- cause safety net benefits are not fully reported in the Census sPM and because the high cost of living in California is not taken into account in the OPM.

• There is wide variation in poverty rates across California, with especially high rates observed in counties with high housing costs, such as Los Angeles County (26.9%) and Orange County (24.3%). By contrast, lower rates tend to be ob- served when housing costs are more moderate, as in Placer County (13.8%) and sacramento County (17.0%).

• immigrant poverty, at nearly 30%, is remarkably high, and over 11 points higher under the CPM than under the OPM.

• The CPM child poverty estimate is 25.1%. Whereas child poverty rates in the CPM slightly exceed those in the OPM, the national sPM rates for children are lower than the correspond- ing OPM estimates. This suggests that Califor- nia’s high cost of living increases child poverty more than our full accounting of safety net ben- efits decreases it.

• The three largest safety net programs targeted toward families with children jointly reduced child poverty rates in California by 12.0 percentage points. This translates into 1.1 million fewer children in poverty.

• safety net programs also substantially reduce the poverty rate for the least educated. The pov- erty rate for those in families with a high school degree or less would have been more than 20 percentage points higher absent the safety net.

2 California povert y me asure

A Portrait of Poverty within California Counties and Demographic Groups

Methodology The CPM is a tool developed by the Stanford Center on Poverty and Inequality and the Public Policy Institute of California to better understand economic disadvan- tage within and across California. It follows in the spirit of the research Supplemental Poverty Measure (SPM) now released each year by the U.S. Census Bureau, with some adjustments to account for underreporting of safety net program benefits and for various factors that are unique to California, such as its large unauthor- ized immigrant population. Both the CPM and the SPM build upon the official poverty measure (OPM) in three important ways:

1) In determining poverty thresholds, a wider range of consumer expenditures is included, and housing costs are adjusted geographically;

2) Non-cash and post-tax transfers, including the Supplemental Nutrition Assistance Program (called “CalFresh” in California) and the Earned Income Tax Credit (EITC), are counted as income; and

3) Non-discretionary expenses, such as medical out- of-pocket costs, child care costs, and work related expenses (including transportation), are subtracted from income before determining an individual’s poverty status.

These key improvements are summarized in Table 1. The CPM follows the SPM by incorporating these improvements on the OPM, but it implements some of them more rigorously by taking advantage of additional data available for California.

The Census Bureau produces SPM estimates of poverty for California by averaging three years of Current Popu- lation Survey data. The SPM is a marked improvement over the official poverty measure developed fifty years ago.2 It suggests that, at 23.5%, California has the most poverty in the nation, a result that attracted widespread attention when it was released. However, the SPM does not account for the underreporting of safety net benefits, that are available to low-income Californians. Because of this underreporting, the SPM may overstate the amount of poverty in California, a possibility that is one of the rationales for constructing the CPM.

We developed the CPM from a variety of data sources, including the Current Population Survey (CPS), the American Community Survey (ACS), and California administrative data.3 We adjust for important factors that may bias estimates, including the substantial underreporting of safety net benefits and the SSI cash- out, which is unique to California.4 Because the CPM is based primarily on the ACS, with its large sample size, it becomes possible to estimate poverty rates for California counties and for relatively small demographic groups.

table 1: Comparison of Poverty Measures

OPM SPM/CPM

Income All pre-tax cash income and transfers

Includes all cash and in kind transfers; based on after-tax income

Expenses N/A Subtracts medical, child care and work-related expenses (including transportation) from income

Threshold Economy Food Plan*3, updated annually for inflation

Based on consumer expenditures on food, clothing, shelter and utilities; includes a small adjustment factor for other necessities

Adjustments Family size and composition Broader definition of family that includes unmarried partners, foster children, and unrelated children under 15; family size and composition; geographic adjustment for housing costs

the stanforD Center on povert y anD inequalit y 3

CPM 99% Confidence Interval

(w/ replicate weights) OPM

Difference (OPM-CPM)

Lower bound Upper bound

California 22.0% 21.6% 22.5% 16.2% 5.7%

Alameda 18.4 16.4 20.4 12.4 6.0

Alpine/Amador/Calaveras/Inyo/Mariposa/Mono/Tuolumne 16.1 11.1 21.1 13.6 2.5

Butte 19.9 15.2 24.5 20.2 -0.4

Colusa/Glenn/Tehama/Trinity 15.7 10.0 21.3 18.5 -3.1

Contra Costa 18.6 15.7 21.5 12.5 6.1

Del Norte/Lassen/Modoc/Siskiyou 20.4 14.1 26.8 22.3 -2.0

El Dorado 13.6 8.9 18.2 10.8 2.8

Fresno 20.2 17.5 22.9 25.3 -5.1

Humboldt 17.3 11.7 22.8 19.0 -2.2

Imperial 22.1 16.4 27.9 26.0 -3.9

Kern 19.2 16.4 21.9 24.4 -5.2

Kings 14.5 8.3 20.6 19.5 -5.5

Lake/Mendocino 19.4 13.6 25.2 21.5 -2.3

Los Angeles 26.9 26.2 27.6 18.2 8.7

Madera 20.5 12.9 27.9 22.6 -2.2

Marin 19.0 14.5 23.5 9.3 9.6

Merced 22.2 17.0 27.4 29.4 -7.5

Monterey/San Benito 24.6 20.2 28.9 15.8 8.6

Napa 25.5 19.1 32.0 12.4 12.9

Nevada/Plumas/Sierra 14.6 9.7 19.6 12.0 2.6

Orange 24.3 22.9 25.6 12.8 11.4

Placer 13.8 10.5 17.1 8.2 5.6

Riverside 20.4 18.4 22.4 15.9 4.3

Sacramento 17.0 14.7 19.4 17.5 -0.5

San Bernardino 19.5 17.4 21.6 18.4 0.9

San Diego 22.7 21.3 24.0 14.9 7.7

San Francisco 23.4 20.2 26.6 12.8 10.6

San Joaquin 18.1 15.3 21.0 17.6 0.4

San Luis Obispo 22.0 17.6 26.3 14.3 7.6

San Mateo 18.4 15.5 21.3 6.7 11.6

Santa Barbara 21.9 18.8 25.1 13.1 8.7

Santa Clara 18.7 16.8 20.5 10.2 8.4

Santa Cruz 22.1 18.0 26.3 12.9 9.1

Shasta 19.0 13.9 24.2 19.7 -0.9

Solano 16.1 12.1 20.1 13.6 2.3

Sonoma 17.3 14.3 20.2 11.7 5.5

Stanislaus 23.1 19.2 27.0 23.3 -0.3

Sutter/Yuba 13.7 9.1 18.2 15.3 -1.7

Tulare 20.2 16.8 23.6 27.2 -7.3

Ventura 21.2 18.3 24.0 11.6 9.4

Yolo 23.6 18.2 28.9 19.9 3.5

table 2: Comparison of OPM and CPM by California County and OPM-CPM Difference

OPM CPM

all People Children Under 6 all Children Working-age adults elderly

35

30

25

20

15

10

5

0

P e

r C

e n

t

16.2%

23.1% 24.9%

14.6%

9.6%

22.0%

25.1% 26.3%

21.4%

18.9%

4 California povert y me asure

A Portrait of Poverty within California Counties and Demographic Groups

County Variation Table 2 reports OPM and CPM poverty rates for each of California’s 58 counties as of 2011.5 An interactive map that compares county rates across the state is available at www.ppic.org/main/mapdetail.asp?i=1396. Because of concerns about sample size, the smallest counties are combined in the ACS public-use data, but even with such combinations our small area estimates still have large margins of error. This should be borne in mind when interpreting the results.

The county-level results reveal that, in many of Califor- nia’s expensive urban centers, there are more people living in poverty than the official poverty measure implies. The poverty rate for San Francisco, for example, nearly doubles, while for Los Angeles it rises by roughly 50%, from 18.2% to 26.9%. California’s three most populous counties, Los Angeles, San Diego, and Orange coun- ties, have some of the highest poverty rates in the state. Although some rural counties also have very high pov- erty rates, the CPM makes it clear that urban poverty in California is more severe—arguably dramatically so— than had been appreciated.

In other counties, we find moderately lower poverty

rates under the CPM than under the OPM. The poverty rate falls from 27.2% (OPM) to 20.2% (CPM) in Tulare County and from 24.4% (OPM) to 19.2% (CPM) in Kern County (both located in California’s Central Valley).

For large urban counties, living costs are higher, and this drives poverty up under the CPM. For counties in which the poverty rate falls, more moderate costs of living and the inclusion of more income and benefits combine to largely account for differences with the OPM.

Demographic Variation In this section, we examine statewide California estimates of poverty, but now take into account demo- graphic variation in those estimates.

As Figure 1 reveals, the statewide CPM estimate is 22.0%, which is substantially higher than the statewide OPM estimate of 16.2%. It is higher mainly because the CPM, unlike the OPM, takes the high cost of living in California into account. Although the CPM estimate is higher than the OPM estimate, it is slightly lower than the 2009-2011 SPM estimate (not shown). The latter dif- ference reflects, in part, the underreporting of safety net benefits in the SPM methodology. The CPM estimate

figUre 1: Comparison of Poverty Measures by age group

White nH black nH Hispanic asian Other native-born immigrant

35

30

25

20

15

10

5

0

P e

r C

e n

t

OPM CPM

9.6%

24.0% 23.4%

11.7%

16.1% 15.4%

18.4%

13.8%

20.1%

32.2%

19.0% 18.5% 19.1%

29.9%

the stanforD Center on povert y anD inequalit y 5

EITC and SNAP, tend to go disproportionately to low- income families with children. It is striking that, even so, child poverty rises slightly under the CPM, whereas it falls in the national SPM estimates.7

Why doesn’t the inclusion of safety net benefits lower the poverty rate for children? The state’s high cost of living, in combination with other major expenses like medical, work, and child care expenses, outweighs the inclusion of more income and benefits under the CPM, resulting in slightly higher estimated child poverty rates.

The increase in the number of poor is yet more dramatic among adults. As shown in Figure 1 and discussed in more detail in the companion publication, the poverty rate for working-aged adults in California is nearly 7 per- centage points higher under the CPM than under the OPM. This translates to an extra 1.6 million adults who count as poor.

The CPM methodology also produces poverty rates almost twice as high among older adults. This differ- ences arises in large part because the CPM, unlike the OPM, subtracts the elderly’s often substantial medical expenses from their income. Similar results are found in

thus implies that the OPM underestimate is far more severe than the SPM overestimate. Because the CPM adjusts at once for the high cost of living in California and for actual safety net use, it provides our best esti- mate to date of poverty in the state.

What does the CPM say about poverty among children, working-age adults, and the elderly? First and foremost, we see that children in California are very often in pov- erty: Figure 1 shows that 25.1% of all children are in poverty and 26.3% percent of all children under age 6 are in poverty. The high poverty rates for young children are of great concern given strong links between young child poverty and health, education, and earnings in later life.6

The differences between the CPM and OPM estimates of child poverty are also revealing. As Figure 1 shows, both the CPM and OPM imply that poverty is highest among children, especially young children. The differ- ence between these measures is relatively small for this age group: The CPM is 2 percentage points higher than the OPM for all children, and only 1.4 percentage points higher than the OPM for young children. The safety net programs included in the CPM measure, such as the

figUre 2: Comparison of Poverty Measures by race/ethnicity and immigration Status

OPM CPM

less than High School

High School

Some College

College or More

60

50

30

10

0

40

20

P e

r C

e n

t

44.7%

25.6%

15.8%

53.9%

33.2%

21.7%

5.5%

9.8%

6 California povert y me asure

A Portrait of Poverty within California Counties and Demographic Groups

national SPM estimates because both the CPM and the SPM subtract medical expenses from income.8

Does the CPM change our understanding of racial dif- ferences in poverty? The black poverty rate is almost 4 percentage points lower under the CPM than the OPM, while the poverty rate for all other racial and ethnic groups is higher under the CPM than the OPM (Figure 2). There are potentially several factors behind this difference, including differential safety net receipt and differences in county of residence, family size and composition. The Hispanic and Asian poverty rates are especially high under the CPM. The striking difference between the OPM and CPM estimates among Hispanics and Asians results from the high cost of living in Cali- fornia, the income-reducing effect of non-discretionary expenses (like medical and work-related expenses), and the exclusion of a relatively large number of undoc- umented immigrants from safety net programs.

The right side of Figure 2 above compares native-born and immigrant poverty. Whereas the native-born pov- erty rate rises by only 3.7 percentage points under the CPM, the immigrant poverty rate rises by over 11 per-

figUre 3: Comparison of Poverty Measures by education

centage points.9 Here again the results are likely driven by the large proportion of the immigrant population that is ineligible for safety net benefits. Of course, these racial and ethnic differences may also result from other factors, including differences in family composition or in county of residence (and associated housing costs).

We next consider gender differences in poverty. Pov- erty rates are higher for women under both the OPM and CPM, although the gap between men and women is somewhat smaller under the CPM (1.2 percentage points) than the OPM (2.1 percentage points). (Data available upon request.) These differences, which are relatively minor, are likely due in part to the participation of single mothers in safety net programs, such as the EITC and CalFresh.

We conclude this section by showing that poverty reaches even into households that are relatively well educated. As Figure 3 shows, 21.7% of those in house- hold units with some college are poor, which is 5.9 percentage points higher than what we find under the OPM calculation.10 While CPM rates are higher than OPM rates for all education groups, the greatest abso- lute increases in poverty are found in the less educated categories. Under the CPM, 53.9% of people in families headed by someone without a high school diploma are in poverty, an increase of 9.2 percentage points relative to the OPM estimate.11 While this is a relatively small group (only about 10 percent of Californians are in such families), CPM poverty rates are also quite high for those in families headed by someone with a high school diploma. We find that approximately one-third of people in such families are in poverty.

The Social Safety Net We next consider whether the safety net is widely pro- tecting Californians from poverty.12 Figure 4 shows how counting safety net benefits yields lower estimated poverty rates under the CPM. We focus on four age groups: young children (under 6), all children, working- age adults, and the elderly. We first show the impact of SNAP, refundable tax credits (EITC and the Child Tax Credit), and CalWORKs (California’s welfare-to-work program), and we then show the combined effect of all

all Children Under 6 all Children Working-age adults elderly

CPM Minus SnaP, tax Credits, & CalWOrKs CPM Minus total Safety net benefits

0

-5

-15

-25

-30

-10

-20

P e

r C

e n

t -14.7% -12.9%

-15.0%

-9.0%

-29.1%

-12.4%

-6.3%

-12.0%

-4.9% -1.2%

the stanforD Center on povert y anD inequalit y 7

figUre 4: the Poverty reducing effect of the Social Safety net in California by age group

major federal and state social safety net programs avail- able in our dataset.13 We focus on CalFresh, tax credits, and CalWORKs because these are arguably the three largest antipoverty programs in California. In our sec- ond comparison, we aim to show what poverty would look like in California absent all safety net benefits that we have considered.

We highlight the findings for children, especially young children, given the importance of safety net programs for them. Families with children are the largest ben- eficiaries of CalFresh, CalWORKs, and refundable tax credits aimed at supporting working families. As shown in Figure 4, an additional 12.0 percent of all California’s children and 12.4 percent of young children would be poor (under the CPM) absent the benefits provided by these social safety net programs. These programs mat- ter less, though still a considerable amount, for poverty rates among adults. Without counting resources from these three programs, poverty rates would be 4.9 per- centage points higher among working age adults.

SNAP, refundable tax credits, and CalWORKS, by

comparison, have little impact on elderly poverty rates (decreasing the rate by 1.2 percentage points), which makes sense given that these programs are much larger for families with children. However, when we remove all the safety net benefits available in our data, the elderly poverty rate soars by over 29 percentage points. This is, of course, largely due to the impact of Social Secu- rity, which is well known to have successfully reduced elderly poverty, especially since its expansion in the early 1970s.14

In Table 3, we consider each safety net program sep- arately, again calculating the poverty rates absent specific programs. As shown here, Social Security reduces poverty rates the most, decreasing the state’s poverty rate by 5.2 percentage points, primarily through reductions in elderly poverty.15 Among children, refund- able tax credits had the largest impact on poverty rates, reducing the poverty rate by 6 percentage points. In interpreting these results, note that the overall reduction in poverty is not the sum of the effects of each safety net program, as many families benefited from more than one program.

8 California povert y me asure

A Portrait of Poverty within California Counties and Demographic Groups

The bottom panel of Table 3 shows how out-of-pocket medical expenses, child care, and taxes affect poverty rates. Absent medical out-of-pocket expenses, poverty rates in California would be over 4 percentage points lower. Absent work and child care expenses, poverty rates in California would be 2.3 percentage points lower. The latter results imply that, were we to adopt programs or policies that lowered these expenses, Cali- fornia poverty rates could be substantially reduced. For more details on these and other effects of California’s safety net programs, please consult the companion publication and the Technical Appendices (www.ppic. org/main/publication.asp?i=1070).

We conclude by examining how the impact of the

safety net varies by gender, race or ethnicity, nativ- ity, and education. In Figure 5, we show that women are slightly more likely to benefit from the inclusion of safety net benefits than are men, as are non-Hispanic blacks relative to other racial and ethnic groups. This figure also shows, again not surprisingly, that persons living in families with less-educated heads benefit more from safety net programs than other Californians. It is striking just how high poverty rates would be for per- sons in these low-education families if our measure did not include safety net benefits. In families with no high school graduate, poverty rates would exceed 70% without the social safety net, a result that reveals the highly compromised position of the less educated and the importance of the safety net in protecting them.

All Californians Children Children Under 6

CPM 22.0% 25.1% 26.3%

Subtractions From Income

CPM Minus Out of Pocket Medical Expenses 17.8 21.1 22.5

CPM Minus Work and Child Care Expenses 19.7 21.9 22.8

CPM Minus Medical, Work, and Child Care Expenses 16.0 18.7 19.8

CPM Minus Federal, State and Payroll Taxes Paid 19.7 22.3 23.5

All Californians Children Children Under 6

CPM 22.0% 25.1% 26.3%

Additions to Income

CPM Minus CalFresh 24.2 29.2 30.6

CPM Minus EITC/CTC 25.3 31.1 32.6

CPM Minus CalWORKs/General Assistance 23.3 27.6 29.0

CPM Minus School Meals 22.6 26.3 27.0

CPM Minus Housing Subsidies 23.4 27.0 27.9

CPM Minus SSI 23.4 26.1 27.0

CPM Minus Social Security 27.2 26.7 27.5

table 3a: California Poverty Without Safety net Programs

table 3b: California Poverty Without Medical and Work-related expenses

0

-5

-15

-25

-10

-20

P e

r C

e n

t

Female Male White Black Hispanic Asian Other Native-born Less

than HS HS

Diploma Some

College College or MoreImmigrant

-13.7%

-12.0%

-10.6%

-23.9%

-15.2%

-8.7%

-11.4%

-13.2% -11.9%

-21.8% -20.6%

-15.3%

-5.1%

the stanforD Center on povert y anD inequalit y 9

figUre 5: the Poverty reducing effect of the Social Safety net in California by gender, race-ethnicity, nativity and educational attainment

figUre 6: the impact of SnaP on PovertyWhile data limitations prevent us from precisely estimating

county-level effects of most programs, it is possible to carry

out a county-level analysis for one of California’s biggest

safety net programs, Calfresh. We use data from the Program

Access index (PAi), calculated by the California food Policy

Advocates, to approximate the percentage of eligible Calfresh

participants who actually participate in the program in their

county (see http://cfpa.net/pai-2013 for details).16

We divided our 41 county or county groups into those with

low, moderate, and high PAi scores. A high-PAi county is one

in which a greater percentage of eligible county residents

participate in the Calfresh program. We then calculated

what the CPM would have been absent our estimated

Calfresh dollars. The results are presented in figure 6. for

low-PAi counties, where Calfresh is reaching relatively few

eligible Californians, the impact of Calfresh is fairly modest,

reducing poverty by 1.2 percentage points. But this grows

to 2.1 percentage points in moderate-PAi counties and to

3.5 percentage points in high-PAi counties. This suggests

that efforts to boost enrollment in Calfresh among eligible

disadvantaged populations could lead to substantial

reductions in poverty (as calculated by the CPM).

An Illustrative County-Level Analysis: The Case of CalFresh

0

-.5

-1.5

-2.5

-3

-3.5

-4

-1

-2

P e

r C

e n

t

low Pai Moderate

Pai High Pai

-1.2%

-2.1%

-3.5%

10 California povert y me asure

A Portrait of Poverty within California Counties and Demographic Groups

Discussion The analyses presented in this brief offer new insights into the landscape of poverty in California. When the Census Bureau released its 2011 report on the SPM, it was widely noted that California had the highest pov- erty rate in the nation.17 But the SPM may overestimate poverty because of CPS underreporting of key safety net programs. We have developed the CPM to address this and other concerns and thereby contribute to Cali- fornia’s poverty monitoring system.

After correcting for underreporting (by applying adminis- trative data), we do indeed find somewhat lower poverty rates than the Census SPM. This result highlights the importance of using administrative data to address underreporting. However, even after administrative data are incorporated and underreporting is mitigated, we still find that over one in five Californians are in poverty, a shockingly high estimate by any calculus.

We also find great variability across counties in the CPM poverty rate. Perhaps not surprisingly, many urban counties had higher poverty rates than had been appre- ciated, a result largely driven by high housing costs in those counties. These new county-level estimates, once built into a regular monitoring system, should prove useful in making local policy decisions.18

Over a quarter of all children in California are in pov- erty. In the country as a whole, child poverty rates are

lower under the SPM than the OPM, as the inclusion of more safety net benefits outweighs the rest of the changes made under the SPM. Although the safety net also delivers many children from poverty in California, the state’s high cost of living results in child poverty rates that remain persistently high. This suggests that renewed attention to child poverty is warranted.

Our analyses likewise reveal that poverty among immigrants is particularly high. The striking differ- ence between the OPM and CPM estimates among immigrants likely stems from the high cost of living in California, the income-reducing effect of non-dis- cretionary expenses (like medical and work-related expenses), and ineligibility for safety net programs among many immigrants. If we want to make headway in reducing poverty in California, the results coming out of the CPM suggest that the immigrant population war- rants special attention.

Finally, we have presented an initial set of analyses doc- umenting how safety net programs reduce poverty rates among Californians, an analysis that the CPM method- ology makes possible. We hope that this tool will allow for prospective policy changes to be considered in light of their implications for poverty. There is vast potential for using the CPM to measure the effects of policies on state and county poverty rates. These include not just the policies we currently have, but also the policies that we might adopt in the future. ■

the stanforD Center on povert y anD inequalit y 11

Endnotes 1. It should be borne in mind that some of our estimates, especially those from small- er counties, are based on small samples. The estimates presented here are in some cases preliminary and are subject to revi- sion.

2. Because our OPM estimate is con- structed using the ACS sample, it differs slightly from Census tabulations (see Technical Appendices (URL)). For SPM and OPM estimates, see Short, Kathleen. “The Research Supplemental Poverty Measure: 2011.” U.S. Census Bureau, Current Popu- lation Reports: P60-244, November 2012. Available at: http://www.census.gov/hhes/ povmeas/methodology/supplemental/re- search/Short_ResearchSPM2011.pdf.

3. The CPS and ACS data used in this brief come both directly from the Census Bureau and from the University of Minnesota’s In- tegrated Public Use Microdata Series: CPS : Miriam King, Steven Ruggles, J. Trent Al- exander, Sarah Flood, Katie Genadek, Mat- thew B. Schroeder, Brandon Trampe, and Rebecca Vick. Integrated Public Use Mi- crodata Series, Current Population Survey: Version 3.0. [Machine-readable database]. Minneapolis: University of Minnesota, 2010. ACS: *Steven Ruggles, J. Trent Alexander, Katie Genadek, Ronald Goeken, Mat- thew B. Schroeder, and Matthew Sobek. Integrated Public Use Microdata Series: Version 5.0 [Machine-readable database]. Minneapolis: University of Minnesota, 2010.

4. Under the cash-out program, people who receive cash assistance from the Sup- plemental Security Income/State Supple- mentary Payment (SSI/SSP) are not eligible for food stamps but instead automatically receive a small cash allowance ($10) each month.

5. We also provide confidence intervals to assist in determining the precision of our estimates. Because the CPM rests on a wide range of imputations that complicate the estimation of confidence intervals, we opted to present the wider 99% confidence intervals instead of the less conservative 95% confidence intervals.

6. See, for example, Jeanne Brooks-Gunn, Greg J. Duncan, and Nancy Maritato, “Poor Families, Poor Outcomes: The Well Being of Children and Youth,” Chapter 1 in Con- sequences of Growing Up Poor, edited by Greg J. Duncan and Jeanne Brooks-Gunn (New York: Russell Sage Foundation, 1997) and “Children at Risk: Consequences for School Readiness and Beyond,” Rand Labor and Population Research Brief. Re- trieved from http://www.rand.org/content/ dam/rand/pubs/research_briefs/2005/ RAND_RB9144.pdf (2005).

7. See http://www.census.gov/hhes/ povmeas/methodology/supplemental/ research/Short_ResearchSPM2011.pdf. Given the potential imprecision in our estimates, stemming mainly from the im- putation of key elements of the CPM, even these small increases may not be signifi- cantly different from zero. Using sample replicates, however, and a conservative 99 percent confidence interval, our analyses suggest that these increases are statisti- cally significant (but such calculations do not factor in error from our imputations).

8. Although not shown in Figure 1, our CPM estimates are slightly lower than Cen- sus SPM estimates for 2011 (see http:// www.census.gov/hhes/povmeas/method- ology/supplemental/research/Short_Re- searchSPM2011.pdf). As noted above, Census typically aggregates three years of data to produce state level estimates, but it is possible to compute one year estimates (which will be reliable for large states like California).

9. We use a definition of immigrant here that includes all foreign-born persons, whether or not they are naturalized citizens or citizens by virtue of their parentage.

10. This calculation reflects the highest level of education obtained by the most educated person in the poverty unit.

11. The baseline OPM rate among the col- lege educated is low enough that even a small absolute change yields a large rela- tive increase in poverty.

12. Because of rounding, there are some- times small differences between the effects reported in the text and the overall rates reported in Table 2.

13. The safety net programs considered here include need-based and other pro- grams: CalFresh, EITC/CTC, CalWORKs, Free and Reduced Price School Breakfast and Lunch programs, General Assistance, Social Security, SSI, and Housing Subsi- dies. This excludes one major program, Unemployment Insurance, because this type of income is not well measured in the ACS (although it is supposed to be cap- tured in the ACS “other income” question). It also excludes some smaller programs like WIC and LIHEAP.

14. See http://www.ssa.gov/policy/docs/ ssb/v66n1/v66n1p1.html and Abell, John D. And Melissa L. Abell. 2004. “Poverty Reduction: Government Transfer Spending vs. Macroeconomic Change.” Journal of Poverty 8(2): 89-109.

15. Since Social Security is a cash transfer, it is included in the OPM, meaning it cannot drive any observed differences between the CPM and OPM.

16. See Shimada, Tia. “Program Access Index 2011: Measuring CalFresh Utilization by County.” Sacramento, CA: California Food Policy Advocates.

17. Short, Kathleen. “ The Research Sup- plemental Poverty Measure: 2011.” U.S. Census Bureau, Current Population Re- ports: P60-244, November 2012. Available at: http://www.census.gov/hhes/povmeas/ methodology/supplemental/research/ Short_ResearchSPM2011.pdf; See also: http://www.huffingtonpost.com/tag/supple- mental-poverty-measure and http://blogs. sacbee.com/capitolalertlatest/2012/11/ californias-poverty-rate-highest-in-us-by- new-federal-measure.html.

18. We also provide confidence intervals to assist in determining the precision of our estimates.

The Stanford Center on Poverty and Inequality monitors and publicizes trends in poverty and inequality, publishes the country’s leading magazine on poverty and inequality, supports research on the causes of poverty and inequality, and examines the effects of public policy on poverty and inequality.

The CPI is generously supported by the Elfenworks Foundation and Stanford University. Partial funding for this research came from a grant to the PPIC from The Walter S. Johnson Foundation and Grant Number AE00101 to the CPI from the U.S. Department of Health and Human Services, Office of the Assistant Secretary for Planning and Evaluation, and awarded by Substance Abuse Mental Health Service Administration. Its contents are solely the responsibility of the authors and do not necessarily represent the official views of the U.S. Department of Health and Human Services (Office of the Assistant Secretary for Planning and Evaluation) or the Substance Abuse Mental Health Service Administration.

Center on Poverty and Inequality Stanford University Building 370, 450 Serra Mall Stanford, CA 94305 650.724.6912 [email protected]

The Stanford Center on Poverty and Inequality