Donate
Independent, objective, nonpartisan research

R 1211SBR

Authors

R 1211SBR

Tagged with:

Publication PDFs

Database

This is the content currently stored in the post and postmeta tables.

View live version

object(Timber\Post)#3742 (44) { ["ImageClass"]=> string(12) "Timber\Image" ["PostClass"]=> string(11) "Timber\Post" ["TermClass"]=> string(11) "Timber\Term" ["object_type"]=> string(4) "post" ["custom"]=> array(5) { ["_wp_attached_file"]=> string(13) "R_1211SBR.pdf" ["wpmf_size"]=> string(7) "2231107" ["wpmf_filetype"]=> string(3) "pdf" ["wpmf_order"]=> string(1) "0" ["searchwp_content"]=> string(75249) "www.ppic.org The Great Recession and Distribution of fncome in California Sarah Bohn ● Eric Schiff Summary T he effects of the Great Recession hafe been felt far and wide. bccording to official mea- sures, the recession ran from December 2007 until June 2009. During that time, California experienced record unemployment, a housing market bust, sizable budget shortfalls, and downturns across nearly all major industries in the state. These problems hafe continued well past the technical end of the recession. California’s families hafe been hit hard by the Great Recession and its aftermath. Family income has declined across the spectrum, with lower incomes seeing the steepest losses (Table 1). The gap between upper- and lower-income families is now wider than efer. bnd the number of families in the middle-income range is shrinking. Specifically, we find: • Total income for the median family in California fell more than 5 percent between 2007 and 2009 (the official recession years) and an additional 6 percent between 2009 and 2010. • bt the lowest income lefel—the 10th percentile—family income fell more than 21 per- cent in total. bt the 90th percentile, family income fell 5 percent. • bfter adjusting for California’s higher cost of lifing, just less than half—47.9 percent— of indifiduals were in families that could be considered middle income in 2010. bs these findings suggest, the Great Recession has brought us to new extremes. These include record high measures of inequality, near-record lows in the proportion of middle- Justin s ullivan/Get ty ima Ges The Great Recession and Distribution of Income in California 2 www.ppic.org income families, and record high unemployment and unemployment duration. Through 2010, past the technical end of the recession, there has been no efidence of recofery in income across the distribution.Unemployment and underemployment—working fewer hours or weeks per year—were hallmarks of the Great Recession, and California is still facing high unemployment numbers. We find that efen for working families, income fell during the Great Recession for the middle of the distribution and below. Underemployment, rather than a decline in wages, appears to hafe drifen this income drop. This suggests that policies that create jobs and promote full- time employment, rather than those that target wage rates, are more likely to be effectife in aiding the recofery of family income. We do not yet know the timing of the recofery from the Great Recession and how that recofery will be shared across the income distribution. If prefious recofery patterns repeat themselfes, it is likely that the lower half of the income distribution will recofer much more slowly than the upper half, potentially allowing already record-high income inequality to per- sist. The erosion of low and middle incomes raises concerns about the equity of economic opportunity in the state. The most important factor drifing the gap between high- and low-income workers is edu - cation. Looking ahead, California may need to find innofatife ways to promote opportunity through education, especially so that middle- and lower-income families are not left behind. Please fisit the report’s publication page to find related resources: www.ppic.org/main/publication.asp?i=965 Table 1. Family income fell in every income catefory between 200b and 2010 Family income (f) Percenbage change 2007 2008 2009 20102007–2009 (official recession) 2007–2010 (acbual peak bo brough) 10th percentile 19,10 017,000 16,200 15,000 –15 . 2–21. 5 25th percentile 34,60034,200 32,400 31,200 –6.4 –10 . 0 Median 68,40066,000 64,700 61,10 0 –5.4 –10 . 7 75th percentile 122,00012 2 , 3 0 0 115 , 6 0 0 112 , 4 0 0 –5.3–7. 9 90th percentile 18 8 , 30 018 7, 5 0 0183,700 17 9,10 0 –2. 5 – 4.9 95th percentile 246,0002 32 ,10 0235,600 226,300 –4.2 –8.0 SOURCE Authors’ calculatiofs from the Curreft bopulatiof Survey of the U.S. Cefsus Bureau. NOTES: Family ifcome is adjusted to 2010 dollars afd formalized to accouft for family size. See Techfical Appefdix A for details. 3 The Great Recession and Distribution of Income in California www.ppic.org Introduction One way to understand the effects of the Great Recessfon on Calffornfa resfdents fs to exabfne overall trends fn fncobe. In thfs report, we focus on descrfbfng fncobe at the fabfly level. Thfs fncludes fncobe frob all sources (fncludfng fnvestbents and governbent sources) but fs predobfnantly cobposed of earnfngs frob ebploybent. Because of thfs, we also closely exabfne the relatfonshfp between the extensfve downturn fn the labor barket and the shfft fn the dfstrfbutfon of fabfly fncobe.As prevfous PPIC reports have shown, changes fn the dfstrfbutfon of fncobe are generally greater fn Calffornfa than fn the rest of the Unfted States. 1 Hfgh-fncobe fabflfes earn bore fn Calffornfa and low-fncobe fabflfes earn less. Over tfbe, hfgh fncobes fn Calffornfa have rfsen sub- stantfally, whereas low fncobes have seen sball declfnes. Because of these trends, the dfvfde between hfgh- and low-fncobe fabflfes has been larger and faster-growfng fn Calffornfa than fn the rest of the natfon. At the sabe tfbe, fewer and fewer fabflfes fall fn the bfddle-fncobe range. The Great Recessfon exacerbated these trends. Cob- pared to the rest of the country, Calffornfa experfenced larger declfnes fn fncobe at the bottob of the dfstrfbu- tfon and sballer declfnes at the top—leadfng to the largest gap between upper and lower fncobes fn at least 30 years. Incobe at the bedfan shrank by bore than 10 percent. And by 2010, just 49.7 percent of Calffornfa’s fabflfes could be consfdered bfddle fncobe, a new low. Unebploybent spfked sharply durfng the Great Reces- sfon, especfally fn Calffornfa. The duratfon of unebploy- bent has also rfsen precfpftously. But we found that even for those who had jobs, bedfan fncobe fell. Thfs appears to have less to do wfth across-the-board declfnes fn wages and bore to do wfth decreases fn the lfkelfhood of both full-tfbe work and overall hours worked. These findfngs suggest that polfcfes that create jobs and probote full-tfbe ebploybent are bore lfkely to be effectfve fn afdfng recov- ery than those that target wage rates. Thfs report first descrfbes the changes fn fncobe dfstrfbutfon fn Calffornfa durfng the Great Recessfon. We then exabfne the effects of unebploybent and underebploybent—labor barket outcobes that drove these changes. Next, we fnvestfgate how fncobe changes have been experfenced across regfons and debographfc groups fn Calffornfa. Ffnally, we gfve context to these changes by cobparfng fncobe trends durfng the Great Recessfon to trends fn prevfous recessfons. Changes in the distribution of income are generaffy greater in Cafifornia than in the rest of the United Statesb Data and methods In this report, we use the bnnual Social and Economic Supple- ment of the CPS data collected by the U.S. Census Bureau and Bureau of Labor Statistics efery March from 1980 to 2011. The CPS data profide a comprehensife picture of what has happened to income on an annual basis through March 2010. We measure income for families rather than indifiduals or households. 2 We assume that the family is the primary unit across which income is shared and that nonrelated indifidu- als in a household do not share income. The bulk of our study describes total family income derifing from all sources— including work, interest on infestments, pensions, unemploy- ment, and welfare—and is measured before tax. 3 In some analyses, we examine family income from work separately. Our total family income measure excludes nonmonetary aspects of family income, such as food assistance, nonpecuni- ary job benefits, or other in-kind transfers. Gifen these cafeats, we proceed with adjusting CPS family income in a number of ways. These adjustments make our income estimates compa- rable ofer time (i.e., by remofing the effect of inflation) and across family size. Except where noted otherwise, all estimates presented can be understood as the 2010 dollar equifalent for a family of four; these adjustments remofe the effects of infla- tion and allow us to compare across families of different sizes. Further details regarding our data and methods may be found in Technical bppendix b. The Great Recession and Distribution of Income in California 4 www.ppic.org However, before turnfng to the central analysfs, we wfll take a bobent to dfscuss our two dfstfnct but equally useful ways of lookfng at fncobe dfstrfbutfon. Tracking Income Distribution Our data allow two dffferent vfews of fncobe dfstrfbutfon fn Calffornfa. The first fnvolves lookfng at changes fn the dfstrfbutfon of fncobe over tfbe—not for partfcular fabflfes but for the overall dfstrfbutfon of fncobe across Calffornfa’s entfre populatfon. 4 In thfs approach, we break the populatfon fnto percentfle groups: The fabfly at the 90th percentfle of fncobe has an fncobe level hfgher than 90 percent of the populatfon, and the fabfly at the 10th percentfle has fncobe hfgher than only 10 percent of the populatfon. In these terbs, the exact bfddle-fncobe, or bedfan, fabfly fs one that falls at the 50th percentfle. Thfs bfddle-fncobe fabfly fs not the sabe every year but fnstead shffts as the fncobe dfstrfbutfon rfses and falls. Exabfnfng the dfstrfbutfon of fncobe fs therefore fbportant to understandfng how the populatfon fs dofng overall. But ft fs also useful to know how bany fabflfes fall fnto each fncobe category. To find thfs out, we define cat- egorfes of fncobe that are roughly constant over tfbe and see how bany fabflfes fall fnto the dffferent groups. To do thfs, we use definftfons of fncobe categorfes fabfl- far to bost readers: low fncobe, bfddle fncobe, and hfgh fncobe. Sfnce the bfddle group fs otherwfse qufte broad, we sobetfbes separate the bfddle-fncobe group fnto thfrds: lower bfddle, central bfddle, and upper bfddle. To define these groups, we use fabfly fncobe cut- offs cobbon to sfbflar research and based on a federal beasure of standard of lfvfng, the federal poverty level of fncobe (FPL). 5 Low fncobe fs defined as at or less than two tfbes the FPL, or $44,200 and below. 6 Mfddle fncobe fs defined as between two and seven tfbes the FPL, or $44,200 to $154,800. 7 Thfs spread fs large because we dfvfde the bfddle-fncobe group fnto three roughly equally sfzed portfons. 8 Hfgh fncobe fs anythfng above $154,800. Impact of the Great Recession The Great Recessfon hft fncobes across the dfstrfbutfon— but certafn fncobe groups felt fts effects bore strongly than others dfd. In thfs sectfon, we detafl overall trends fn fncobe durfng the Great Recessfon, place these trends fnto a long-terb context, and consfder the shfftfng sfze of each fncobe class over tfbe. Chanfes in Income Distribution durinf the Great Recession and Beyond Wage and salary fncobe for the bedfan fabfly fn Calf - fornfa fell bore than 5 percent durfng the Great Recessfon (Table 2). Declfnes below the bedfan were even larger— at the 10th percentfle, fncobe fell bore than 15 percent Measurinf the Great Recession In this report, we obserfe the Great Recession’s effect through its two official years, 2008 and 2009, as well as the first year af ter, 2010. 9 Suitable data are not yet afailable for 2011. In some tabulations, we compare the Great Recession to the income peak in 2007, immediately beforehand. This gifes an initial measure of the seferity of the decline from peak to trough. In other tabulations, we compare the official two years of the recession (2008 and 2009) to the two years immediately pre - ceding it (2006 and 2007). These gife a measure of the seferity of the decline from the recent peak period to the period of the current recession. The Great Recession hit incomes across the distribution—but certain income groups feft its effects more strongfy than others didb 5 The Great Recession and Distribution of Income in California www.ppic.org between 2007 and 2009. Above the bedfan, fabfly fncobe also decreased durfng the recessfon but by only a fractfon of that abount, and the 90th percentfle fell about 2 percent. Although the recessfon officfally ended fn 2009, fncobes contfnued to fall fn the year followfng. Between 2009 and 2010, bedfan fabfly fncobe fell another 5 percent—the sabe declfne experfenced over the full two years of the recessfon. Incobes both above and below the bedfan also contfnued to fall fnto 2010. The rate of the declfne across the dfstrfbutfon was at or above the rate experfenced durfng the recessfon, on a per year basfs. By that beasure, there fs no evfdence of a slowfng of the fncobe effects of thfs recessfon. The hfghest fncobe level we can consfstently beasure fs at the 95th percentfle. 10 As Table 2 shows, the 95th percentfle of fncobe fn Calffornfa—beanfng fabflfes that have fncobe hfgher than 95 percent of the populatfon—also fell durfng the recessfon. The 95th percentfle appeared to rebound by 2009 but took another hft fn 2010. Thus, even the top end of the fncobe dfstrfbutfon does not yet appear to be fn recovery. However, the declfnes experfenced at the top of the fncobe dfstrfbutfon are over three tfbes smaller than those experfenced at the lowest end of the dfstrfbutfon. Cobpared to the effects fn the rest of the Unfted States, the Great Recessfon’s effects fn Calffornfa are sobewhat bfxed. Ffrst, note that fabfly fncobe levels fn Calffornfa were hfgher for all cutpofnts above the bedfan, as well as the 10th percentfle, before the recessfon. Despfte larger declfnes between 2007 and 2010, fncobe fn all categorfes above the bedfan—the 75th, 90th, and 95th percentfles— fn Calffornfa were stfll hfgher than fn the rest of the Unfted States by 2010. The sabe fs not true, however, for the lowest cutpofnt of the dfstrfbutfon. The 10th percentfle fell 56 percent bore fn Calffornfa than fn the rest of the natfon, brfngfng ft lower than the natfonal level by 2010. Table 2. Family income fell further in California than in the rest of the United States California Family income (f) Percenbage change 2007 2008 2009 20102007–2009 2009–2010 10th percentile 19,10 017,00016,200 15,000–15 . 2 –7. 4 25th percentile 34,60034,200 32,400 31,200 –6.4–3.8 Median 68,40066,000 64,700 61,10 0–5.4–5.6 75th percentile 122,00012 2 , 3 0 0 115 , 6 0 0112 , 4 0 0 –5.3–2.7 90th percentile 18 8 , 30 018 7, 5 0 0183,700 17 9,10 0 –2. 5–2. 5 95th percentile 246,0002 32 ,10 0235,600 226,300 –4.2–3.9 resb of bhe u nibed Sbabes Family income (f) Percenbage change 2007 2008 2009 20102007–2009 2009–2010 10th percentile 18 ,9 0 018,200 17,00016 , 30 0–10 . 3 –4.2 25th percentile 3 7, 9 0 036,500 35,30034,200 – 6.9–3.2 Median 70,5006 7, 8 0 06 6 ,10 065,800 –6.3–0.3 75th percentile 115 , 9 0 0111, 9 0 0111, 5 0 011 0 , 5 0 0 –3.8–1. 0 90th percentile 170,600164,800 165,10 016 4, 4 0 0 –3.2–0.4 95th percentile 212 , 5 0 0204,200 204,700203,800 –3.7–0.5 SOURCE Authors’ calculatiofs from the Curreft bopulatiof Survey of the U.S. Cefsus Bureau. NOTES: Family ifcome is adjusted to 2010 dollars afd formalized to accouft for family size. See Techfical Appefdix A for details. The Great Recession and Distribution of Income in California 6 www.ppic.org Lonf-Term Chanfes in Income Distribution To put these fncobe dfstrfbutfon changes fnto a larger context, we wfll now exabfne a longer perfod: 1980–2010 (Ffgure 1). Thfs figure shows the percentage change fn fncobe at several pofnts fn the fncobe dfstrfbutfon fn each year cobpared to the base year of 1980. 11 All fncobe levels have experfenced sfgnfficant peaks and valleys over thfs tfbe perfod. The 50th percentfle reached fts 30-year peak fn 2003, wfth the bedfan fabfly earnfng 12 percent bore fncobe than fn 1980. After the bost recent low fn 2004, the bedfan began to recover but was hft agafn by the Great Recessfon. The declfne between 2007 and 2010 cobpletely reversed—and bore—the recovery frob the prevfous recessfon. By 2010, the bedfan fabfly earned about 1 percent less than the bedfan fabfly fn 1980. However, despfte declfnes durfng the Great Reces- sfon, bedfan fabfly fncobe fs stfll hfgher than ft was fn the lows of the recessfons of the 1980s and 1990s. The sabe cannot be safd for fncobe below the bedfan. Not only dfd the Great Recessfon strfp away any gafns fn fncobe at the 10th and 25th percentfles that followed the bust of the dot-cob bubble, but ft also pushed fncobes at these levels to near-record lows. By 2010, fabflfes at the 10th percentfle had fncobes roughly 24 percent lower than the 10th percentfle dfd fn 1980, and fabflfes at the 25th per- centfle had fncobes 12 percent lower. The 10th and 25th percentfles have not yet fallen to the lows of the 1990s recessfon, but by 2010 there fs no evfdence that fncobes have yet troughed fn the Great Recessfon. At the other end of the spectrub, the 90th percentfle saw a declfne frob fts 2006 peak. However, the gafns at the 90th percentfle over the past three decades bean that despfte the Great Recessfon, the 90th percentfle of fncobe was stfll 34 percent hfgher fn 2010 than fn 1980. Incobe declfnes at thfs level are also buch less severe than the declfnes experfenced at lower pofnts fn the dfstrfbutfon. Notably for the 90th percentfle, the Great Recessfon has not as of yet strfpped away the recovery bade after 2004. The 75th percentfle of fncobe saw larger declfnes than the 90th percentfle durfng the Great Recessfon, brfngfng ft to a level last seen fn the late 1990s. However, over the longer terb, fncobe at the 75th percentfle fs stfll substantfally hfgher than ft was fn prevfous decades. By 2009, the 75th percentfle was earnfng over 18 percent bore than fn 1980. Currently, declfnes fn the lower fncobe levels durfng the Great Recessfon appear sfbflar fn severfty to those felt fn the early 1990s. But the steepness of the recent declfnes outpaced that of the early 1990s. It rebafns to be seen ff lower fncobe levels wfll fall further fn 2011 and beyond. By 2010, families at the 10th perceftile had ifcomes roubhly 24 perceft lower thaf if 1980. l ucy n icholson/Reute Rs Figure 1. Family income moves wifh fhe businessw cycle 90th percentile 75th percentile Median 25th percentile 10th percentile –f0 –20 –10 0 10 20 f0 40 50 1980 1982 1984 1986 1988 1990 1992 1994 1996 1998 2000 2002 2004 2006 2008 2010 Percentage change in famsily income (base f 198b) SOURCE: Authors’ calculatfons from thb Currbnt Populatfon Survby of thb U.S. Cbnsus Burbau. NOTES: Famfly fncomb fs adjustbd to 2010 dollars and lnormalfzbd to account for famfly sfzb. Sbb Tbchnfcal Appbndfx A for dbtafls. Shadbd rbg fons dbnotb rbcb ssfonary pbrfods as mbasur bd by pbaks and troughs fn fnco mb lbvbls. 7 The Great Recession and Distribution of Income in California www.ppic.org The Growinf Income Gap In Calffornfa, the gap between lower- and upper-fncobe fabflfes has been larger than fn the rest of the natfon for bany decades and has tended to fncrease fn recessfonary perfods. The Great Recessfon fs no exceptfon. A cobbon way to exabfne thfs gap fs to look at the ratfo of fncobe for fabflfes at the top of the dfstrfbutfon to fabflfes at the bottob. Here, we present two standard fncobe ratfos: the ratfo of fncobe at the 90th relatfve to the 10th percentfle (the “90/10 ratfo”) and at the 75th relatfve to the 25th percentfle (the “75/25 ratfo”). 12 The forber fs a bore extrebe beasure of hfgh versus low fncobe, whereas the lat - ter fs less so because the 75th and 25th percentfles are closer to the bfddle of the dfstrfbutfon. These beasures are useful for understandfng gaps between rfch and poor, for exabple, as beasured by shffts fn the overall dfstrfbutfon of fncobe. Durfng the Great Recessfon fn Calffornfa, the 90/10 ratfo jubped to fts hfghest level ever, 11.9, fn 2010 (Ffg- ure 2). 13 Thfs beans that fabflfes at the 90th percentfle (where only a tenth of the populatfon does better fn terbs of fncobe) had fncobe 11.9 tfbes hfgher than fabflfes at the 10th percentfle (where only a tenth of the populatfon does worse). The dfsparfty between hfgh and low fncobes durfng the Great Recessfon even exceeded the gap experf- enced durfng the long and severe recessfon of the bfd-1990s, durfng whfch the 90/10 ratfo reached 11.0 (fn 1997). In the rest of the country, the 90/10 ratfo also grew to a new hfgh durfng the Great Recessfon, to 10.1 by 2010. But thfs ratfo rebafned buch sballer fn the natfon as a whole than fn Calffornfa alone. The gap between fncobe levels was less volatfle toward the bfddle of the dfstrfbutfon. The 75/25 ratfo rebafned fafrly steady, at roughly 3.6 fn Calffornfa and 3.2 fn the rest of the Unfted States fn 2010, both up just one-tenth of a pofnt frob 2007. How Bif Is Each Income Group in California? We now turn to our second vfew of fncobe fn Calffornfa, whfch holds categorfes of fncobe constant and asks how bany fabflfes fall fnto each category. Here, we show three fncobe categorfes—low, bfddle, and hfgh—over tfbe (Ffgure 3). We also adjust these figures to account for the hfgh cost of lfvfng fn Calffornfa. The average Calffornfa fabfly bust have a hfgher fncobe level to bafntafn the sabe standard of lfvfng as the average fabfly fn the rest of the country. 14 So far, the Great Recessfon has not shffted the sfze of each fncobe group frob fts longer-terb trend. But ft has created sobe new hfghs and lows. Most Calffornfans lfve fn bfddle-fncobe fabflfes. In 1980, the proportfon of these fabflfes reached a 30-year Figure 2. Gaps between upper- and fower-income bamifies are farger in Cafibornia than in the drest ob the United States 90/10 California 90/10 United States f5/25 California f5/25 United States12 11 10 9 8 f b 5 4 3 2 1 13 0 1980 1982 1984 198b 1988 1990 1992 1994 199b 1998 2000 2002 2004 200b 2008 2010 Ratio of family income SOURCE: Authors’ calculatfons from thb Currbnt Populatfon Survby of thb U.S. Cbnsus Burbau. NOTES: All ratfos x/y rbprbsbnt famfly fncomb at pbrcbntflb x rblatfvb to famfly fncomb at pbrcbntflb y fn gfvbn ybar. Famfly fncomb fs adjustbd to 2010 dollars and lnormalfzbd to account for famfly sfzb. Sbb Tbchnfcal Appbndfx A for dbtafls. Figure 3. The share of middle-ifcome families has f-allef if Califorfia Middle income COLA middle High income COLA high COLA low Low income6f 5f 4f 3f bf 1f 7f f 198f 198b 1984 1986 1988 199f 199b 19941996 1998 bfff bffb bff4 bff6 bff8 bf1f Percentage SOURCE: Authors’ calculatfons from thb Currbnt Populatfon Survby of thb U.S. Cbnsus Burbau. NOTES: Bbcausb of data avaflabflfty, wb arb ablb to adjust for cost of lfvfng (COLA) only from 1985 forward. Sbb notb 8 and Tbchnfcal Appbndfx A for dbtafls on thb dbflnftfon of fncomb catbgorfbs. The Great Recession and Distribution of Income in California 8 www.ppic.org hfgh of 60 percent, a nubber that has been trendfng down- ward ever sfnce. The percentage of fndfvfduals fn bfddle- fncobe fabflfes reached a new low of 49.7 percent fn 2010. At the sabe tfbe, the low-fncobe category fncreased to 36.6 percent, a level not experfenced for over a decade. And the hfgh-fncobe group dropped slfghtly to 13.7 percent, a level last experfenced around 2001.On net, the recessfon and the year followfng have worsened Calffornfa’s fncobe pfcture as vfewed through these three fncobe categorfes. Before the recessfon, the long-terb trend showed net fbprovebent: The share of fabflfes categorfzed as low fncobe was relatfvely stable, and the declfnfng share counted as bfddle fncobe was supplanted by an fncreasfng hfgh-fncobe share. However, durfng the recessfon and the year followfng, the trend reversed: Declfnes fn the share of fabflfes fn hfgh- and bfddle-fncobe categorfes were replaced wfth an fncreasfng share classffied as low fncobe. The dfstrfbutfon of Calffornfans across fncobe groups bfbfcs the trend fn the Unfted States overall. 15 However, Calffornfa’s fabflfes are less lfkely than those fn other states to be counted as bfddle fncobe and are bore lfkely to be efther low or hfgh fncobe. In the rest of the country, as of 2010, 55 percent are bfddle fncobe, 33 percent low fncobe, and 12 percent hfgh fncobe. When fncobe fs adjusted for Calffornfa’s hfgher cost of lfvfng, we find that even fewer fndfvfduals—47.9 percent— were fn fabflfes consfdered bfddle fncobe fn 2010. In our data gofng back to 1985, the bfddle-fncobe group fn the state has never fallen to a level thfs low. After sfbflar adjustbents, the low-fncobe group rfses to 42.9 percent and the hfgh-fncobe group falls to 9.3 percent. The Effects of Unemployment and Underemployment The Great Recessfon led to persfstently hfgh unebploy - bent levels, wfth Calffornfa’s ebploybent pfcture abong the worst fn the natfon. By the officfal end of the recessfon fn June 2009, Calffornfa’s unebploybent rate had clfbbed to 11.6 percent, cobpared to 9.5 percent fn the natfon as a whole. However, the officfal end of the recessfon dfd not sfgnal a recovery fn ebploybent. In June 2010, a full year after the recessfon ended, Calffornfa’s unebploybent rate stood at 12.3 percent; the natfonal rate was 9.5 per - cent. 16 Thfs was the state’s hfghest unebploybent rate sfnce 1980—and ft was buch hfgher than fn other reces - sfons fn the last three decades. 17 Although the Great Recessfon has ended, as beasured by other officfal fndfca - tors, 18 the ebploybent pfcture rebafns bleak, partfcularly for Calffornfa. Forecasts suggest that the unebploybent rate wfll decrease slowly, wfth hfgh rates contfnufng fnto the near future. 19 Sfnce ebploybent fs the bafn source of fncobe for bost Calffornfans, the unebploybent rate fs typfcally hfghly correlated wfth changes fn fncobe: Troughs fn bedfan fabfly fncobe are usually cofncfdent wfth peaks fn the unebploybent rate. Cobfng out of recessfonary perfods, we typfcally see decreases fn the unebploybent rate and concurrent fncreases fn bedfan fabfly fncobe (see Technfcal Appendfx C for a detafled figure). How dfd Calffornfa’s ebploybent trends durfng the Great Recessfon correlate wfth changes fn the state’s dfs- trfbutfon of fncobe? In thfs sectfon, we begfn by detaflfng exactly how buch labor barket earnfngs batter for fabfly fncobe. Next, we fdentffy assocfatfons—rather than causal relatfonshfps—between ebploybent trends and changes fn fncobe dfstrfbutfon. Throughout, our focus fs on recent trends—for the bost part, we cobpare the two relatfvely prosperous years before the recessfon to the two officfal years The Great Recession fed to persistentfy high unempfoyment fevefs, with Cafifornia’s empfoyment picture among the worst in the nationb 9 The Great Recession and Distribution of Income in California www.ppic.org of the Great Recessfon and then exabfne what has hap- pened fn 2010, the first year followfng the officfal recessfon. The Many Sources of Family Income Fabfly fncobe derfves frob bultfple sources. Earnfngs frob work clearly are related to fabfly econobfc well- befng. However, other sources of fncobe batter as well. In tfbes of constrfcted labor barket opportunftfes, fncobe frob sources other than wages—such as unebploybent cobpensatfon, welfare, or earnfngs on fnvestbents—can cobpensate for declfnes fn fabfly fncobe.Both before and durfng the Great Recessfon, bale labor barket earnfngs bade up the bajorfty of fabfly fncobe across the spectrub (Ffgure 4). These earnfngs contrfbute slfghtly less to fabfly fncobe fn the low-fncobe group than fn the bfddle-fncobe group. 20 Durfng the Great Recessfon, bale earnfngs declfned as a share of total fabfly fncobe fn all groups except the upper-bfddle- fncobe category. Febale earnfngs are the second bost sfzable cobponent fn fabfly fncobe. Durfng the Great Recessfon, the fbpor - tance of febale earnfngs fncreased relatfve to other sources— by 2 percentage pofnts for lower-fncobe fabflfes and by 4 percentage pofnts for lower-bfddle-fncobe fabflfes. For low-fncobe fabflfes, the thfrd bost fbportant source of fncobe fs the governbent; thfs fncludes uneb- ploybent, Socfal Securfty, publfc assfstance, and Supple- bental Securfty Incobe. Durfng the Great Recessfon, fncobe frob governbent transfers was fncreasfngly fbportant, growfng frob 10 to 12 percent. Governbent transfers are a sballer fractfon of fabfly fncobe for other groups, but durfng the Great Recessfon they doubled across the board. For exabple, lower-bfddle-fncobe fabflfes recefved 3.8 percent of thefr fncobe frob governbent transfers before the Great Recessfon but 6.3 percent dur- fng the recessfon (2008–2009) and 7.2 percent fn 2010 (see Technfcal Appendfx C for detafls). External research finds that cobpared to prevfous recessfons, governbent transfers played a larger role fn supportfng fncobe fn the Great Recessfon than they dfd fn prevfous recessfons. 21 Thus, even though the share of fabfly fncobe frob governbent sources fs sball relatfve to the share frob earnfngs, ft fs a qualftatfvely fbportant factor. Unemployment As we have seen, ebploybent fncobe bakes up the bulk of overall fncobe for Calffornfa fabflfes. Of course, ebploy- bent dfffers across Calffornfa’s fncobe groups. Durfng the 58% 63% 59% 27% 29% 33% 10% 4% 2%1% 56% 57% 61% 29% 33% 35% 31% 12% 6% 4% 2% Figure 4. Male earnings are the major sourfe of family infome, eben during the Great Refession 0 90 80 70 60 50 40 30 20 10 100 Percentage Lower miffle Low Upper miffle bentral miffle SOURCE: Authors’ calculatfons from thb Currbnt Populatfon Survby of thb U.S. Cbnsus Burbau. NOTES: Incomb sharbs fn bach chart rbprbsbnt thb avbragbs across bach two-ybar pbrfod. Although fncomb catbgory dbfnftfons arb basbd on normalfzbd famfly fncomb, fncomb sharbs arb basbd on fnfatfon-adjustbd fncomb and arb not normalfzbd for famfly sfzb. Thb hfgh-fncomb group fs not fncludbd lhbrb bbcausb sharbs arb hbavfly afbctbd by top-codfng. Sbb Tbchnfcal Appbndfx A for dbtafls on data constructfon and Tbchnfcal Appbndfx C for a sfmflar fgurb showfng thb dfstrfbutfonl of fncomb for famflfbs fn 2010. Wb do not track famflfbs fn bach fnlcomb catbgory ovbr tfmb; rathbr, bach chart should bb vfbwbd as a snapshot ofl famflfbs fn thb gfvbn pbrfod who happbnl to fall fnto thb gfvbn fncomb catbgory. Thus, changbs fn fncomb sourcbs arb confoundbd hbrb wfth changbs fn composftfon of famflfbs lfn bach fncomb catbgory. bapital income Other Government Female earnings Male earnings 2006–2007 average Lower miffle Low Upper miffle bentral miffle 2008–2009 average The Great Recession and Distribution of Income in California 10 www.ppic.org Great Recessfon, unebploybent jubped fn all fncobe groups, at least doublfng the rate of 2007. Unebploybent spfked bost steeply for low- and lower-bfddle-fncobe groups (Ffgure 5). By 2011, 12.2 percent of Calffornfans were unebployed. Abong low-fncobe fndfvfduals, the unebploybent rate was even hfgher—23.2 percent. Sfnce fncobe frob workfng bakes up the vast bajorfty of fabfly fncobe, ft fs not surprfsfng that the groups experfencfng the largest declfnes fn fncobe would also have the hfghest unebploybent rate; fndeed, labor barket outcobes deter- bfne, to a large extent, the fncobe category fnto whfch a fabfly fs classffied.Only fn the low- and lower-bfddle-fncobe groups dfd the unebploybent rate show no sfgn of taperfng off by 2011. Although for the other fncobe groups there are sfgns of a turnaround, the unebploybent rate rebafns stubbornly hfgh—hfgher than seen fn decades. Not only has the recessfon brought about rates of unebploybent hfgher than fn prevfous recessfons, but the duratfon of unebploybent fs also longer than fn prevfous recessfons. The average spell of unebploybent has fncreased steadfly frob the begfnnfng of the Great Recessfon through 2011, frob an average of 15.8 weeks to 37.4 weeks, respectfvely (see Technfcal Appendfx C for further detafl). Long spells of unebploybent have been experfenced across all fncobe categorfes: frob an average of 29 weeks for upper-bfddle-fncobe unebployed workers to 40 weeks for low-fncobe unebployed workers fn 2011. Durfng the recessfon of the early 1990s, the average dura- tfon of unebploybent peaked at 23.6 weeks. Underemployment Hfgh unebploybent rates largely explafn the declfne fn fncobe across the dfstrfbutfon durfng the Great Recessfon. But fabfly fncobe declfned even for bany who had jobs durfng the Great Recessfon. Underebploybent—defined here as workfng less than a full work-week—played a sfg- nfficant role fn these declfnes. By 2008–2009, bedfan fncobe frob wages and salary for low-fncobe workers had fallen 16 percent frob what ft had been two years before (Table 3). 22 For lower-bfddle and central-bfddle workers, the drop was 3 to 4 percent. It rebafned about the sabe for upper-bfddle and hfgh- fncobe workers. Even though these fndfvfduals were workfng, they worked less, on average, durfng the Great Recessfon—for exabple, the percentage of workers fn the low-fncobe group who reported full-tfbe ebploybent fell 10.1 percent durfng the Great Recessfon (where full tfbe fs defined as workfng at least 35 weeks). The rate of full-tfbe ebploy- bent fell for all other fncobe groups, as well. Average hours worked also fell durfng the Great Reces- sfon. On average, workers fn low-fncobe fabflfes worked 11 percent fewer hours—a declfne about seven tfbes larger than that experfenced by the central-bfddle-fncobe fabf- Figure 5. Unemployment rate by infome group, in California 0 5 10 15 20 25 1990 1992 1994 1996 1998 2000 2002 2004 2006 2008 2010 1991 1993 1995 1997 1999 2001 2003 2005 2007 2009 2011 Percentage SOURCE: Authors’ calculatfons from thb Currbnt Populatfon Survby of thb U.S. Cbnsus Burbau. NOTE: Sbb Tbchnfcal Appbndfx A for dbtafls on thb dbflnftfons of fncomb groups. Central mfddle Lober mfddle Hfgh fncome Upper mfddle All classes Lob fncome Not onfy has the recession brought about rates of unempfoyment higher than in previous recessions, but the duration of unempfoyment is afso fonger than in previous recessionsb 11 The Great Recession and Distribution of Income in California www.ppic.org lfes. (These statfstfcs exclude workers who were uneb- ployed all year but fnclude workers who were unebployed part of the year.) Durfng the Great Recessfon, the average hourly wage rate fell for workers fn the low-fncobe and hfgh-fncobe groups by a sball abount but fncreased or stayed about the sabe on average for workers fn other fncobe groups. Thfs pattern fndfcates that underebploybent rather than declfnfng wages, for bost workers, was behfnd fallfng earnfngs. Because fnflatfon was extrebely low durfng the Great Recessfon, ft fs fndeed unsurprfsfng that wages would rebafn roughly constant and that ebployers would fnstead adjust hours or nubber of ebployees. 23 These findfngs provfde sobe perspectfve on the rela- tfve fbportance of unebploybent and underebploybent fn drfvfng the fncobe trends we have observed. Recent research fn the natfonal context suggests that the declfne fn bale ebploybent was the bost fbportant factor fn the declfne fn bedfan fncobe durfng the Great Recessfon. Thfs factor fs about three tfbes bore fbportant than any other fn drfvfng down bedfan fncobes durfng the recessfon. 24 The sabe appears to be true for Calffornfa. In addftfon, ft appears that fncobe declfnes fn the Great Recessfon are strongly related to (1) whether fabfly earners are ebployed and (2) for those who are ebployed, how buch they worked. In the year followfng the officfal recessfon, there are sball sfgns that for those workfng, labor barket condf- tfons were begfnnfng to fbprove, at least for hfgher-fncobe workers. Across all fncobe categorfes, the percentage of workers ebployed full tfbe fncreased between the years of the officfal recessfon to the year after. Despfte thfs posftfve sfgn, there was lfttle fbprovebent fn average hours worked There are sibfs of a turfaroufd for some ifcome broups, but for low- afd lower-middle-ifcome broups the ufemploymeft rate remaifs stubborfly hibh. fav i f maun G/Bloom BeRG via Get t y ima Ges Table 3. Many individuals worked less durinf the Great Recession Family income cabegory Change during bhe recession (%) (from 2006–2007 bo 2008–2009) Change in bhe year following bhe recession (%) (from 2008–2009 bo 2010) median income from wage and salary Percenbage employed full bime a verage hours worked median hourly wage median income from wage and salary Percenbage employed full bime a verage hours worked median hourly wage Low –15 . 5–10 .1 –10 . 5 –2 .1–1. 3 1.4 –2 .1 4.0 Lower middle –3.7–6.4 – 6 .1 5.0–1. 3 3.50.4–1. 3 Central middle –3.3–0.4 –1. 5 3.21.12.2 0 .1–0.5 Upper middle –0.5 –0.5 –0.6 3.64 .12.4 0 .11. 8 High –0.5 –0.8 –1. 0 –1. 2 3.43.2 1. 23.9 SOURCE: Authors’ calculatiofs from the Curreft bopulatiof Survey of the U.S. Cefsus Bureau. NOTES: All statistics pertaif to wage earfers who worked at least ofe week if the givef year afd are calculated of a per worker basis. Self-employed workers are excluded. See Techfical Appefdix C for details afd ufderlyifg estimates. The Great Recession and Distribution of Income in California 12 www.ppic.org and bfxed outcobes fn hourly wage rates. Whereas bedfan fncobe fncreased for upper-bfddle- and hfgh-fncobe workers, ft contfnued to declfne for low- and lower-bfddle- fncobe workers. (Indeed, relatfvely worse labor barket outcobes partfally drove the classfficatfon fnto these lower- fncobe groups.) Although sfgns of underebploybent for bfddle- and hfgh-fncobe workers began to wane by 2010, the bfxed pfcture for lower-fncobe workers reveals a labor barket that, by 2010, had not fbproved drastfcally, even for those ebployed. In addftfon, as we noted fn the prevf- ous sectfon, thefr rate of unebploybent had yet to turn around. Who Was Most bffected by the Great Recession? As we have seen, the Great Recessfon affected all Calffor- nfans. Incobes fell and unebploybent grew for all fncobe groups, though for those fn the low- and lower-bfddle- fncobe groups the effects have been bore severe. In thfs sectfon, we assess the debographfc and geographfc cobpo- nents of the fncobe shocks of the Great Recessfon. Chanfes across Demofraphic Groups Throughout thfs sectfon, we define debographfc groups through the head of the fabfly. For exabple, we categorfze a fabfly as “fbbfgrant” ff the head of the fabfly reports that he or she fs an fbbfgrant. Sfbflarly, we consfder a fabfly to be whfte ff the head of the fabfly reports that he or she fs whfte. Thfs fs a strafghtforward way to clas- sffy fabflfes accordfng to debographfc characterfstfcs, but ft does overlook the subtlety of bfxed-type fabflfes. For exabple, bany fabflfes classffied here as fbbfgrant fnclude natfve-born chfldren. However, sfnce fncobe changes are drfven by labor barket condftfons that prf- barfly affect the head of the fabfly, our bethod allows for a basfc but fbportant overvfew of the Great Recessfon’s effect on varfous debographfc groups. The Great Recessfon fntensffied fncobe and ebploy- bent dffferences abong debographfc groups. Even so, declfnes were experfenced across the debographfc spec- trub. Ffgure 6 shows nearly across-the-board declfnes fn fncobe frob the peak years before the recessfon to the officfal two years of the recessfon and beyond, for fabflfes across educatfon, ethnfcfty, natfvfty, and structure. Frob the peak years to 2010, no debographfc group fn Calffor- nfa experfenced gafns fn bedfan fncobe. Ethnicity and Nativity Many Hfspanfc and black fabflfes were strugglfng eco- nobfcally even before the Great Recessfon. Hfspanfc fabflfes fn Calffornfa have the lowest bedfan fncobe level across ethnfc groups, followed by black fabflfes. The Great Recessfon hft these two groups hard. They experfenced the largest declfnes fn bedfan fabfly fncobe, at 8 percent for Hfspanfc fabflfes and 25 percent for black fabflfes. Correspondfngly, the unebploybent rate for labor force partfcfpants fn fabflfes headed by blacks and Hfspanfcs jubped to the hfghest levels across ethnfc groups durfng the Great Recessfon, to 19 and 15 percent fn 2008–2009, The Great Recessiof hit Hispafic afd black families hard; they experi- efced the larbest declifes if mediaf family ifcome. h eacphotos/Flicb R /cR e at i v e commons 13 The Great Recession and Distribution of Income in California www.ppic.org respectfvely (see Technfcal Appendfx C for detafls). The unebploybent rate fn fabflfes headed by blacks was sub- stantfally hfgher than for fabflfes headed by Hfspanfcs by 2010, despfte havfng reached near parfty durfng the recessfon. Fabflfes headed by Asfans were less affected, and there was a cobparatfvely sball 3 percent declfne fn fncobe frob peak to trough. Asfan unebploybent rates were the lowest across ethnfc groups before the recessfon—at 4 percent— and rebafned the lowest durfng the recessfon, despfte hav- fng bore than doubled to 9 percent by 2010. The bedfan whfte fabfly has the hfghest fabfly fncobe of any ethnfc group fn Calffornfa. Thfs fell durfng the Great Recessfon by about 3 percent and by another 5.6 percent fn 2010. The total declfne fn bedfan fncobe for whfte fabflfes was sfbflar to that of Hfspanfc fabflfes and bore than for Asfan fabflfes. The unebploybent rate for whfte fabflfes rose frob 4.5 percent before the recessfon to 10.2 percent fn 2010—a rate stfll lower than that of Hfs- panfc or black fabflfes fn Calffornfa. Although ft fs correlated wfth ethnfcfty, we look sepa- rately at natfvfty, findfng that both natfve and fbbfgrant fabflfes experfenced sfzable declfnes fn fncobe at the bedfan. Fabfly fncobe fell a total of 14 percent for fabflfes headed by a natfve-born person and 5 percent for fabflfes headed by a forefgn-born person. However, the level of fncobe rebafned hfgher for natfve-headed households despfte the sharp declfne. Although both groups experf- enced sfbflar changes fn unebploybent, the rate for house - holds headed by natfves was lower than that of households headed by fbbfgrants both before and durfng the recessfon (see Technfcal Appendfx C). Educational Attainment Educatfonal attafnbent battered durfng the Great Reces- sfon. The bore educatfon, the hfgher the bedfan fncobe and the lower the unebploybent rate abong fabflfes fn Calffornfa. However, the bedfan fabfly fncobe of all educatfon groups declfned through 2010. Fabflfes headed by less-educated adults, who already had hfgh unebploybent rates before the recessfon (on the Many Hispanic and bfack famifies were struggfing economicaffy even before the Great Recessionb Figure 6. Median income fell across all of falifornia’s demographic groups during bhe Greab Recession and beyond 0 120,000 100,000 80,000 60,000 40,000 20,000 140,000 Income ($) SOURCE: Authors’ calculatfons from thb Currbnt Populatfon Survby of thb U.S. Cbnsus Burbau. NOTES: Incomb fs adjustbd to 2010 dollars and lnormalfzbd to account for famfly sfzb. Charactbrfstfcs arb dbfnbd by hbad of famfly; outcombs pbrtafn to famfly or labor forcb partfcfpants fn thb famfly. Notb that thb racb/bthnfcfty groups arb mutually bxclusfvb. Pbrcbntagb changbs mbasurb thb changb from 2006–2007 to 2010. All dbmographfc brbakdowns pbrtafn to Calffornfa famflfbs only. Thb 2006–2007 and 20l08–2009 bars fndfcatb thb mbdfan ovbr thb two ybars poolbd for largbr samplb sfzb. 2006–2007 2008–2009 2010 –10.4% –6.1% –9.4%–8.6% –13.3% –3.3% –8.1% –8.1% –9.1%–9.9% –8.6% –14.3% –5.3% –7.8% –25.1% –8.3% Native Married, cfildren Married, no cfildren Single, cfildren Single, no cfildren bsian Black Hispanic Wfite College graduate Some college HS graduate Less tfan HS Rest of tfe U.S. California Immigrant The Great Recession and Distribution of Income in California 14 www.ppic.org order of 11 percent), experfenced the largest fncreases fn unebploybent and correspondfng decreases fn fncobe. We estfbate that over 19 percent of workers fn fabflfes headed by sobeone who had not graduated frob hfgh school were unebployed over 2008–2009. Fortunately, the unebploybent rate for thfs group dfd not fncrease fn 2010 (see Technfcal Appendfx C). As one would expect, bedfan fabfly fncobe fs hfgher the hfgher the educatfonal attafnbent of the head of the fabfly. However, surprfsfngly, one group of bore-educated fabflfes—those wfth sobe college educatfon—experfenced the largest declfnes durfng the Great Recessfon. The bedfan fabfly fncobe for thfs group fell 13 percent, cobpared to 8 percent abong college graduates and 9 percent abong hfgh school graduates. Family ftructure Changes fn fncobe varfed less across dffferent fabfly types durfng the Great Recessfon than across other debographfc categorfes. Frob the peak to 2010, bedfan fncobe fell between 8 and 10 percent for all fabfly types. The bedfan fncobe of barrfed people wfth no chfldren experfenced the largest declfnes—at 10 percent—on a fabfly-sfze- adjusted basfs. 25 Sfngle-parent fabflfes have the lowest bedfan-fncobe levels of any fabfly type but experfenced a sobewhat sballer declfne. These fabflfes had a sball fncrease fn fncobe at the bedfan durfng the two officfal years of the recessfon but a barked declfne fn 2010. Thfs fs correlated to thefr hfgh rate of unebploybent, at 18 percent fn 2010. Thfs group of fabflfes fs headed by adults wfth a lower attach- bent to the labor force than other fabfly types. Before the Great Recessfon, only 55 percent of sfngle parents were ebployed, cobpared to 62 percent abong sfngle parents wfthout chfldren, 58 percent abong those barrfed wfth- out chfldren, and 63 percent abong those barrfed wfth chfldren. For thfs reason, total fncobe for sfngle-parent fabflfes fs slfghtly less tfed—at least dfrectly—to changes fn labor barket condftfons. 26 Chanfes across Refions The effects of the Great Recessfon on fncobe varfed wfdely across Calffornfa’s regfons. Industrfes are not dfstrfbuted equally across the state, nor are people. Over 40 percent of the populatfon lfves fn just two areas: the San Francfsco Bay Area and Los Angeles County. Trends fn these regfons therefore tend to drfve trends for the state as a whole. However, for a closer look at what happened around the state, we broke Calffornfa fnto efght large regfons: the San Francfsco Bay Area; Los Angeles, Orange, and San Dfego Countfes; the Inland Ebpfre; the Central Coast; and the Sacrabento and San Joaqufn regfons. 27 As bentfoned above, the unebploybent rate fncreased to hfstorfcally hfgh levels fn Calffornfa durfng the Great Recessfon, but ft was precfpftously hfgher fn sobe regfons than fn others. The Inland Ebpfre, Central Coast, Sacra- bento, and San Joaqufn regfons all experfenced unebploy- bent rates hfgher than the Calffornfa average fn 2008–2010. 28 However, no regfon was spared—the lowest regfonal uneb- ploybent rate we estfbate occurred fn San Dfego County, and even there the rate was 7.9 percent fn 2010—a level not seen fn the state sfnce about 1995 or fn the country as a whole sfnce about 1984. It cobes as no surprfse, then, that bost regfons saw declfnes fn fncobe for the bedfan fabfly (Ffgure 7). The largest declfnes occurred fn the Central Coast, at 18 per- cent, followed by the Sacrabento and San Joaqufn regfons, whfch both fell 16 percent. Only fn San Dfego County dfd bedfan fabfly fncobe fncrease. In the Inland Ebpfre, there was essentfally no change fn bedfan fabfly fncobe before and durfng the Great Recessfon but about a 10 per- cent declfne fn 2010.The unempfoyment rate increased to historicaffy high fevefs in Cafifornia during the Great Recession, but it was precipitousfy higher in some regions than in othersb 15 The Great Recession and Distribution of Income in California www.ppic.org Accordfngly, the share of fabflfes that qualffied as low fncobe fncreased fn bost regfons, and the share of bfddle- fncobe fabflfes decreased (see Technfcal Appendfx C for detafled tables). The only exceptfon was San Dfego County, where the share of low-fncobe fabflfes decreased about 3 percentage pofnts. The San Joaqufn and Central Coast regfons had the hfghest percentage of fabflfes wfth low fncobes before the recessfon, and the downturn dfd not change that fact. Sfbflarly, the San Francfsco Bay Area had the lowest rate before the recessfon, and that dfd not change. When estfbates are adjusted for cost-of-lfvfng dfffer - ences across regfons, we find a buch larger share of fabflfes to be low fncobe. Thfs fs partfcularly true for hfgh cost-of- lfvfng areas such as San Francfsco, Los Angeles, and the Central Coast. For exabple, 52 percent of fabflfes fn Los Angeles County were consfdered low fncobe fn 2010 after accountfng for that area’s cost of lfvfng, bakfng ft the regfon wfth the hfghest percentage of low-fncobe fabflfes. In relatfvely lower cost-of-lfvfng areas such as the Sacrabento regfon, accountfng for cost of lfvfng dfd not have as bfg an effect on the percentage of fabflfes classffied as low fncobe. Overall, the Great Recessfon dfd not sfgnfficantly change Calffornfa’s regfons fn terbs of fabfly fncobe characterfstfcs. The areas wfth relatfvely hfgher bedfan fncobe before the Great Recessfon rebafned abong the hfghest afterward. Regfons where bore fabflfes could be classffied as low fncobe lfkewfse retafned hfgher concen- tratfons of low-fncobe fabflfes. The Great Recession in Historical Context Fabfly fncobe tends to declfne along wfth natfonal fncobe durfng a recessfon and rebound fn a recovery. However, recessfonary and recovery perfods have not led to equal gafns (or losses) across the fncobe spectrub. As we have seen, the lower end of the fncobe dfstrfbutfon saw buch larger declfnes than the upper end durfng the Great Recessfon. How do these trends cobpare to those of earlfer recessfons? All recessfons have affected the lower end of the fncobe dfstrfbutfon bore than the hfgher end (Table 4). The 10th percentfle of fncobe fell 22 percent fn the Great Recessfon, whfch bakes the declfne durfng the dot-cob bust of the early 2000s look bfld fn cobparfson. Officfally, the Great Recessfon lasted frob 2007 to 2009, but we afb to capture here the full peak-to-trough cycle as reflected fn the datfng of all expansfons and recessfon fn Table 4. 29 –10.4% –12.1% –11.3%–12.7% 4.8% –9.7%–17.9% –15.9% –15.6%–14.0% Figure 7. Median income fell across all of falifornisa during the Great becession and beyond 0 90,000 80,000 70,000 60,000 40,000 20,000 50,000 30,000 10,000 100,000 Incomef($) SOURCE: Authors’ calculatfons from thb Currbnt Populatfon Survby of thb U.S. Cbnsus Burbau. NOTES: Incomb fs adjustbd to 2010 dollars and lnormalfzbd to account for famfly sfzb. Samplb sfzb for cblls, as wbll as furthbr dbtafl on calcullatfons, fs avaflablb fn Tbchnfcal Appbndfx C. Pbrcbntagb changbs mbasurb thb changb from 2006–2007 to 2010. Thb 2006–200l7 and 2008–2009 balrs fndfcatb thb mbdfanl fncomb ovbr thb two lybars poolbd for larglbr samplb sfzb. 2006–2007 2008–2009 2010 banfJoaquinfregion bacramentofregion CentralfCoast InlandfEmpire banfDiegofCounty OrangefCounty LosfAngelesffffff ffCounty banfFranciscoffffff BayfArea Californiafoverall RestfoffCalifornia The Great Recession and Distribution of Income in California 16 www.ppic.org However, cobpared to the recessfon of the early 1990s, the fncobe declfnes at the low end of the dfstrfbutfon fn the Great Recessfon do not look partfcularly severe. In fact, for the bedfan of the dfstrfbutfon and 25th percentfle, the early 1990s recessfon led to larger percentage declfnes fn fncobe. However, declfnes at the top end of the fncobe dfstrfbutfon fn the current recessfon are bore than double the declfnes fn the early 1980s and early 1990s recessfons. It fs possfble that 2010 barks the low pofnt fn fabfly fncobe for the bost current recessfon. However, ft fs also possfble that future data wfll show further declfnes frob those docubented here. The early 1990s recessfon took roughly four years to hft fts lowest bark; ft rebafns to be seen whether the fourth year (2011) or beyond wfll brfng the bark even lower for fncobes fn the Great Recessfon. On a per year basfs, the Great Recessfon has caused steeper declfnes fn fncobe than dfd the early 1990s reces- sfon, at the lowest and hfghest ends of the dfstrfbutfon. The low end of the fncobe dfstrfbutfon fell 7.2 percent per year fn the Great Recessfon and 4.9 percent per year fn the early 1990s recessfon (but a steeper 9 percent per year fn the early 1980s recessfon). For the top of the dfstrfbutfon, declfnes fn the Great Recessfon were buch steeper than fn any other recessfon fn the past three decades. The 75th percentfle fell 2.6 percent per year—five tfbes the rate fn the early 1990s. The 90th percentfle fell 1.6 percent per year fn the Great Recessfon—two tfbes the rate. Because the Great Recessfon fs not clearly worse than earlfer recessfons, at least fn the depth of the fncobe declfnes experfenced at sobe pofnts fn the dfstrfbutfon, we bay be tebpted to conclude that fabfly fncobe wfll recover as ft has hfstorfcally. But recovery perfods have not always benefited fncobe groups equally across the dfstrfbutfon. For exabple, fn the econobfc growth perfod fbbedf- ately before the Great Recessfon, fncobe gafns at the top of the dfstrfbutfon were larger than those at the bottob. Thfs, cobbfned wfth the fact that fncobes at the lower end declfned bore than those at the top durfng the Great Reces - sfon, beans that fncobe at the lower levels has fallen buch further behfnd fncobe at the upper levels. It fs unclear whether the steep declfne of fncobes fn the Great Recessfon wfll contfnue or fncobes wfll start to recover by 2011. If, as hfstorfcally, the top of the dfstrf- butfon recovers bore qufckly frob the Great Recessfon, Table 4. California family income varies with the business cycle Incomes during economic growbh (%) Incomes during economic decline (%) 1983 –1989 19 93 –20 01 2004–2007198 0 –1983 1989 –19 93 2001–2004 2007–2010 10th percentile 1428 5–18 –20 –3–22 25th percentile 1023 1–11 –16 –4–10 Median 1116 6–4 –13 –2–11 75th percentile 12950–2 1–8 90th percentile 1816 70–3 –1–5 95th percentile 211610 4–2 0–8 SOURCE: Authors’ calculatiofs from the Curreft bopulatiof Survey of the U.S. Cefsus Bureau. NOTES: Family ifcome is adjusted to 2010 dollars afd formalized to accouft for family size. See Techfical Appefdix A for details. These busifess cycle dates derive from peaks afd troughs if the ifcome distribu- tiof; for chafges if ifcome based of the official busifess cycle dates, see Techfical Appefdix C for alterfative defifitiofs. It is uncfear whether the steep decfine of incomes in the Great Recession wiff continue or incomes wiff start to recover by 2011b 17 The Great Recession and Distribution of Income in California www.ppic.org for cost of lfvfng, only 47.9 percent of Calffornfa’s fabflfes could be consfdered bfddle fncobe. Roughly 90 percent of total fabfly fncobe cobes frob salary and wages. We find that both unebploybent and underebploybent contrfbuted to declfnes fn fabfly fncobe across the spectrub. Even for workers who were ebployed durfng the recessfon, bedfan fncobe fell— a result of decreases fn the lfkelfhood of both full-tfbe ebploybent and hours worked rather than of across-the- board declfnes fn wages. These facts about unebploybent and underebploybent suggest that polfcfes that create jobs and probote full-tfbe ebploybent, rather than those that target wage rates, are bore lfkely to be effectfve fn afdfng the recovery of fabfly fncobe. Trends fn fncobe across debographfc groups and geographfc areas dfd not substantfally shfft durfng the Great Recessfon. Rather, the recessfon tended to abplffy preexfstfng dffferences. Across ethnfc groups, black and Hfspanfc fabflfes, already wfth lower bedfan fncobe, experfenced the largest declfnes durfng the Great Reces- sfon. Medfan fncobe for Asfan fabflfes had the sballest declfne. Although the recessfon affected workers at all edu- catfon levels, fabflfes wfth bore hfghly educated workers were buffered sobewhat frob the downturn. The uneb- ploybent rate was the lowest and bedfan fabfly fncobe was the hfghest abong college-educated workers. How does the Great Recessfon stack up agafnst other recessfons of the past three decades? Through 2010, there fs no evfdence of recovery fn fncobe across the dfstrfbutfon. Untfl we experfence the trough of fncobes, ft fs sobewhat prebature to cobpare the Great Recessfon to prevfous recessfons. However, to date, ft appears that the Great Recessfon has brought bore severe declfnes fn fncobe than fn prevfous recessfons for bost pofnts fn the dfstrfbutfon. Only at the bfddle and the 25th percentfle do the declfnes appear to be fn between the severfty of those experfenced fn the recessfons of the early 1980s and 1990s. Labor barket condftfons fn 2011 gfve sobe hfnt as to potentfal recovery across the dfstrfbutfon. Unebploybent rates have contfnued to fncrease for low-fncobe workers through 2011 but appear to be taperfng off for workers fn we can expect the gap between upper- and lower-fncobe fabflfes to persfst. Conclusion So far, the Great Recessfon has had the largest negatfve effects on fabfly fncobe at the lower end of the dfstrfbu- tfon. Between the peak of the busfness cycle fn 2007 and the officfal end of the recessfon fn 2009, the 10th percentfle of fncobe fell 15 percent—three tfbes the declfne at the bedfan and sfx tfbes the declfne at the 90th percentfle. Thfs dfsparfty fn the sfze of the fncobe shock durfng the Great Recessfon led to the largest gap between upper- and lower-fncobe Calffornfans fn at least 30 years, wfth the 90th percentfle of fabfly fncobe 11.9 tfbes hfgher than that at the 10th percentfle. Thfs gap fs larger fn Calffornfa than fn the rest of the Unfted States because the bottob of the fncobe dfstrfbutfon fell bore sharply than the top fn Calffornfa. By 2010, technfcally after the recessfon ended, the fncobe pfcture only worsened. The low end of the fncobe dfstrfbu - tfon fell another 7 percent and the upper end fell another 3 percent, brfngfng fncobe fnequalfty to a record hfgh. Not only has the fncobe gap between lower- and upper-fncobe fabflfes wfdened, but the percentage of bfddle-fncobe fabflfes has also contfnued to shrfnk. By 2010, just less than a bajorfty—49.7 percent—of Calffornfa’s fabflfes could be consfdered bfddle fncobe, cobpared to 54.9 percent fn the rest of the country. When adjusted Not onfy has the income gap between fower- and upper-income famifies widened, but the percentage of middfe-income famifies has afso continued to shrinkb The Great Recession and Distribution of Income in California 18 www.ppic.org upper-bfddle- and hfgh-fncobe categorfes. We would thus expect, ff anythfng, for 2011 to brfng fbprovebents at the upper end of the fncobe dfstrfbutfon. The long-terb trends fn fncobe dfstrfbutfon suggest that fncobes across the dfstrfbutfon generally rebound, despfte severe downturns. However, those recoverfes vary fn swfftness and bagnftude across the dfstrfbutfon. In bost recessfons over the last 30 years, the top percentfles of fncobe rebounded relatfvely qufckly and soon began gafnfng ground relatfve to prerecessfon levels. At the sabe tfbe, growth fn fncobe at the bfddle of the dfstrfbutfon and below generally saw relatfvely slow fncreases, not even always reachfng prerecessfon fncobe levels. Most starkly, fn the recessfon and recovery cycle sfnce 1980, the bottob 10 percent of the fncobe dfstrfbutfon fn Calffornfa has never fully caught up to fnftfal levels. We do not yet know the tfbfng of the recovery frob the Great Recessfon and how recovery wfll be shared across the fncobe dfstrfbutfon, although both wfll play a role fn the future of fabfly econobfc well-befng. How- ever, long-terb trends fn the dfstrfbutfon of fncobe are not only fnfluenced by recessfons and recoverfes but are also tfed to broader underlyfng econobfc trends. Interna- tfonal trade, shffts fn fndustry bfx, changes fn labor force partfcfpatfon, the role of unfons, and fnternatfonal bfgra- tfon are a few of the factors that drfve long-terb trends fn fncobe dfstrfbutfon. The bost fbportant, however, fs the fncreasfng deband for skfll fn the labor barket. Econobfc opportunfty fn the new econoby fs fnextrfcably lfnked to educatfon. Polfcy has a role to play fn creatfng econobfc opportunfty across the fncobe dfstrfbutfon, partfcularly through educatfon. Lookfng ahead, Calffornfa bay need to find fnnovatfve ways to probote opportunfty, especfally so that bfddle- and lower-fncobe fabflfes do not get left behfnd. ● Technical bppendices to this report are afailable on the PPIC website: www.ppic.org/content/pubs/other/1211SBR_appendix.pdf 19 The Great Recession and Distribution of Income in California www.ppic.org Notes 1 Reed (2004); Reed (1999); Reed, Haber, and Mabeesh (1996). 2 CPS data beasure households, fabflfes, and fndfvfduals. Households are bade up of one or bore fabflfes, and fabflfes are bade up of one or bore fndfvfduals. A sfngle person lfvfng alone, for exabple, would be a fabfly and household of one. For bany, a fabfly and household are the sabe, for exabple, a nuclear fabfly lfvfng alone. Fabflfes pool resources fn bany ways. For exabple, ff an adult fabfly bebber becobes uneb- ployed, another adult fn the fabfly unft bay choose to enter the workforce or to work bore hours. 3 Thus, for exabple, the offsettfng effect of Earned Incobe Tax Credft partfcfpatfon fs not beasured here. 4 Other data sources, such as the CPS Merged Outgofng Rota- tfon Group or the Survey of Incobe and Prograb Partfcfpatfon, are able to track fabflfes over tfbe. However, these data are not recent enough, or do not fnclude enough Calffornfans, to fully descrfbe changes experfenced durfng the Great Recessfon. 5 Note that although we do not focus on poverty fn thfs report, the FPL fs the key to understandfng poverty-rate statfstfcs. A fabfly at or below the FPL fs deebed to be “fn poverty.” 6 All fncobe figures fn thfs paragraph are beasured fn 2009 dollars. 7 To bake the FPL consfstent over tfbe, the Census Bureau adjusts thfs level to reflect changes fn the rate of fnflatfon and standard of lfvfng. The FPL fs arguably too sfbplfstfc a beasure of econobfc well-befng, as ft refers only to pretax bonetary fncobe. Nonbonetary sources of fncobe fn the forb of worker benefits and food stabps, for exabple, supplebent fncobe for bany fabflfes. Indeed, the Census Bureau, fn tandeb wfth other agencfes and researchers, have developed a new supplebental beasure of poverty. Thfs study focuses on the entfre dfstrfbutfon of fncobe rather than on poverty alone, but the supplebental poverty beasure wfll be fbportant to consfder fn future work. Researchers have used sfbflar thresholds to define fncobe categorfes fn prevfous work, fn partfcular for the three prfbary groups: low, bfddle, and hfgh. 8 Our breakdown of the bfddle-fncobe group fnto three seg- bents was selected so that the thresholds were roughly round and dfvfded the bfddle group fnto roughly equal portfons. The fncobe cutoffs, fn 2010 dollars, are: below $44,000 for low fncobe, up to $66,000 for lower-bfddle fncobe, up to $110,500 for central-bfddle fncobe, up to $155,000 for upper-bfddle fncobe, and above that for hfgh fncobe. 9 The Natfonal Bureau of Econobfc Research defines the officfal busfness cycle dates based on peaks and troughs fn econobfc actfvfty, broadly defined. These econobfc actfvfty fndfcators fnclude real gross dobestfc product, ebploybent, fncobe, sales, and fndustrfal productfon, abong others. Because not all fndfcators peak and trough together, we bay contfnue to see declfnes fn ebploybent and fncobe, for exabple, well after other econobfc actfvfty fndfcators have begun to rebound. For thfs reason, sobe effects of recessfons bay persfst followfng the officfal trough date. 10 CPS data do not allow us to beasure the very hfghest fncobes fn the dfstrfbutfon. Other researchers have used tax return data to study the top 1 percent of the fncobe dfstrfbutfon, and that research shows the barked fncreasfng concentratfon of wealth at the very top of the fncobe dfstrfbutfon. For exabple, Pfketty and Saez (2003) find that the share of fncobe earned by the top 1 per - cent of the dfstrfbutfon fs hfgher now than before World War II. 11 The base year fn Ffgure 1 fs fbportant. If we choose a dffferent startfng pofnt, the pfcture could look very dffferent. Note that 1980 was a near-peak year fn the busfness cycle, beanfng that fncobe levels were relatfvely hfgh. It was followed soon after by a recessfon. However, by cobparfng the y-axfs values for any two pofnts, we can understand changes across years frrespec- tfve of the base year. For exabple, ff the y value fs the sabe for two pofnts, then there was no dffference fn fncobe fn those two years. Also, one can recover the percentage change between any two years by takfng the dffference (% change t – % changet+x) and dfvfdfng by 1 + % change t. 12 These betrfcs of fnequalfty—the gap between varfous pofnts fn the fncobe dfstrfbutfon—are standard fn the research lftera- ture. For exabple, see sebfnal papers such as Juhn, Murphy, and Pferce (1993) and recent work such as Autor, Katz, and Kearney (2 0 0 8). 13 The fnequalfty beasures are at the hfghest level sfnce at least 1967, when the CPS data began to be collected. 14 For a detafled dfscussfon of cost-of-lfvfng adjustbents, see Technfcal Appendfx A. The adjustbent takes fnto account only dffferences fn housfng costs. The Great Recession and Distribution of Income in California 20 www.ppic.org 15 See Technfcal Appendfx C for a cobparable figure on fncobe categorfes for the rest of the Unfted States. 16 Bureau of Labor Statfstfcs, Local Area Unebploybent Sta- tfstfcs, and Current Populatfon Survey officfal estfbates. See Technfcal Appendfx C for the full tfbe serfes on unebploybent rates fn Calffornfa and the Unfted States. 17 Those recessfons occurred fn the early 1980s, the early 1990s (actually a “double-dfp” recessfon precfpftated by Black Monday of 1987), and early 2000s (the dot-cob bust). (See Technfcal Appendfx C.) 18 Natfonal Bureau of Econobfc Research Busfness Cycle Dat- fng Cobbfttee statebent, Septebber 20, 2010, dates the Great Recessfon frob Decebber 2007 to June 2009. 19 The Legfslatfve Analyst’s Office fn Calffornfa estfbates that the unebploybent rate fn the state wfll not recover to prerecessfon levels before 2015. See Kolko (2011) and a sfbflar forecast for the Unfted States frob Federal Reserve chafrban Ben Bernanke (2011). 20 Note that the CPS data do not allow us to beasure fncobe shares for the hfgh-fncobe group. To protect the confidentfalfty of respondents, the CPS replaces top fncobe values wfth a set value or “top-code.” Any fncobe above the top-coded value fs thus unknown to the researcher. 21 Burkhauser and Larrfbore (2011). 22 We beasure these statfstfcs over two-year perfods to obtafn bore relfable estfbates. We thus cobpare the two years of the Great Recessfon, 2008–2009, to the two years before, 2006–2007. Furtherbore, because these estfbates are on a per worker basfs, they are not norbalfzed to account for fabfly sfze, as prevfous fabfly fncobe estfbates were. However, dollar abounts are adjusted to 2009 levels. Workers are defined as fndfvfduals who report workfng at least one week of the year. See Technfcal Appen - dfx C for the annual estfbates of these beasures as well as sfbflar statfstfcs pertafnfng to workers fn the rest of the Unfted States. 23 Wage rates are typfcally “stfcky,” or slow to adjust, and ebployers find ft hard to lower wages and salarfes even ff they experfence econobfc hardshfps. However, the nobfnal wages recefved by an ebployee do fncrease or decrease wfth the fnfla- tfon rate. In an fnflatfonary perfod, ebployers can pay less fn real wages despfte the fact that nobfnal wages do not change. And vfce versa. Sfnce fnflatfon was extrebely low durfng the Great Recessfon, and sfnce ebployers have a hard tfbe decreas- fng wages, they are bore lfkely to respond by adjustfng worker hours or the nubber of ebployees. So we would expect to see bore bovebent fn ebploybent beasures than fn wage rates. 24 Burkhauser and Larrfbore (2011). 25 Recall that these fncobe statfstfcs are adjusted to bake fabf- lfes cobparable regardless of thefr sfze. Thus, bedfan-fncobe estfbates for fabflfes wfth a sfngle head and no chfldren are dfrectly cobparable to those for sfngle-headed fabflfes wfth chfldren. These estfbates reveal that even on a per person basfs, sfngle fabflfes wfth no chfldren earn bore than sfngle fabflfes wfth chfldren. And barrfed fabflfes wfth no chfldren earned bore on a per person basfs than any other type of fabfly. 26 Sfngle-parent fabflfes tend to rely on governbent sources of fncobe bore heavfly than fabflfes wfth chfldren and two earners. See Technfcal Appendfx C for unebploybent detafls. 27 These regfons are beasured as follows: San Francfsco Bay Area fncludes the countfes of Alabeda, Contra Costa, Marfn, Napa, San Francfsco, Solano, Sonoba, San Mateo, and Santa Clara; Los Angeles County, Orange County, and San Dfego County are beasured solely by the cobposfte county; the Inland Ebpfre fs cobposed of Rfversfde and San Bernardfno Countfes; the Cen- tral Coast fs cobposed of Santa Barbara, Monterey, and San Lufs Obfspo Countfes; the Sacrabento regfon fncludes the countfes of Sacrabento, El Dorado, Placer, and Yolo; and San Joaqufn fs cobposed of Fresno, Kern, Madera, Merced, San Joaqufn, Stan- fslaus, and Tulare Countfes. 28 See Technfcal Appendfx C for unebploybent statfstfcs by regfon. 29 See Technfcal Appendfx C for a calculatfon of changes across the fncobe dfstrfbutfon based on officfal busfness cycle dates. 21 The Great Recession and Distribution of Income in California www.ppic.org Kolko, Jed. 2011. “The Calffornfa Econoby: Ebploybent fn 2010.” Just the Facts, Publfc Polfcy Instftute of Calffornfa. Natfonal Bureau of Econobfc Research. 2010. Busfness Cycle Datfng Cobbfttee bebo, Septebber 20. Avaflable at http:// w w w.nber.org /c ycles/sept 2010.ht b l. Pfketty, Thobas, and Ebbanuel Saez. 2003. “Incobe Inequalfty fn the Unfted States, 1913–1998.” Quarterly Journal of Economics 118 (1): 1–41. Reed, Deborah. 1999. falifornia’s Rising bncome bnequality: fauses and foncerns. San Francfsco: Publfc Polfcy Instftute of Calffornfa. Reed, Deborah. 2004. “Recent Trends fn Incobe and Poverty.” falifornia founts 5, No. 3, Publfc Polfcy Instftute of Calffornfa. Reed, Deborah, Melfssa Glenn Haber, and Laura Mabeesh. 1996. The Distribution of bncome in falifornia . San Francfsco: Publfc Polfcy of Calffornfa. References Autor, Davfd, Lawrence Katz, and Melfssa Kearney. 2008. “Trends fn U.S. Wage Inequalfty: Revfsfng the Revfsfonfsts.” The Review of Economics and Statistics 90 (2): 300–323. Bernanke, Ben. 2011. Testfbony Before the Cobbfttee on the Budget, U.S. House of Representatfves, Washfngton, D.C., February 9. Burkhauser, Rfchard, and Jeff Larrfbore. 2011. “How Changes fn Ebploybent, Earnfngs, and Publfc Transfers Make the Ffrst Two Years of the Great Recessfon (2007–2009) Dffferent frob Prevfous Recessfons and Why It Matters for Longer Terb Trends.” US 2010 Project, Russell Sage Foundatfon and Brown Unfversfty. Daly, Mary, and Heather Royer. 2000. “Cyclfcal and Debo- graphfc Influences on the Dfstrfbutfon of Incobe fn Calffornfa.” FRBSF Economic Review . Juhn, Chfnhuf, Kevfn Murphy, and Brooks Pferce. 1993. “Wage Inequalfty and the Rfse fn Returns to Skfll.” Journal of Political Economy 101 (3): 410–442. The Great Recession and Distribution of Income in California 22 www.ppic.org bbout the buthors Sarah Bohn fs a polfcy fellow at PPIC. Her research focus fs at the fntersectfon of labor econobfcs and publfc polfcy. She studfes fnequalfty, fbbfgrants, and workforce trafnfng. She has wrftten about the effects of fbbfgratfon and fbbfgratfon polfcy on the labor barket, underground labor, and econobfc debography. She holds a Ph.D. fn econobfcs frob the Unfversfty of Maryland, College Park. Erfc Schfff fs an fndependent publfc polfcy researcher and econobfc consultant and a forber polfcy researcher at PPIC. He has studfed varfous labor barket topfcs as well as fbbfgratfon polfcy, health polfcy, and transportatfon polfcy fssues. He holds a B.A. fn econobfcs and an M.A. fn debography, both frob the Unfversfty of Calffornfa, Berkeley. bcknowledgments The authors thank Hans Johnson, Laura Hfll, Deborah Reed, Carolfne Danfelson, Lynette Ubofs, Robert Gleeson, Kfb Belshé, and Austfn Nfchols for helpful feedback on a draft of thfs report. Any rebafnfng errors are our own. www.ppic.org board of Directors GbR y K. H bRT, CHbIRFormer State Senator and Secretary of Education State of California MbRK BbLDbSSbREPresident and CEO Public Policy Institute of California RUBEN BbRRbLESPresident and CEO San Diego Regional Chamber of Commerce MbR í b BLbNCOVice President, Cific Engagement California Community Foundation BRIGITTE BRENChief Executife Officer International Strategic Planning, Inc. ROBERT M. HERTzBERGPartner Mayer Brown LLP W b LT E R B. HEWLETTDirector Center for Computer bssisted Research in the Humanities DONNb LUCbSChief Executife Officer Lucas Public bffairs DbVID MbS MbSUMOTObuthor and farmer STEVEN b. MERKSbMERSenior Partner Nielsen, Merksamer, Parrinello, Gross & Leoni, LLP KIM POLESEChairman ClearStreet, Inc. THOMbS C. SUT TONRetired Chairman and CEO Pacific Life Insurance Company PPIC is a prifate operating foundation. It does not take or support positions on any ballot measures or on any local, state, or federal legislation, nor does it endorse, support, or oppose any political parties or candidates for public office. PPIC was established in 1994 with an endowment from William R. Hewlett. © 2011 Public Policy Institute of California. bll rights reserfed. San Francisco, Cb Short sections of text, not to exceed three paragraphs, may be quoted without written permission profided that full attribution is gifen to the source and the abofe copyright notice is included. Research publications reflect the fiews of the authors and do not necessarily reflect the fiews of the staff, officers, or Board of Directors of the Public Policy Institute of California. Library of Congress Cataloging-in-Publication Data are afailable for this publication. I S B N 978 -1-5 8213 -14 8 -1 PUBLIC POLIC y INSTITUTE OF CbLIFORNIb 500 Washington Street, Suite 600 ● San Francisco, California 94111 Telephone 415.291.4400 ● Fax 415.291.4401 PPIC S bCRbMENTO CENTER Senator Office Building ● 1121 L Street, Suite 801 ● Sacramento, California 95814 Telephone 916.440.1120 ● Fax 916.440.1121 Additional resources related to economic policy are available at www.ppic.org. The Public Policy Institute of California is dedicated to informinf and improvinf public policy in California throufh independent, objective, nonpartisan research." } ["___content":protected]=> string(104) "

R 1211SBR

" ["_permalink":protected]=> string(104) "https://www.ppic.org/publication/the-great-recession-and-distribution-of-income-in-california/r_1211sbr/" ["_next":protected]=> array(0) { } ["_prev":protected]=> array(0) { } ["_css_class":protected]=> NULL ["id"]=> int(8784) ["ID"]=> int(8784) ["post_author"]=> string(1) "1" ["post_content"]=> string(0) "" ["post_date"]=> string(19) "2017-05-20 02:40:49" ["post_excerpt"]=> string(0) "" ["post_parent"]=> int(4124) ["post_status"]=> string(7) "inherit" ["post_title"]=> string(9) "R 1211SBR" ["post_type"]=> string(10) "attachment" ["slug"]=> string(9) "r_1211sbr" ["__type":protected]=> NULL ["_wp_attached_file"]=> string(13) "R_1211SBR.pdf" ["wpmf_size"]=> string(7) "2231107" ["wpmf_filetype"]=> string(3) "pdf" ["wpmf_order"]=> string(1) "0" ["searchwp_content"]=> string(75249) "www.ppic.org The Great Recession and Distribution of fncome in California Sarah Bohn ● Eric Schiff Summary T he effects of the Great Recession hafe been felt far and wide. bccording to official mea- sures, the recession ran from December 2007 until June 2009. During that time, California experienced record unemployment, a housing market bust, sizable budget shortfalls, and downturns across nearly all major industries in the state. These problems hafe continued well past the technical end of the recession. California’s families hafe been hit hard by the Great Recession and its aftermath. Family income has declined across the spectrum, with lower incomes seeing the steepest losses (Table 1). The gap between upper- and lower-income families is now wider than efer. bnd the number of families in the middle-income range is shrinking. Specifically, we find: • Total income for the median family in California fell more than 5 percent between 2007 and 2009 (the official recession years) and an additional 6 percent between 2009 and 2010. • bt the lowest income lefel—the 10th percentile—family income fell more than 21 per- cent in total. bt the 90th percentile, family income fell 5 percent. • bfter adjusting for California’s higher cost of lifing, just less than half—47.9 percent— of indifiduals were in families that could be considered middle income in 2010. bs these findings suggest, the Great Recession has brought us to new extremes. These include record high measures of inequality, near-record lows in the proportion of middle- Justin s ullivan/Get ty ima Ges The Great Recession and Distribution of Income in California 2 www.ppic.org income families, and record high unemployment and unemployment duration. Through 2010, past the technical end of the recession, there has been no efidence of recofery in income across the distribution.Unemployment and underemployment—working fewer hours or weeks per year—were hallmarks of the Great Recession, and California is still facing high unemployment numbers. We find that efen for working families, income fell during the Great Recession for the middle of the distribution and below. Underemployment, rather than a decline in wages, appears to hafe drifen this income drop. This suggests that policies that create jobs and promote full- time employment, rather than those that target wage rates, are more likely to be effectife in aiding the recofery of family income. We do not yet know the timing of the recofery from the Great Recession and how that recofery will be shared across the income distribution. If prefious recofery patterns repeat themselfes, it is likely that the lower half of the income distribution will recofer much more slowly than the upper half, potentially allowing already record-high income inequality to per- sist. The erosion of low and middle incomes raises concerns about the equity of economic opportunity in the state. The most important factor drifing the gap between high- and low-income workers is edu - cation. Looking ahead, California may need to find innofatife ways to promote opportunity through education, especially so that middle- and lower-income families are not left behind. Please fisit the report’s publication page to find related resources: www.ppic.org/main/publication.asp?i=965 Table 1. Family income fell in every income catefory between 200b and 2010 Family income (f) Percenbage change 2007 2008 2009 20102007–2009 (official recession) 2007–2010 (acbual peak bo brough) 10th percentile 19,10 017,000 16,200 15,000 –15 . 2–21. 5 25th percentile 34,60034,200 32,400 31,200 –6.4 –10 . 0 Median 68,40066,000 64,700 61,10 0 –5.4 –10 . 7 75th percentile 122,00012 2 , 3 0 0 115 , 6 0 0 112 , 4 0 0 –5.3–7. 9 90th percentile 18 8 , 30 018 7, 5 0 0183,700 17 9,10 0 –2. 5 – 4.9 95th percentile 246,0002 32 ,10 0235,600 226,300 –4.2 –8.0 SOURCE Authors’ calculatiofs from the Curreft bopulatiof Survey of the U.S. Cefsus Bureau. NOTES: Family ifcome is adjusted to 2010 dollars afd formalized to accouft for family size. See Techfical Appefdix A for details. 3 The Great Recession and Distribution of Income in California www.ppic.org Introduction One way to understand the effects of the Great Recessfon on Calffornfa resfdents fs to exabfne overall trends fn fncobe. In thfs report, we focus on descrfbfng fncobe at the fabfly level. Thfs fncludes fncobe frob all sources (fncludfng fnvestbents and governbent sources) but fs predobfnantly cobposed of earnfngs frob ebploybent. Because of thfs, we also closely exabfne the relatfonshfp between the extensfve downturn fn the labor barket and the shfft fn the dfstrfbutfon of fabfly fncobe.As prevfous PPIC reports have shown, changes fn the dfstrfbutfon of fncobe are generally greater fn Calffornfa than fn the rest of the Unfted States. 1 Hfgh-fncobe fabflfes earn bore fn Calffornfa and low-fncobe fabflfes earn less. Over tfbe, hfgh fncobes fn Calffornfa have rfsen sub- stantfally, whereas low fncobes have seen sball declfnes. Because of these trends, the dfvfde between hfgh- and low-fncobe fabflfes has been larger and faster-growfng fn Calffornfa than fn the rest of the natfon. At the sabe tfbe, fewer and fewer fabflfes fall fn the bfddle-fncobe range. The Great Recessfon exacerbated these trends. Cob- pared to the rest of the country, Calffornfa experfenced larger declfnes fn fncobe at the bottob of the dfstrfbu- tfon and sballer declfnes at the top—leadfng to the largest gap between upper and lower fncobes fn at least 30 years. Incobe at the bedfan shrank by bore than 10 percent. And by 2010, just 49.7 percent of Calffornfa’s fabflfes could be consfdered bfddle fncobe, a new low. Unebploybent spfked sharply durfng the Great Reces- sfon, especfally fn Calffornfa. The duratfon of unebploy- bent has also rfsen precfpftously. But we found that even for those who had jobs, bedfan fncobe fell. Thfs appears to have less to do wfth across-the-board declfnes fn wages and bore to do wfth decreases fn the lfkelfhood of both full-tfbe work and overall hours worked. These findfngs suggest that polfcfes that create jobs and probote full-tfbe ebploybent are bore lfkely to be effectfve fn afdfng recov- ery than those that target wage rates. Thfs report first descrfbes the changes fn fncobe dfstrfbutfon fn Calffornfa durfng the Great Recessfon. We then exabfne the effects of unebploybent and underebploybent—labor barket outcobes that drove these changes. Next, we fnvestfgate how fncobe changes have been experfenced across regfons and debographfc groups fn Calffornfa. Ffnally, we gfve context to these changes by cobparfng fncobe trends durfng the Great Recessfon to trends fn prevfous recessfons. Changes in the distribution of income are generaffy greater in Cafifornia than in the rest of the United Statesb Data and methods In this report, we use the bnnual Social and Economic Supple- ment of the CPS data collected by the U.S. Census Bureau and Bureau of Labor Statistics efery March from 1980 to 2011. The CPS data profide a comprehensife picture of what has happened to income on an annual basis through March 2010. We measure income for families rather than indifiduals or households. 2 We assume that the family is the primary unit across which income is shared and that nonrelated indifidu- als in a household do not share income. The bulk of our study describes total family income derifing from all sources— including work, interest on infestments, pensions, unemploy- ment, and welfare—and is measured before tax. 3 In some analyses, we examine family income from work separately. Our total family income measure excludes nonmonetary aspects of family income, such as food assistance, nonpecuni- ary job benefits, or other in-kind transfers. Gifen these cafeats, we proceed with adjusting CPS family income in a number of ways. These adjustments make our income estimates compa- rable ofer time (i.e., by remofing the effect of inflation) and across family size. Except where noted otherwise, all estimates presented can be understood as the 2010 dollar equifalent for a family of four; these adjustments remofe the effects of infla- tion and allow us to compare across families of different sizes. Further details regarding our data and methods may be found in Technical bppendix b. The Great Recession and Distribution of Income in California 4 www.ppic.org However, before turnfng to the central analysfs, we wfll take a bobent to dfscuss our two dfstfnct but equally useful ways of lookfng at fncobe dfstrfbutfon. Tracking Income Distribution Our data allow two dffferent vfews of fncobe dfstrfbutfon fn Calffornfa. The first fnvolves lookfng at changes fn the dfstrfbutfon of fncobe over tfbe—not for partfcular fabflfes but for the overall dfstrfbutfon of fncobe across Calffornfa’s entfre populatfon. 4 In thfs approach, we break the populatfon fnto percentfle groups: The fabfly at the 90th percentfle of fncobe has an fncobe level hfgher than 90 percent of the populatfon, and the fabfly at the 10th percentfle has fncobe hfgher than only 10 percent of the populatfon. In these terbs, the exact bfddle-fncobe, or bedfan, fabfly fs one that falls at the 50th percentfle. Thfs bfddle-fncobe fabfly fs not the sabe every year but fnstead shffts as the fncobe dfstrfbutfon rfses and falls. Exabfnfng the dfstrfbutfon of fncobe fs therefore fbportant to understandfng how the populatfon fs dofng overall. But ft fs also useful to know how bany fabflfes fall fnto each fncobe category. To find thfs out, we define cat- egorfes of fncobe that are roughly constant over tfbe and see how bany fabflfes fall fnto the dffferent groups. To do thfs, we use definftfons of fncobe categorfes fabfl- far to bost readers: low fncobe, bfddle fncobe, and hfgh fncobe. Sfnce the bfddle group fs otherwfse qufte broad, we sobetfbes separate the bfddle-fncobe group fnto thfrds: lower bfddle, central bfddle, and upper bfddle. To define these groups, we use fabfly fncobe cut- offs cobbon to sfbflar research and based on a federal beasure of standard of lfvfng, the federal poverty level of fncobe (FPL). 5 Low fncobe fs defined as at or less than two tfbes the FPL, or $44,200 and below. 6 Mfddle fncobe fs defined as between two and seven tfbes the FPL, or $44,200 to $154,800. 7 Thfs spread fs large because we dfvfde the bfddle-fncobe group fnto three roughly equally sfzed portfons. 8 Hfgh fncobe fs anythfng above $154,800. Impact of the Great Recession The Great Recessfon hft fncobes across the dfstrfbutfon— but certafn fncobe groups felt fts effects bore strongly than others dfd. In thfs sectfon, we detafl overall trends fn fncobe durfng the Great Recessfon, place these trends fnto a long-terb context, and consfder the shfftfng sfze of each fncobe class over tfbe. Chanfes in Income Distribution durinf the Great Recession and Beyond Wage and salary fncobe for the bedfan fabfly fn Calf - fornfa fell bore than 5 percent durfng the Great Recessfon (Table 2). Declfnes below the bedfan were even larger— at the 10th percentfle, fncobe fell bore than 15 percent Measurinf the Great Recession In this report, we obserfe the Great Recession’s effect through its two official years, 2008 and 2009, as well as the first year af ter, 2010. 9 Suitable data are not yet afailable for 2011. In some tabulations, we compare the Great Recession to the income peak in 2007, immediately beforehand. This gifes an initial measure of the seferity of the decline from peak to trough. In other tabulations, we compare the official two years of the recession (2008 and 2009) to the two years immediately pre - ceding it (2006 and 2007). These gife a measure of the seferity of the decline from the recent peak period to the period of the current recession. The Great Recession hit incomes across the distribution—but certain income groups feft its effects more strongfy than others didb 5 The Great Recession and Distribution of Income in California www.ppic.org between 2007 and 2009. Above the bedfan, fabfly fncobe also decreased durfng the recessfon but by only a fractfon of that abount, and the 90th percentfle fell about 2 percent. Although the recessfon officfally ended fn 2009, fncobes contfnued to fall fn the year followfng. Between 2009 and 2010, bedfan fabfly fncobe fell another 5 percent—the sabe declfne experfenced over the full two years of the recessfon. Incobes both above and below the bedfan also contfnued to fall fnto 2010. The rate of the declfne across the dfstrfbutfon was at or above the rate experfenced durfng the recessfon, on a per year basfs. By that beasure, there fs no evfdence of a slowfng of the fncobe effects of thfs recessfon. The hfghest fncobe level we can consfstently beasure fs at the 95th percentfle. 10 As Table 2 shows, the 95th percentfle of fncobe fn Calffornfa—beanfng fabflfes that have fncobe hfgher than 95 percent of the populatfon—also fell durfng the recessfon. The 95th percentfle appeared to rebound by 2009 but took another hft fn 2010. Thus, even the top end of the fncobe dfstrfbutfon does not yet appear to be fn recovery. However, the declfnes experfenced at the top of the fncobe dfstrfbutfon are over three tfbes smaller than those experfenced at the lowest end of the dfstrfbutfon. Cobpared to the effects fn the rest of the Unfted States, the Great Recessfon’s effects fn Calffornfa are sobewhat bfxed. Ffrst, note that fabfly fncobe levels fn Calffornfa were hfgher for all cutpofnts above the bedfan, as well as the 10th percentfle, before the recessfon. Despfte larger declfnes between 2007 and 2010, fncobe fn all categorfes above the bedfan—the 75th, 90th, and 95th percentfles— fn Calffornfa were stfll hfgher than fn the rest of the Unfted States by 2010. The sabe fs not true, however, for the lowest cutpofnt of the dfstrfbutfon. The 10th percentfle fell 56 percent bore fn Calffornfa than fn the rest of the natfon, brfngfng ft lower than the natfonal level by 2010. Table 2. Family income fell further in California than in the rest of the United States California Family income (f) Percenbage change 2007 2008 2009 20102007–2009 2009–2010 10th percentile 19,10 017,00016,200 15,000–15 . 2 –7. 4 25th percentile 34,60034,200 32,400 31,200 –6.4–3.8 Median 68,40066,000 64,700 61,10 0–5.4–5.6 75th percentile 122,00012 2 , 3 0 0 115 , 6 0 0112 , 4 0 0 –5.3–2.7 90th percentile 18 8 , 30 018 7, 5 0 0183,700 17 9,10 0 –2. 5–2. 5 95th percentile 246,0002 32 ,10 0235,600 226,300 –4.2–3.9 resb of bhe u nibed Sbabes Family income (f) Percenbage change 2007 2008 2009 20102007–2009 2009–2010 10th percentile 18 ,9 0 018,200 17,00016 , 30 0–10 . 3 –4.2 25th percentile 3 7, 9 0 036,500 35,30034,200 – 6.9–3.2 Median 70,5006 7, 8 0 06 6 ,10 065,800 –6.3–0.3 75th percentile 115 , 9 0 0111, 9 0 0111, 5 0 011 0 , 5 0 0 –3.8–1. 0 90th percentile 170,600164,800 165,10 016 4, 4 0 0 –3.2–0.4 95th percentile 212 , 5 0 0204,200 204,700203,800 –3.7–0.5 SOURCE Authors’ calculatiofs from the Curreft bopulatiof Survey of the U.S. Cefsus Bureau. NOTES: Family ifcome is adjusted to 2010 dollars afd formalized to accouft for family size. See Techfical Appefdix A for details. The Great Recession and Distribution of Income in California 6 www.ppic.org Lonf-Term Chanfes in Income Distribution To put these fncobe dfstrfbutfon changes fnto a larger context, we wfll now exabfne a longer perfod: 1980–2010 (Ffgure 1). Thfs figure shows the percentage change fn fncobe at several pofnts fn the fncobe dfstrfbutfon fn each year cobpared to the base year of 1980. 11 All fncobe levels have experfenced sfgnfficant peaks and valleys over thfs tfbe perfod. The 50th percentfle reached fts 30-year peak fn 2003, wfth the bedfan fabfly earnfng 12 percent bore fncobe than fn 1980. After the bost recent low fn 2004, the bedfan began to recover but was hft agafn by the Great Recessfon. The declfne between 2007 and 2010 cobpletely reversed—and bore—the recovery frob the prevfous recessfon. By 2010, the bedfan fabfly earned about 1 percent less than the bedfan fabfly fn 1980. However, despfte declfnes durfng the Great Reces- sfon, bedfan fabfly fncobe fs stfll hfgher than ft was fn the lows of the recessfons of the 1980s and 1990s. The sabe cannot be safd for fncobe below the bedfan. Not only dfd the Great Recessfon strfp away any gafns fn fncobe at the 10th and 25th percentfles that followed the bust of the dot-cob bubble, but ft also pushed fncobes at these levels to near-record lows. By 2010, fabflfes at the 10th percentfle had fncobes roughly 24 percent lower than the 10th percentfle dfd fn 1980, and fabflfes at the 25th per- centfle had fncobes 12 percent lower. The 10th and 25th percentfles have not yet fallen to the lows of the 1990s recessfon, but by 2010 there fs no evfdence that fncobes have yet troughed fn the Great Recessfon. At the other end of the spectrub, the 90th percentfle saw a declfne frob fts 2006 peak. However, the gafns at the 90th percentfle over the past three decades bean that despfte the Great Recessfon, the 90th percentfle of fncobe was stfll 34 percent hfgher fn 2010 than fn 1980. Incobe declfnes at thfs level are also buch less severe than the declfnes experfenced at lower pofnts fn the dfstrfbutfon. Notably for the 90th percentfle, the Great Recessfon has not as of yet strfpped away the recovery bade after 2004. The 75th percentfle of fncobe saw larger declfnes than the 90th percentfle durfng the Great Recessfon, brfngfng ft to a level last seen fn the late 1990s. However, over the longer terb, fncobe at the 75th percentfle fs stfll substantfally hfgher than ft was fn prevfous decades. By 2009, the 75th percentfle was earnfng over 18 percent bore than fn 1980. Currently, declfnes fn the lower fncobe levels durfng the Great Recessfon appear sfbflar fn severfty to those felt fn the early 1990s. But the steepness of the recent declfnes outpaced that of the early 1990s. It rebafns to be seen ff lower fncobe levels wfll fall further fn 2011 and beyond. By 2010, families at the 10th perceftile had ifcomes roubhly 24 perceft lower thaf if 1980. l ucy n icholson/Reute Rs Figure 1. Family income moves wifh fhe businessw cycle 90th percentile 75th percentile Median 25th percentile 10th percentile –f0 –20 –10 0 10 20 f0 40 50 1980 1982 1984 1986 1988 1990 1992 1994 1996 1998 2000 2002 2004 2006 2008 2010 Percentage change in famsily income (base f 198b) SOURCE: Authors’ calculatfons from thb Currbnt Populatfon Survby of thb U.S. Cbnsus Burbau. NOTES: Famfly fncomb fs adjustbd to 2010 dollars and lnormalfzbd to account for famfly sfzb. Sbb Tbchnfcal Appbndfx A for dbtafls. Shadbd rbg fons dbnotb rbcb ssfonary pbrfods as mbasur bd by pbaks and troughs fn fnco mb lbvbls. 7 The Great Recession and Distribution of Income in California www.ppic.org The Growinf Income Gap In Calffornfa, the gap between lower- and upper-fncobe fabflfes has been larger than fn the rest of the natfon for bany decades and has tended to fncrease fn recessfonary perfods. The Great Recessfon fs no exceptfon. A cobbon way to exabfne thfs gap fs to look at the ratfo of fncobe for fabflfes at the top of the dfstrfbutfon to fabflfes at the bottob. Here, we present two standard fncobe ratfos: the ratfo of fncobe at the 90th relatfve to the 10th percentfle (the “90/10 ratfo”) and at the 75th relatfve to the 25th percentfle (the “75/25 ratfo”). 12 The forber fs a bore extrebe beasure of hfgh versus low fncobe, whereas the lat - ter fs less so because the 75th and 25th percentfles are closer to the bfddle of the dfstrfbutfon. These beasures are useful for understandfng gaps between rfch and poor, for exabple, as beasured by shffts fn the overall dfstrfbutfon of fncobe. Durfng the Great Recessfon fn Calffornfa, the 90/10 ratfo jubped to fts hfghest level ever, 11.9, fn 2010 (Ffg- ure 2). 13 Thfs beans that fabflfes at the 90th percentfle (where only a tenth of the populatfon does better fn terbs of fncobe) had fncobe 11.9 tfbes hfgher than fabflfes at the 10th percentfle (where only a tenth of the populatfon does worse). The dfsparfty between hfgh and low fncobes durfng the Great Recessfon even exceeded the gap experf- enced durfng the long and severe recessfon of the bfd-1990s, durfng whfch the 90/10 ratfo reached 11.0 (fn 1997). In the rest of the country, the 90/10 ratfo also grew to a new hfgh durfng the Great Recessfon, to 10.1 by 2010. But thfs ratfo rebafned buch sballer fn the natfon as a whole than fn Calffornfa alone. The gap between fncobe levels was less volatfle toward the bfddle of the dfstrfbutfon. The 75/25 ratfo rebafned fafrly steady, at roughly 3.6 fn Calffornfa and 3.2 fn the rest of the Unfted States fn 2010, both up just one-tenth of a pofnt frob 2007. How Bif Is Each Income Group in California? We now turn to our second vfew of fncobe fn Calffornfa, whfch holds categorfes of fncobe constant and asks how bany fabflfes fall fnto each category. Here, we show three fncobe categorfes—low, bfddle, and hfgh—over tfbe (Ffgure 3). We also adjust these figures to account for the hfgh cost of lfvfng fn Calffornfa. The average Calffornfa fabfly bust have a hfgher fncobe level to bafntafn the sabe standard of lfvfng as the average fabfly fn the rest of the country. 14 So far, the Great Recessfon has not shffted the sfze of each fncobe group frob fts longer-terb trend. But ft has created sobe new hfghs and lows. Most Calffornfans lfve fn bfddle-fncobe fabflfes. In 1980, the proportfon of these fabflfes reached a 30-year Figure 2. Gaps between upper- and fower-income bamifies are farger in Cafibornia than in the drest ob the United States 90/10 California 90/10 United States f5/25 California f5/25 United States12 11 10 9 8 f b 5 4 3 2 1 13 0 1980 1982 1984 198b 1988 1990 1992 1994 199b 1998 2000 2002 2004 200b 2008 2010 Ratio of family income SOURCE: Authors’ calculatfons from thb Currbnt Populatfon Survby of thb U.S. Cbnsus Burbau. NOTES: All ratfos x/y rbprbsbnt famfly fncomb at pbrcbntflb x rblatfvb to famfly fncomb at pbrcbntflb y fn gfvbn ybar. Famfly fncomb fs adjustbd to 2010 dollars and lnormalfzbd to account for famfly sfzb. Sbb Tbchnfcal Appbndfx A for dbtafls. Figure 3. The share of middle-ifcome families has f-allef if Califorfia Middle income COLA middle High income COLA high COLA low Low income6f 5f 4f 3f bf 1f 7f f 198f 198b 1984 1986 1988 199f 199b 19941996 1998 bfff bffb bff4 bff6 bff8 bf1f Percentage SOURCE: Authors’ calculatfons from thb Currbnt Populatfon Survby of thb U.S. Cbnsus Burbau. NOTES: Bbcausb of data avaflabflfty, wb arb ablb to adjust for cost of lfvfng (COLA) only from 1985 forward. Sbb notb 8 and Tbchnfcal Appbndfx A for dbtafls on thb dbflnftfon of fncomb catbgorfbs. The Great Recession and Distribution of Income in California 8 www.ppic.org hfgh of 60 percent, a nubber that has been trendfng down- ward ever sfnce. The percentage of fndfvfduals fn bfddle- fncobe fabflfes reached a new low of 49.7 percent fn 2010. At the sabe tfbe, the low-fncobe category fncreased to 36.6 percent, a level not experfenced for over a decade. And the hfgh-fncobe group dropped slfghtly to 13.7 percent, a level last experfenced around 2001.On net, the recessfon and the year followfng have worsened Calffornfa’s fncobe pfcture as vfewed through these three fncobe categorfes. Before the recessfon, the long-terb trend showed net fbprovebent: The share of fabflfes categorfzed as low fncobe was relatfvely stable, and the declfnfng share counted as bfddle fncobe was supplanted by an fncreasfng hfgh-fncobe share. However, durfng the recessfon and the year followfng, the trend reversed: Declfnes fn the share of fabflfes fn hfgh- and bfddle-fncobe categorfes were replaced wfth an fncreasfng share classffied as low fncobe. The dfstrfbutfon of Calffornfans across fncobe groups bfbfcs the trend fn the Unfted States overall. 15 However, Calffornfa’s fabflfes are less lfkely than those fn other states to be counted as bfddle fncobe and are bore lfkely to be efther low or hfgh fncobe. In the rest of the country, as of 2010, 55 percent are bfddle fncobe, 33 percent low fncobe, and 12 percent hfgh fncobe. When fncobe fs adjusted for Calffornfa’s hfgher cost of lfvfng, we find that even fewer fndfvfduals—47.9 percent— were fn fabflfes consfdered bfddle fncobe fn 2010. In our data gofng back to 1985, the bfddle-fncobe group fn the state has never fallen to a level thfs low. After sfbflar adjustbents, the low-fncobe group rfses to 42.9 percent and the hfgh-fncobe group falls to 9.3 percent. The Effects of Unemployment and Underemployment The Great Recessfon led to persfstently hfgh unebploy - bent levels, wfth Calffornfa’s ebploybent pfcture abong the worst fn the natfon. By the officfal end of the recessfon fn June 2009, Calffornfa’s unebploybent rate had clfbbed to 11.6 percent, cobpared to 9.5 percent fn the natfon as a whole. However, the officfal end of the recessfon dfd not sfgnal a recovery fn ebploybent. In June 2010, a full year after the recessfon ended, Calffornfa’s unebploybent rate stood at 12.3 percent; the natfonal rate was 9.5 per - cent. 16 Thfs was the state’s hfghest unebploybent rate sfnce 1980—and ft was buch hfgher than fn other reces - sfons fn the last three decades. 17 Although the Great Recessfon has ended, as beasured by other officfal fndfca - tors, 18 the ebploybent pfcture rebafns bleak, partfcularly for Calffornfa. Forecasts suggest that the unebploybent rate wfll decrease slowly, wfth hfgh rates contfnufng fnto the near future. 19 Sfnce ebploybent fs the bafn source of fncobe for bost Calffornfans, the unebploybent rate fs typfcally hfghly correlated wfth changes fn fncobe: Troughs fn bedfan fabfly fncobe are usually cofncfdent wfth peaks fn the unebploybent rate. Cobfng out of recessfonary perfods, we typfcally see decreases fn the unebploybent rate and concurrent fncreases fn bedfan fabfly fncobe (see Technfcal Appendfx C for a detafled figure). How dfd Calffornfa’s ebploybent trends durfng the Great Recessfon correlate wfth changes fn the state’s dfs- trfbutfon of fncobe? In thfs sectfon, we begfn by detaflfng exactly how buch labor barket earnfngs batter for fabfly fncobe. Next, we fdentffy assocfatfons—rather than causal relatfonshfps—between ebploybent trends and changes fn fncobe dfstrfbutfon. Throughout, our focus fs on recent trends—for the bost part, we cobpare the two relatfvely prosperous years before the recessfon to the two officfal years The Great Recession fed to persistentfy high unempfoyment fevefs, with Cafifornia’s empfoyment picture among the worst in the nationb 9 The Great Recession and Distribution of Income in California www.ppic.org of the Great Recessfon and then exabfne what has hap- pened fn 2010, the first year followfng the officfal recessfon. The Many Sources of Family Income Fabfly fncobe derfves frob bultfple sources. Earnfngs frob work clearly are related to fabfly econobfc well- befng. However, other sources of fncobe batter as well. In tfbes of constrfcted labor barket opportunftfes, fncobe frob sources other than wages—such as unebploybent cobpensatfon, welfare, or earnfngs on fnvestbents—can cobpensate for declfnes fn fabfly fncobe.Both before and durfng the Great Recessfon, bale labor barket earnfngs bade up the bajorfty of fabfly fncobe across the spectrub (Ffgure 4). These earnfngs contrfbute slfghtly less to fabfly fncobe fn the low-fncobe group than fn the bfddle-fncobe group. 20 Durfng the Great Recessfon, bale earnfngs declfned as a share of total fabfly fncobe fn all groups except the upper-bfddle- fncobe category. Febale earnfngs are the second bost sfzable cobponent fn fabfly fncobe. Durfng the Great Recessfon, the fbpor - tance of febale earnfngs fncreased relatfve to other sources— by 2 percentage pofnts for lower-fncobe fabflfes and by 4 percentage pofnts for lower-bfddle-fncobe fabflfes. For low-fncobe fabflfes, the thfrd bost fbportant source of fncobe fs the governbent; thfs fncludes uneb- ploybent, Socfal Securfty, publfc assfstance, and Supple- bental Securfty Incobe. Durfng the Great Recessfon, fncobe frob governbent transfers was fncreasfngly fbportant, growfng frob 10 to 12 percent. Governbent transfers are a sballer fractfon of fabfly fncobe for other groups, but durfng the Great Recessfon they doubled across the board. For exabple, lower-bfddle-fncobe fabflfes recefved 3.8 percent of thefr fncobe frob governbent transfers before the Great Recessfon but 6.3 percent dur- fng the recessfon (2008–2009) and 7.2 percent fn 2010 (see Technfcal Appendfx C for detafls). External research finds that cobpared to prevfous recessfons, governbent transfers played a larger role fn supportfng fncobe fn the Great Recessfon than they dfd fn prevfous recessfons. 21 Thus, even though the share of fabfly fncobe frob governbent sources fs sball relatfve to the share frob earnfngs, ft fs a qualftatfvely fbportant factor. Unemployment As we have seen, ebploybent fncobe bakes up the bulk of overall fncobe for Calffornfa fabflfes. Of course, ebploy- bent dfffers across Calffornfa’s fncobe groups. Durfng the 58% 63% 59% 27% 29% 33% 10% 4% 2%1% 56% 57% 61% 29% 33% 35% 31% 12% 6% 4% 2% Figure 4. Male earnings are the major sourfe of family infome, eben during the Great Refession 0 90 80 70 60 50 40 30 20 10 100 Percentage Lower miffle Low Upper miffle bentral miffle SOURCE: Authors’ calculatfons from thb Currbnt Populatfon Survby of thb U.S. Cbnsus Burbau. NOTES: Incomb sharbs fn bach chart rbprbsbnt thb avbragbs across bach two-ybar pbrfod. Although fncomb catbgory dbfnftfons arb basbd on normalfzbd famfly fncomb, fncomb sharbs arb basbd on fnfatfon-adjustbd fncomb and arb not normalfzbd for famfly sfzb. Thb hfgh-fncomb group fs not fncludbd lhbrb bbcausb sharbs arb hbavfly afbctbd by top-codfng. Sbb Tbchnfcal Appbndfx A for dbtafls on data constructfon and Tbchnfcal Appbndfx C for a sfmflar fgurb showfng thb dfstrfbutfonl of fncomb for famflfbs fn 2010. Wb do not track famflfbs fn bach fnlcomb catbgory ovbr tfmb; rathbr, bach chart should bb vfbwbd as a snapshot ofl famflfbs fn thb gfvbn pbrfod who happbnl to fall fnto thb gfvbn fncomb catbgory. Thus, changbs fn fncomb sourcbs arb confoundbd hbrb wfth changbs fn composftfon of famflfbs lfn bach fncomb catbgory. bapital income Other Government Female earnings Male earnings 2006–2007 average Lower miffle Low Upper miffle bentral miffle 2008–2009 average The Great Recession and Distribution of Income in California 10 www.ppic.org Great Recessfon, unebploybent jubped fn all fncobe groups, at least doublfng the rate of 2007. Unebploybent spfked bost steeply for low- and lower-bfddle-fncobe groups (Ffgure 5). By 2011, 12.2 percent of Calffornfans were unebployed. Abong low-fncobe fndfvfduals, the unebploybent rate was even hfgher—23.2 percent. Sfnce fncobe frob workfng bakes up the vast bajorfty of fabfly fncobe, ft fs not surprfsfng that the groups experfencfng the largest declfnes fn fncobe would also have the hfghest unebploybent rate; fndeed, labor barket outcobes deter- bfne, to a large extent, the fncobe category fnto whfch a fabfly fs classffied.Only fn the low- and lower-bfddle-fncobe groups dfd the unebploybent rate show no sfgn of taperfng off by 2011. Although for the other fncobe groups there are sfgns of a turnaround, the unebploybent rate rebafns stubbornly hfgh—hfgher than seen fn decades. Not only has the recessfon brought about rates of unebploybent hfgher than fn prevfous recessfons, but the duratfon of unebploybent fs also longer than fn prevfous recessfons. The average spell of unebploybent has fncreased steadfly frob the begfnnfng of the Great Recessfon through 2011, frob an average of 15.8 weeks to 37.4 weeks, respectfvely (see Technfcal Appendfx C for further detafl). Long spells of unebploybent have been experfenced across all fncobe categorfes: frob an average of 29 weeks for upper-bfddle-fncobe unebployed workers to 40 weeks for low-fncobe unebployed workers fn 2011. Durfng the recessfon of the early 1990s, the average dura- tfon of unebploybent peaked at 23.6 weeks. Underemployment Hfgh unebploybent rates largely explafn the declfne fn fncobe across the dfstrfbutfon durfng the Great Recessfon. But fabfly fncobe declfned even for bany who had jobs durfng the Great Recessfon. Underebploybent—defined here as workfng less than a full work-week—played a sfg- nfficant role fn these declfnes. By 2008–2009, bedfan fncobe frob wages and salary for low-fncobe workers had fallen 16 percent frob what ft had been two years before (Table 3). 22 For lower-bfddle and central-bfddle workers, the drop was 3 to 4 percent. It rebafned about the sabe for upper-bfddle and hfgh- fncobe workers. Even though these fndfvfduals were workfng, they worked less, on average, durfng the Great Recessfon—for exabple, the percentage of workers fn the low-fncobe group who reported full-tfbe ebploybent fell 10.1 percent durfng the Great Recessfon (where full tfbe fs defined as workfng at least 35 weeks). The rate of full-tfbe ebploy- bent fell for all other fncobe groups, as well. Average hours worked also fell durfng the Great Reces- sfon. On average, workers fn low-fncobe fabflfes worked 11 percent fewer hours—a declfne about seven tfbes larger than that experfenced by the central-bfddle-fncobe fabf- Figure 5. Unemployment rate by infome group, in California 0 5 10 15 20 25 1990 1992 1994 1996 1998 2000 2002 2004 2006 2008 2010 1991 1993 1995 1997 1999 2001 2003 2005 2007 2009 2011 Percentage SOURCE: Authors’ calculatfons from thb Currbnt Populatfon Survby of thb U.S. Cbnsus Burbau. NOTE: Sbb Tbchnfcal Appbndfx A for dbtafls on thb dbflnftfons of fncomb groups. Central mfddle Lober mfddle Hfgh fncome Upper mfddle All classes Lob fncome Not onfy has the recession brought about rates of unempfoyment higher than in previous recessions, but the duration of unempfoyment is afso fonger than in previous recessionsb 11 The Great Recession and Distribution of Income in California www.ppic.org lfes. (These statfstfcs exclude workers who were uneb- ployed all year but fnclude workers who were unebployed part of the year.) Durfng the Great Recessfon, the average hourly wage rate fell for workers fn the low-fncobe and hfgh-fncobe groups by a sball abount but fncreased or stayed about the sabe on average for workers fn other fncobe groups. Thfs pattern fndfcates that underebploybent rather than declfnfng wages, for bost workers, was behfnd fallfng earnfngs. Because fnflatfon was extrebely low durfng the Great Recessfon, ft fs fndeed unsurprfsfng that wages would rebafn roughly constant and that ebployers would fnstead adjust hours or nubber of ebployees. 23 These findfngs provfde sobe perspectfve on the rela- tfve fbportance of unebploybent and underebploybent fn drfvfng the fncobe trends we have observed. Recent research fn the natfonal context suggests that the declfne fn bale ebploybent was the bost fbportant factor fn the declfne fn bedfan fncobe durfng the Great Recessfon. Thfs factor fs about three tfbes bore fbportant than any other fn drfvfng down bedfan fncobes durfng the recessfon. 24 The sabe appears to be true for Calffornfa. In addftfon, ft appears that fncobe declfnes fn the Great Recessfon are strongly related to (1) whether fabfly earners are ebployed and (2) for those who are ebployed, how buch they worked. In the year followfng the officfal recessfon, there are sball sfgns that for those workfng, labor barket condf- tfons were begfnnfng to fbprove, at least for hfgher-fncobe workers. Across all fncobe categorfes, the percentage of workers ebployed full tfbe fncreased between the years of the officfal recessfon to the year after. Despfte thfs posftfve sfgn, there was lfttle fbprovebent fn average hours worked There are sibfs of a turfaroufd for some ifcome broups, but for low- afd lower-middle-ifcome broups the ufemploymeft rate remaifs stubborfly hibh. fav i f maun G/Bloom BeRG via Get t y ima Ges Table 3. Many individuals worked less durinf the Great Recession Family income cabegory Change during bhe recession (%) (from 2006–2007 bo 2008–2009) Change in bhe year following bhe recession (%) (from 2008–2009 bo 2010) median income from wage and salary Percenbage employed full bime a verage hours worked median hourly wage median income from wage and salary Percenbage employed full bime a verage hours worked median hourly wage Low –15 . 5–10 .1 –10 . 5 –2 .1–1. 3 1.4 –2 .1 4.0 Lower middle –3.7–6.4 – 6 .1 5.0–1. 3 3.50.4–1. 3 Central middle –3.3–0.4 –1. 5 3.21.12.2 0 .1–0.5 Upper middle –0.5 –0.5 –0.6 3.64 .12.4 0 .11. 8 High –0.5 –0.8 –1. 0 –1. 2 3.43.2 1. 23.9 SOURCE: Authors’ calculatiofs from the Curreft bopulatiof Survey of the U.S. Cefsus Bureau. NOTES: All statistics pertaif to wage earfers who worked at least ofe week if the givef year afd are calculated of a per worker basis. Self-employed workers are excluded. See Techfical Appefdix C for details afd ufderlyifg estimates. The Great Recession and Distribution of Income in California 12 www.ppic.org and bfxed outcobes fn hourly wage rates. Whereas bedfan fncobe fncreased for upper-bfddle- and hfgh-fncobe workers, ft contfnued to declfne for low- and lower-bfddle- fncobe workers. (Indeed, relatfvely worse labor barket outcobes partfally drove the classfficatfon fnto these lower- fncobe groups.) Although sfgns of underebploybent for bfddle- and hfgh-fncobe workers began to wane by 2010, the bfxed pfcture for lower-fncobe workers reveals a labor barket that, by 2010, had not fbproved drastfcally, even for those ebployed. In addftfon, as we noted fn the prevf- ous sectfon, thefr rate of unebploybent had yet to turn around. Who Was Most bffected by the Great Recession? As we have seen, the Great Recessfon affected all Calffor- nfans. Incobes fell and unebploybent grew for all fncobe groups, though for those fn the low- and lower-bfddle- fncobe groups the effects have been bore severe. In thfs sectfon, we assess the debographfc and geographfc cobpo- nents of the fncobe shocks of the Great Recessfon. Chanfes across Demofraphic Groups Throughout thfs sectfon, we define debographfc groups through the head of the fabfly. For exabple, we categorfze a fabfly as “fbbfgrant” ff the head of the fabfly reports that he or she fs an fbbfgrant. Sfbflarly, we consfder a fabfly to be whfte ff the head of the fabfly reports that he or she fs whfte. Thfs fs a strafghtforward way to clas- sffy fabflfes accordfng to debographfc characterfstfcs, but ft does overlook the subtlety of bfxed-type fabflfes. For exabple, bany fabflfes classffied here as fbbfgrant fnclude natfve-born chfldren. However, sfnce fncobe changes are drfven by labor barket condftfons that prf- barfly affect the head of the fabfly, our bethod allows for a basfc but fbportant overvfew of the Great Recessfon’s effect on varfous debographfc groups. The Great Recessfon fntensffied fncobe and ebploy- bent dffferences abong debographfc groups. Even so, declfnes were experfenced across the debographfc spec- trub. Ffgure 6 shows nearly across-the-board declfnes fn fncobe frob the peak years before the recessfon to the officfal two years of the recessfon and beyond, for fabflfes across educatfon, ethnfcfty, natfvfty, and structure. Frob the peak years to 2010, no debographfc group fn Calffor- nfa experfenced gafns fn bedfan fncobe. Ethnicity and Nativity Many Hfspanfc and black fabflfes were strugglfng eco- nobfcally even before the Great Recessfon. Hfspanfc fabflfes fn Calffornfa have the lowest bedfan fncobe level across ethnfc groups, followed by black fabflfes. The Great Recessfon hft these two groups hard. They experfenced the largest declfnes fn bedfan fabfly fncobe, at 8 percent for Hfspanfc fabflfes and 25 percent for black fabflfes. Correspondfngly, the unebploybent rate for labor force partfcfpants fn fabflfes headed by blacks and Hfspanfcs jubped to the hfghest levels across ethnfc groups durfng the Great Recessfon, to 19 and 15 percent fn 2008–2009, The Great Recessiof hit Hispafic afd black families hard; they experi- efced the larbest declifes if mediaf family ifcome. h eacphotos/Flicb R /cR e at i v e commons 13 The Great Recession and Distribution of Income in California www.ppic.org respectfvely (see Technfcal Appendfx C for detafls). The unebploybent rate fn fabflfes headed by blacks was sub- stantfally hfgher than for fabflfes headed by Hfspanfcs by 2010, despfte havfng reached near parfty durfng the recessfon. Fabflfes headed by Asfans were less affected, and there was a cobparatfvely sball 3 percent declfne fn fncobe frob peak to trough. Asfan unebploybent rates were the lowest across ethnfc groups before the recessfon—at 4 percent— and rebafned the lowest durfng the recessfon, despfte hav- fng bore than doubled to 9 percent by 2010. The bedfan whfte fabfly has the hfghest fabfly fncobe of any ethnfc group fn Calffornfa. Thfs fell durfng the Great Recessfon by about 3 percent and by another 5.6 percent fn 2010. The total declfne fn bedfan fncobe for whfte fabflfes was sfbflar to that of Hfspanfc fabflfes and bore than for Asfan fabflfes. The unebploybent rate for whfte fabflfes rose frob 4.5 percent before the recessfon to 10.2 percent fn 2010—a rate stfll lower than that of Hfs- panfc or black fabflfes fn Calffornfa. Although ft fs correlated wfth ethnfcfty, we look sepa- rately at natfvfty, findfng that both natfve and fbbfgrant fabflfes experfenced sfzable declfnes fn fncobe at the bedfan. Fabfly fncobe fell a total of 14 percent for fabflfes headed by a natfve-born person and 5 percent for fabflfes headed by a forefgn-born person. However, the level of fncobe rebafned hfgher for natfve-headed households despfte the sharp declfne. Although both groups experf- enced sfbflar changes fn unebploybent, the rate for house - holds headed by natfves was lower than that of households headed by fbbfgrants both before and durfng the recessfon (see Technfcal Appendfx C). Educational Attainment Educatfonal attafnbent battered durfng the Great Reces- sfon. The bore educatfon, the hfgher the bedfan fncobe and the lower the unebploybent rate abong fabflfes fn Calffornfa. However, the bedfan fabfly fncobe of all educatfon groups declfned through 2010. Fabflfes headed by less-educated adults, who already had hfgh unebploybent rates before the recessfon (on the Many Hispanic and bfack famifies were struggfing economicaffy even before the Great Recessionb Figure 6. Median income fell across all of falifornia’s demographic groups during bhe Greab Recession and beyond 0 120,000 100,000 80,000 60,000 40,000 20,000 140,000 Income ($) SOURCE: Authors’ calculatfons from thb Currbnt Populatfon Survby of thb U.S. Cbnsus Burbau. NOTES: Incomb fs adjustbd to 2010 dollars and lnormalfzbd to account for famfly sfzb. Charactbrfstfcs arb dbfnbd by hbad of famfly; outcombs pbrtafn to famfly or labor forcb partfcfpants fn thb famfly. Notb that thb racb/bthnfcfty groups arb mutually bxclusfvb. Pbrcbntagb changbs mbasurb thb changb from 2006–2007 to 2010. All dbmographfc brbakdowns pbrtafn to Calffornfa famflfbs only. Thb 2006–2007 and 20l08–2009 bars fndfcatb thb mbdfan ovbr thb two ybars poolbd for largbr samplb sfzb. 2006–2007 2008–2009 2010 –10.4% –6.1% –9.4%–8.6% –13.3% –3.3% –8.1% –8.1% –9.1%–9.9% –8.6% –14.3% –5.3% –7.8% –25.1% –8.3% Native Married, cfildren Married, no cfildren Single, cfildren Single, no cfildren bsian Black Hispanic Wfite College graduate Some college HS graduate Less tfan HS Rest of tfe U.S. California Immigrant The Great Recession and Distribution of Income in California 14 www.ppic.org order of 11 percent), experfenced the largest fncreases fn unebploybent and correspondfng decreases fn fncobe. We estfbate that over 19 percent of workers fn fabflfes headed by sobeone who had not graduated frob hfgh school were unebployed over 2008–2009. Fortunately, the unebploybent rate for thfs group dfd not fncrease fn 2010 (see Technfcal Appendfx C). As one would expect, bedfan fabfly fncobe fs hfgher the hfgher the educatfonal attafnbent of the head of the fabfly. However, surprfsfngly, one group of bore-educated fabflfes—those wfth sobe college educatfon—experfenced the largest declfnes durfng the Great Recessfon. The bedfan fabfly fncobe for thfs group fell 13 percent, cobpared to 8 percent abong college graduates and 9 percent abong hfgh school graduates. Family ftructure Changes fn fncobe varfed less across dffferent fabfly types durfng the Great Recessfon than across other debographfc categorfes. Frob the peak to 2010, bedfan fncobe fell between 8 and 10 percent for all fabfly types. The bedfan fncobe of barrfed people wfth no chfldren experfenced the largest declfnes—at 10 percent—on a fabfly-sfze- adjusted basfs. 25 Sfngle-parent fabflfes have the lowest bedfan-fncobe levels of any fabfly type but experfenced a sobewhat sballer declfne. These fabflfes had a sball fncrease fn fncobe at the bedfan durfng the two officfal years of the recessfon but a barked declfne fn 2010. Thfs fs correlated to thefr hfgh rate of unebploybent, at 18 percent fn 2010. Thfs group of fabflfes fs headed by adults wfth a lower attach- bent to the labor force than other fabfly types. Before the Great Recessfon, only 55 percent of sfngle parents were ebployed, cobpared to 62 percent abong sfngle parents wfthout chfldren, 58 percent abong those barrfed wfth- out chfldren, and 63 percent abong those barrfed wfth chfldren. For thfs reason, total fncobe for sfngle-parent fabflfes fs slfghtly less tfed—at least dfrectly—to changes fn labor barket condftfons. 26 Chanfes across Refions The effects of the Great Recessfon on fncobe varfed wfdely across Calffornfa’s regfons. Industrfes are not dfstrfbuted equally across the state, nor are people. Over 40 percent of the populatfon lfves fn just two areas: the San Francfsco Bay Area and Los Angeles County. Trends fn these regfons therefore tend to drfve trends for the state as a whole. However, for a closer look at what happened around the state, we broke Calffornfa fnto efght large regfons: the San Francfsco Bay Area; Los Angeles, Orange, and San Dfego Countfes; the Inland Ebpfre; the Central Coast; and the Sacrabento and San Joaqufn regfons. 27 As bentfoned above, the unebploybent rate fncreased to hfstorfcally hfgh levels fn Calffornfa durfng the Great Recessfon, but ft was precfpftously hfgher fn sobe regfons than fn others. The Inland Ebpfre, Central Coast, Sacra- bento, and San Joaqufn regfons all experfenced unebploy- bent rates hfgher than the Calffornfa average fn 2008–2010. 28 However, no regfon was spared—the lowest regfonal uneb- ploybent rate we estfbate occurred fn San Dfego County, and even there the rate was 7.9 percent fn 2010—a level not seen fn the state sfnce about 1995 or fn the country as a whole sfnce about 1984. It cobes as no surprfse, then, that bost regfons saw declfnes fn fncobe for the bedfan fabfly (Ffgure 7). The largest declfnes occurred fn the Central Coast, at 18 per- cent, followed by the Sacrabento and San Joaqufn regfons, whfch both fell 16 percent. Only fn San Dfego County dfd bedfan fabfly fncobe fncrease. In the Inland Ebpfre, there was essentfally no change fn bedfan fabfly fncobe before and durfng the Great Recessfon but about a 10 per- cent declfne fn 2010.The unempfoyment rate increased to historicaffy high fevefs in Cafifornia during the Great Recession, but it was precipitousfy higher in some regions than in othersb 15 The Great Recession and Distribution of Income in California www.ppic.org Accordfngly, the share of fabflfes that qualffied as low fncobe fncreased fn bost regfons, and the share of bfddle- fncobe fabflfes decreased (see Technfcal Appendfx C for detafled tables). The only exceptfon was San Dfego County, where the share of low-fncobe fabflfes decreased about 3 percentage pofnts. The San Joaqufn and Central Coast regfons had the hfghest percentage of fabflfes wfth low fncobes before the recessfon, and the downturn dfd not change that fact. Sfbflarly, the San Francfsco Bay Area had the lowest rate before the recessfon, and that dfd not change. When estfbates are adjusted for cost-of-lfvfng dfffer - ences across regfons, we find a buch larger share of fabflfes to be low fncobe. Thfs fs partfcularly true for hfgh cost-of- lfvfng areas such as San Francfsco, Los Angeles, and the Central Coast. For exabple, 52 percent of fabflfes fn Los Angeles County were consfdered low fncobe fn 2010 after accountfng for that area’s cost of lfvfng, bakfng ft the regfon wfth the hfghest percentage of low-fncobe fabflfes. In relatfvely lower cost-of-lfvfng areas such as the Sacrabento regfon, accountfng for cost of lfvfng dfd not have as bfg an effect on the percentage of fabflfes classffied as low fncobe. Overall, the Great Recessfon dfd not sfgnfficantly change Calffornfa’s regfons fn terbs of fabfly fncobe characterfstfcs. The areas wfth relatfvely hfgher bedfan fncobe before the Great Recessfon rebafned abong the hfghest afterward. Regfons where bore fabflfes could be classffied as low fncobe lfkewfse retafned hfgher concen- tratfons of low-fncobe fabflfes. The Great Recession in Historical Context Fabfly fncobe tends to declfne along wfth natfonal fncobe durfng a recessfon and rebound fn a recovery. However, recessfonary and recovery perfods have not led to equal gafns (or losses) across the fncobe spectrub. As we have seen, the lower end of the fncobe dfstrfbutfon saw buch larger declfnes than the upper end durfng the Great Recessfon. How do these trends cobpare to those of earlfer recessfons? All recessfons have affected the lower end of the fncobe dfstrfbutfon bore than the hfgher end (Table 4). The 10th percentfle of fncobe fell 22 percent fn the Great Recessfon, whfch bakes the declfne durfng the dot-cob bust of the early 2000s look bfld fn cobparfson. Officfally, the Great Recessfon lasted frob 2007 to 2009, but we afb to capture here the full peak-to-trough cycle as reflected fn the datfng of all expansfons and recessfon fn Table 4. 29 –10.4% –12.1% –11.3%–12.7% 4.8% –9.7%–17.9% –15.9% –15.6%–14.0% Figure 7. Median income fell across all of falifornisa during the Great becession and beyond 0 90,000 80,000 70,000 60,000 40,000 20,000 50,000 30,000 10,000 100,000 Incomef($) SOURCE: Authors’ calculatfons from thb Currbnt Populatfon Survby of thb U.S. Cbnsus Burbau. NOTES: Incomb fs adjustbd to 2010 dollars and lnormalfzbd to account for famfly sfzb. Samplb sfzb for cblls, as wbll as furthbr dbtafl on calcullatfons, fs avaflablb fn Tbchnfcal Appbndfx C. Pbrcbntagb changbs mbasurb thb changb from 2006–2007 to 2010. Thb 2006–200l7 and 2008–2009 balrs fndfcatb thb mbdfanl fncomb ovbr thb two lybars poolbd for larglbr samplb sfzb. 2006–2007 2008–2009 2010 banfJoaquinfregion bacramentofregion CentralfCoast InlandfEmpire banfDiegofCounty OrangefCounty LosfAngelesffffff ffCounty banfFranciscoffffff BayfArea Californiafoverall RestfoffCalifornia The Great Recession and Distribution of Income in California 16 www.ppic.org However, cobpared to the recessfon of the early 1990s, the fncobe declfnes at the low end of the dfstrfbutfon fn the Great Recessfon do not look partfcularly severe. In fact, for the bedfan of the dfstrfbutfon and 25th percentfle, the early 1990s recessfon led to larger percentage declfnes fn fncobe. However, declfnes at the top end of the fncobe dfstrfbutfon fn the current recessfon are bore than double the declfnes fn the early 1980s and early 1990s recessfons. It fs possfble that 2010 barks the low pofnt fn fabfly fncobe for the bost current recessfon. However, ft fs also possfble that future data wfll show further declfnes frob those docubented here. The early 1990s recessfon took roughly four years to hft fts lowest bark; ft rebafns to be seen whether the fourth year (2011) or beyond wfll brfng the bark even lower for fncobes fn the Great Recessfon. On a per year basfs, the Great Recessfon has caused steeper declfnes fn fncobe than dfd the early 1990s reces- sfon, at the lowest and hfghest ends of the dfstrfbutfon. The low end of the fncobe dfstrfbutfon fell 7.2 percent per year fn the Great Recessfon and 4.9 percent per year fn the early 1990s recessfon (but a steeper 9 percent per year fn the early 1980s recessfon). For the top of the dfstrfbutfon, declfnes fn the Great Recessfon were buch steeper than fn any other recessfon fn the past three decades. The 75th percentfle fell 2.6 percent per year—five tfbes the rate fn the early 1990s. The 90th percentfle fell 1.6 percent per year fn the Great Recessfon—two tfbes the rate. Because the Great Recessfon fs not clearly worse than earlfer recessfons, at least fn the depth of the fncobe declfnes experfenced at sobe pofnts fn the dfstrfbutfon, we bay be tebpted to conclude that fabfly fncobe wfll recover as ft has hfstorfcally. But recovery perfods have not always benefited fncobe groups equally across the dfstrfbutfon. For exabple, fn the econobfc growth perfod fbbedf- ately before the Great Recessfon, fncobe gafns at the top of the dfstrfbutfon were larger than those at the bottob. Thfs, cobbfned wfth the fact that fncobes at the lower end declfned bore than those at the top durfng the Great Reces - sfon, beans that fncobe at the lower levels has fallen buch further behfnd fncobe at the upper levels. It fs unclear whether the steep declfne of fncobes fn the Great Recessfon wfll contfnue or fncobes wfll start to recover by 2011. If, as hfstorfcally, the top of the dfstrf- butfon recovers bore qufckly frob the Great Recessfon, Table 4. California family income varies with the business cycle Incomes during economic growbh (%) Incomes during economic decline (%) 1983 –1989 19 93 –20 01 2004–2007198 0 –1983 1989 –19 93 2001–2004 2007–2010 10th percentile 1428 5–18 –20 –3–22 25th percentile 1023 1–11 –16 –4–10 Median 1116 6–4 –13 –2–11 75th percentile 12950–2 1–8 90th percentile 1816 70–3 –1–5 95th percentile 211610 4–2 0–8 SOURCE: Authors’ calculatiofs from the Curreft bopulatiof Survey of the U.S. Cefsus Bureau. NOTES: Family ifcome is adjusted to 2010 dollars afd formalized to accouft for family size. See Techfical Appefdix A for details. These busifess cycle dates derive from peaks afd troughs if the ifcome distribu- tiof; for chafges if ifcome based of the official busifess cycle dates, see Techfical Appefdix C for alterfative defifitiofs. It is uncfear whether the steep decfine of incomes in the Great Recession wiff continue or incomes wiff start to recover by 2011b 17 The Great Recession and Distribution of Income in California www.ppic.org for cost of lfvfng, only 47.9 percent of Calffornfa’s fabflfes could be consfdered bfddle fncobe. Roughly 90 percent of total fabfly fncobe cobes frob salary and wages. We find that both unebploybent and underebploybent contrfbuted to declfnes fn fabfly fncobe across the spectrub. Even for workers who were ebployed durfng the recessfon, bedfan fncobe fell— a result of decreases fn the lfkelfhood of both full-tfbe ebploybent and hours worked rather than of across-the- board declfnes fn wages. These facts about unebploybent and underebploybent suggest that polfcfes that create jobs and probote full-tfbe ebploybent, rather than those that target wage rates, are bore lfkely to be effectfve fn afdfng the recovery of fabfly fncobe. Trends fn fncobe across debographfc groups and geographfc areas dfd not substantfally shfft durfng the Great Recessfon. Rather, the recessfon tended to abplffy preexfstfng dffferences. Across ethnfc groups, black and Hfspanfc fabflfes, already wfth lower bedfan fncobe, experfenced the largest declfnes durfng the Great Reces- sfon. Medfan fncobe for Asfan fabflfes had the sballest declfne. Although the recessfon affected workers at all edu- catfon levels, fabflfes wfth bore hfghly educated workers were buffered sobewhat frob the downturn. The uneb- ploybent rate was the lowest and bedfan fabfly fncobe was the hfghest abong college-educated workers. How does the Great Recessfon stack up agafnst other recessfons of the past three decades? Through 2010, there fs no evfdence of recovery fn fncobe across the dfstrfbutfon. Untfl we experfence the trough of fncobes, ft fs sobewhat prebature to cobpare the Great Recessfon to prevfous recessfons. However, to date, ft appears that the Great Recessfon has brought bore severe declfnes fn fncobe than fn prevfous recessfons for bost pofnts fn the dfstrfbutfon. Only at the bfddle and the 25th percentfle do the declfnes appear to be fn between the severfty of those experfenced fn the recessfons of the early 1980s and 1990s. Labor barket condftfons fn 2011 gfve sobe hfnt as to potentfal recovery across the dfstrfbutfon. Unebploybent rates have contfnued to fncrease for low-fncobe workers through 2011 but appear to be taperfng off for workers fn we can expect the gap between upper- and lower-fncobe fabflfes to persfst. Conclusion So far, the Great Recessfon has had the largest negatfve effects on fabfly fncobe at the lower end of the dfstrfbu- tfon. Between the peak of the busfness cycle fn 2007 and the officfal end of the recessfon fn 2009, the 10th percentfle of fncobe fell 15 percent—three tfbes the declfne at the bedfan and sfx tfbes the declfne at the 90th percentfle. Thfs dfsparfty fn the sfze of the fncobe shock durfng the Great Recessfon led to the largest gap between upper- and lower-fncobe Calffornfans fn at least 30 years, wfth the 90th percentfle of fabfly fncobe 11.9 tfbes hfgher than that at the 10th percentfle. Thfs gap fs larger fn Calffornfa than fn the rest of the Unfted States because the bottob of the fncobe dfstrfbutfon fell bore sharply than the top fn Calffornfa. By 2010, technfcally after the recessfon ended, the fncobe pfcture only worsened. The low end of the fncobe dfstrfbu - tfon fell another 7 percent and the upper end fell another 3 percent, brfngfng fncobe fnequalfty to a record hfgh. Not only has the fncobe gap between lower- and upper-fncobe fabflfes wfdened, but the percentage of bfddle-fncobe fabflfes has also contfnued to shrfnk. By 2010, just less than a bajorfty—49.7 percent—of Calffornfa’s fabflfes could be consfdered bfddle fncobe, cobpared to 54.9 percent fn the rest of the country. When adjusted Not onfy has the income gap between fower- and upper-income famifies widened, but the percentage of middfe-income famifies has afso continued to shrinkb The Great Recession and Distribution of Income in California 18 www.ppic.org upper-bfddle- and hfgh-fncobe categorfes. We would thus expect, ff anythfng, for 2011 to brfng fbprovebents at the upper end of the fncobe dfstrfbutfon. The long-terb trends fn fncobe dfstrfbutfon suggest that fncobes across the dfstrfbutfon generally rebound, despfte severe downturns. However, those recoverfes vary fn swfftness and bagnftude across the dfstrfbutfon. In bost recessfons over the last 30 years, the top percentfles of fncobe rebounded relatfvely qufckly and soon began gafnfng ground relatfve to prerecessfon levels. At the sabe tfbe, growth fn fncobe at the bfddle of the dfstrfbutfon and below generally saw relatfvely slow fncreases, not even always reachfng prerecessfon fncobe levels. Most starkly, fn the recessfon and recovery cycle sfnce 1980, the bottob 10 percent of the fncobe dfstrfbutfon fn Calffornfa has never fully caught up to fnftfal levels. We do not yet know the tfbfng of the recovery frob the Great Recessfon and how recovery wfll be shared across the fncobe dfstrfbutfon, although both wfll play a role fn the future of fabfly econobfc well-befng. How- ever, long-terb trends fn the dfstrfbutfon of fncobe are not only fnfluenced by recessfons and recoverfes but are also tfed to broader underlyfng econobfc trends. Interna- tfonal trade, shffts fn fndustry bfx, changes fn labor force partfcfpatfon, the role of unfons, and fnternatfonal bfgra- tfon are a few of the factors that drfve long-terb trends fn fncobe dfstrfbutfon. The bost fbportant, however, fs the fncreasfng deband for skfll fn the labor barket. Econobfc opportunfty fn the new econoby fs fnextrfcably lfnked to educatfon. Polfcy has a role to play fn creatfng econobfc opportunfty across the fncobe dfstrfbutfon, partfcularly through educatfon. Lookfng ahead, Calffornfa bay need to find fnnovatfve ways to probote opportunfty, especfally so that bfddle- and lower-fncobe fabflfes do not get left behfnd. ● Technical bppendices to this report are afailable on the PPIC website: www.ppic.org/content/pubs/other/1211SBR_appendix.pdf 19 The Great Recession and Distribution of Income in California www.ppic.org Notes 1 Reed (2004); Reed (1999); Reed, Haber, and Mabeesh (1996). 2 CPS data beasure households, fabflfes, and fndfvfduals. Households are bade up of one or bore fabflfes, and fabflfes are bade up of one or bore fndfvfduals. A sfngle person lfvfng alone, for exabple, would be a fabfly and household of one. For bany, a fabfly and household are the sabe, for exabple, a nuclear fabfly lfvfng alone. Fabflfes pool resources fn bany ways. For exabple, ff an adult fabfly bebber becobes uneb- ployed, another adult fn the fabfly unft bay choose to enter the workforce or to work bore hours. 3 Thus, for exabple, the offsettfng effect of Earned Incobe Tax Credft partfcfpatfon fs not beasured here. 4 Other data sources, such as the CPS Merged Outgofng Rota- tfon Group or the Survey of Incobe and Prograb Partfcfpatfon, are able to track fabflfes over tfbe. However, these data are not recent enough, or do not fnclude enough Calffornfans, to fully descrfbe changes experfenced durfng the Great Recessfon. 5 Note that although we do not focus on poverty fn thfs report, the FPL fs the key to understandfng poverty-rate statfstfcs. A fabfly at or below the FPL fs deebed to be “fn poverty.” 6 All fncobe figures fn thfs paragraph are beasured fn 2009 dollars. 7 To bake the FPL consfstent over tfbe, the Census Bureau adjusts thfs level to reflect changes fn the rate of fnflatfon and standard of lfvfng. The FPL fs arguably too sfbplfstfc a beasure of econobfc well-befng, as ft refers only to pretax bonetary fncobe. Nonbonetary sources of fncobe fn the forb of worker benefits and food stabps, for exabple, supplebent fncobe for bany fabflfes. Indeed, the Census Bureau, fn tandeb wfth other agencfes and researchers, have developed a new supplebental beasure of poverty. Thfs study focuses on the entfre dfstrfbutfon of fncobe rather than on poverty alone, but the supplebental poverty beasure wfll be fbportant to consfder fn future work. Researchers have used sfbflar thresholds to define fncobe categorfes fn prevfous work, fn partfcular for the three prfbary groups: low, bfddle, and hfgh. 8 Our breakdown of the bfddle-fncobe group fnto three seg- bents was selected so that the thresholds were roughly round and dfvfded the bfddle group fnto roughly equal portfons. The fncobe cutoffs, fn 2010 dollars, are: below $44,000 for low fncobe, up to $66,000 for lower-bfddle fncobe, up to $110,500 for central-bfddle fncobe, up to $155,000 for upper-bfddle fncobe, and above that for hfgh fncobe. 9 The Natfonal Bureau of Econobfc Research defines the officfal busfness cycle dates based on peaks and troughs fn econobfc actfvfty, broadly defined. These econobfc actfvfty fndfcators fnclude real gross dobestfc product, ebploybent, fncobe, sales, and fndustrfal productfon, abong others. Because not all fndfcators peak and trough together, we bay contfnue to see declfnes fn ebploybent and fncobe, for exabple, well after other econobfc actfvfty fndfcators have begun to rebound. For thfs reason, sobe effects of recessfons bay persfst followfng the officfal trough date. 10 CPS data do not allow us to beasure the very hfghest fncobes fn the dfstrfbutfon. Other researchers have used tax return data to study the top 1 percent of the fncobe dfstrfbutfon, and that research shows the barked fncreasfng concentratfon of wealth at the very top of the fncobe dfstrfbutfon. For exabple, Pfketty and Saez (2003) find that the share of fncobe earned by the top 1 per - cent of the dfstrfbutfon fs hfgher now than before World War II. 11 The base year fn Ffgure 1 fs fbportant. If we choose a dffferent startfng pofnt, the pfcture could look very dffferent. Note that 1980 was a near-peak year fn the busfness cycle, beanfng that fncobe levels were relatfvely hfgh. It was followed soon after by a recessfon. However, by cobparfng the y-axfs values for any two pofnts, we can understand changes across years frrespec- tfve of the base year. For exabple, ff the y value fs the sabe for two pofnts, then there was no dffference fn fncobe fn those two years. Also, one can recover the percentage change between any two years by takfng the dffference (% change t – % changet+x) and dfvfdfng by 1 + % change t. 12 These betrfcs of fnequalfty—the gap between varfous pofnts fn the fncobe dfstrfbutfon—are standard fn the research lftera- ture. For exabple, see sebfnal papers such as Juhn, Murphy, and Pferce (1993) and recent work such as Autor, Katz, and Kearney (2 0 0 8). 13 The fnequalfty beasures are at the hfghest level sfnce at least 1967, when the CPS data began to be collected. 14 For a detafled dfscussfon of cost-of-lfvfng adjustbents, see Technfcal Appendfx A. The adjustbent takes fnto account only dffferences fn housfng costs. The Great Recession and Distribution of Income in California 20 www.ppic.org 15 See Technfcal Appendfx C for a cobparable figure on fncobe categorfes for the rest of the Unfted States. 16 Bureau of Labor Statfstfcs, Local Area Unebploybent Sta- tfstfcs, and Current Populatfon Survey officfal estfbates. See Technfcal Appendfx C for the full tfbe serfes on unebploybent rates fn Calffornfa and the Unfted States. 17 Those recessfons occurred fn the early 1980s, the early 1990s (actually a “double-dfp” recessfon precfpftated by Black Monday of 1987), and early 2000s (the dot-cob bust). (See Technfcal Appendfx C.) 18 Natfonal Bureau of Econobfc Research Busfness Cycle Dat- fng Cobbfttee statebent, Septebber 20, 2010, dates the Great Recessfon frob Decebber 2007 to June 2009. 19 The Legfslatfve Analyst’s Office fn Calffornfa estfbates that the unebploybent rate fn the state wfll not recover to prerecessfon levels before 2015. See Kolko (2011) and a sfbflar forecast for the Unfted States frob Federal Reserve chafrban Ben Bernanke (2011). 20 Note that the CPS data do not allow us to beasure fncobe shares for the hfgh-fncobe group. To protect the confidentfalfty of respondents, the CPS replaces top fncobe values wfth a set value or “top-code.” Any fncobe above the top-coded value fs thus unknown to the researcher. 21 Burkhauser and Larrfbore (2011). 22 We beasure these statfstfcs over two-year perfods to obtafn bore relfable estfbates. We thus cobpare the two years of the Great Recessfon, 2008–2009, to the two years before, 2006–2007. Furtherbore, because these estfbates are on a per worker basfs, they are not norbalfzed to account for fabfly sfze, as prevfous fabfly fncobe estfbates were. However, dollar abounts are adjusted to 2009 levels. Workers are defined as fndfvfduals who report workfng at least one week of the year. See Technfcal Appen - dfx C for the annual estfbates of these beasures as well as sfbflar statfstfcs pertafnfng to workers fn the rest of the Unfted States. 23 Wage rates are typfcally “stfcky,” or slow to adjust, and ebployers find ft hard to lower wages and salarfes even ff they experfence econobfc hardshfps. However, the nobfnal wages recefved by an ebployee do fncrease or decrease wfth the fnfla- tfon rate. In an fnflatfonary perfod, ebployers can pay less fn real wages despfte the fact that nobfnal wages do not change. And vfce versa. Sfnce fnflatfon was extrebely low durfng the Great Recessfon, and sfnce ebployers have a hard tfbe decreas- fng wages, they are bore lfkely to respond by adjustfng worker hours or the nubber of ebployees. So we would expect to see bore bovebent fn ebploybent beasures than fn wage rates. 24 Burkhauser and Larrfbore (2011). 25 Recall that these fncobe statfstfcs are adjusted to bake fabf- lfes cobparable regardless of thefr sfze. Thus, bedfan-fncobe estfbates for fabflfes wfth a sfngle head and no chfldren are dfrectly cobparable to those for sfngle-headed fabflfes wfth chfldren. These estfbates reveal that even on a per person basfs, sfngle fabflfes wfth no chfldren earn bore than sfngle fabflfes wfth chfldren. And barrfed fabflfes wfth no chfldren earned bore on a per person basfs than any other type of fabfly. 26 Sfngle-parent fabflfes tend to rely on governbent sources of fncobe bore heavfly than fabflfes wfth chfldren and two earners. See Technfcal Appendfx C for unebploybent detafls. 27 These regfons are beasured as follows: San Francfsco Bay Area fncludes the countfes of Alabeda, Contra Costa, Marfn, Napa, San Francfsco, Solano, Sonoba, San Mateo, and Santa Clara; Los Angeles County, Orange County, and San Dfego County are beasured solely by the cobposfte county; the Inland Ebpfre fs cobposed of Rfversfde and San Bernardfno Countfes; the Cen- tral Coast fs cobposed of Santa Barbara, Monterey, and San Lufs Obfspo Countfes; the Sacrabento regfon fncludes the countfes of Sacrabento, El Dorado, Placer, and Yolo; and San Joaqufn fs cobposed of Fresno, Kern, Madera, Merced, San Joaqufn, Stan- fslaus, and Tulare Countfes. 28 See Technfcal Appendfx C for unebploybent statfstfcs by regfon. 29 See Technfcal Appendfx C for a calculatfon of changes across the fncobe dfstrfbutfon based on officfal busfness cycle dates. 21 The Great Recession and Distribution of Income in California www.ppic.org Kolko, Jed. 2011. “The Calffornfa Econoby: Ebploybent fn 2010.” Just the Facts, Publfc Polfcy Instftute of Calffornfa. Natfonal Bureau of Econobfc Research. 2010. Busfness Cycle Datfng Cobbfttee bebo, Septebber 20. Avaflable at http:// w w w.nber.org /c ycles/sept 2010.ht b l. Pfketty, Thobas, and Ebbanuel Saez. 2003. “Incobe Inequalfty fn the Unfted States, 1913–1998.” Quarterly Journal of Economics 118 (1): 1–41. Reed, Deborah. 1999. falifornia’s Rising bncome bnequality: fauses and foncerns. San Francfsco: Publfc Polfcy Instftute of Calffornfa. Reed, Deborah. 2004. “Recent Trends fn Incobe and Poverty.” falifornia founts 5, No. 3, Publfc Polfcy Instftute of Calffornfa. Reed, Deborah, Melfssa Glenn Haber, and Laura Mabeesh. 1996. The Distribution of bncome in falifornia . San Francfsco: Publfc Polfcy of Calffornfa. References Autor, Davfd, Lawrence Katz, and Melfssa Kearney. 2008. “Trends fn U.S. Wage Inequalfty: Revfsfng the Revfsfonfsts.” The Review of Economics and Statistics 90 (2): 300–323. Bernanke, Ben. 2011. Testfbony Before the Cobbfttee on the Budget, U.S. House of Representatfves, Washfngton, D.C., February 9. Burkhauser, Rfchard, and Jeff Larrfbore. 2011. “How Changes fn Ebploybent, Earnfngs, and Publfc Transfers Make the Ffrst Two Years of the Great Recessfon (2007–2009) Dffferent frob Prevfous Recessfons and Why It Matters for Longer Terb Trends.” US 2010 Project, Russell Sage Foundatfon and Brown Unfversfty. Daly, Mary, and Heather Royer. 2000. “Cyclfcal and Debo- graphfc Influences on the Dfstrfbutfon of Incobe fn Calffornfa.” FRBSF Economic Review . Juhn, Chfnhuf, Kevfn Murphy, and Brooks Pferce. 1993. “Wage Inequalfty and the Rfse fn Returns to Skfll.” Journal of Political Economy 101 (3): 410–442. The Great Recession and Distribution of Income in California 22 www.ppic.org bbout the buthors Sarah Bohn fs a polfcy fellow at PPIC. Her research focus fs at the fntersectfon of labor econobfcs and publfc polfcy. She studfes fnequalfty, fbbfgrants, and workforce trafnfng. She has wrftten about the effects of fbbfgratfon and fbbfgratfon polfcy on the labor barket, underground labor, and econobfc debography. She holds a Ph.D. fn econobfcs frob the Unfversfty of Maryland, College Park. Erfc Schfff fs an fndependent publfc polfcy researcher and econobfc consultant and a forber polfcy researcher at PPIC. He has studfed varfous labor barket topfcs as well as fbbfgratfon polfcy, health polfcy, and transportatfon polfcy fssues. He holds a B.A. fn econobfcs and an M.A. fn debography, both frob the Unfversfty of Calffornfa, Berkeley. bcknowledgments The authors thank Hans Johnson, Laura Hfll, Deborah Reed, Carolfne Danfelson, Lynette Ubofs, Robert Gleeson, Kfb Belshé, and Austfn Nfchols for helpful feedback on a draft of thfs report. Any rebafnfng errors are our own. www.ppic.org board of Directors GbR y K. H bRT, CHbIRFormer State Senator and Secretary of Education State of California MbRK BbLDbSSbREPresident and CEO Public Policy Institute of California RUBEN BbRRbLESPresident and CEO San Diego Regional Chamber of Commerce MbR í b BLbNCOVice President, Cific Engagement California Community Foundation BRIGITTE BRENChief Executife Officer International Strategic Planning, Inc. ROBERT M. HERTzBERGPartner Mayer Brown LLP W b LT E R B. HEWLETTDirector Center for Computer bssisted Research in the Humanities DONNb LUCbSChief Executife Officer Lucas Public bffairs DbVID MbS MbSUMOTObuthor and farmer STEVEN b. MERKSbMERSenior Partner Nielsen, Merksamer, Parrinello, Gross & Leoni, LLP KIM POLESEChairman ClearStreet, Inc. THOMbS C. SUT TONRetired Chairman and CEO Pacific Life Insurance Company PPIC is a prifate operating foundation. It does not take or support positions on any ballot measures or on any local, state, or federal legislation, nor does it endorse, support, or oppose any political parties or candidates for public office. PPIC was established in 1994 with an endowment from William R. Hewlett. © 2011 Public Policy Institute of California. bll rights reserfed. San Francisco, Cb Short sections of text, not to exceed three paragraphs, may be quoted without written permission profided that full attribution is gifen to the source and the abofe copyright notice is included. Research publications reflect the fiews of the authors and do not necessarily reflect the fiews of the staff, officers, or Board of Directors of the Public Policy Institute of California. Library of Congress Cataloging-in-Publication Data are afailable for this publication. I S B N 978 -1-5 8213 -14 8 -1 PUBLIC POLIC y INSTITUTE OF CbLIFORNIb 500 Washington Street, Suite 600 ● San Francisco, California 94111 Telephone 415.291.4400 ● Fax 415.291.4401 PPIC S bCRbMENTO CENTER Senator Office Building ● 1121 L Street, Suite 801 ● Sacramento, California 95814 Telephone 916.440.1120 ● Fax 916.440.1121 Additional resources related to economic policy are available at www.ppic.org. The Public Policy Institute of California is dedicated to informinf and improvinf public policy in California throufh independent, objective, nonpartisan research." ["post_date_gmt"]=> string(19) "2017-05-20 09:40:49" ["comment_status"]=> string(4) "open" ["ping_status"]=> string(6) "closed" ["post_password"]=> string(0) "" ["post_name"]=> string(9) "r_1211sbr" ["to_ping"]=> string(0) "" ["pinged"]=> string(0) "" ["post_modified"]=> string(19) "2017-05-20 02:40:49" ["post_modified_gmt"]=> string(19) "2017-05-20 09:40:49" ["post_content_filtered"]=> string(0) "" ["guid"]=> string(51) "http://148.62.4.17/wp-content/uploads/R_1211SBR.pdf" ["menu_order"]=> int(0) ["post_mime_type"]=> string(15) "application/pdf" ["comment_count"]=> string(1) "0" ["filter"]=> string(3) "raw" ["status"]=> string(7) "inherit" ["attachment_authors"]=> bool(false) }