How Financial Factors Outside of a Defined Contribution Plan Can Impact Retirement Readiness: An Examination of Public-Sector Participants

Workers’ finances can face many challenges over their careers, including irregular expenses, which are sometimes quite large. How workers deal with covering these expenses and how they affect other aspects of their financial goals are ripe areas of analysis, particularly with respect to retirement preparations. This study builds on prior work done by the Employee Benefit Research Institute (EBRI) and J.P. Morgan Asset Management focused on 401(k) plan participants’ behavior when faced with irregular expenses. This analysis examines the behavior of public-sector defined contribution (DC) plan participants on the tradeoff between credit card debt and a plan loan. Key findings from the study:

A monthly unfunded spending spike is defined as a spike at least 25 percent above the previous 12 months’ median spending that cannot be funded by the household’s income and available cash reserves in that month. In this study, 29 percent of the household observations were found to have had at least one month where an unfunded spending spike occurred.
On a dollar basis, among those with incomes of $150,000 or less, 60 percent of the household observations had spikes not covered by income and cash reserves larger than $2,500 aggregated over the year, and 82 percent had spending not covered by income alone above this threshold.
The likelihood of experiencing a spike increased with the spending ratio and beginning-of-the-year credit card utilization. In contrast, the likelihood of a spike decreased as gross income increased. However, nearly one-quarter of the households with incomes of $100,000 or more had a spike, so these spikes do not only occur among those with lower incomes.
These spending spikes have a clear impact on the likelihood of public-sector DC plan participants taking a plan loan and increasing their credit card debt in the year of the spike. Of those with a spending spike in the analysis year, 7.0 percent took a new plan loan and 31.7 percent increased their credit card debt, compared with 2.7 percent and 25.9, respectively, of those without a spending spike in that same year.
Households are more likely to take on additional credit card debt before taking the plan loan, as approximately 37–47 percent of those with credit card utilization of >0–79 percent increased their credit card debt, while less than 8 percent took a new plan loan with that level of credit card utilization. However, when credit card utilization reached 80 percent or more, the likelihood of increasing credit card debt decreased to 22.4 percent, while the increasing trend of taking a new plan loan went up by nearly twice the amount it had before the increase to 80–100 percent at 11.5 percent.

This research found that, like private-sector DC plan participants, public-sector DC plan participants who lack income and cash reserves to support a spending spike are likely to end up with more credit card debt. This higher debt can have a long-lasting impact on retirement security, since higher credit card utilization is correlated with lower DC plan contributions and account balances, even when controlling for income. Thus, the availability of emergency savings to cover spending spikes can be a critical factor in preventing or stalling a cycle of increasing debt that can significantly impact retirement readiness, wherever the individual works.

Figure 4 Figure Figure 6 17 Appendix Figure 2 Appe Figur ndix e Fig 13 ure 1 Figure 11 Percentage of Households With Spending Spikes of 25 Percent or More Than the Distribution of Account Balances, by Initial Credit Card Utilization Distribution of Annual Spending Amounts Above Income and Cash Reserves Summary Statistics of Probit Variables Likelihood of Increasing Credit Card Debt and Taking a New Plan Loan, F igure 7, Distribution of Annual Sp en M dio ng d e Alm R oe us nu tsl ts Ab a on ve d In Va co rm iae b lan e d D e Ca sc sh ri p Rti eser onv ses and Above Income ................. 10 In P C U C c o h r r nfund ver las o e oo d b e kin sit it p D en N g a C R e e t at dar a t e id n I s t— g h d n ults S e co s J p p U P pim e M o ke tt nd o il en e s o r iz g — c n tan ian in al at Th Fac l g y Ch bion ie ls o S as n a it sp g e o ct-and ik rh r B lia te s an ise c to al iA s k, n b ’g ff s DC f N ac er I ie m .m v A. tc o ed pP rt p ac ( ing Ch ilan ac n d t ep o p as ft ro even e) c t L so rhe i ed o tn s ed er an t it iFin n ve P a cg m s ar r s o o o d anc 6 ru b 6 d n s ab t eb tm al fial ilr tlio ,lil n i o m h g it n igh B a y tU h e c er .e o S yc havior . Ch f c lh e rT ed o as ou aki fe is ti eh n d cc at ar ng o rea la d d s fs u o a iw tn ril g iis P tz h al at dlan e ar a io b ib n tes r to iL h s a an at o as d an dr c s an an o wc ag g i at seies g ed onf i tfw fh iic n at ian tan h i tslc lo y in al w et er Figure 1 Figure 2 Craig Copeland is Director of Wealth Benefits Research at the Employee Benefit Research Institute (EBRI). Michael How Financial Factors Outside of a Defined Contribution Plan Incidence of an Unfunded Spending Spike, by Beginning of Year Credit Card Utilization (Households With Incomes of $150,000 or Less) Median Spending of the Prior 12 Months Not Covered by Income and Cash Reserves 5 by Unfunded Spending Spike Occurrence See Copeland, Conrath, and Carson (2023). ic sm D er on e pvic tac m r 10 ib es o t0% u g rtet o iirn o f a ic n rp an lem ur h d at yiien c n es t ax gC t . c r e h S h ead sa p ec ec r an a kin in ic d fe ite c g s dal ,s ed r ,ils y, s avin w u ti c h tc th er is o e ge n s o av ver s , f iter n th a ve ken tag h es e e tSa m io n cu en o d m t n ivid ttp b s re il,b u e fc u o al rt red ie o w n tio th rre c k a ar s tpe . d ayc s o ,f h an tec ho dk s F le o iis n an w d aiep s t n h .c o Ch n is a oilas t ed c C re’ ed h .s a T is tr h c a c ial s c ar e te in dan c ru o is tm d iti l ie w z cat is sdo ie o o u n f r rc eac w th e as ie s h Sa u 6 al s .l1 ed o m w p p er s fo lfe c ro en r a t, Conrath is the Chief Retirement Strategist and Head of the Retirement Insights Strategy Team for J.P. Morgan Asset In T Cr h ed e ad h itd o icu tar is oeh n d tu oot l d ila isz c at w ro iit os h n s -siss pec ia kes tnio o n th h al ad er an ilm oal w pys er ori t sian ,n c ato p m fac ro eb s to ian tr rin eg d d h ret es igh er ser io m n is np cian n en g d b w in eh get u r sh at eer dio ts a o tDC s hee an p th lh an o eu ip s m eh ar ptac o ic ld itp so an w f t var it h tak o io ues u t ss a p fac ip kes ltan o r( slF oo ian g nu . trh e Ae s 8) . Model: P(new plan loan)=f(credit card utilization, spending spike, new mortgage, FCan igu50 re% 8,I Mm edpac ian Demt Re ographtir ic aneme d Financnt R ial Factoead rs, by i Unes nfunded s : S An pendin E g Sxa pike mina Occurrenction e ................................ of Public 11 - Variable Average c co om mp par r 8eh 0ed %c en al w s ciiu ve th lat 5 vie in .7 gw p t h er oe fc h en sp oen t u s fo eh drin o tg h ld o r at s fie n i o an w si.t c h es. 2 0In –5 t9 h ip ser an cen alys t u ist,i lithe zat io Ch n as an ed d 4 at .8 a p ser am cen ple t r fo ef rer th en osce ed w 11.1% ih th er 80 ein– i1 s0 r0est per riccten edt to 100% increase in credit card debt, demographics) 6 M s p Ch rran o ed ob w ag ab itn c e iilar n m it y d F en i o g utfu t. i a rlS ie zh p at 2 ar ar , io o t4 in n c 3 i .p an Car 7an d ps t er ro tev c n akin en o islt vin g ao R a fg et t p h clr ie an red eh m io lto en uc an sar eh t .d ST o tb r h ld at al is seg an al had ic ls o etw sn o sw o n f er o rtevo h re e al al Jll vin .s fP. ac o g b M t o o c o rtrs rh ed g ex an hitig a cm h A ar er sis n d et ed at b al M tth an o an e b c ag b e eeg sc e o m at in nen n ttrh io n te lg lR ed b an et eg ifd ro iem n ren n tio en n d g so t ee o fIn ft h w s th e ig h e h year icy th sea o rn o , es f 15.3% For more information on the PRRL Database and the findings from the database, see Gropper, Michael, “The State of Public Gross Income ($10,000) $8.5 90% D S Ae TA ctor PRIVA P CYarti : JPMocip rg 20.7% an Ch ant ase s has a number of security protocols in place which are designed to ensure all No Spike 401 S( pi k k)e Account Balance Age F igure 9, Changes in R evo l vin g Cr ed i t Car d De b t an d Cas h Balances of Consistently Employed Households During and u the tili zhat ou iosn eh (o Filg du s rie n 1 26 0)1. 9C –o 2r0 respo 21 wn hd oi n ug sle y, Ch thas e e m as ed itan hei ac r p cr oim unar t y b al ban anc kin e g d ec inrsease titutid o n fr,o an m d$ 2 th 6ei ,0r8 t4 o tfal or h th ou os se ehw oiltdh scpren edd iti n cg ar d Strategy team. Alex Nobile is Retirement Insights Strategist for J.P. Morgan Asset Management. Matt Petersen is h t w h ave h e ils e pa 1 i ke 6 s.t 4 at an p isal er tiys ccal en is ly ,t b s hiu ad gtn ag ifcirce ed an w itta ic s m ar n pd o ac tu tst. iig lPr iz nat io fib ico ian tn o to n ly fl y 8 d0 al if– flo er 1w 0en 0 s t p f.o e r S r c p ten w ec oti f.o i c T p al h tilio y, sn cs trh ed fe o rim t tc h ed ar e id an d ep u g ten ir lio zd at ss en i o in tn c var o pm lays iab e o lan e, f t h iin m o s tp h eo is w r tc an itas ht e a ro tsakin lp ei kin e g a 40% Sector DC Plans: 2021,” PRRL Research Study (PRRL September 28, 2023). 45% Account Balance $64,631 71.3% September 5, 2024 • No. 618 c u< s3 to 0m 90er %G d rat osas is I n kept com ce o n — fi dT en hit si al is an and e s sec tim u 1at r 1e. .2 e % R bea ass ed on o ab n ln e et p hin ys cio cm al e ,L el ed ss ec esc tth rra o ib n n ed i$ c2 an ,ab 00d o 0 ve pr ow ced ithu trh al e saaf dd eg itiu oar n d osf 2 ar e1 s.e t0 i m % us at eed d t hat After the Pandemic ................................................................ The Probit P................................ rocedure .............................................. 12 u th tr ilo izu at gih o n al o l fp ay 0 p m er en cen t m t ec toh $ an 15 is ,4 m 23 s (fin or c lt uh do in se g w sel itec h c t rc ed red it ic t ar an dd u d tieb lizat it ic oar n d o ft r8 an 0– s1 ac 0t0io p ner s,c el en ec t t(rFoig nu icr e p ay 17m ).en Fu t rttrhan ers m ac otrie, on s, 45.1% E t w lo h xas an e ec p $ u o ar 5 tri 9 ve tn i,co 9 ip tDi 1 an 7 trakin ec vs t tt.akin o g $ r 6 a o9 f g l, o 7 ta an h 0e 2 p. l N an fT o at h r e ilto o h p an no ral s o,e b A as a w sb silio tes lh ic to y isat u o a tif o vail an t akin so ab pfi kG ig le. io ta y ver T o n hfe e n w c m Figure m r ed p en ed lan itti an De c lar o 3 f an sd ip n en b ed io s d r a C rin o sos g w n u itr m n rat ig b ed iu o ct oit w o u on as l d b Ad e f0 oa .m r9 c f7 e iu n n fito sh ctre tr iat top hn o ar o rs os te fi.c s w T ip p h ian ec tih s o t ifIs su ic ts t o p u a ar e ts ake B a pm r ike ief et a, e w prlas s an By Craig Copeland, Michael Conrath, Sharon Carson, Alex Nobile, and Matt Pete 36.6% rsen 80 70% % Age 44.1 7 30–39 31.6% $2,000–$4,999 6.5% ar Se ee desig Lucfas n ed ed , er L o tal o ri ,ic n Joa cm o ck m p V ly ean w an D itd er h F h fed ed ei,er er Kal al el l y In stH an sa uh d ran n ar , d c Je e s O tC oo h p n , r tan r oit b d L ec uttii o van in a sd S Ac al lim otn i(t eF n ac ICA , “ cT ess )h te a t 3 xo es % p er D fois frfer otn hen al e h c in e: ofu o W sreh m ha at otl L id o.eads n . Th ter o e H iar ghe ers everal 30.9% (n=5,755) c 2h 0ec .7 k p an er c den cPer a t s o h f cent tp h ay ese m age en p ar t s t )of i c an ip Hou an d s t s o u seho w rc ites h no lds of r in ev cW o om lvin ith e g i n Mont c clru ed diin thly g c ar w d ag Spend deeb intc h o a ing m d e, a S 25 b o al c an iPer al c S e ec cent o fu r$ i t1 y, 0 or More 0 an , 0 0 n0 u itoy, r m Th pen oran e si,o c n th o sm ,e et par c. ed ca n written with assistance from the Institute’s research and editorial staff. Any views expressed in this report are those of lw co oian tm hp in tar o t ed h pe ay w DC ex ith p p en l1 an .2 s,9 es s fu o n cro h tt h as coo sver te he w ed iac th c b a o yu sin n ptc i ke. o bm al e an T h oc e re c m (as a ed h si er an reies s c er r ed o ves f id t. u cIn m ar m d fac y u ttvi,lar iz th at iab ei o m le ns e s d fw o ian er r var ec rp ed io ar u itts i c cb u ar al lar d an ly uc th e ilii g ztat h hrer iesh o n a m o olfo d tn sh ) g o , s tth e ho e ts aki a eg n eg Tenure 8.6 Figu40 re% 35 1% 0, Incidence of an Unfunded Spending Spike, by Spending Ratio ................................................................. 12 29% with spikes 80% $5,000–$9,999 32.5% 7.8% 40–49 24.2% Retirement Spending?” Employee Benefit Research Institute & J.P. Morgan Asset Management Research Collaboration key contr o ls an d p o l ic ies in plac e w h i c h ar e designed to ensure customer data is s af e, s ec ure, and anonymous: (1) 70% b w e ith li n1 ked 1.1 tp oer th cen e PR t oR f LtMed hDat oseab ian was it h e Spen .c F red or itm di coar ng re d in u of fto i lirth zm at e at ioiPrior 12 Mo n o no fab 8o 0u –t1 0 Ch 0 as pnt er e,c hs en visit tNo . tht e Co follo vw ered ing wby ebs In ite: co me o a w th fip e tth lh an au e a tp s lh o p ar o an ik rts ie c,w iat p an as an d t h 6 t s e 6 (h a .b 6 o eg s up er ld er in i e n n csen o i n t o g tfb c o e do fuas ccu m m thc p m e trar b iy b sed p ed var ike w tio ab iyear t h tlh e2 e s5 an o f.o 5 ffr al i p cvar y er er sis cs i,o en at u trs tu 2 fs ag o 7 tees, r.e 2 t h tp h oer r o sesh e rc en o wto h h tl o 2 d er v 8 s A s d .). v s 6 ,i d e p % 1 g ra o .n r9 n o ge o s s p to s er t number r ake isn c en o co f m a tE fB p e, o of R lr an I, an s th pik E ld o oB s an t e R e h s Ie . w :- E 2 t itR en hF ou ,u ro te r a o t h s f p ei th ik re e s.t a ff. 31.7% 35.8% Un 50–59 funded Spending Spike — An unfund 1ed 9.4 s % pending spike is det $er 10 m ,0in 00 ed –$ t 1o 4 ,o 9c 9c 9ur when a household’s m 6o .4n % thly Analysis of Maximum Likelihood Parameter Estimates H avaio lab w 60 le % atFin https:anci //am.jpmo al F rgan.act com/uor s/ens /as set O-ut mans ag i em de of ent/mod a /insigh De ts/rf eine tiremen d C t-insigh ontr ts/the-ibut 3-differen ion ce-whP at-lan Befo r e J.P. Morgan Asset Management (JPMAM) receives the d at a, all selected data is highly aggregated and all unique 38.0% Figure 11, Incidence of an Unfunded Spending Spike, by Beginning of Year Credit Card Utilization ............................ 13 h Itnt tpsr:o / 25 /d w %uc wwt .cion hase. com/digital/resources/about-chase. Neither EBRI nor EBRI-ERF lobbies or takes positions on specific policy proposals. EBRI invites comment on this participant with their current employer (a series of dummy variables for various tenure thresholds). Also, this analysis 70% 59.2% 35% 35.6% $15,000–$19,999 6.0% s p60+ ending is at least 25 percent or more than 13 .t 6h % e previous 12 months’ median monthly spending, and this spending 60 30 % % leads-to-higher-retirement-savings-rates and VanDerhei, Jack, and Kelly Hah n , “ In Data There Is Truth: Understanding How T idh en e tsifaim abele trien nfo drsm h at olid o nw , h in en clu ld oio nkin g n ga at m es, dif fac erc en ou t n in t cnou m m e bler evel s, s ad . F do re r sex sea sm , d pat le, es am ofo b nig rt h th an osd e m Soakin cial g S ec $1u 0r0 it,y 00 n0 u m orb m ero s,r eis , Can Also amoI nm g thpac e houst Re eholds wtir ith aeme spendint R ng spikeead , higheri nes credit cs ar:d An utilizat E ionxa wasmina correlated tion with a o hif gh P er ubl likelihoioc d - Variable Inciden c e al r Wo esear lorw ker s cfs h o’. r f itn han e u ces niq c uan e var faciab e m lean s oy uc ts hial de len ofg e th se o DC ver p th lan eir t o car beer e test s, ied nc las ud iw ng el li,r rieg ncu lular din ex g p th en e sces red , iw t h ciar ch d ar ute ilis zo at m io et n im ates th q eu ite $20,000–$49,999 21.9% Figure 12, Incidence of an Unfunded Spending Spike, by Gross Income................................................................... 13 Figure 8 cannot be covered by the household’s incom e and cash reserves (checking and savings accounts), i.e., it is not funded Households Actually Support Spending in Retiremen Stt,a”n d Ea BrR dI Issue B 95% C rio ef n,f n ide on . 5 ce3 1 (Emp C h lio -yee Benefit Research Institute, June D r th em at e o m av e p ed d riiv .an (acy 2 b ) al J PM an is c fAM u e ll d y h ec as p rease r o pt ue t ct d p rfe irv d oac m . N y $ o p 3 r6 p o,er t2o6 sco 1 on lfs al o r iln y t h p id o lac en se e tiw ffio a itr b h l ie tn so i n r e c fo r sed r ear mit at c c h io ar er nd s i .s u R tciese o liz nat tar ai io n cn h ed er to w s $iar t1h8 e in , 7 o t 4 bh 6 lie g fat d oat red ta h o tan o s ed u s w al e it l th s h p e 8en 0d– at d 1in a 0 g 0 21.8% 25.9% of taking a plan loan. Of the hou tb sa eh l5olds with a spike having no revolvin 2g 7 .c 5r % edit card debt at the beginning of the year, 50 % 17.1% 90% experience this spending 60% b S lar eg ge ie nctor .50 n H in % ogw o w f P t oh re ker arti ye s ar d, eal tcip h e wicth han ant cogve erin is n g th te hese revo e lvin xpen g csre ed s an it cd ar h do w bal tan hec y ea ffrfo ec m t o th th e er beg asin pn ec in ts g o off tth he e iry fear inan to c ial th e g o en ald s ar ofe t he Gross Income $50,000–$99,999 13.3% 30% Median Demographic and FinanciaE l rF roa rctors, by L iU mn itsfunded Sp Squ ea n re ding Spike Occurrence by immediate liquid asP sa et ra sm . e Ye tear r tw DF o isE t st hie m a test te year of whether a spending spike 7 occursP . r > ChiSq 24, 2021) for more info 24.1% rmation about the EBRI/JPMorgan Asset Management research collaboration. Average of 3 months with this spending 25% s p an o er ld el c en s ya fvin to u r tg ap il iat zpat trro iio b ved n u t(es F riese gan urar al e y c 1z h 8 ed )an . F id n u ar rtthh e is er orm b ese lo igrar at e,ed cth h e ar no p e ter kept to cen re t-c ag io dm en e p w tliet fity h el an b y al y an an in od c ne iym vi s dlo e us u al s s .rtep h an re s $en 20t,ed 00 0 in i n th cre ease data. d f(r3 o) m J PM thoAM se w do ites h tbal20 21.9% F 4i.g 8u p re er 1 c3 en , tLikel took iho a op dl an of lIn oan crea , cs oim ng p ar Cred ed iw t Car ith 1 d1 De .5 b pter an cen d t T akin of thgo s ae N h ew avin Plg an a L so pan ike , w by it h U n cr fu ed nid t ed car Sdp en utid liz in at gi o Sn p iof ke 80–100 Special thanks to Mike Cross and Katya Chegaeva for their contributions to this paper. . 29.9% year ripe ,ar m ea or st g oag f an e al pa ys ym is,en par tst s ictu ar lar ted ly d wu itrh in rgespec the ye t tar o r(et =i1 r em if a en ntew 20.0% pr ep mo ar r $tat 1 g0 ag io 0,n 0 e) s 0., 0 In to hr e m m san op ren ey c da in sg es, rat w io o,r ker and s ’ so pn ike ly soo 1c7 u c.ru 2 cr % e ren ofc e $20,000–$29,999 5.1% Intercept 1 -2.2984 0.1248 -2.543 -2.0538 339.09 <.0001 By Cra 20% ig Copeland, Michael Conrath, Sharon Carson, Alex Nobile, and Matt Petersen n n 8 o ot u al ti50 llo iz % w at ito hn e (p3u 7b .7 lic p at er iocn en otf) an toy t th ib n o afs lo 5e r 0 m wat ith io 80 n ab –1o0u 0t p an er c in en ditv iu dtu ilal izat oiro en n (t5 it2 1 y. 3 .2 .An 3 p % er y c den ata t).p oint included in any publication Occur40 ren %ce ......................................................................................................................................................... 14 percent (Figure 15). Interestingly, it appears that the households are more likely to take on additional credit card debt Data privacy of customers and contractual relationships with recordkeepers have been carefully protected, and no data was $30,000–$49,999 25.3% ( s= ig F1 n a ic iffti c o tan r het a savin spike gs o ic sc an urr ed em). p l( oS yee men Ap t-pben asd eid x r Fet igiu rem resen 1 tan sav d 2 in g fo sr S psp lu an ik m e ,m typi aryc al stlat y ia s t4 ic0s1 , (c k) om pp lan let e o rvar oth iN ab er o l S e dp e d if k e ie n f ied nit ions, and Cre d i 40 t % Card Utilizat tb io aln 5 — This is1 measu -0.r 4ed 35 9 by th 0e .1r 1at 61io o-f0 t.h 66 e 3r 6evo-lvin 0.20 g8 c 3redit 1 c4 ar .0d 9 balances 0. 0in 00 t2he last month of the 25% 12.3% tbal100 17.2% based 20 o% n customer data may only reflect aggregate information. (4) The data is stored on a secure server and can be 25.4% t bref anosfer re trakin ed to g JtPM he oprlg an an l o Aan sse , tas M an apag premen oximat t. el EB y R37 I h –as 47 no per ac cc eess nt o t fo t per h osso e n w al itlh y ic den redtiitf iab car le d in ufto ilr iz ma atito io nn o . f >0–79 percent $50,000–$74,999 26.7% Credit Card/Limit Ratio-Y1 Suggested citation: Copeland, Craig, Michael Conrath, Sharon Carson, Alex Nobile, and Matt Petersen, “How p co rA o ng b te riitb ru esu tion lt s(.DC ) ) plan tb. alT 2h 0us, some 1DC p0 lan .04 p 9ar 5 tici0 p.an 09t 6s1 cou -0 ld .1 t3ak 88e a l0 o.an 23 7f8 rom42 th0 e. 2 p7 lan or ad 0j.u 6s 0t6 3 their contribution 43 s to year to the credit limit on those 17.3% cards. A ratio of 0 percent means that the household has no revolving credit card debt, Figure 14, Average Amount of New Plan Loans and Increases in Credit Card Debt, by Unfunded Spending Spike 30% 40% age20 11.2% L aco cess nged itu od nly inal under S a stm rictp slec e uC rio ty ns pro tc red uc ut res. ion R esearchers are not permitted to export the dat 56.4% a outside of J.P. Morgan increased their credit card debt, while less than 8 percent took a new 0% plan loan with that level of credit car 43 d. 7 u% tilization. $75,000–$99,000 15.2% Gross Wage Income $59,917 $69,702 F 9 inancial Factors Outts bia dle 5 0of a Defin 1ed Con0t.r2 ib 3u 9tion0 Pl .1an 039 Can 0I.m 03p 5ac 2 t R0 et .4 ir4em 27ent Re 5ad .29iness: An 0. 0 E2x1a5mination of Public- these plans, while others could access credit outside of a plan o r use some combination of all three to cover unusually w Oc hcilu 20 e rr % 1en 00c e p er ................................ cen21.7% t means that the ................................ household has used th ................................ e full allowable credit................................ on their credit card(s......................... ). 14 See the appendix in VanDerhei, Jack, and Kelly Hahn, “In Data There Is Truth: Understanding How Households Actually 48.9% age40 A T A G L A N C 24.E 2% 12.1% 30 % Chase’s (JPMC) systems. 43.7% The system complies with all JPMC Inform >at 0% io–n 1 9 T% echnology Risk Management re1 q4 u.i7r% ements $100,000 or more 27.7% T he un 15 iq %ue financial factors outside of the plan that had a highly statistically significant impact on a DC plan participant Spending Ratio tbal100 1 0.0735 0.1085 -0.1392 0.2861 1.29 0.46 0.4984 0.97 Secto20 r Par % ticipants,” E B RI Issue Brief, no. 618 (September 5, 2024). 14.7% high expenses. Balancing these decisions is a key component of participants’ financial well-being. 30% 19.1% Support Spending in Retirement,” EBRI Issue Brief, no. 531 (Employee Benefit Research Institute, June 24, 2021) for an 15% age50 19.4% In this study, spending and public-sector DC plan data from 2019–2021 at the household level are examined. In order How ever , when credit card utilization reached 80 percent, the likelihood of increasing credit card debt decreased to f Fo ig r utr hee 1 m 5o , n Per itoc ren ingt ag ane d osfec Tu hroitsy e o Wi f d th at U a.n f(u 5n ) d Jed PM AM Spen pd ro in vid g S es piv kes alu ab Wh 2l0 e o% iT n –o s 5io 9 gk % h tas N to e w p oPl lic an ym Lo akan er so, rb In uscir nease esses d 1 , Cr 7 an .e 9d % d itf in Ca an rd ci al Credit Card UtilizationA B de gg riin nc ning of Ye 1ar 0.0135 0.0052 0.0033 0.023727.2% 6.7 0.0097 1.9% N t Wo akin ew rker g P a s lan ’ pfli an n L an o la o can es n — c in an c Pll u an fd ac e le d o an m cran ed dat y it a c c h ar ar al d e len u otn g ille iy zs at av oio ver ai nl ab an th le d ei at r th c e ar ye oeer ar cc u en sr,r d en in,c c slu e od o an in f g y a ip s rp r ar eg ike. tiu cil p ar Tan h ex et h w pien h go hs er es, has t hw e nh o ci c ro ed hu t ar is t te c an ar so d dm i nu g et t illio im zan at esi o q nu , ite exampl 15%e schematic of how the overlap of the Chase data with data from an EBRI database is determined. age60 13.6% 10.2% to create this household view, the following steps were taken to merge the spending data from Chase and the public- 22.4 percent, while the increasing trend of taking a new plan loan w 60en %– t 7u 9p % by nearly 4 percentage points (77 .3 .6 % percent De Teb nt u,r 10 e b % y W In ith it iE al m Cr plo ed ye itr Card Utilization .................................................................................................................... 15 prC o rf e es d is t io Cn aal rds U , tb iliu za t tt io hn ese En din osfi g Yh etasr cannot come at the expense o f c onsumer p1r7 iv .7 ac %y. We take every precaution to 0 en .0s % ure Copyr 20 ig %ht Informaage20 tion: This rep1ort is- 0 c.o 0p 1yr 35ighted 0.1 2 b8 y9 the E -0 m .2p 6lo 6yee 0B .2en 39e 1fit Resear 0.01ch Instit0 u.t9e 16 (E 6BRI). You may copy, lt T ar hhe g ise. m s o tH u rd e oy w l ikel b w uo iy lr d ker a s n os ew n dpeal rp io lan rw w ilto o han r k co d w vo er as nie n tg b aken yt h tese h.e In E e m x ad p plen d oiyee ts io en s ,B an ten hd e ef h oio tc w c Ruesear rtr hen ey cc e ah fo f In ec f a s t to s itp tu h en ter e d (as iE nB g pR ec sI) pt is ke an o d fw tJas h .P. ei p r M ofs io n itran ig vel an ciy al A as s gs o set o alc si at ar ed e b alanc 20 e %at the end of year one but has a balance at the end of year two is considered to have a new loan in year two. 10% ten2 21.1% 10 sector DC plan data from the PRRL Database to create the full data sample: 15.2% 80%–100% 16.4% Less than 2 years 21.1% to S 1ee 1. 5G p ro er ppe cen r,t M ) fic ro hm ae lc , “ red Thit e St car ad te uotf iliPu zat bilo icn Se ofc t6o0r– D 17 7C 9 Pl per ancs: en 2t0 t2o1 ,c” red PRiR t L car Re dsea utirlc iz hat St ioun dy o f( P 80 RR –L 10 S0 ep 1 pter em cber ent 2 . T 8h , 2 e 0li2 kel 3) ih fo oro d thC e rc eo din t fC id aen rd c Ree v an olv din s gage40 ec Bu alr ait ny ceo B f e og uin r n 1 ac ing c o ou f n 0 Y.te 1 a h 1ro 9 7 lders0 ’ .p 0r 8ivat 76 e in -f 0o .0 rm 52ation 0.. 2 913 $1,592 1.87 0.1718 $196 print, or download this report solely for personal and noncommercial use, provided that all hard copies retain any and w Mian th ag an e m inen creased t focu slied kel o ihno o 4d 0 1o(fk) ta p kin lan g p aar ptlan icip lan oan ts.’ b eh Coavio nseq r u w en hen tly, 7.0% fac thed e f iw nan ithc iial rr eg fac utlar or se x op uen tsid se es o . f In wh p at ar itsic kn ular ow , n ch w an itg hes in r In ip e ad ar deas ition o , fan an yal pys aritsic , ip par anttic w ulh ar ol y has wi ta h h respec igher to u to ts rtet anirdem inge n lotan pr ep balar an at cie on at s. tT he his en sd tu o dfy year build tsw o on t h pan rio rt h w ey or k had do n at e tbhye ten he d ten5 27.4% Figu10 re% 16, Mean Contribution Rates, by Initial Credit Card Utilization ................................ 10.2%....................................... 16 0% 2–4 years 21.0% o thfC e trakin pu edbl it g iC c a -a se rdp clR an to erv nu o lo lvan im ng be age50 d Bir d a a l an n no ce dt H i n E ocln den rd ease o1 f, S Y ar e ba yah r 0 m .,0 S o6rt1 e e3 ven than B 0 a 2.ss .05 9 , a 8 pe nr d C c-en 0.r1 ai t3 ag g C 08e op po el 0 in an .2 ts 5d, 3in 5 “ $ 1 4 an ,0 31 5 y ( 7k o 0 )t P .h 3er 9 lan i n Asstset anc 0 A e .5 l l3 o o1 f c 6 at inic orneas , Ac ed co u cn red t it$0 card 10% 20.5% all copyright and9.6% other applicable notices contained therein, and you may cite or quote small portions of the report E in th m e cp rp ed lo lan yiee t c ar ar B e en di m u etp fiilo ti zrR at tan esear iotn , ,i f4 c 0. n h 00% o 1In t( k) tsh t ie p t u lm an te o s c (E to B n im R trp I) ib o u r an ttan io dn tJ s ,. ,P. c an o n M d so i/d o re g rr an 4 at 0 i1 o As ( nk) s set ip n l M an DC an lo ag pan lan e m upen sar e ttw ifco er ip ce an useed tx s’a m d oec n in 4 ed is0 io 1 af n (k) st er t o p ltan th aese ke par a ptp ar iclian tpic an ip lo tan a s’n t. s of year 10 o %ne is also considered to have taken a new loan in 20% year –59% two. 80%–100% ten10 17.2% 1) Usi10 n5% g % the unique participant/customer identifier (not personally identifiable information) in each dataset, the 9.0% Spending-to-Income Ratio 5–9 years 27.4 6.7% % This research paper was produced through a collaboration between the Employee Benefit Research Institute (EBRI), Balances, and Loan Ac age60 tivity in 2022,”1 EBRI- Issu 0.054 e9 Brief 0,. n 11 o9 . 5606-,0 an .28 d 9I 1CI R0 es .1 ea 79 r4 ch Persp 0.ec 21tive, vol.0 3 .6 04 , n 62o. 3 (April 2024) for F uitg ilu izrat e i 1 o 7 n ,. Di s tri b u t i o n of Ac c ou n t B al an c es, b y In iti al Cr ed i t Car d U t ili z at i o n ................................ ............................ 17 2.7% 8.3% provided that you do so verbatim and with proper citation. Any use beyond the scope of the foregoing requires EBRI’s ex beh per avio ien r cw e h aen si g fac nife ic dan wti ts hp en irreg din utg lear n “2 s e 0 px ike pen .” sT es. his T an his al an ysial s ys exis am ex in aes mi n th es e t bh eh 1e 3avi .b 2eh % oa r vio of rp o ufb lp icu-b sec lic- to sec r DC tor p dlef anin p ed ar tcic oin ptan ribtu s to ionn Median-$26,084 Figure Me dian -$ 2 16 1,1 0 3 M edian -$15,423 indivi5% duals in both sets of data are established. These individ 3.6% uals with both the spending and the saving data are then <0.80 7.6% 21.2% 10–19 years 17.2% t hSo e N urc at 0% e: io Es nal tim As ates s f oc rom iat th tie on e P n2 o RR f G L D ovata ern ba m 1 se en an t d s De 0.ele 0f1 in ct 1ed 7Ch C as o 0 e d n .1 tr ata 2i1 b6 . F utior o-n m 0 .ore Ad 22 in 6 m 7for inim st ati rat on 0o ., s 2 rs 5 ee ( th NAG e D0 DC ata .01 S A) ou , an rced s b J 0 ox ..P. 9 in 2 3 M th 7o e t rg ex an t. Asset t he 401(k) plan loan number. Only a few of the demographic factors show a statistically significant impac1.7% t on taking a new plan loan. Those with Spending Ratio — This is the ratio of total annual spending to annual net income. 0.9% prior express permission. For per inm crids2 sions, please contact EBRI at perm2 is 6s.i3o% ns@ebri.org. 0.6% t (h DC e )t rp ad lan eo p ffar btet icw ipeen ants c o red n tih t e c ar trd ad deo ebft f an bet dw a een plan cr ed loan it .c ard debt and a plan loan. Key findin 8 gs from the study: Mean Above Inc Con ome an tribu d Resetio rvesn Rates, by Initial Credit Card AUtilization bove Income F girg ou urp eed 18 i,n t Di o sh tr oiu bs ueh tioo nl d osf u Ac sic no gu Ch nt as Bal e’an s m ces et h fo ord Tfh oo r sdeet Wi ertm h in In in cg o m mes e0m .8 ob 0 fer –$ 0s 1 . 9 0 o 4 0 f ,a 0 0h0o u os r eh Mo orld e,. b T y hIn e iu tin al it Cr ofed ob its 1er Car 2.vat 2d % ion in 20 or more years 13.2% 0% ten5 1 -0.0937 0.102 -0.2936 0.1063 0.84 0.3587 Manag0% ement. J.P. Morgan Asset Management is the brand for the asset management business of JPMorgan Chase & Less Than $20,000 $20,000–$99,999 $100,000 or more account balances of less than $5,000 were less likely to have taken a loan than those with balances of $5,000–$19,999, 11 newmtg 5.5% Certain public-sec In to crr ea p slean Cr s ed po it Ca nso rd Deb rs htave a 401(k) plan tha Ne t w w P as lane Lo sta an blished before they were no A l ny o A ng ctier on allowed in the U tilization................................ 0 ................................ 1 2 ................................ 3 0.95–1.04................................ 4 5 ............................ 6 or Mo 23.4 re% 17 this s0% 7. tu 0% dy is the household. The number of people in these households may not truly reflect the exact household size, $1–$1,000 >$1,000–$2,500 >$2,500–$4,000 ten10 1 0.2324 0.1076 0.0216 0.4433 4.67 0.0307 T Co e.n an urd e i— ts a Tfh fiis li at ises th w e o nru ld m w bier de. o f years that the DC plan participant of the household has been with their current Report availability: This report is available on the internet at www.ebri.org w Alh th ilo e •u tg h 5% h oA sDC e m w op in tlh an th b l ylal ou an an nfs c u es n ar d e o ed fa $ss5 p o0 en u ,0 rc d 0e i0 no – g$ f s9 lp eak 9ike ,9ag 9 i9 se w d fef er roie m n ed m ret o a ris e rem a likel sen py ike t tso avin at h ave least gs tiaken f 2t5 h ey p er a ar c lo en ean n t o ab tth an p oai ve d th tb e hac e $ 5 k p ,r0 ievi n0 0 fo u –u l$ l s1 w 1 9h2 ,en 9 m 9 9a o n bt al ha sn ’ ce Spk 28.7% Source: Estimate 0% s from the PRRL Database> a0% nd s –e 19% lect Chase data. For more 20% inform –59% ation, see the Data Sourc60% es bo – x79% in the text. 80%–100% While the median revolving credit card balance at the end of the year was higher for those with a spike than for those public sector. These plans were grandfathered to be continued to be offered, but they follow the Internal Revenue Service Source: Estimates from PRRL Database and select Chase data. For more information, se 1 e. 0 th5 e– Da 1. ta 4 S 9ources box in the text. 23.1% Source: ESo stim urc ates from e: Estimate the PR s from the R PRR L Database and select L Database and select Chase data. For more information, see the Data Sources box in the text. as the household size tec nan 20 only be ap 1 prox0im .04 at 3ed 8 bas 0.ed 128 o 3n th -0 e .2n 0u 7m 6 ber0 o .2f9 u 5n 2ique in 0d .1 ivid 2 uals w0h .7 o3 h 2ave 8 Chase accounts. Source: Estimates from the PRRL Database and select Chase data. For more information, see the Data Sources box in the text. 2 employer. Source: Estimates from the PRRL Dat N ae bw as eP ala nd n sL elo ec at n Chase data. For more information, see th3 e. 9 Dat % a Sources box in the text. Ap g par roptuen ic pi.p d In an ixm tF ad ed ilg eav d u ian irte ies o s n 1 p ,,t en h tM h ei o d o rd is n e e el g m w R tp h ie tlo at h syer u tlc en tan s, u an tn h ro es ey d t V b w d e ar o i t fh iab p u r n th o ld e ei vid ed De r e c b s u fy c r lex r rtien h pie t bit o i h le in to m y su p ................................ tsh eh lo at yer o c ld an ’o sf iln ead 1c 0o –m 1t9 o e year h an igd h s er avail ................................ w per ar a etb ic m le ip oc at ra es i o h lin kel r es an yer dt o ves c o take n ................... tir nib t lo u ha tat in o s n m sto .hn a A t nh s 20 . a without a spike, this median balance was lower than thFigure e median 15 balance at the beginning of the year. This was not the (IRS) code on these plans, not the rules exclusively in the Employee Retirement Income Security Act (ERISA). 1.50–1.99 10.5% Chase data. For more information, see the Ccutb6.1% 1 0.4589 0.0865 0.2894 0.6283 28.16 <.0001 As an example, if only one spouse has a Chase account, this will be considered a one-person household. This EBRI and NAGDCA are not affiliated with JPMorgan Chase & Co. or any of its affiliates or subsidiaries. r th esu ose lt,w piltan h 2 l- o4 an year s arse oafn t en opu tiro en , w inh silo em an e p in ucbrlease ic-sec in to g r rD oC ssp in lan cosm . R e esu was lt sc o frro rel mat th ed e P wuitbhli c a R het igih rer em len ikel t iR ho ese odar oc f ht akin Labg a In this study, 29 percent of the household observations were found to have had at least one month where an Percentage of Those With Unfunded Spending 2.00 Spikes or more Who Took a New Plan r esult for the study focusing on 401(k) plans, where the median end-of-year balance was higher for those 9 w .6 it% h spikes. Data Sources box 6.0% in the text. Figure 10 12 incrd2 1 0.1158 0.0695 -0.0204 0.252 2.78 0.0957 Appendix Figure 2, Summary Statistics of Probit Variables ...................................................................................... 21 household unit observation necessitates the defining of speci5.7% fic data variables. In future research, the cause of the spike will be examined to see if the effects are different depending on the cause of the 18 Source: Estimates from the PRRL Database and select C Figure 5 hase data. Table of Contents p (PR lanR L) loan Dat u.nfab u n ad sed e c Lo r so ps en an s-d sec ior Increased ng ti o sn pal ike s toucd ciu ersr ed sCredit h.o w theCard inciden Figure Debt ce of, lo by 18 an s Init anial d thCredit e loan am Card ounts Utilization across specific participant However, this result is driven by the decrease in credit card debt during 2020, when the impact of the pandemic was Incid NEWMence TG of 1 an Unf 0.217 un 8 ded 0.1 2Spend 26 -0.0 ing 226 Spike, 0.4582by Spend 3.15 ing Ratio 0.0758 Figure 14 spike. 0% For more information, see the Data So Figure 7 urces box in the text. Household Spend Demo ing gr a Relativ phicse to the Median of the Prior 12 Months of Spending — 3 Distribution of Account Balances for Those With Incomes of $100,000 or More, dem• o grap Oh ni c a fd ac oltlo ar rs b b as utis n , o atm in o nrg eg tar hods e to w tih th e ip nar cotm icies pan otfs $ ’ 1 o5 ver 0,0 al0l 0 fi n oan r les ces s, o 6r0 p p oer ten cen tial t o reas f tho e nh so fu osreh takin old g o b th se erlvat oan iosn .s Introduction ................................ ................................................................ Source: Estimates from ................................ the PRRL Database and select .......................... 4 the strongest. In factS , ptk he Chase dat 1 a sho 0w .4ed 132 a sig 0n .0 if6ic 6an 4 t d0 ec .2l8 in 3e 1 in r0 ev .5o 4l3vin 3 g deb 38t .7 an 4 d a sig< n.i0 fi0c0 an 1 t increase in cash 2) In order to ens0 ure that the dat 1a sample Ne ow P nlyl2 ainn L colu an d Figure es hous12 eh 3 Inco rea ldss e Cr wh edi er t Ca e t rh d4 Deb e Ch t ase data h 5ave all or th 6 e or Mo maj reority of 60% Average Amount of New Plan Loans and Increases in Credit Card Debt, Distribution of Annual Spending Amounts Above Income Percentage Above Specific Thresholds of Median Monthly Spending 13 As shown in Figure 1, the household particby ipan Init ts w ial ereCredit widely Card distribu Utilization ted across ages, incomes, and tenures with their 5.0% Chase data. For more information, see the 56.6% As a po hiad nt osf priefe kesr en nocte c , t oh ver e $ ed 15 b 0,y0 0 in0c o inm co eme an d th c resh asho r ld e sis ertves he l ilmi artg s eet r t f ho an r pa $2 rt,i5 ci0 p0 an ag ts t gr oeg par atted ak e o ver i4n .8 t % hte h t ea year x-pref , an errd ed 8,2 in - balances from 2019 Incid to 202ence 0 that of wer an e coUnf nsidun erab ded ly o uSpend tside wh ing at w Spike, as seen by in t hG e ross years Inco directme ly after 2020 (Figure 9). their spending, filters are applied to the households to meet the full (majority) spending criteria. These filters include Household So De urcem : Es otig mraap tes h fro ic ms th ................................ e PRRL Databy base Un and sfu elecnd t Cha ................................ ed se d aSpen ta. For modi re ing nform Spike ation, se ................................ e thO e Dat ccu a Sorence urces box in the tex................................ t. ....... 6 and Cash Reserves and Above Income The an 50 al %ysis presented here, which links public-sector DC plan data and banking data, builds on the cross-sectional c C uo rrnc entlus emp ion loyer . For example, 11.2 percent were less than age 3 D0 ata Sources b , 24.2 perc ox en in the tex t were t.ages 40–49, and 13.6 percent 15 plan50 eme % rgency savings vehicle under SECURE 2.0, and $2,500 is the maximum they can have saved in the account. percent had V sa pren iab dlien g D en fo inti tc io on ver s ed Abo b vy e I inc nc om om e e al Ao bon ve e I ab ncom ove e an td Ca his sthh Res resh erv oes ld. 100% 46.5% Furthermore, overall revolving credit card debt in the economy decreased by 0.3 percent. When the observation years but are not limited to: all 12 month (H s ouse of sp holds endin W g ith Inc data, ome hous s eh of More olds w Tha ith n $ spen 150 d,0 in0g 0 )more than 50 percent of their 19 S E PR pnd en RLd no rin esu gt S e ltp ss ike t o seval ................................ uate the impact of ................................ financial factors outsid ................................ e of the plan, such as................................ overall spending level .................... s and debt 7 were ages 60 or older. For incomes, 5.1 percent had incomes of $20,000–$29,999 and 27.7 percent had incomes of Revolving consumer tcbra ed l5= it1 g ir f e aw cco bu y nan t ba aver lanceag is el erss ate $5o ,f 0 0 30.3 percent per year during the study period from 2019–2021. • The likelihood of experiencing a spike increased with the spending ratio and beginning-of-the-year credit card 14 were split between 2020 and 2021, the end-of-year credit card debt decreased in 2020 but increased in 2021 among 45.2% estimat 45ed % gross income, and households with credit card spending outside of Chase of less than 30 percent of their This would be expected since the threshold for the spikes was 25 percent, so meeting that requirement for higher-income 4 No Spike Spike 4.0% $1 5090 6,0 %% 00 90% accumulation, on behavior inside the plan. It also follows th 20.9% e methodology of the prior EBRI and J.P. Morgan Asset $100,000 or more. Ju tb sa t l2 o0 ver =1 io f n ae cco -f if utn h t b (2 a1 la.n 1 ce p er is c $e 2nt) 0,00o 0f- $ t4 h9 e ,9 p9ar 9ticipants had tenures wi 47.7% th their current employers of 45% Ho wever 100%25 , i% or t grew by 5.7 percent in 2021 and by 7.6 percent in 2022. In the 2024 Retirement C 23.4% onfidence Survey (RCS), Unfunded u tS ilip zen atid oin 9n.g In S pcio kes ntr’ as Im t,p tac he t o likel n Fiih no an od c io alf B a eh spavi ikeo rd ................................ ecreased as gross inco ................................ me increased. Howev....................... er, nearly one- 11 those with spikes in the respective years. Thus, the pandemic closures and government support that was provided S over pe aln l d sping endi nS gp . ikes households would result in2 mu 8.3% ch higher spike amounts than for those with lower incomes. 5 1 More Management study examining the impact of these same factors on behavior inside 401(k) plans. tbal50=1 if account balance is $50,000-$99,999 less Co pel than an dt,w C or a year ig, M s ic an hael d 1 C 3o .2 n rp at er hc , a en ntd Sh had ar 2 o0 n 29% o Crar m so on re , “ ye Ho ar ws F . inancial Factors Outside of a 401(k) Plan Can Impact 60 per40 cen % t of workers said that their debt is a problem, and nearly half of workers said debt is negatively impacting quarter of the households with incomes of $100,000 or more had a spike, so these spikes do not only occur d uring 2020 appears to have offset some of the overall impact of the spike on credit card debt, but the larger credit 80% 40.9% $14,015 S Cr ped enid t iCar ng d su U rg tie lis z at can ion p an layd h DC av oPl c an on Lo a an hos u ................................ sehold’s finances and ................................ possibly lead to the ne ................................ ed to access more fun .............. ds. Thus, 15 a 15 90% tbal100=1 if acco 20unt balance is $100,000 or more Board of Governors of the Federal Reserve System, Consumer Credit G.19, Retir $ 40 eme 14 % ,00n 0t Readiness,” EBRI Issue Brief, no. 591 (Employee Benefit Research Institute, September 7, 2023). their ability to save for retirement. Thus, it is clear that debt is an issue for many Americans. This study builds on the 3) Once am theo se ngh o th uo ss eh e o w ld its h ar loe wier den intcifoied mes. , at least one of the individuals in the household is also a DC plan participant card balances of those with a spike vs. those without one still held. study 3o .0f % irregular spending spikes can provide insight into DC plan participants’ financial decisions, including taking a This st 35 ud %y is part of a joint effort between the Employee Benefit Research Institu 37.1% te (EBRI) and J.P. Morgan Asset (omitted variable account balances $5,000-$19,999) As far as financial factors, 21.0 percent of the household participants had DC plan account balances of less than $2,000 h Pr ttopb s: it70 / /R w % esu wwl.tfed s oer n a Flac rese tor rs ve A .g ffo ec v/triel ng e ase thes /Pr g1 o9 b/ab cuirlrit en y to /fdefau Takin lt.g h ta m Pl . an Loan ................................................................. 18 p rior J.P. Morgan/EBRI study that looked at the link between spending, credit card debt, and 401(k) plan loans among 2 4 •0 % These spending spikes have a clear impact on the likelihood of public-sector DC plan participants taking a plan where a loan is available in their plan. The demographic and financial characteristics of the person identified as the DC For example, see Holden, Sarah, and Jack VanDerhei, “Contribution Behavior of 401(k) Plan74% Participants.” EBRI Issue Brief 80% plan loan and/or increasing credit card debt. A spending spike could be a result of an unexpected expense, e.g., a car 50% or Management to deliver Ad gd rat inc- a- gd ro riss ven in co res mear e in c$h 1 0 to ,0 0 b0 et ster understand how the financial factors that exist outside defined and 6.5 percent had balances of $2,000–$4,999, while 17.2 percent had balances of $100,000 or more (Figure 2). Over 24.4% 33.6% 35% 35.6% 16 $12,000 The privat likel e-siec hoto od r DC of e p xlp an er ip en arctiin cig p an a s tp s ike to d in et crer ease mind e w if it th h et h se a m sp een lin dks in gar re atfio o uan ndd abm eg oin ng n ip nu gb -of lic--the sec-tyear or DC c rp ed lan it c par ard ti cuip tian liza ts ti on plan partlio Mo can ip re an an t d ar ie n ctrheasin ose u gs ed thei in r c th red e an it c al ar ys di sd.eb t in the year of the spike. Of those with a spending spike in the Co T nh cese lu 30 si pa % on r................................ ticipants were not likely ................................ to both increase c12 redi................................ t card debt and take a n ................................ ew plan loan in the an.......................... alysis year, as only 18 no. 238 (October 2001). Available at https://www.ebri.org/publ36.0% ications/research-publications/issue- 50.9% 60% 23% repair, or an expected ag ex e2p 0= en 1 s ife a , gseu c is hl ea ss s a tha vac n 3at 0ion. contribution plans that face DC plan participants impact their retirement preparat $10,994 ions. Thus, the aim is to provide two-fifths (43.7 percent) of these households had no revolving credit card debt at the end of year one, while 16.4 an (Fid g ufr oeus n10 d th an atd t h 11 e )s. aIn m e co rn eltat ras iotn , sth hie p sli kel exiis hto a om d o on f ga ts hpeis ke e w dec orrker ease s a ds as w el glr. oT ss h ese inco lm ink es i n bcet rease ween d s (F pien gud re in g 12 an ). d debt 2. 70 0% %analysis year, 7.0 percent took a new plan loan and 31.7 percent increased their credit card debt, compared 2.1 percent of those with spikes and 0.9 percent of those without spikes did both. br Foir e fs/ uncfo un ntd en ed t/ fu spllen /co dn in trgibu sp 34.0% tiio kes n-beh (spav en io dri-nof g -s 4u 0r 1g (es k)- pl no an t -cp oar ver ticed ipan by ts -in 1c 5o 4m ; M e uan nnd el c l,as Ah lic ira es H er .,ves Ann ),i ka 29 Su per ndcén en, a t o nfd C hoauts heh erio nle d age40=1 if age is 40-49 u En nid qn uo e tes fac ................................ t-based insights to hel ................................ p build a stronger ret................................ irement system by pol................................ icymakers, plan spons............................ ors, and plan 21 percent were using 80–100 percent of their credit card limit. Over two-fifths (43.2 percent) of the households had 25% 30% 4) Since the status of many of the variables must be known 28.1% at the beginning and the end of the study year, these suggest 50% that retirem 26.4% ent planning is not wholly different by place of employment, even where benefits availability may $10,000 with 2.7 percent and 25.9, respectively, of those without a spending spike in that same year. As Tay n lo ort, ed “W , h a at m D on ette hrlm y iu nn es fu4 n0 d1 e(d k )s P par entd iciin pg at s io pn ike an id s C do en fitn ried bu ta io sn a s?” sp C ike RRat W loea rks in t g 25 Pap per erc , n en ot. 2 ab 0o 0ve 0-1 t2h.e Cp hr es evi tno uu t H s 1 ill2 , M mA o:n ths’ observations were found to have had at least one month where an unfunded spending spike occurred. In addition, 7 17 30% age50=1 if age is 50-59 pr o In vid a er dis ff. e rent specification using a simple ordinary least squares (OLS) regression of the same variables, a spike occurrence 78.7% spend 60 in % g ratios of 1.05 or more, while one-third (33.4 percent) had ratios of less than 0.95. These spending spikes have a clear impact on the likelihood of DC plan participan 25.7% ts taking a plan loan and increasing households must have two contiguous years in the sample to be included 58% . Thus, each instance of a household having be dissimilar, but part of a broader holistic financial planning journey where all factors need to be incorporated. In fact, • 7 Ho 5% o us reholds are more likely to take on additional credit card debt before taking the plan loan, as approximately Center for Retirement Research at Boston College, December 2000. Available at https://crr.bc.edu/wp- median spending that a gc ean 60n =o 1t ifb a eg f eu in s d 6ed 0 o b r y old th ere household’s income and available cash reserves in that month. percent of household observations had three or more months of these spikes (Figure 4). Overall, the a 22.4% verage number 20% 40% and higher beginning-of-the-year credit card utilization are also strongly associated with increases in the likelihood of taking a More 23.5% their 25 1. c % r 0% edit card debt in the year of the spike. Of those with a spending spike in the analysis year, 7.0 21 percent took a two contiguous years of complete data during 2019–2021 is an observation for this analysis. This results in 5,755 19% participating in a budget webinar has been found to be associated with higher DC plan contributions. Programs to content/u37 plo –ad 47s/ p 2er 00 c0 en (/o 1tm 2 o /itw ft ep td h _ o v 2a s 0re 0 ia 0 w b-l1 ie t2 h a . p g ce df red s ; 3an i0 t -c 3 dar 9 U )d S G uten ilizeat ral io A n co cfo u >n0 t– in7 g9 O pff er icc een , “t 4 0 in 1c (r keas ) Pen eds itohnei P rl an cred s: L ito can ar P d rd oeb vistio , n ws hile less Fig Separ ur $8 at ,0e el 00s y, spending that cannot be funded by the household’s income alone is discussed. o Sf p ec these ifical sly, pike pu sb am lic-o sn ec gt o th r oDC se h pavin lan p gar th tie cm ipan wtas s w tw ho o . 22.2% experience spending spikes are compared with those who do not Public50 -s% ector DC plan participants in most cases are also covered by a defined benefit plan, and the DC plan is loan. new plan loan and 31.7 percent increased their credit card debt, compared with 2.7 percent and 25.9, respectively, of observations from 3,709 unique households. 15% help w 30it % h workers’ overall finances — for example, financial wellness benefits — could be indispensable. The decision to Enhance tPa han rtic 8 ip at per ioc nen tbu en t t 2t M = o1 o ay k if A a teff n nu ec erw et In w piltc an ho cu me lo rran e Sne tc w e um ir tih p ty lo t fo h ye at rr S ils ev olme. eel ss o ”t f hLe a cn rt ed t2 e r ye i tR a cep rar so dr tu , G tiliA zO at /iH oE nH . S H-o 98 w-ever 5 (O,c t w oh be en r 1 c9 red 97) it. W car as dh ington, experience them in terms of their credit card utilization and DC plan loan use. First, the households with participants considered a supplemental plan to the DB plan. On the other h 16and, for private-sector participants, the 11.5% 401(k) plan, in Figure 1, Demographic Characteristics of the Sample ............................................................................................... 7 18 20 20% % 52.2% those without a spending spike in that same year (Figure 13). This same relative result was found among the private- B S a p y tS iake kes an imy i l ar a c m an f peasu iln an b de i n l ro q gs e, an u fo ih t e io s ru t lar td s e hep eh n e g5e d = en oem 1 r ld el d i fen at m o tg eo itr v n n aph u e nt ro h e to tliiy c s ji fa u n s 5s p c -c t 9 o en t o m o ye n d re s a in w . rw s g Wh h e at irse i lg h e fo en ap 2 u9 n er p d i en p al er n lsy t c ien h h ne i g tth OL h o e ly fS p tvar r lh an egr e i ab h b ess o ulu e t s i.o o eh In n n a o tfh ls d ac e w o ttel ,b o l9 s t.al er 0 v p fiat er nan ic oen n cs ial t h o ad p f rto h sfe p ile ih kes o ou f s toh eh fe 2 op 5 ld ar p ter icc ip en an t t. DC: U40 S G %eneral Accounting Office. A 17.7% vailable at www.gao.gov/assets/hehs-98-5.pdf. utilization reached 80 percent or more, the likelihood of increasing credit card debt decreased to 22.4 percent, $6,000 16.2% who have a spending spike are identified. Once identified, th 43.2% e 46% impact of the spending spikes on credit card debt and many cases, would be the only retirement plan available through their workplace. Even with the supplemental nature of 0.0% 20 10 % % 100% or 7.6% sector participants, but th 37.7% e percentages increasing debt and taking plan loans among the private-sector participants ten10=1 if tenure is 10-19 years observations were found to have had at least one month where their spending was 25 percent or more of their median o Firg lu ar re g er 2,, F 2i3 n an per cic al en Ch t h ar ad ac s ter pikes istic s o fo 5 f 0 th p e er Sc am enp t lo er ................................ larger and 19 percen................................ t had spikes of 75 perc ................................ ent or larger (Figure 5 ...... ). 7 19 while the increasin 0% g trend of taking a new plan loan 20 – w 59% ent up by nearly twice the amo 80 u– n 100% t it had before the 13.5% 3 More plan See loB an oar ud o sagf G e ar ove e rtn ho en rs as of stess he F ed ed . er al R 5.6 es % erve System, Consumer Credit G.19, the public-sector DC plans, public-sector DC plan participants were less likely to take a loan from the plan when eligible For the most recent results, see Gropper, Michael, “The State o 5.2% f Public Sector DC Plans: 2021” PRRL Research Study, (PRRL 15% 15% 4.8% w Gier ven e 30 lar th %g e er im . p act of participants’ overall finances on the need for a plan loan, it appears clear that prohibiting plan loans Data Definition tes n 2 0=1 if tenure is 20 or more years spending in the prior 12 months and was not covered by that month’s income (Figure 3). Thirty-one percent of the On a dollar basis, 60 percent of the household observations had spikes not covered by income and cash reserves larger 105% %increase to 80–100 percent at 11.5 percent. 10 https://www.federalreserve.gov/releases/g19/current/default.htm. $4,000 $3,564 t Sh ept ane tmbe hose r 2 in 8 ,t 2 h0 e2 p 3r)i vat and e T sh ec eph tora si — t10.6% ,7 S am peric ta en an t d M vs. i1 c5 h ael per G cr en opper t. ,In “A t h Lio sn sgtiu tu ddi y,n o alver An a 9l6 ys p is er oc f en Cotn o sifs t th en et p Pa ub rtliic ci-p san ectts o ri nD tC he Figure 3, Percentage of Households With Monthly Spending 25 Percent or More Than the Median Spending of the Prior Source: Estimates from the PRRL Database and select Chase data. For more information, see the Data Sources box in the text. would not necessarily (oim mitp te ro dv ve arp iaar blt eic tie pnan ure ts s ’ orfet 2i-r4 em yeen ars) t security. Without the option of taking a plan loan, participants hous 10 eh %olds had four or more $2,977 months where their spending was 25 percent or more of the median spending of the prior than $2,500 aggregated over the year, and 82 percent had spending not covered by income alone above this threshold Data Sources 11 Public20 R% etirement Research Lab Database, 2019-2021,” PRRL Research Study, (PRRL August 8, 2024). 1 p 202 lan M o pn ar th tis c ip Nan ot tCo s w ver hoed w er by e In inc a om ple an ................................ with a loan option av ................................ ailable were in a 457 ................................ plan or a non-ERISA 4......................... 01(k) plan. 8 Thos 10 e %with a spending spike not only had greater likelihoods of taking a new plan loan and increasing credit card debt, Spending — Total sccu pen tb d-icr ng e dis it ca the rd an utin lizu aal tio s nu am t t h oe f b th ee g in m nio nn g to hfl y ths ep st en ud dy in (g re v co ap lvit nu gr ed bal atn hce ro/u lim gh it) credit and debit cards, Empl 0% o 0% yee Benefit Research Institute and Greenwald Research, Retirement Confidence, 2024 RCS Fact Sheet #1 (Employee would seek loans outside the plan to fill spending gaps, an 13 d those loans may have terms more expensive than those of This research found that, like private-sector DC plan participants, public-sector DC plan participants who lack income 21% 12 months, with an average of three months for those having these spikes. Given that9.5% nearly all of the observations among those with incomes of $150,000 or less (Figure 6). Fifty-nine percent of those with spending not covered by 200% or 0.0% 0% 12.7%>0%–19% 20% 20%– –59% 59% 60%–79% 80%–100% 80%–100% 4 but th $2e ,0av 00 erage amoiu ncr ntd o 2= f 1 th ife re lo van olv in an g d cr e th die t ca avrer d d ag eb et iin ncr cre ease ased ib ny cm red ore it tc har and $ 1 d0 eb 0 t d u w rier nge t h al es st o uld ar y g ye er a rthan they were for electronic payment transactions, Chase checks, and cash across 10 specific spending categories: apparel & services, Benefit Research Institute, April 29, 2024). More a Ip nl a an pr loio an r s . t udy using the PRRL Database and Chase banking data, spending ratios of the public-sector DC plan participants and cash reserves to support a spending spike are likely to end up with more credit card debt. This higher debt can P h Fiad RRL gu rte h iD s 4 ,at i rPer reg ab cu as en lar e tly ag — h e iT g oh h f e Ho s PR pen uR seh d Li n Dat ogld , s ab s iWi ga nst ie fh ic iS an s pan en t d d o ifip fn er tg -i en n S p c cio es kes ll aib n o o p fr ar at 25 ito i c Per n ip am an cen to b n t eh g o rp avi M ub o o lr irc e s rT b et h as ian red em t h o en e n tM t h ped ils an is an p sen p S op d nen is no gd r sin n , o g tth o c efo ver thee P dr ib oy r income alone had totals 8% larger than $7,500. For households with incomes of more than $150,000, 79 percent could not 10% 4.8% Median-$36,261 those 5% without a spenn dein wg m s tg p=ik 1e if. m Th oro ts ge a gw e ip th a ym spe ike nts s st ha ad rte a d Mn e d dian aver urin-g $ 2 ag t9 h ,6 e e 9 st 1ou ud ty stye anad ring new balance M,edian or ad -$18,746 ditional balance for education, entertainment, food & beverage, health care, housing, transportation, travel, charitable contributions, and who also had a DB plan were compared with those who only had a DC plan. Also, the spending ratios were compared among h 1 212 ave Moan lto hn sg N -lo ats tCo ing ver im ed pac by t o In nc roet mir e ean men d Cas t sec hu R rie ty, sersves ince ................................ higher credit card ut................................ ilization is correlated w ............................. ith lower DC plan 9 Employee Benefit Research Institute (EBRI), and the National Association of Government Defined Contribution income alone were not found. However, this definition is an important measure to con4.4% sider for policymakers and plan f u n “F di eslp d en of dD in rea g s ms pike ? M s eas over ur i$ n2 g t ,5h 0e 0 Iw mpa ith ctth o ei f F r iin nan com cia el W anel dl bei cas nh g In res iter iaves tives , an ond 4 9 03 1 (p k) er Pclen an tUt co iliu zla dt io nn o,t” dEoB s Ro I Issu withe tB hrei ief r , no. 8.2% 0% Source: Estimates from the PRRL Database and select Chase data. For more information, see the Data Sources box in the text. spk=1 if a spending spike occurred during the study year T thhois s e resear havincgh a folo uan nd itn hcat rease , like fr po rm ivat th ee -s p ec rito orr year DC p , lo an f $p1ar 4,t0 ic1ip 5an vst.s ,$ 1 p0 u,b9li9 c4 -s fec ort o th r o D sC e p wlan ith opuar t ta ics ip pan iket s( F w ig hu or e lac 14 k )i.n c To hm e e other. $0 2.6% those w 0% ith DB plans by thei 14r level of tenure. DC-plan-only participants were found to spend less relative their income, on 0% <0.80 10% 20% 30% 40% 50% 60% 70% 80% 90 2.% 00 or More 100% contributions and account balan0.80 ces, – 0.94 even when co 0.95 ntr –o 1.04 lling for incom 1.05 e. –T 1.49 hus, the avail 1.50 abi –l1.99 ity of emergency savings to s Ad po m nis nois rs tr at aso trh s ey (Ng AG rap DCA) ple .w T ith he t h de at L eap a ssb Tha p as ro e np $ irn 2iat 0 c,0 lu e 0d 0seis z ed o at f ae m fr $o 2 e0 m r,0 g 0 en t0w -$c o 9y 9h ,9 ru eser 99 ndrve eds an fod r $ 1ts 0h i0x e ,0 t y 0w 0 - o s oe rr m kin ven org e 4 p5 o7 p(u bl) at , i4 o0 n1 , (b a) ec , au 40s 1e ( k) w,h a at n d is income alone (Figure 7). 554 (Employee Benefit Research Institute, March 10, 2022). Increase in Credit Card Debt New Loan Amount Figur0% e 5, Spending Relative Abotvo e It nh ce om M e a ed ndi Re ans o erf v etshe Prior 12 Months of Spending — Per Abc oen ve Itn ag coe m eAbove Specific Thresholds aver and ag cas eh i n reser crease vesi n t oc rsed upip t o cr ar t d a d sp eb en t d fo in r gt h so psike e w ar ith e la ikel spy ik teo w en as d $ u3 p, 5 w 6i4 th c m om or pear ced red w iti tchar $d2, d 9eb 77t .f o Trh t is h o hs ig e hw er it h do eb utt a can sp ike. average, than those with a DB plan, while the tenure of the DB plan participants did not appear to have an impact on the cover sSo pen urcd e: in Esg ti m sap teike s frosm c PRR an Lb Da e ta ab a csreit aic ndal s ef le ac ct t Cha or si en d ap tar. even For motre in ig nfo o rm r as tit oal n, l sien eg th e a Dat cyc a So leu rc of e si n boc xr iea n ths ei n teg xt. debt that can significantly 403(b) defined contribution (DC) plans; over 3.0 million retirement accounts across 2.5 million state, county, city, and held outside of the DC plan is not known by plan sponsors. $20,000 Source: – Es $29,999 timates from the PRR $30,000 L Datab– as $49,999 e and select Chase d$50,000 ata. For m –o $74,999 re information, see th$75,000 e Data So –u $99,000 rces box in the text. $100,000 or more Source: Estimates from the PRRL Database and select Chase data. For more information, see the Data Sources box in the text. Source: Estimate Sso fro urm ce th:e E PRR stim L Dat ate as bafsre o a m nd ts he e le P ct R Cha RL se D da a ta ta . Fo ba r se mo re a n ind fo se rmale tioct n, s C eh e a thse e Dat da a So ta.u rces box in the text. of Median Monthly Spending .................................................................................................................................. 9 Income — Since all the spending data are at the household level, the income used in this study is also at the have a long-lasting im $1 p–ac $1,000 t on retire >m $1en ,00t 0 –s $2,500 ecurity, sin>c$2 e ,50 hig 0–h $4,000 er credit car >$4 d,00 ut 0i– li$7,500 zation is coMo rrel re Tha ated n $ w 7,50 ith 0 lower DC plan amount spent relative to income. See Copeland, Craig, Kelly Hahn, and Matt Petersen, “S 6 pending and Saving Behavior of impact retirement readiness, wherever the individual works. subdivision government employees; and $170 billion in assets as of year-end 2021. For more information, see the Data Sources box in the text. household level. There are two income values used in this study from the Chase data. Source: Estimates from the PRRL Database and select Chase data. For more information, see the Data Sources box in the text. c ontribuSo tio urc nes : Es an tim d a te ac s fro co mu th n et PRR bal Lan Dac taes, base even and se lew ct h Cha en s ec do ata n. tFo ror lm lin ore g inffo orm r ian tio cno , m seee. th eThus Data So , utrc he e s b a ovail x in tab he te ilxit t.y of emergency savings to Public-Sector Defined Contribution Plan Participants,” EBRI Issue Brief, no. 570 (September 19, 2022). Figure 6, Distribution of Annual Spending Amounts Above Income and Cash Reserves.............................................. 10 e e e e e e e e e e e e e e e e e e e e eb b b b b b b b b b b b b b b b b b b b br r r r r r r r r r r r r r r r r r r r riiiiiiiiiiiiiiiiiiiii.....................o o o o o o o o o o o o o o o o o o o o or r r r r r r r r r r r r r r r r r r r rg g g g g g g g g g g g g g g g g g g g g IIIIIIIIIIIIIIIIIIIIIs s s s s s s s s s s s s s s s s s s s ss s s s s s s s s s s s s s s s s s s s su u u u u u u u u u u u u u u u u u u u ue e e e e e e e e e e e e e e e e e e e e B B B B B B B B B B B B B B B B B B B B Bri ri ri ri ri ri ri ri ri ri ri ri ri ri ri ri ri ri ri ri riA e e e e e e e e e e e e e e e e e e e e efffffffffffffffffffff re • • • • • • • • • • • • • • • • • • • • • s e Se Se Se Se Se Se Se Se Se Se Se Se Se Se Se Se Se Se Se Se Se arc p p p p p p p p p p p p p p p p p p p p pttttttttttttttttttttth e e e e e e e e e e e e e e e e e e e e e m m m m m m m m m m m m m m m m m m m m m reb b b b b b b b b b b b b b b b b b b b b pe e e e e e e e e e e e e e e e e e e e e or r r r r r r r r r r r r r r r r r r r rrt 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 f,,,,,,,,,,,,,,,,,,,,,r o 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2m 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 02 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 t4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 h e • • • • • • • • • • • • • • • • • • • • • E N N N N N N N N N N N N N N N N N N N N NBR o o o o o o o o o o o o o o o o o o o o o..................... 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 I 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 Ed 8 8 8 8 8 8 8 8 8 8 8 8 8 8 8 8 8 8 8 8 8 ucation and Research Fund © 2024 Employee Benefit Research Institute 15 12 21 14 13 18 17 10 22 16 20 19 11 8 5 9 6 4 2 7 3