the inefficiencies of existing retirement savings incentives

8/20/2019 The Inefficiencies of Existing Retirement Savings Incentives

1/18

1 Center for American Progress | The Inef fic iencies of Existing Reti rement Savin gs I ncentives

The Inefficiencies of Existing

Retirement Savings IncentivesBy Christian E. Weller and Teresa Ghilarducci October 29, 2015

America’s middle class aces a growing reiremen crisis. More han hal o all working-

age households are expeced o be a risk o having o cu back heir sandard o living

ofen making painul adjusmenswhen hey reire.1 Tere are several reasons or he

ever-larger looming crisis, bu people’s inabiliy o save enough money is a key obsacle

o achieving more reiremen securiy. On average, Americans need o save beween 10percen and 20 percen o heir salaries each year ouside o Social Securiy o ensure a

secure reiremen.2 Ye nearly one-hird o working-age Americans have no reiremen

savings or pension,3 and less han hal o all privae-secor workers paricipaed in a

reiremen plan a work in 2013, he las year or which daa are available.4

Te growing reiremen crisis resuls, in par, rom inefficien savings incenives embed-

ded in he U.S. ax code. Households ha need he mos help saving or reiremen

receive he leas assisance rom he muliude o savings incenives. Te ederal govern-

men and several sae governmens use he ax code o encourage people o save. Tese

savings incenives ypically come in he orm o ax advanages and vary by ype o sav-ings, such as individual reiremen accouns, or IRs; 401(k) plans; or oh IRs.

Tese ax incenives, however, all shor o allowing workers o secure adequae reire-

men savings. Firs, exising savings incenives can be overwhelming and incredibly

complex. People need o undersand which savings plans are available; how much hey

and heir employer can conribue o he various plans; how long o keep heir money

in ax-advanaged savings; and how heir decisions inerac wih curren and uure ax

raes or personal income and capial income.

Second, savings incenives ofen benefi higher-income earners more han middle- and

lower-income earners. Higher-income earners ace higher ax raes and hus enjoy

greaer ax breaks rom exising savings incenives; hey can beter ake advanage o

maximum conribuions o muliple reiremen plans because hey have more income.

Higher-income earners are more likely han lower-income earners o have a reiremen

plan hrough heir employerwhich come wih more savings incenives han oher

reiremen plans. Higher-income earners also earn a higher ne o ax rae o reurn and


2/18


pay lower ees, so even i a high- and low-income earner save exacly he same amoun

o money, he higher earner will accumulae more reiremen asses.5 All o hese acors

end up boosing savings he mos or higher-income earners, who arguably need he

leas assisance o save or reiremen.

Tird, even as he savings incenives ail o prepare households adequaely or reire-

men, he public loses ou on increasingly large amouns o ax revenue ha oherwise would have been colleced wihou hese ax breaks. Federal and sae governmens

orgo a subsanial amoun o ax revenue o creae incenives mean o help people save

or reiremen bu in realiy produce litle addiional savings. Te ederal governmen

alone annually orgoes more han $100 billion in personal income ax revenue due o

reiremen savings incenives.6 And sae governmens wih income axes urher lose

ou on subsanial ax revenueabou $20 billion, according o one esimae rom

researchers a Te New Schoolas hey generally offer he same ax breaks on sae

income axes as he ederal governmen.7

o be clear, savings incenives are no a bad idea. In he Unied Saes, however, heexising ax srucure has ailed o adequaely prepare mos people or reiremen. Tis

issue brie will illusrae he link beween he reiremen crisis and savings incenives

and urher examine several key elemens wihin his relaionship:

• Exising reiremen savings incenives are inefficien, as hey are unnecessarily com-

plex and skewed in avor o higher-income earners.

• Tese savings incenives exacerbae inequiies in a sysem ha heavily relies on

employer-based reiremen benefis such as 401(k) plans, as access o employer-based

plans is unevenly disribued and as such plans offer greaer ax advanages han non-employer plans such as IRs.

• Lower-income earners receive less o a benefi rom exising savings incenives han

higher-income earners.

• Wih litle help available rom reiremen savings incenives, a growing share o house-

holds is inadequaely prepared or reiremen.

• Inadequae reiremen savings are unevenly disribued. Te reiremen savings shor-

all is especially pronounced among lower-income households, communiies o color,

and single women.

Tis issue brie highlighs he need or policymakers o address he realiy o he grow-

ing reiremen crisis. Amid inacion, a growing number o Americans will spend heir

golden years in povery.8 More reirees will sruggle o pay heir bills, rely more and

more on help rom relaives and riends, and simulaneously increase demand on public


3/18


saey ne programs.9 ax reorm could play an inegral par in addressing he loom-

ing shorall in reiremen savings, largely by simpliying savings incenives and beter

argeing incenives o hose who ruly need help preparing or reiremen.

Savings incentives in the tax code

Exising reiremen savings incenives reduce axable income or ederal and sae

income ax calculaions.10 Alhough employee and employer conribuions o radiional

reiremen savings plans such as 401(k) plans and IRs11 are par o an employee’s

income, hey are ypically no subjec o ederal and sae personal income axes

unil money is disribued rom such plans. Te axable amoun o personal income

is reduced overall when conribuions are made o hese plans, creaing an incenive

o save. Te money in a ax-advanaged reiremen savings accoun hen accumulaes

wihou households paying personal income ax on earnings. Personal income axes are

due, however, when people evenually wihdraw money rom heir reiremen savings o

spend in reiremen.12

Households can liquidae some o heir reiremen savings accouns beore reiremen,

wihin limis. Curren rules or 401(k) plans, or insance, allow or so-called hardship

wihdrawals, which include medical emergencies, prevenion o evicion or oreclosure,

uiion paymens, purchase o a primary residence, uneral expenses, and some expenses

or repairs on a primary residence. However, hardship wihdrawals can be aken only

while sill working or an employer.13 Te employee has o pay income axes and, ypi-

cally, a 10 percen excise ax on any hardship wihdrawal.14 People can also premaurely

wihdraw money rom an IR, as long as hey pay he associaed income and, ofen,

excise axes.15

In shor, he ax code includes financial hurdles o wihdrawing moneyrom reiremen accouns premaurely.

Households may alernaively access some o heir reiremen savings by aking ou a

loan rom heir reiremen savings accouns. Tey could, or insance, borrow rom heir

own 401(k) plans. Tere are limiaions similar o hose imposed on hardship wih-

drawals, hough here are no immediae ax penalies or aking such loans. Loans rom

401(k) plans are consequenly more prevalen han wihdrawals.16 An employee sill has

o work or an employer o ake ou a loan rom his or her own 401(k) plan, however,

and ypically has o repay a loan wihin 90 days afer losing a job. Oherwise, ax penal-

iesincome axes plus a 10 percen excise axapply.17

Limied access o reiremen savings prior o reiremen is a mixed blessing. Having

some access o savings prior o reiremen may increase savings or some people, since i

increases flexibiliy. A he same ime, however, sudies have shown ha prereiremen

wihdrawals can lead o reiremen savings depleion.18


4/18


Complexity of existing incentives slows savings

eiremen savings incenives are ofen complex and conusing o he exen ha even

households who would like o ake advanage o hem may be deerred rom doing

so. eiremen savings incenives alone orm a complex web o differen opions and

varying incenives ha is difficul or mos households o navigae.19 o maximize all

available reiremen savings incenives, or insance, people need o have exensiveknowledge o he separae ax reamens o varying reiremen savings plans. o receive

he maximum ax benefis, people ideally mus deermine:

• Wheher heir employer offers a reiremen benefi a work such as a 401(k) plan or

a defined benefi, or DB, pension and wheher and how much money hey and heir

employer can conribue o such plans

• How much money hey will need o save or hings oher han reiremensuch as

emergencies, healh care, and heir children’s educaiono avoid premaurely wih-

drawing money rom reiremen savings accouns and acing ax penalies

• Which reiremen savings ouside he employer-based sysem are available o hem,

such as IRs

• How o compare heir marginal ax raes in he curren ax year wih he esimaed ax

raes or he res o heir careers and in reiremen o deermine which savings opions

are more advanageous rom a ax perspecive

• How hey wan o spend heir savings in reiremen o ge he maximum ax benefis

rom savings incenivesor insance, how much money hey wan o wihdraw regu-larly and how much money hey would like o leave o heir children

Giving households some choice o savings opions is no in and o isel a flaw o savings

incenives. Behavioral economics, however, has shown ha overwhelming consumers

wih excessive choices is effecively he same as providing no choice a all. Muliple sud-

ies show ha oo many choices in key decisions can overwhelm and rusrae consum-

ers, resuling in consumers relying on heurisicseducaed guesseswhen making

complex choices20 or making no choice a all.21 People may even absain rom choos-

ing anyhing ou o a ear o making a choice ha could end up damaging he financial

well-being o he household, such as limiing access o savings in an emergency.22 Many

people are simply unaware o or misundersand he complicaed ax rules under which

hey could access heir reiremen savings beore reiremen.


5/18


Existing savings incentives skew toward higher-income earners

eiremen savings incenives offer less assisance o lower-income households han o

higher-income ones. As a resul, households ha presumably need more help o save

or reiremen ge less suppor rom he ax code han hose households ha are already

able o adequaely prepare or reiremen.

Te link beween income and he value o reiremen savings incenives is complex. Tis

secion discusses some o hese complexiies and provides some illusraive calculaions

o show how exising savings incenives are skewed oward higher-income earners.

Consider he mechanism used o encourage conribuions o a reiremen savings

accoun. Households deduc heir reiremen savings conribuions rom heir curren

axable income, hus reducing he amoun o income subjec o axaion. Te ederal ax

code is progressive: Higher-income earners pay higher marginal axeshe axes due on

heir las dollar earnedhan lower-income earners. Because higher-income earners ace

higher marginal income axes han lower-income earners, hey have a sronger incenivehan lower-income earners o reduce heir axable income wih a ax deducion.23

Te highes ax brackeor hose annually making more han $406,750 individually

or $457,600 joinly in 2014is 39.6 percen.24 Earners in his ax bracke would lower

wha hey owe on heir ederal income axes by 39.6 cens or each dollar conribued

o an eligible 401(k) or IR. Lower-income earners, by comparison, may ace a mar-

ginal ax rae o 10 percen and hus save only 10 cens in curren-year income axes or

each dollar hey conribue o a reiremen savings accoun. Moreover, people do no

pay personal income axes on earnings in heir savings unil he money is wihdrawn.

Higher-income earners again benefi more rom his ax advanage han lower-incomeearners because hey ace higher marginal ax raes, so hey save more money beore

hey wihdraw heir savings.

Exising savings incenives disproporionaely avor higher-income earners. Figure 1

illusraes he unequal disribuion o ax incenives. I shows he esimaed amoun o

ne pension conribuions and earnings on reiremen accouns as a share o afer-ax

income, by income percenile in 2013. Te daa show ha he conemporaneous ax

benefi as a share o income increases as incomes increase.

Higher-income earners benefi significanly more rom hese savings incenives han

lower-income earners, relaive o heir incomes. Households in he op fifh o he

income disribuion, on average, receive savings incenives equal o an esimaed 3.1

percen o heir income, almos wice as much as he 1.8 percen or households in he

second-highes fifh o he income disribuion.25 Meanwhile, households in he lowes

fifh o he income disribuion receive only a racion o hose benefis, wih an average

o 0.4 percen o average ax income.


6/18


Te picure ges less sraighorward when looking a a household’s enire financial liespanrom saving and invesing o ulimaely wihdrawing heir money. Accouning

or some real-lie aspecs o reiremen savings shows ha higher-income earners may

benefi even more rom hese savings incenives han suggesed by a simple comparison

o marginal ax raes a he ime o conribuions o reiremen plans.

Le’s sar wih he basic ax complicaion. Households are supposed o pay axes in he

uure, when hey wihdraw heir savings. Because households are supposed o bu ofen

do noor insance, because hey experience a decline in marginal ax raes when hey

reireevenually pay axes on he money in heir reiremen savings, ax expers reer

o hese savings incenives as ax deerrals, raher han ax breaks or ax shelers.26

Bueven afer accouning or he ac ha people evenually could pay axes on heir savings,

high-income earners sill benefi more rom ax incenives han lower-income ones, as

he simulaions below show.

Te logic o he ax deerral argumen goes as ollows: Because he ax code reas

wihdrawals as personal income, ax paymens upon wihdrawal will vary wih per-

sonal income ax raes in he same way ha he iniial ax benefis varied wih income.

Higher-income earners will pay higher income axes in he uure han lower-income

earners when hey wihdraw heir unds because he ax code is progressive, wih higher

marginal ax raes or higher incomes han or lower incomes. Fuure ax paymens upon

wihdrawal will parially offse he iniial ax benefis. Due o he progressiveness o

personal income axes, he offseting effec upon wihdrawal is larger or higher-income

earners han or lower-income earners. High-income earners, hereore, receive larger

ax benefis upron bu also pay higher axes in he uure.

FIGURE 1

Net pension contributions and earnings as a share of after-tax income

By income percentile in 2013

Source: Congressional Budget Office, “The Distribution of Major Tax Expenditures in the Individual Tax System” (2013), table 2, available at

http://www.cbo.gov/sites/default/files/43768_DistributionTaxExpenditures.pdf. All figures are in percent of income.

0%

1%

2%

3%

Bottomquintile

Secondquintile

Middlequintile

Fourthquintile

Topquintile

81st to 90thpercentile

91st to 95thpercentile

96th to 99thpercentile

Top 1

percent

0.4% 0.7% 1.2% 2.0% 2.0% 2.7% 1.7%1.6%0.8%


7/18


Tis is no o say ha here are no ax benefis and ax revenue los when boh iniial ax

benefis and uure ax paymens are couned. Addiionally, he ne ax benefisax

benefis minus ax paymensare ypically much larger or higher-income earners

han or lower-income ones, as he calculaions below illusrae. Te main benefi is ha

households can generae invesmen earnings on a larger invesmen han would be he

case wihou iniial ax incenives. Households benefi rom he power o compounded

ineres on a larger iniial invesmen by deerring axes. Higher-income earners sillge more value rom his ax-ree compounded ineres effec. Compared wih lower-

income earners, hey ge o keep and inves more money ha oherwise would have

gone o he governmen.

Real-world circumstances magnify tax benefits for higher-income earners

Higher-income earners can also benefi rom ax deerral due o imporan real-lie

aspecs o saving or reiremen. Tese circumsances exacerbae he inequaliies already

inheren o he exising savings incenives.27

More opportunities to save

Te ax benefis are greaer i households have more opporuniies o save on a ax-

advanaged basis. Having such an opporuniy ofen depends on wheher one has access

o a reiremen plansuch as a 401(k) plana work, since such plans allow or more

annual conribuions han nonemployer-based reiremen savings plans such as IRs.

Higher-income earners are more likely o have access o an employer-sponsored reire-

men savings plan han lower-income earners.28

Higher-income earners are also morelikely o ake advanage o addiional reiremen savings plans, such as IRs, due o

addiional sources o income, including sel-employmen.29

Greater potential for decline in marginal tax rates

A household’s marginal ax raes can decline or wo reasons: (1) when people reire

because heir income decreases; and (2) because older households are able o enjoy

addiional ax breaks no available o younger households.30 Lower uure marginal ax

raes mean lower uure ax paymens, hereby generaing ewer offses o balance ou

he iniial ax benefis and greaer overall ax benefis over one’s lieime. Households

wih high incomes will have high marginal ax raes when hey conribue o heir savings

accouns and during heir careers when hey inves heir savings. Imporanly, hough,

higher iniial marginal ax raes se or higher-income earners have more room o all

han lower iniial marginal ax raes se or lower-income earners. Due o he increased

poenial or changes in income level, and hereore, in marginal ax rae, higher-income

earners may have a larger ax benefi over ime han lower-income households.


8/18


Earlier opportunities for saving

Tird, he benefi rom ax deerrals in reiremen savings accouns depends o some

degree on ax-ree compounding o ineresearnings on invesmens buil up over ime

wihou being subjec o axaion. Te longer ineres compounds, he larger he ne ben-

efi o he household. Higher-income earners may sar o save earlier and wihdraw sav-

ings laer in lie han would be he case or lower-income households,31 simply becausehey have more income and hence more flexibiliy o save or boh reiremen and oher

goals, such as emergencies, a down paymen on a home, and children’s educaion.

Tax-advantaged inheritance

Lasly, he ne ax benefisax benefis minus ax paymensrom saving in a ax-

advanaged reiremen accoun increase when no all money is wihdrawn rom a

reiremen savings accoun and hen passed on o heirs.32 Higher-income households are

more likely han lower-income households o leave money or heir heirs in heir IRs.33

Illustrating the workings of existing savings incentives

A ew simplified calculaions or a range o hypoheical households wih differen

income levels, accouning or hese real-lie circumsances, illusrae he link beween

he benefis o savings incenives and income disribuion. Te calculaions here ocus

on wo ypes o reiremen savings plans: an IR and a 401(k) plan.

Assumptions and metrics

o make he calculaions manageable, a ew assumpions are necessary. Firs, as he

baseline scenario, a household conribues $5,500 annually o an IRhe maximum

allowable amoun in 2015.34 Te hypoheical households inves his conribuion or 25

years beore wihdrawing all o he conribuions and he reurns earned on hem and

paying personal income axes on hese wihdrawals.35 Te simulaions evenually vary

he invesmen ime horizon.

Second, he value o he ax incenives depends on a household’s marginal ax raes a

hree differen poins in ime. Tese include he household’s marginal ax rae when i

makes he conribuion. Tis consiues a ax benefi because he conribuion is no

subjec o ederal income axes. Te value o he ax incenive also includes he marginal

ax rae during he invesmen period as anoher ax benefi, since he income earned on

he invesmens is no axed during ha ime. Finally, because wihdrawals are subjec o


9/18


income ax, i includes he marginal ax rae when he money is wihdrawn as an offse-

ing ax burden, since all hree marginal ax raesa conribuion, during invesmen,

and upon wihdrawalare se equal o 25 percen a he beginning. Te calculaions

laer change some o hese marginal raes in he simulaions.

Tird, he value o he ax incenives depends o some degree on he rae o reurn ha

he household can earn on is conribuions. Here i is assumed ha he household canearn a nominal rae o reurn o 6 percen on average or he 25 years during which he

money is invesed. Tis assumpion is made across all o he ollowing simulaions.

Te calculaions below use wo separae measures o make he calculaions somewha

inuiive while allowing comparisons beween households wih differen characeris-

ics. Te firs measure, in nominal dollars, shows he oal dollars in ax benefis ha a

household receives rom conribuing o a reiremen savings accoun on a ax-advan-

aged basis. Nominal dollars are he amoun o money invesed in a reiremen savings

accounon which he ineres earned is unaxedminus he axes ha are due a he

end o he 25-year period when all money is wihdrawn.36

Te second measure is somehing called he ne presen value o he deerral benefi.37

Tis measure allows or a proper comparison beween households wih differen mar-

ginal ax raes a separae poins in ime and varying invesmen horizons, so ha he

invesmen horizon is no longer consan a 25 years. Te ne presen value calculaion

adds he ax benefis o deducing he ax conribuion rom income axes and he ax

benefis rom no paying ederal income axes on he conribuions during he inves-

men period and hen subracs he income axes paid on he wihdrawal when he

enire conribuionsplus he accumulaed raes o reurnare wihdrawn. I adjuss

each ax benefi and ax burden by a process called discouning, which makes uure benefis and paymens comparable o oday’s ax benefis. All amouns are hen compa-

rable o each oher, regardless o how shor or how long he ime periods during which

axes have been deerred.

Tese adjused amouns o ax benefis and ax burdens are he amouns ha he house-

hold would have o se asideor, in he case o ax burdens, receiveoday ha would

amoun, ogeher wih he expeced ineres raehe discoun raeo he uure

dollar amouns calculaed as nominal ax benefis and ax burdens. Te calculaions here

assume ha he discoun rae is equal o 6 percen.38 Te sum o he wo adjused ax ben-

efis, minus he uure ax burden, shows he oal value o deerring axes ino he uure.


10/18


Simulation results

able 1 shows some o hese key assumpions and hen presens he simulaion resuls.

Te firs panel in able 1 shows he nominal values, and he second panel shows he ne

presen values under a range o scenarios.

Te nominal value calculaions in able 1 show ha he ax benefi o deerral increases wih marginal ax raes. Te firs hree examples assume ha he household maxes ou

on heir conribuions o an IR. A low-income earner wih a marginal ax rae o 10

percen receives a ax benefi o $2,810 over 25 years, no accouning or he discoun

rae, or saving $5,500 in 2015. A middle-income earner wih a ax rae o 25 percen,

meanwhile, receives a ne ax benefi o $5,306, or almos wice as much as a low-

income earner. A high-income earner wih a marginal ax rae o 39.6 percen receives a

benefi o $6,168, or 119 percen more han he low-income household.

Te nominal value calculaions in able 1 urher show ha high-income earners can

receive muliple imes he benefi ha low-income earners receive because hey ofenhave more opporuniies o save in ax-advanaged reiremen accouns and because

heir marginal ax raes have more room o all as hey reire. Boh larger conribuion

amouns and declining marginal ax raes increase he ne ax benefis o deerring ax

paymens ino he uure or high-income earners. A high-income earning household

deerring $18,000 in 2015he maximum employee conribuion o a 401(k) plan39

and who iniially has a marginal ax rae o 39.6 percen ha alls o 25 percen in

reiremen can receive a ax benefi o $21,132 on an iniial deerral. Such a high-income

household will receive seven and a hal imes he ax benefi ha a low-income house-

hold will receive.

I bears repeaing ha high-income earners receive his addiional benefi no jus

because hey can save more due o higher incomes bu also because hey have more

opporuniies o receive ax benefis based on increased access o employer-sponsored

reiremen plans and a higher chance o lower marginal ax raes in he uure.

Te botom panel o calculaions in able 1 shows he ne presen value numbers or he

same scenarios, in addiion o one more simulaion ha assumes a longer invesmen

period o 35 years insead o 25 years. As in he firs panel, he ax benefi increases wih

income because o greaer ax benefis o deerring each dollar. Tese benefis are due o

more ax-advanaged opporuniies o save money or reiremen, as well as he possibil-

iy o declining marginal ax raes. Te addiional calculaion also shows ha invesing

or an addiional decade urher increases he ax benefis. A high-income earner wih

a marginal ax rae invesing or 35 years receives a benefi o $1,819 insead o $1,437

afer 25 yearsan increase o 26.5 percen. Tis addiional benefi is due o compound-

ing ineres over ime; higher-income households generally have more ime o inves

because hey are more likely o have addiional savings o cover oher expenses.


11/18


Te lessons rom hese simulaions are clear. Firs, households mus navigae a com-plicaed sysem o rules o maximize heir ax benefis. Tey need o undersand which

savings plans hey have available, how much hey can conribue, how long o keep

heir money in a ax-advanaged asse, and how heir decisions inerac wih curren

and uure ax raes or personal income and capial income. Second, higher-income

earners can benefi much more rom savings incenives because hey are more likely o

have access o reiremen savings plans, since hey are more likely o work or employers

ha offer reiremen savings plans and more likely o have muliple sreams o income.

Higher-income earners also pay higher ax raes and ofen wai longer beore wihdraw-

ing heir money.40

Current savings incentives are ineffective and inefficient

I would be one hing i incenives benefied high-income earners he mos bu suc-

ceeded in reducing reiremen insecuriy overall. Unorunaely, ha has no been he

case. Te overwhelming evidence suggess ha a subsanial share o households is

TABLE 1

Simulated net tax benefits of tax deferral under varying assumptions

Tax rate at

deferral

Tax rate at

withdrawal

Deferral benefit per

every $1 invested;

net present value

calculation only

Total tax

deferral benefit

Ratio of total benef

to baseline benefit

net present value

calculation only

Nominal dollars

Baseline scenario: $5,500 deferred 25.0% 25.0% N/A $5,306 N/A

Low-income earner: $5,500 deferred 10.0% 10.0% N/A $2,811 N/A

High-income earner: $5,500 deferred 39.6% 39.6% N/A $6,168 N/A

High-income earner: $18,000 deferred 39.6% 39.6% N/A $20,187 N/A

High-income earner: $18,000 deferred,

marginal tax rate declines39.6% 25.0% N/A $21,132 N/A

Net present value dollars

Baseline scenario: $5,500 deferred 25.0% 25.0% 22.5% $1,236 N/A

Low-income earner: $5,500 deferred 10.0% 10.0% 11.9% $655 53.0%

High-income earner: $5,500 deferred 39.6% 39.6% 26.1% $1,437 116.2%

High-income earner: $18,000 deferred 39.6% 39.6% 26.1% $4,704 380.4%


marginal tax rate declines39.6% 25.0% 40.7% $2,240 181.2%


35-year deferral period39.6% 39.6% 33.1% $1,819 147.2%

Notes: Benefits from tax deferral are calculated as net present value. The discount rate is equal to the government interest rate, which is set equal to 6 percent nominally. All tax rates are marginal tax rates. The

deferral period is 25 years, unless otherwise stated. “N/A” stands for not applicable.

Source: Author’s calculations based on Peter Brady, “The Tax Benefits and Revenue Costs of Tax Deferral” (Washington: Investment Company Institute, 2012), available at https://www.ici.org/pdf/ppr_12_tax_

benefits.pdf.


12/18


already inadequaely prepared or reiremen, even hough he ederal governmen is

orgoing increasing amouns o ax revenue o incenivize households o save or reire-

men. Wihou subsanial reorm, an ever-larger share o American households will be

inadequaely prepared and will no be able o pay or is living sandard in reiremen.

Increasing numbers of Americans are not prepared for retirement

esearchers ineresed in undersanding wheher households are on rack o save

enough money or reiremen ulimaely wan o link individual savings in he presen o

he income ha people will need in he uure. Economiss ofen use an approach ha

atemps o measure wheher people will have he resources o susain heir qualiy o

lie posreiremen. Tis reiremen income adequacy approach defines adequacy as a

minimum reiremen income relaive o people’s prereiremen earnings.41

A household is considered adequaely prepared or reiremen i is expeced reire-

men income is greaer han a minimum share, such as 75 percen o is earnings beorereiremen.42 Tis raio o reiremen income o prereiremen earnings is also known as

he replacemen rae. I measures he share o prereiremen earnings a household can

replace wih he income i can expec o receive rom Social Securiy; defined benefi

pensions; and privae savings. In esimaes ha use 75 percen as a hreshold, people are

considered adequaely prepared or reiremen i heir expeced income during reire-

men is projeced o be a leas 75 percen as large as heir income beore reiring.

Te evidence on reiremen income adequacy generally shows ha a large share o

householdsespecially among communiies o color, single women, and households

wih less educaionis ill prepared o mainain is sandard o living in reiremen. Andhe share has grown over long periods o ime, according o mos sudies ha provide

longer views.43 One such measure is he Cener or eiremen esearch a Boson

College’s Naional eiremen isk Index, or NI.44 Te NI measures he share o

working-age households ha has no ye reached he ull reiremen age and ha is a

risk o being unable o mainain is sandard o living in reiremen based on expeced

income rom Social Securiy, DB pensions, and individual savings, including money in

401(k) plans, IRs, and housing. Te NI esimaes ha 52 percen o working-age

households were a risk o no being able o mainain heir sandards o living in reire-

men in 2013, up rom 31 percen in 1983. Ta is, he share o households inadequaely

prepared or reiremen is large and increasing.45

Imporanly, he NI shows shoralls in reiremen savings less severe han hose

recorded by oher researchers. For insance, a repor by he Naional Insiue on

eiremen Securiy, or NIS, finds ha 65 percen o households ell shor o heir sav-

ings arges in 2010 using savings levels recommended by he financial service indusry


13/18


and he same daa as he NI.46 Te NI, by conras, ound ha in 2010, only 53

percen o working-age households were a risk o no being able o mainain heir san-

dard o living in reiremen.47

Even sudies ha have idenified a lower share o households han he NI as being

inadequaely prepared or reiremen have ound evidence o increasing reiremen

income inadequacy. A widely ciedand comparaively opimisicassessmen oreiremen income adequacy by researchers a he Universiy o Wisconsin in 2006

ound ha hose born beween 1931 and 1941based on daa rom he Universiy

o Michigan’s Healh and eiremen Sudyhad only a 16 percen chance o alling

below heir opimal savings arge.48 An updae o his research in a working paper in

2009beore he ull effec o he Grea ecession was elound ha 26 percen o

households were inadequaely prepared or reiremen.49 Sudies ha break down daa

by age, meanwhile, find ha younger generaions are worse prepared or reiremen han

older cohors.50

Esimaes or he share o households inadequaely prepared or reiremen also vary wih household characerisics. Te respecive shares end o be greaer among com-

muniies o color, single women, and hose wih less educaion han among whie

households, single men, and households wih more educaion. In his research on wealh

inequaliy, New York Universiy Proessor Edward Wolff offers breakdowns or house-

holds beween he ages o 47 and 64 years.51 Wolff’s research shows ha 51 percen

o households beween hese ages in 2010 were unable o replace 75 percen o heir

prereiremen income in reiremen.52 Te relevan share or non-Hispanic whies is 45

percen, compared wih 60 percen or Arican Americans and Hispanics combined.

Based on 2010 daa, 59 percen o single women can expec o have o cu back on heirliving expenses once hey reire, compared wih only 51 percen o single men. Finally,

households wih less han 12 years o schoolinghose wihou a high school diploma

or GEDhave an esimaed 61 percen chance o alling shor o mainaining heir san-

dards o living in reiremen, while only 43 percen o households wih 16 years or more

o schoolinghose wih a leas a college degreemay have o cu back on consump-

ion in reiremen.53

The federal government sacrifices billions of dollars in revenue for the benefit of

higher-income earners

Exising savings incenives conribue o a pronounced imbalance in who benefis

rom hem and who does no, while cosing he public billions o dollars. In 2013

alone, he ederal governmen orwen abou $137 billion in ax revenue rom ax

breaks or reiremen savingsmoney ha he ederal governmen would have col-

leced had i no been or he special ax reamens o reiremen benefis. Moreover,


14/18


only 7 percen o his orgone revenue wen o he botom 40 percen o earners, while

66 percen accrued o he op 20 percen o earners. O his 66 percen, close o hree-

quarers o he ax incenive wen o he op 10 percen o earners.54* Imporanly,

his unequal disribuion o ax incenives is a snapsho o only one year’s ax benefis

offered o households or saving or reiremen.

Despie he unequal benefis accorded o high earners hrough he exising, unequalincenives, here does no appear o be much o an offseting macroeconomic effec in he

orm o increased personal savings. esearch shows ha higher-income earners, on aver-

age, largely replace nonax-advanaged savings wih ax-advanaged savings. Tis means

ha higher-income earners would save similar amouns absen he savings incenives. 55

Te ederal governmen spen as much as $92 billion in fiscal year 2013 on reiremen

savings incenives or he op quinile o earners alone, wihou acually increasing per-

sonal savings beyond where savings would have been already.56 Te exising ax-advan-

aged reiremen savings are valuable ax breaks, especially or higher-income earners, bu

hey do litle o advance reiremen income securiy or many lower-income households.

Conclusion

Te need o save or reiremen ouside Social Securiy has increased over ime. Social

Securiy’s reiremen age is increasing, defined benefi pensions have declined, and labor

and financial markes have become riskier. Households need o save more money o

proec hemselves rom hese risks.57 Te ax code offers a number o savings incenives

o help people save more or reiremen, bu hese incenives are complex and skewed

oward higher-income earners.

Tese incenives are ulimaely inefficien. Te ederal governmen and several sae

governmens lose subsanial amouns o ax revenue wihou semming he ide o

rising reiremen income insecuriy. Te resuling economic insecuriy as people age is

expeced o be severe among lower-income households, single women, and communi-

ies o color. Addressing he shorcomings o exising savings incenives in he ax code

would be a welcome sep oward middle-class reiremen income securiy.

Chrisian E. Weller is a Senior Fellow a he Cener for American Progress and a profes-

sor of public policy a he McCormack Graduae School of Policy and Global Sudies a he

Universiy of Massachusets, Boson. eresa Ghilarducci is he Bernard L. and Irene Schwarz

chair in economic policy analysis in he Deparmen of Economics and he direcor of he

Schwarz Cener for Economic Policy Analysis a Te New School in New York.

* Correction, August 9, 2016: Tis issue brief has been updaed o reflec ha he auhors

looked specifically a he reiremen incenives included in he referenced daa.


15/18


Appendix

Te calculaion o he nominalno inflaion adjused and no adjused or he ime

value o moneyax benefi ollows work by Peer Brady o he Invesmen Company

Insiue, as well as oher researchers’ developmen o i:58

Equaion 1: B Dnominal =(1 + r c ) (1 – 0 ) – (1 – 00 )(1 + r c(1 – 0d )

Te variable “ B Dnominal” reers o he ne nominal ax benefi or each dollar on which ax

paymen is deerred ino he uure. Te oher relevan variables in his equaion are

defined as ollows:

00: he marginal income ax rae when he iniial conribuion is made

0d: he marginal income ax rae during he deerral period, when he money is invesed

0

: he marginal income ax rae when money is wihdrawn

r c: he rae o reurn earned on he invesmens during he deerral period

: he lengh o ime o deerral

Te equaion is he difference beween he afer-ax disribuion afer “ ” periods rom

a ax-advanaged reiremen savings accoun minus he afer-ax disribuion rom a

axable accoun. Te difference beween he money received rom a ax-advanaged

accoun and a axable accoun logically has o be he ne ax benefi in nominal dollars.

Te calculaion o he ne presen value o he ne ax benefi o ax deerral is also aken

rom Brady’s work, as well as oher researchers’ developmen o i:59

Equaion 2:

Equaion 2 calculaes he ne presen value o a dollar ha has been deerred rom pay-

ing axes, “R D PV .” And “r g ” denoes he governmen ineres rae, which is equal o he

discoun rae. All oher variables are defined as above.

Equaion 2 has hree separae pars on he righ-hand side. Te firs, “ 00 ,” is he ax

benefi rom deducing he iniial conribuion rom axable income in he firs year.

Te second par, afer he summaion sign, shows he ax benefis rom earning ax-ree

compounded ineres on he invesmen over he deerral period, “ .” And he hird par

afer he minus sign shows he ax burden ha he household has o pay when wihdraw-

ing money rom he ax-advanaged asse.

T g

T c

T T

nn

g

ndcc

d PV D

r

r t

r

t r t r t t R

)1(

)1(

)1(

] )1( 1 )( 1[( 0

1

10

0000

0+

+−

+

−+−+=

=

−


16/18


Endnotes

1 For the next 20 years, 10,000 Americans will turn 65 years oldevery day, according to the Bureau of the Census. By 2025,the number of Americans in their adult prime, ages 34 to 55,will increase by only 3.6 million, while the number of peopleages 65 to 74 will increase by more than 10 million. Thoseturning 65 in the next few decades face a higher risk of down-ward mobility than any cohort since the end of World War II,when Social Security, Old Age Assistance, employer pensions,

and the economy were expanding. Of the 18 million workersages 55 to 64 in 2015, 1.7 million live on incomes defined aspoor or near poor. This means that 1.7 million people—whoare working—live on less than $11 per day for food and $500per month for rent. When this cohort retires at age 65, thenumber of poor or near poor will increase to 4.3 million dueto no or diminished retirement assets, increased Medicarepremiums, and their sole reliance on Social Security. Withoutaction to address this crisis, the number of 65-year-olds peryear who are poor or near poor will increase by 146 percentbetween 2013 and 2022. What goes unmeasured is the as-of-yet unspoken financial burden that the increasing number ofpoor and near-poor elderly will place on taxpayers and mu-nicipal budgets. See Teresa Ghilarducci and Zachary Knauss,“More Middle Class Workers Will Be Poor Retirees” (New York:Schwartz Center for Economic Policy Analysis, 2015), availableat http://www.economicpolicyresearch.org/images/docs/retirement_security_background/Downward_Mobility.pdf .

2 Alicia H. Munnell, Anthony Webb, and Wenliang Hou, “How

Much Should People Save?” (Boston: Center for RetirementResearch at Boston College, 2014), available at http://crr.bc.edu/briefs/how-much-should-people-save/.

3 Keith Miller, David Madland, and Christian E. Weller, “The Re-ality of the Retirement Crisis” (Washington: Center for Amer-ican Progress, 2015), available at https://www.american-progress.org/issues/economy/report/2015/01/26/105394/the-reality-of-the-retirement-crisis/.

4 Craig Copeland, “Employment-Based Retirement Plan Par-ticipation: Geographic Differences and Trends, 2013” (Wash-ington: Employee Benefit Research Institute, 2014), availableat http://www.ebri.org/pdf/briefspdf/EBRI_IB_405_Oct14.RetPart.pdf.

5 Teresa Ghilarducci and Adam Hayes, “401(K) Tax Policy Cre-ates Inequality” (New York: Schwartz Center for EconomicPolicy Analysis, 2015), available at http://www.economicpol-icyresearch.org/images/docs/research/retirement_security/

Hayes_Ghilarducci_Policy_Note_1.9.15_FINAL.pdf .

6 See Congressional Budget Office, “The Distribution of Major Tax Expenditures in the Individual Tax System” (2013), avail-able at https://www.cbo.gov/publication/43768. See Figure5 in this source for supplemental data.

7 Teresa Ghilarducci and others, “Retirement Savings TaxExpenditures: The Need for Refundable Tax Credits” (NewYork: Schwartz Center for Economic Policy Analysis and De-partment of Economics, The New School for Soc ial Research,2015), available at http://www.economicpolicyresearch.org/images/docs/retirement_security_background/Retire-ment_Savings_Tax_Expenditures.pdf .

8 Ghilarducci and Knauss, “More Middle Class Workers will bePoor Retirees.”

9 See Christian E. Weller and David Madland, “Keep Calmand Muddle Through” (Washington: Center for AmericanProgress, 2014), available at https://www.americanprogress.

org/issues/economy/report/2014/08/06/95222/keep-calm-and-muddle-through/.

10 Not all states have personal income taxes. Those that do,though, often follow the lead o f the federal governmentand do not subject contributions to tax-advantaged retire-ment savings accounts to state income taxes if they arenot subject to federal income taxation. For a more detaileddiscussion of the link between federal and state income taxtreatment of tax-advantaged savings, see Ghilarducci andothers, “Retirement Savings Tax Expenditures: The Need forRefundable Tax Credits.”

11 This applies only to non-Roth type plans. Roth-type retire-ment savings plans require people to pay taxes before theycontribute, but earnings on their savings and withdrawalsfrom such savings plans are tax free. See Internal RevenueService, “Publication 590-A: Contributions to IndividualRetirement Arrangements (IRAs)” (2014), available at https://www.irs.gov/publications/p590a/.

12 Roth IRAs and Roth 401(k)s receive a different tax advan-tage. Contributions to these types of retirement savingsplans occur after a taxpayer has paid income taxes, butinvestment gains and withdrawals from these savings ac-counts are tax free. See ibid.

13 For details on potential exceptions to this excise tax, seeInternal Revenue Service, “Topic 558 - Additional Tax onEarly Distributions from Retirement Plans Other Than IRAs,”available at http://www.irs.gov/taxtopics/tc558.html (lastaccessed October 2015).

14 Ibid.

15 Internal Revenue Service, “Retirement Plans FAQs regardingHardship Distributions: Are hardship distributions allowedfrom an IRA?”, available at http://ww w.irs.gov/Retirement-Plans/Retirement-Plans-FAQs-regarding-Hardship-Distribu-tions#7 (last accessed October 2015).

16 On pension loans, see Christian E. Weller and JeffreyWenger, “Easy Money or Hard Times? Health and 401(k)Loans,” Contemporary Economic Policy 30 (1) (2012): 29–42.On withdrawals, see Robert Argento, Victoria Bryant, andJohn Sabelhaus, “Early Withdrawals from Retirement Ac-counts During the Great Recession.” Working Paper 2013-2(Federal Reserve Board Divisions of Research and Statisticsand Monetary Affairs, 2013).

17 See, for instance, Weller and Wenger, “Easy Money or Hard Times?”

18 Alicia H. Munnell and Anthony Webb, “The Impact of Leak-ages on 401(K)/IRA Assets” (Chestnut Hill, MA: Center forRetirement Research, 2015), available at http://crr.bc.edu/wp-content/uploads/2015/01/IB_15-2.pdf .

19 Pamela Perun and C. Eugene Steuerle, “Reality Testing forPension Reform.” Working Paper 2004-10 (Pension ResearchCouncil, 2004).

20 Shlomo Benartzi and Richard H. Thaler, “Heuristics andBiases in Retirement Savings Behavior,” Journal of EconomicPerspectives21 (3) (2007): 81–104, available at http://wolf-web.unr.edu/homepage/pingle/Teaching/BADM%20791/Week%209%20Behavioral%20Microeconomics/Benartzi- Thaler%20Biased%20Savings%20Behavior.pdf .

21 Gary R. Mottola and Stephen P. Utkus, “Can There Be TooMuch Choice in a Retirement Savings Plan?” (Valley Forge PA:Vanguard Center for Retirement Research, 2006), available athttp://www.403bwise.com/pdf/vcrr_choice_study.pdf .

22 Sheena Sethi-Iyengar and others, “How Much Choice is TooMuch? Contributions to 401(k) Retirement Plans.” In OliviaS. Mitchell and Stephen P. Utkus, eds., Pension Design andStructure: New Lessons from Behavioral Finance (New York:Oxford University Press, 2004); Ilona Babenko and Rik Sen,“Money Left on the Table: An Analysis of Participation inEmployee Stock Purchase Plans,”Review of Financial Studies 27 (12) (2014): 3658–3698.

23 See Teresa Ghilarducci, When I’m Sixty-Four: The Plot AgainstPensions and the Plan to Save Them (Princeton, NJ: PrincetonUniversity Press, 2008).

24 Internal Revenue Service, “1040 Tax Tables 2014” (2014),available at http://www.irs.gov/pub/irs-pdf/i1040tt.pdf .

http://www.economicpolicyresearch.org/images/docs/retirement_security_background/Downward_Mobility.pdfhttp://www.economicpolicyresearch.org/images/docs/retirement_security_background/Downward_Mobility.pdfhttp://crr.bc.edu/briefs/how-much-should-people-save/http://crr.bc.edu/briefs/how-much-should-people-save/https://www.americanprogress.org/issues/economy/report/2015/01/26/105394/the-reality-of-the-retirement-crisis/https://www.americanprogress.org/issues/economy/report/2015/01/26/105394/the-reality-of-the-retirement-crisis/https://www.americanprogress.org/issues/economy/report/2015/01/26/105394/the-reality-of-the-retirement-crisis/http://www.ebri.org/pdf/briefspdf/EBRI_IB_405_Oct14.RetPart.pdfhttp://www.ebri.org/pdf/briefspdf/EBRI_IB_405_Oct14.RetPart.pdfhttp://www.economicpolicyresearch.org/images/docs/research/retirement_security/Hayes_Ghilarducci_Policy_Note_1.9.15_FINAL.pdfhttp://www.economicpolicyresearch.org/images/docs/research/retirement_security/Hayes_Ghilarducci_Policy_Note_1.9.15_FINAL.pdfhttp://www.economicpolicyresearch.org/images/docs/research/retirement_security/Hayes_Ghilarducci_Policy_Note_1.9.15_FINAL.pdfhttps://www.cbo.gov/publication/43768http://www.economicpolicyresearch.org/images/docs/retirement_security_background/Retirement_Savings_Tax_Expenditures.pdfhttp://www.economicpolicyresearch.org/images/docs/retirement_security_background/Retirement_Savings_Tax_Expenditures.pdfhttp://www.economicpolicyresearch.org/images/docs/retirement_security_background/Retirement_Savings_Tax_Expenditures.pdfhttps://www.americanprogress.org/issues/economy/report/2014/08/06/95222/keep-calm-and-muddle-through/https://www.americanprogress.org/issues/economy/report/2014/08/06/95222/keep-calm-and-muddle-through/https://www.americanprogress.org/issues/economy/report/2014/08/06/95222/keep-calm-and-muddle-through/https://www.irs.gov/publications/p590a/https://www.irs.gov/publications/p590a/http://www.irs.gov/taxtopics/tc558.htmlhttp://crr.bc.edu/wp-content/uploads/2015/01/IB_15-2.pdfhttp://crr.bc.edu/wp-content/uploads/2015/01/IB_15-2.pdfhttp://wolfweb.unr.edu/homepage/pingle/Teaching/BADM%20791/Week%209%20Behavioral%20Microeconomics/Benartzi-Thaler%20Biased%20Savings%20Behavior.pdfhttp://wolfweb.unr.edu/homepage/pingle/Teaching/BADM%20791/Week%209%20Behavioral%20Microeconomics/Benartzi-Thaler%20Biased%20Savings%20Behavior.pdfhttp://wolfweb.unr.edu/homepage/pingle/Teaching/BADM%20791/Week%209%20Behavioral%20Microeconomics/Benartzi-Thaler%20Biased%20Savings%20Behavior.pdfhttp://wolfweb.unr.edu/homepage/pingle/Teaching/BADM%20791/Week%209%20Behavioral%20Microeconomics/Benartzi-Thaler%20Biased%20Savings%20Behavior.pdfhttp://www.403bwise.com/pdf/vcrr_choice_study.pdfhttp://www.irs.gov/pub/irs-pdf/i1040tt.pdfhttp://www.irs.gov/pub/irs-pdf/i1040tt.pdfhttp://www.403bwise.com/pdf/vcrr_choice_study.pdfhttp://wolfweb.unr.edu/homepage/pingle/Teaching/BADM%20791/Week%209%20Behavioral%20Microeconomics/Benartzi-Thaler%20Biased%20Savings%20Behavior.pdfhttp://wolfweb.unr.edu/homepage/pingle/Teaching/BADM%20791/Week%209%20Behavioral%20Microeconomics/Benartzi-Thaler%20Biased%20Savings%20Behavior.pdfhttp://wolfweb.unr.edu/homepage/pingle/Teaching/BADM%20791/Week%209%20Behavioral%20Microeconomics/Benartzi-Thaler%20Biased%20Savings%20Behavior.pdfhttp://wolfweb.unr.edu/homepage/pingle/Teaching/BADM%20791/Week%209%20Behavioral%20Microeconomics/Benartzi-Thaler%20Biased%20Savings%20Behavior.pdfhttp://crr.bc.edu/wp-content/uploads/2015/01/IB_15-2.pdfhttp://crr.bc.edu/wp-content/uploads/2015/01/IB_15-2.pdfhttp://www.irs.gov/taxtopics/tc558.htmlhttps://www.irs.gov/publications/p590a/https://www.irs.gov/publications/p590a/https://www.americanprogress.org/issues/economy/report/2014/08/06/95222/keep-calm-and-muddle-through/https://www.americanprogress.org/issues/economy/report/2014/08/06/95222/keep-calm-and-muddle-through/https://www.americanprogress.org/issues/economy/report/2014/08/06/95222/keep-calm-and-muddle-through/http://www.economicpolicyresearch.org/images/docs/retirement_security_background/Retirement_Savings_Tax_Expenditures.pdfhttp://www.economicpolicyresearch.org/images/docs/retirement_security_background/Retirement_Savings_Tax_Expenditures.pdfhttp://www.economicpolicyresearch.org/images/docs/retirement_security_background/Retirement_Savings_Tax_Expenditures.pdfhttps://www.cbo.gov/publication/43768http://www.economicpolicyresearch.org/images/docs/research/retirement_security/Hayes_Ghilarducci_Policy_Note_1.9.15_FINAL.pdfhttp://www.economicpolicyresearch.org/images/docs/research/retirement_security/Hayes_Ghilarducci_Policy_Note_1.9.15_FINAL.pdfhttp://www.economicpolicyresearch.org/images/docs/research/retirement_security/Hayes_Ghilarducci_Policy_Note_1.9.15_FINAL.pdfhttp://www.ebri.org/pdf/briefspdf/EBRI_IB_405_Oct14.RetPart.pdfhttp://www.ebri.org/pdf/briefspdf/EBRI_IB_405_Oct14.RetPart.pdfhttps://www.americanprogress.org/issues/economy/report/2015/01/26/105394/the-reality-of-the-retirement-crisis/https://www.americanprogress.org/issues/economy/report/2015/01/26/105394/the-reality-of-the-retirement-crisis/https://www.americanprogress.org/issues/economy/report/2015/01/26/105394/the-reality-of-the-retirement-crisis/http://crr.bc.edu/briefs/how-much-should-people-save/http://crr.bc.edu/briefs/how-much-should-people-save/http://www.economicpolicyresearch.org/images/docs/retirement_security_background/Downward_Mobility.pdfhttp://www.economicpolicyresearch.org/images/docs/retirement_security_background/Downward_Mobility.pdf


17/18


25 The value of these savings incentives as a share of after-taxincome drops for the top 1 p ercent to 2.6 percent. Thisrelative decline at the very top of the income distributionreflects very high incomes and some limits on savingsincentives. Counting all itemized deductions, the top earn-ers still receive a larger share of income than lower-incomeearners. See, for instance, Tax Policy Center, “2013 Table T13-0099 Tax Benefit of All Itemized Deductions; Distribu-tion of Federal Tax Change by Cash Income Percentile, 2015,”March 1, 2013, available at http://www.taxpolicycenter.org/numbers/displayatab.cfm?DocID=3857.

26 For some illustrative examples and a review of other

relevant research, see, for instance, Peter Brady, “The TaxBenefits and Revenue Costs of Tax Deferral” (Washington:Investment Company Institute, 2012), available at https://www.ici.org/pdf/ppr_12_tax_benefits.pdf.

27 This discussion includes only some of the most importantreal-life aspects but leaves out additional ones such asthe interaction between federal and state income taxes,differential taxes on varying forms of withdrawals suc h asannuities and self-managed withdrawals.

28 Copeland, “Employment-Based Retirement Plan Participation.”

29 Christian Weller, Retirement on the Rocks: Why AmericansCan’t Get Ahead and How New Savings Policies Can Help (New York: Palgrave Macmillan, forthcoming). Higher-income households typically receive more diverse incomestreams—wages in addition to self-employment income—than is the case for lower-income households. See ChristianE. Weller and Jeffrey B. Wenger, “Income Diversification as

Insurance in an Increasingly Risky World: Identifying PolicyGoals.” In Christian E. Weller, ed., Inequality, Uncertainty andOpportunity: The Varied and Growing Role of Finance in LaborRelations (Ithaca, NY: ILR Press, 2015).

30 Internal Revenue Service, “Publication 501: Exemptions,Standard Deduction, and Filing Information” (2014), availableat http://www.irs.gov/publications/p501/. Those ages 65 andolder are able to take a higher standard deduction—specifi-cally, $1,550 more individually and up to $2,400 more jointly.

31 Gary R. Mottola, “The Financial Capability of Young Adults—AGenerational View” (Washington: FIRNA Investor EducationFoundation, 2014), available at http://www.usfinancialcapa-bility.org/downloads/FinancialCapabilityofYoungAdults.pdf ;Karen Smith, Mauricio Soto, and Rudolph G. Penner, “HowSeniors Change Their Asset Holdings During Retirement”(Washington: Urban Institute, 2009), available at http://www.nber.org/2009rrc/3.1%20Smith,%20Soto,%20Penner.pdf ;James M. Poterba, Steven F. Venti, and David A. Wise, “401(k)Plans and Future Patterns of Retirement Saving,” The American

Economic Review 88 (2) (1998): 179–184.

32 Stewart E. Sterk and Melanie B. Leslie, “Accidental Inheri-tance: Retirement Accounts and the Hidden Law o f Succes-sion,” New York University Law Review 89 (2014): 165–237,available at http://www.nyulawreview.org/sites/default/files/pdf/NYULawReview-89-1-Sterk-Leslie.pdf . The tax codetries to counter this possibility by requiring that householdswithdraw a minimum share of their savings i n IRAs—forinstance, no later than the year after they turn 70.5 yearsold. See Internal Revenue Service, “Retirement Topic - Re-quired Minimum Distributio ns (RMDs),” available at http://www.irs.gov/Retirement-Plans/Plan-Participant,-Employee/Retirement-Topics-Required-Minimum-Distributions-%28RMDs%29 (last accessed October 2015).

33 Higher-income households tend to have more assets heldin retirement accounts than lower-income households andto experience a lower rate of decumulation from these ac-counts, starting at a later age. See, for example, Smith, Soto,

and Penner, “How Seniors Change Their Asset HoldingsDuring Retirement.” See also Congressional Budget Office,“Will the Demand for Assets Fall When Baby Boomers Re -tire?” (2009), available at http://www.cbo.gov/sites/default/files/09-08_baby-boomers.pdf .

34 Internal Revenue Service, “Retirement Topics - IRA Contribu-tion Limits,” available at https://www.irs.gov/Retirement-Plans/Plan-Participant,-Employee/Retirement-Topics-IRA-Contribution-Limits (last accessed October 2015).

35 The exact length of investment—or tax deferral—does notmaterially affect the conclusions of this discussion.

36 All calculations here assume that taxes are paid upon with-drawal and that all money is w ithdrawn at the end of theinvestment period. Households can reap additional tax ben-efits by delaying the payout and by choosing from a rangeof payout options, such as annuities. For additional detailson this issue, see Brady, “The Tax Benefits and Revenue Costsof Tax Deferral.”

37 The Appendix contains the exact formula, taken from ibid.

38 There is a substantial debate in the literature over the exactdiscount rate to use for retirement calculations. For a discus-sion of discount rates for defined benefit pension plans,

see Christian E. Weller and Dean Baker, “Smoothing theWaves of Pension Funding: Could Changes in Funding RulesHelp Avoid Cyclical Under-funding?”, The Journal of PolicyReform 8 (2) (2005): 131–151. We choose 6 percent here asthe discount rate to make our results comparable to others,specifically those generated by Brady, “The Tax Benefits andRevenue Costs of Tax Deferral.”

39 Internal Revenue Service, “IRS Announces 2015 Pension PlanLimitations; Taxpayers May Contribute up to $18,000 to their401(k) plans in 2015,” Press release, October 23, 2014, avail-able at https://www.irs.gov/uac/Newsroom/IRS-Announces-2015-Pension-Plan-Limitations-1.

40 Annamaria Lusardi and Olivia Mitchell, “Financial Literacyand Retirement Preparedness: Evidence and Implicationsfor Financial Education Programs.” Working Paper 2007/15(Center for Financial Studies, 2007), available at http://www.econstor.eu/bitstream/10419/25516/1/527633305.PDF.

41 This approach is anchored in a neoclassical economic theoryknown as the life cycle hypothesis, whereby people smooththeir consumption over their life cycle. This implies that theymaintain their preretirement consumption in retirementand that their total lifetime consumption relates to theirtotal lifetime income. Recent applications of the life cyclehypothesis to the question of retirement income adequacyinclude Eric M. Engen, William G. Gale, and Cori E. Uccello,“The Adequacy of Household Saving,” Brookings Papers onEconomic Activity 1999 2 (1999): 65–165, available at http://www.ssc.wisc.edu/~scholz/Teaching_742/EGU.pdf; JohnKarl Scholz, Ananth Seshadri, and Surachai Khitatrakun, “AreAmericans Saving ‘Optimally’ for Retirement?”, Journal ofPolitical Economy 114 (4) (2006): 607–643, available at https://www.ssc.wisc.edu/~scholz/Research/Optimality.pdf.

42 See, for instance, B. Douglas Bernheim, “The Adequacy ofPersonal Retirement Saving.” In David A. Wise, ed., Facing the

Age Wave (Stanford, CA: Hoover Institute Press, 1997); CharlesD. Ellis, Alicia H. Munnell, and Andrew D. Eschtruth, FallingShort: The Coming Retirement Crisis and What to Do About It

(New York: Oxford University Press, 2014); Alan L. Gustman and Thomas L. Steinmeier, “Effects of Pensions on Savings: AnalysisWith Data From the Health and Retirement Study,”Carnegie-Rochester Conference Series on Public Policy 50 (1) (1999):271–324; Peter Henle, “Recent Trends in Retirement BenefitsRelated to Earnings,” Monthly Labor Review 95 (6) (1972): 12–20;Engen, Gale, and Uccello, “The Adequacy of Household Sav-ing”; Lusardi and Mitchell, “Financial Literacy and RetirementPreparedness; James F. Moore and Olivia S. Mitchell, “ProjectedRetirement Wealth and Saving Adequacy.” In Olivia S. Mitchell,P. Brett Hammond, and Anna M. Rappaport, eds., Forecast-ing Retirement Needs and Retirement Wealth (Philadelphia:University of Pennsylvania Press, 2000); Alicia H. Munnell,Francesca Golub-Sass, and Anthony Webb, “What Moves theNational Retirement Risk Index? A Look Back and an Update”(Boston: Center for Retirement Research at Boston College,2007), available at http://crr.bc.edu/briefs/what-moves-the-national-retirement-risk-index-a-look-back-and-an-update/; Nari Rhee, “The Retirement Savings Crisis: Is it Worse ThanWe Think?” (Washington: National Institute on Retirement

Security, 2013), available at http://www.nirsonline.org/index.php?option=com_content&task=view&id=768&Itemid=48; Mark J. Warshawsky and John Ameriks, “How Prepared areAmericans for Retirement?” In Mitchell, Hammond, andRappa-port, eds., Forecasting Retirement Needs and Retirement Wealth;Jack VanDerhei, ”Retirement Savings Shortfalls: Evidence fromEBRI’s Retirement Security Projection Model” (Washington:Employee Benefit Research Institute, 2015), available at http://www.ebri.org/pdf/briefspdf/EBRI_IB_410_Feb15_RSShrtfls.pdf ;Christian E. Weller and Edward N. Wolff, Retirement Security: TheCrucial Role of Social Security (Washington: Economic PolicyInstitute, 2005), available at http://www.epi.org/publication/book_retirement_income/. The replacement ratio—retirementincome to preretirement income—is less than 100 percent,since the income needs of retirees are likely to be lower thanthose of workers because they no longer need to save forretirement and they pay fewer taxes, have no work-relatedexpenses, have smaller families, and do not have mortgages.

http://www.taxpolicycenter.org/numbers/displayatab.cfm?DocID=3857http://www.taxpolicycenter.org/numbers/displayatab.cfm?DocID=3857https://www.ici.org/pdf/ppr_12_tax_benefits.pdfhttps://www.ici.org/pdf/ppr_12_tax_benefits.pdfhttp://www.irs.gov/publications/p501/http://www.usfinancialcapability.org/downloads/FinancialCapabilityofYoungAdults.pdfhttp://www.usfinancialcapability.org/downloads/FinancialCapabilityofYoungAdults.pdfhttp://www.nber.org/2009rrc/3.1%20Smith,%20Soto,%20Penner.pdfhttp://www.nber.org/2009rrc/3.1%20Smith,%20Soto,%20Penner.pdfhttp://www.nyulawreview.org/sites/default/files/pdf/NYULawReview-89-1-Sterk-Leslie.pdfhttp://www.nyulawreview.org/sites/default/files/pdf/NYULawReview-89-1-Sterk-Leslie.pdfhttp://www.irs.gov/Retirement-Plans/Plan-Participant,-Employee/Retirement-Topics-Required-Minimum-Distributions-%28RMDs%29http://www.irs.gov/Retirement-Plans/Plan-Participant,-Employee/Retirement-Topics-Required-Minimum-Distributions-%28RMDs%29http://www.irs.gov/Retirement-Plans/Plan-Participant,-Employee/Retirement-Topics-Required-Minimum-Distributions-%28RMDs%29http://www.irs.gov/Retirement-Plans/Plan-Participant,-Employee/Retirement-Topics-Required-Minimum-Distributions-%28RMDs%29http://www.cbo.gov/sites/default/files/09-08_baby-boomers.pdfhttp://www.cbo.gov/sites/default/files/09-08_baby-boomers.pdfhttps://www.irs.gov/Retirement-Plans/Plan-Participant,-Employee/Retirement-Topics-IRA-Contribution-Limitshttps://www.irs.gov/Retirement-Plans/Plan-Participant,-Employee/Retirement-Topics-IRA-Contribution-Limitshttps://www.irs.gov/Retirement-Plans/Plan-Participant,-Employee/Retirement-Topics-IRA-Contribution-Limitshttps://www.irs.gov/uac/Newsroom/IRS-Announces-2015-Pension-Plan-Limitations-1https://www.irs.gov/uac/Newsroom/IRS-Announces-2015-Pension-Plan-Limitations-1http://www.econstor.eu/bitstream/10419/25516/1/527633305.PDFhttp://www.econstor.eu/bitstream/10419/25516/1/527633305.PDFhttp://www.ssc.wisc.edu/~scholz/Teaching_742/EGU.pdfhttp://www.ssc.wisc.edu/~scholz/Teaching_742/EGU.pdfhttps://www.ssc.wisc.edu/~scholz/Research/Optimality.pdfhttps://www.ssc.wisc.edu/~scholz/Research/Optimality.pdfhttp://crr.bc.edu/briefs/what-moves-the-national-retirement-risk-index-a-look-back-and-an-update/http://crr.bc.edu/briefs/what-moves-the-national-retirement-risk-index-a-look-back-and-an-update/http://www.nirsonline.org/index.php?option=com_content&task=view&id=768&Itemid=48http://www.nirsonline.org/index.php?option=com_content&task=view&id=768&Itemid=48http://www.ebri.org/pdf/briefspdf/EBRI_IB_410_Feb15_RSShrtfls.pdfhttp://www.ebri.org/pdf/briefspdf/EBRI_IB_410_Feb15_RSShrtfls.pdfhttp://www.epi.org/publication/book_retirement_income/http://www.epi.org/publication/book_retirement_income/http://www.epi.org/publication/book_retirement_income/http://www.epi.org/publication/book_retirement_income/http://www.ebri.org/pdf/briefspdf/EBRI_IB_410_Feb15_RSShrtfls.pdfhttp://www.ebri.org/pdf/briefspdf/EBRI_IB_410_Feb15_RSShrtfls.pdfhttp://www.nirsonline.org/index.php?option=com_content&task=view&id=768&Itemid=48http://www.nirsonline.org/index.php?option=com_content&task=view&id=768&Itemid=48http://crr.bc.edu/briefs/what-moves-the-national-retirement-risk-index-a-look-back-and-an-update/http://crr.bc.edu/briefs/what-moves-the-national-retirement-risk-index-a-look-back-and-an-update/https://www.ssc.wisc.edu/~scholz/Research/Optimality.pdfhttps://www.ssc.wisc.edu/~scholz/Research/Optimality.pdfhttp://www.ssc.wisc.edu/~scholz/Teaching_742/EGU.pdfhttp://www.ssc.wisc.edu/~scholz/Teaching_742/EGU.pdfhttp://www.econstor.eu/bitstream/10419/25516/1/527633305.PDFhttp://www.econstor.eu/bitstream/10419/25516/1/527633305.PDFhttps://www.irs.gov/uac/Newsroom/IRS-Announces-2015-Pension-Plan-Limitations-1https://www.irs.gov/uac/Newsroom/IRS-Announces-2015-Pension-Plan-Limitations-1https://www.irs.gov/Retirement-Plans/Plan-Participant,-Employee/Retirement-Topics-IRA-Contribution-Limitshttps://www.irs.gov/Retirement-Plans/Plan-Participant,-Employee/Retirement-Topics-IRA-Contribution-Limitshttps://www.irs.gov/Retirement-Plans/Plan-Participant,-Employee/Retirement-Topics-IRA-Contribution-Limitshttp://www.cbo.gov/sites/default/files/09-08_baby-boomers.pdfhttp://www.cbo.gov/sites/default/files/09-08_baby-boomers.pdfhttp://www.irs.gov/Retirement-Plans/Plan-Participant,-Employee/Retirement-Topics-Required-Minimum-Distributions-%28RMDs%29http://www.irs.gov/Retirement-Plans/Plan-Participant,-Employee/Retirement-Topics-Required-Minimum-Distributions-%28RMDs%29http://www.irs.gov/Retirement-Plans/Plan-Participant,-Employee/Retirement-Topics-Required-Minimum-Distributions-%28RMDs%29http://www.irs.gov/Retirement-Plans/Plan-Participant,-Employee/Retirement-Topics-Required-Minimum-Distributions-%28RMDs%29http://www.nyulawreview.org/sites/default/files/pdf/NYULawReview-89-1-Sterk-Leslie.pdfhttp://www.nyulawreview.org/sites/default/files/pdf/NYULawReview-89-1-Sterk-Leslie.pdfhttp://www.nber.org/2009rrc/3.1%20Smith,%20Soto,%20Penner.pdfhttp://www.nber.org/2009rrc/3.1%20Smith,%20Soto,%20Penner.pdfhttp://www.usfinancialcapability.org/downloads/FinancialCapabilityofYoungAdults.pdfhttp://www.usfinancialcapability.org/downloads/FinancialCapabilityofYoungAdults.pdfhttp://www.irs.gov/publications/p501/https://www.ici.org/pdf/ppr_12_tax_benefits.pdfhttps://www.ici.org/pdf/ppr_12_tax_benefits.pdfhttp://www.taxpolicycenter.org/numbers/displayatab.cfm?DocID=3857http://www.taxpolicycenter.org/numbers/displayatab.cfm?DocID=3857


18/18

43 For a discussion of a sample of the most influential studiesin this debate, see Miller, Madland and Weller, “The Realityof the Retirement Crisis.”

44 See Weller and Madland, “Keep Calm and Muddle Through.”

45 Alicia H. Munnell, Wenliang Hu, and Anthony Webb, “NRRIUpdate Shows Half Still Falling Short” (Boston: Center forRetirement Research at Boston College, 2014).

46 Rhee, “The Retirement Savings Crisis.”

47 Munnell, Webb, and Golub-Sass, “The National Retirement

Risk Index: An Update.”

48 Scholz, Seshadri, and Khitatrakun, “Are Americans Saving‘Optimally’ for Retirement?”

49 William G. Gale, John Karl Scholz, and Ananth Seshadri,“Are All Americans Saving ‘Optimally’ for Retirement?”Working Paper (Brookings Institution and University ofWisconsin-Madison, 2009), available at https://www.ssc.wisc.edu/~scholz/Research/Are_All_Americans_v6.pdf .

50 Munnell, Webb, and Golub-Sass, “The National RetirementRisk Index: An Update”; Rhee, “The Retirement SavingsCrisis”; Gale, Scholz, and Seshadri, “Are All Americans Saving‘Optimally’ for Retirement?”; Scholz, Seshadri, and Khitatr-akun, “Are Americans Saving ‘Optimally’ for Retirement?”;Barbara A. Butrica, Karen E. Smith, and Howard M. Iams,“This is Not Your Parents’ Retirement: Comparing RetirementIncome Across G enerations,” Social Security Bulletin 72 (1)(2012), available at http://www.ssa.gov/policy/docs/ssb/

v72n1/v72n1p37.html; VanDerhei, ”Retirement SavingsShortfalls.” The research by the Employee Benefit ResearchInstitute—another widely cited and somewhat optimisticassessment—specifically finds that the share of Gen Xers—defined by the following source as households betweenthe ages of 40 and 49 in 2014—was comparable to thoseof Late Boomers—defined as households between theages of 50 and 59 in 2014—but that the savings shortfallstend to be larger for younger cohorts, in l arge part becauseof higher future health care costs. See Jack Vanderhei,“What Causes EBRI Retirement Readiness Ratings TM toVary: Results from the 2014 Retirement Security ProjectionModel” (Washington: Employee Benefit Research Institute,2014), available at http://www.ebri.org/pdf/briefspdf/EBRI_IB_396_Feb14.RRRs.pdf .

51 Wolff’s research relies on the same data as the Center forRetirement Research, but there are some key methodologi-cal differences, as we discuss in the Appendix.

52 This share is comparable to the 53 percent of householdsshown as being at risk of not being able to maintain theirstandard of living in retirement in the National RetirementRisk Index. And both the NRRI and Wolff’s research rely onthe Federal Reserve’s Survey of Consumer Finances.

53 All data in this paragraph are taken from Edward Wolff,“Household Wealth Trends, 1983 to 2010,” Oxford Review ofEconomic Policy 30 (1) (2014): 21–43.

54 See Congressional Budget Office, “The Distribution of Major Tax Expenditures in the Individual Tax System.” See Figure 2in this source for supplemental data.

55 Eric M. Engen and William G. Gale, “The Effects of 401(k)Plans on H ousehold Wealth.” Working Paper 8032 (NationalBureau of Economic Research, 2000); Daniel J. Benjamin,“Does 401(k) Eligibility Increase Saving?”, Journal of PublicEconomics 87 (5) (2003): 1259–1290.

56 From supplemental data in Congressional Budget Office,“The Distribution of Major Tax Expenditures in the IndividualIncome Tax System,” figure 2. The Congressional Budget Of-fice found that the difference in revenues collec ted from thetop quintile of taxpayers between the current tax systemand a system in which contributions to retirement accountswere taxed as ordinary income and in which investmentearnings in retirement accounts were taxed as ordinaryinvestment income was $92 billion.

57 See Weller, Retirement on the Rocks; Christian E. Weller, “Mak-ing Sure Money Is Available When We Need It” (Washington:Center for American Progress, 2013), available at https://www.americanprogress.org/wp-content/uploads/2013/03/WellerMiddleClass.pdf.

58 Brady, “The Tax Benefits and Revenue Costs of Tax Deferral”;David Love, “What Can the Life Cycle Model Tell Us About401(k) Contributions and Participation?” (Williamstown, MA:Williams College, 2006), available at http://projects.vassar.edu/lamacro/web/Love.pdf ; Geng Li and Paul Smith, “Bor-rowing From Yourself: 401(k) Loans and Household BalanceSheets.” Working Paper 2008-42 (Federal Reserve Bo ard Divi-sions of Research and Statistics and Monetary Affairs, 2008).

59 Ibid.
https://www.ssc.wisc.edu/~scholz/Research/Are_All_Americans_v6.pdfhttps://www.ssc.wisc.edu/~scholz/Research/Are_All_Americans_v6.pdfhttp://www.ssa.gov/policy/docs/ssb/v72n1/v72n1p37.htmlhttp://www.ssa.gov/policy/docs/ssb/v72n1/v72n1p37.htmlhttp://www.ebri.org/pdf/briefspdf/EBRI_IB_396_Feb14.RRRs.pdfhttp://www.ebri.org/pdf/briefspdf/EBRI_IB_396_Feb14.RRRs.pdfhttps://www.americanprogress.org/wp-content/uploads/2013/03/WellerMiddleClass.pdfhttps://www.americanprogress.org/wp-content/uploads/2013/03/WellerMiddleClass.pdfhttps://www.americanprogress.org/wp-content/uploads/2013/03/WellerMiddleClass.pdfhttp://projects.vassar.edu/lamacro/web/Love.pdfhttp://projects.vassar.edu/lamacro/web/Love.pdfhttp://projects.vassar.edu/lamacro/web/Love.pdfhttp://projects.vassar.edu/lamacro/web/Love.pdfhttps://www.americanprogress.org/wp-content/uploads/2013/03/WellerMiddleClass.pdfhttps://www.americanprogress.org/wp-content/uploads/2013/03/WellerMiddleClass.pdfhttps://www.americanprogress.org/wp-content/uploads/2013/03/WellerMiddleClass.pdfhttp://www.ebri.org/pdf/briefspdf/EBRI_IB_396_Feb14.RRRs.pdfhttp://www.ebri.org/pdf/briefspdf/EBRI_IB_396_Feb14.RRRs.pdfhttp://www.ssa.gov/policy/docs/ssb/v72n1/v72n1p37.htmlhttp://www.ssa.gov/policy/docs/ssb/v72n1/v72n1p37.htmlhttps://www.ssc.wisc.edu/~scholz/Research/Are_All_Americans_v6.pdfhttps://www.ssc.wisc.edu/~scholz/Research/Are_All_Americans_v6.pdf

the inefficiencies of existing retirement savings incentives

Documents