Contributions Í O O P E R S M I T H S U M U T K A
Tax-Efficient RetirementWithdrawal Planning Using aLinear Programming Model
by Lewis W. C o o p e r s m i t h , Ph.D. , and A lan R. S u m u t k a , (PA
Lewis W. Coopersmilh, Ph.D., is an associate professor of
managetnent sciences at Rider University in Lawrenceville.
New ¡ersey, and provides coafulting to industry and
government on forecasting and market dynamic;. He can
be reached at [email protected]
Alan R. Sumutka, CPA, is an associate professor of
accounting at Rider University in /.awrenceuille, New
Jersey, and the owner of Alan R. Sumutka, CPA.
w ith the U.S. populationaging and baby boomersreaching retirement age,
attention has focused on how nesteggs can best provide income duringretirement. Recent articles and paperssuch as Davis (2009) and Seibert andMeredith (2010) discuss policies andprocedures used in retirement planning.They include distribution (or decumula-tion) plans that consist of sequencingwithdrawals from retirement savingsneeded to satisfy a desired lifestyle.
Distribution planning is simple if onehas only tax-deferred accounts and/or apension. However, nest eggs often con-sist of various accounts (taxable savings,tax-deferred savings. Roth IRAs) thatoften produce different rates of return(ROR) and tax liabilities. For example,stocks generally produce a higherROR than bonds and are often a highpercent of tax-efficient taxable savingsportfolios, which provide low tax ratesfor qualified dividends and long-term
Executive Summary
A common rule (CR) for with-
drawing retirement savings is to
withdraw taxable savings before
tax-deferred savings, but this strat-
egy can inflate required minimum
distributions (RMDs) and reduce tax
efficiency and wealth. However, tax-
efficient (TE) withdrawal schemes
can determine withdrawals that
maximize the final total account
balance over a retirement horizon.
We consider identical scenarios
(for example, initial wealth, living
expenses excluding federal taxes.
Social Security, tax deductions), but
use different withdrawal methods
(TE using a linear programming
spreadsheet model versus CR) to
determine withdrawals and federal
capital gains. In contrast, bonds oftenprovide a lower ROR than stocks; if heldin taxable savings, the interest is taxedannually at ordinary income tax rates.Thus, bonds are often a high percent oftax-deferred savings portfolios in whichthe interest is sheltered from taxationuntil withdrawn and then taxed entirelyas ordinary income.
When the retirement nest eggincludes both taxable and tax-deferred
taxes while satisfying RMDs over a
25-year planning horizon (to age 90).
We compare the final total account
balances for various combinations
of taxable rates of return (ROR), tax-
deferred ROR, initial taxable savings,
and itemized deductions.
Results show that TE models can
significantly outperform CR when
taxable ROR is greater than tax-
deferred ROR, initial taxable wealth
is greater than 10 percent of total
retirement wealth, and itemized
deductions are greater than the
standard deduction. For a realistic
combination of these conditions,
total remaining account balances
for TE are shown to be more than 16
percent higher than CR.
savings, a common rule (CR) forsequencing withdrawals is to spendtaxable savings before withdrawingtax-deferred savings, allowing tax-deferred earnings to grow continuouslyin a tax-sheltered environment. It canbe shown that CR usually providesthe highest growth in wealth over aone-year period. However, even over aone-year horizon, it is better to retaintaxable savings when the taxable ROR is
so JOURNAL OF FINANCIAL PLANNING | September 2011 www.FPAnet.org/Journal
CoîTtnbutions
higher than the tax-deferred ROR plusthe marginal tax on the tax-deferredwithdrawal (see the online appendix, AMarginal Analysis, at www.FPAnet.org/Journal/SeptllAppendix). Reichensteinnotes that the latter scenario shouldrarely occur (see Butteil (2010a)).The most likely exception would arisewhen the marginal tax rate is zero (taxdeductions exceed income) and theROR for taxable savings (for example,8 percent) is greater than the ROR ontax-deferred savings (for example, 4percent). However, when RORs for tax-able and tax-deferred savings are close,complex aspects of the tax laws suchas tax brackets and required minimumdistributions (RMDs) after age 7O'/2may make it more tax efficient andfinancially advantageous to withdrawfrom tax-deferred accounts earlier thanrequired to avoid less tax efficiency laterin retirement.
We consider a withdrawal plan taxefficient (TE) if it includes all of thefollowing:
• Consideration of more than oneannual withdrawal plan over aretirement horizon, where eachplan differs by the sequenceof withdrawals from differentaccounts and each account has adifferent tax treatment
• A realistic calculation of taxes foreach plan
• Selection of the plan that is bestwith respect to some performancemeasure (for example, final totalaccount balance)
CR plans make withdrawals fromtaxable savings until they are exhausted,unless RMDs mandate early withdrawalsfrom tax-deferred savings. As tax-deferredsavings increase early in retirement, futureRMDs may produce excessive withdrawalstaxed at higher than expected tax rates. Incontrast, TE plans withdraw from eithertaxable or tax-deferred savings to generatethe maximum total remaining accountbalance at the end of 25 years (or what-
ever the retirement horizon). TE plansmay withdraw from tax-deferred savingsearlier than needed and transfer additionalamounts to taxable savings to avoid futurelarge tax payments. TE plans tend tocreate annual taxable income levels taxedat similar marginal tax brackets, reducefuture excessive RMDs, bolster the moretax-efficient taxable savings, and enhanceoverall wealth over long periods.
The importance of tax efficiency isemphasized by Ed Slott, who states, "Ifyou really W2mt to attract the retire-ment market, you have to address thetax planning issues involved. It's thenumber one, biggest factor that willseparate people fiom their money" (seeButteil (2010a)). Guyton (2010) alsoemphasizes the need for "tax efficiencyon both a short- and long-term basis."Coopersmith, Sumutka, and Arvesen(2009) use mathematical optimization(as discussed below) to achieve taxefficiency for retirement savings thatinclude taxable and tax-deferred savings.Feedback from this earlier paper sug-gested comparing TE and CR plans fora wide range of variables. In this paperwe address this objective and morethoroughly evaluate when TE modelsprovide significant practical benefits
over CR plans.'I
Literature ReviewPrior literature on tax efficiencycompares withdrawal strategies undera variety of conditions. Ragsdale, Seila,and Little (1994) refer to a withdrawalplan as "heuristic" when it is based ona subjective rule. CR is an example ofa heuristic plan considered in manypapers. A TE plan is "mathematicallyoptimal" when it provides the bestoutcome for all possible plans. Math-ematical optimization, althoughdesirable, may not always be achievahlequickly. When a spreadsheet is used todetermine a TE plan, a mathematicallyoptimal model runs fast when a cell ina spreadsheet (for example, final total
account balance) is selected as the per-formance measure and a time-efficientmathematical optimization method, forexample linear programming (LP), isused to optimize that cell. However, thepractical use of mathematical optimiza-tion diminishes rapidly with increasingrun time when various "non-linear"aspects of the tax code are introduced(for example, the exact taxable amountof Social Security) or when the perfor-mance measure is not a single cell (forexample, longevity, which is the lengthof time before funds are exhausted).
Ragsdale, Seila, and Little (1994)examine mathematically optimalTE withdrawals using an LP model.Subsequent papers that use heuristic TEmethods have expressed concern thatthe Ragsdale model is not applicable tothe current tax code. This concern is notvalid because optimization models caneasily be adapted to changes in federaltax and estate laws. This is demon-strated by the LP model of Coopersmith,Sumutka, and Arvesen (2009), whichincludes taxable savings and SocialSecurity and allows for tax deductions(standard deduction and exemptions)and tax brackets. The LP model used inour current research runs in less than aminute, making it quite time efficientfor generating the many alternativeresults discussed below.
Many recent papers use heuristicmethods of retirement withdrawalplanning, but not all assess tax impact.Spitzer and Singh (2006) considerinitial wealth divided between taxableand tax-deferred savings and comparelongevity for various portfolio scenariosassuming a fiat rate of taxation. Horan(2006) includes multiple tax brackets,deductions, and exemptions, but notRMDs, to evaluate six rules for sequenc-ing withdrawals from Roth and tradi-tional IRA accounts; total remainingaccount balance is used as a comparativemeasure. Van Harlow and Feinschreiber(2006) distribute wealth among various
www.FPAnet.org/Journal September 2011 I JOURNAL OF FINANCIAL PLANNING SI
Contributions C O O P E R S H I T H I S U M U I K A
account types and consider federal taxesto develop five guidelines for maintain-ing a "personal withdrawal 'hierarchy'that is tax efficient." Bernachhi (2008)includes RMDs, tax brackets, and CRplans to determine the mix of taxableand tax-deferred savings at the start ofretirement that results in the highestfinal total remaining account balance.Spitzer (2008) studies the impact ofRMDs on average balance remainingfor four withdrawal cases. He computesapproximate federal taxes and considers abroad range of portfolio possibilities. Weuse Spitzer's terminology and expand onhis research design as described below.
Research DesignTwo withdrawal plans are comparedwith respect to the total account balanceremaining at the end of 25 years:
1. CR: At the start of each year,taxable savings are withdrawn tosatisfy expenses and taxes; if not
sufficient, tax-deferred savingsare also withdrawn. When RMDsapply, excess tax-deferred with-drawals are transferred to taxablesavings at year end.
2. TE: An LP model similar tothat employed in Coopersmith,Sumutka, and Arvesen (2009) isused to determine tax-efficientannual withdrawals from taxableand tax-deferred savings. With-drawals used to satisfy specifiedliving expenses and tiixes aremade at the start of each year.If additional withdrawals fromtax-deferred savings are needed forRMDs or to achieve tax efficiencyover the longer planning horizon,the excess is transferred to taxablesavings at year-end.
The CR and TE withdrawalplans compared in this researchare identical with respect to thefollowing:
FPA.FINANCIALPLANNINGASSOCIATION
FPA would like to recognize itsNATIONAL SPONSOR:
AmeritradeInstitutional
When the right people come together, THEY CAN REACH NEW HEIGHTS.
• Withdrawals are for a marriedcouple, both age 65.
• Total savings (taxable plus tax-deferred) are initially $1 million.
• 25 years of withdrawals arecalculated, ending with the coupleat age 90; we consider this a suit-able horizon for comparing the twowithdrawal planning methods.
• Annual living expenses (excludingfederal taxes) are fixed initiallyat Social Security ($20,000) plusa percent of total savings. Forexample, $50,000 initial totalannual living expenses would bedrawn: $20,000 from Social Secu-rity plus $30,000 (or 3 percent)from the $1 million savings. Weuse initial withdrawal rates of 3percent, 4 percent, and 5 percent oftotal savings. Past literature listedin Salter and Evensky (2008) hasfound these rates to have relativelylow sustciinability risk, definedas the probability of exhaustingretirement funds during the plan-ning horizon. Because we compareCR and TE plans with the sameannual initial withdrawal rates,we leave the evaluation of differ-ences between CR and TE relativeto sustainability risk for futureresearch.
• Federal taxes, if any, require addi-tional vkdthdrawals from savings.
• 2010 tax laws are used in determin-ing all tax issues.
• Interest from tiixable savings andwithdrawals from tax-deferred sav-ings are taxable as ordinary income;the couple earns no dividends orcapital gains.
• 85 percent of Social Security istaxed.
• RMDs are determined as prescribedby federal law and must be satisfied.
• Tax deductions of a standarddeduction for a married couplewho files jointly ($12,500) and twoexemptions ($3,650 for each) are
52 JOURNAL OF FINANCIAL PLANNING | September 2011 www.FPAnet.org/Journal
Contributions C O O P £ R S M I I H I S U M U I K A
Table 1 : CR Withdrawal Plan: Initial Taxable Savings = $200K, Initial— — — Withdrawal = $40K,Taxable ROR = 7%,Tax-Deferred ROR = 5%
Expenses Income Sources/Withdrawals
r\ge( i6
6/
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
858687888990
1 ^Living
Expenses($000)
60.0
61.2
62.4
63.7
64.9
66.2
67.6
68.9
70.3
71.7
73.1
74.6
76.1
77.6
79.2
80.8
82.4
84.0
85.7
87.4
89.2
90.9
92.8
94.6
96.5
FederalIncomeTaxes($000)
0.8
0.6
0.4
0.1
4.1
4.2
4.4
4.7
4.9
5.2
5.96.9
7.1
7.2
7.4
7.5
7.7
7.8
8.0
8,1
8.3
8.5
8.6
8.8
9.0
Total($000)60.8
61.8
62.8
63.8
69.0
70.4
72.0
73.6
75.1
76.9
79.0
81.5
83.2
84.8
86.6
88.3
90.1
91.8
93.7
95.5
97.5
99.4
101.4
103.4
105.5
SocialSecurity
($000)
20.0
20.4
20.8
21.2
21.6
22.1
22.5
23.0
23.4
23.9
24.4
24.9
25.4
25.8
26.4
26.9
27.5
28.0
28.6
29.1
29.7
30.3
30.9
31.5
32.2
Taxable($000)40.8
41.4
42.0
42.6
47.4
48.3
42.7
32.1
22.3
13.2
4.9
Deferred($000)
6.8
18.5
29.5
39.8
49.8
56.7
57.8
59.0
60.1
61.3
62.6
63.8
65.1
66.4
67.7
69.1
70.5
71.9
73.3
AdditionalTax-
DeferredWithdrawals
for RMD($000)
35,5
37.2
32.2
22.3
13.2
4.9
RMD($000)
35.5
37.2
39.0
40.8
42.7
44.7
46.7
48.5
50.2
51.8
53.4
55.0
56.7
58.4
60.2
61.6
62.9
64.2
65.4
66.5
66.9
subtracted from income to arrive attaxable income.Annual cost of living adjustments of 2percent are applied to living expenses(before federal taxes). Social Security,deductions, and tax brackets.Federal income taxes are based ontaxable income and are computed aspart of the process of calculating thewithdrawal amounts for each plan:» For CR, the computation of
federal taxes for a given tax-able income is formulated in aspreadsheet.
» For TE, the computation offederal taxes is built into theLP model as in Coopersmith,Sumutka, and Arvesen (2009).
The comparative measure of wealthis the total account balance remain-
ing (in taxable plus tax-deferredsavings) at the end of 25 years.
Scenarios are compared for combina-tions of:
• Taxable ROR: 5 percent, 6 percent,7 percent
• Tax-deferred ROR: 5 percent, 6percent, 7 percent
• Percent of initial total wealthin taxable savings: 0 percent, 5percent, 10 percent, 15 percent, 20percent, 25 percent
• Amount of deductions, excludingexemptions: standard deduction,1.25 X standard, 1.50 x standard,1.75 X standard, 2.00 x standard
ResultsTo illustrate the difference between TE andCR plans, we consider the case in which:
• Initial total savings of $1 million is$200,000 taxable plus $800,000tax-deferred
• Initial annual withdrawal is$40,000 (4 percent of initial totalsavings)
• Taxable ROR is 7 percent; tax-deferred ROR is 5 percent
The use of LP ensures that the TEplan never has a lower comparativeremaining total account balance after25 years. Tables 1 and 2 illustrate thediffering withdrawal streams for CR andTE plans, respectively; federal taxes andRMDs are shown for each plan. Table 3compares annual account balances over25 years of withdrawals for CR and TE.As will be demonstrated, TE generates$100,000 more in total remainingaccount balance after 25 years thanCR (a 14 percent improvement) andprovides additional financial and estateplanning advantages.
Table 1 illustrates that, to satisfy totalrequired expenses, CR withdraws onlyfrom taxable savings through age 71.Additional amounts must be withdrawnfrom tax-deferred savings from ages 70to 75 and transferred to taxable savingsat year-end to satisfy RMDs. This actioncauses a spike in taxes from $100 at age69 to $4,100 at age 70. Taxable savingsare depleted at age 76, after which allwithdrawals must be made from lesstax-efficient, tax-deferred savings.
In contrast. Table 2 illustrates thewithdrawal pattern of the TE plan:
• From ages 66 through 72, TEwithdraws only from tax-deferredsavings ($49,400 to $55,600,respectively) to satisfy totalrequired expenses
• From ages 66 through 72, TEwithdraws additional amounts atyear-end from tax-deferred savings($7,400 to $700, respectively)then transferred to taxable savingsto provide tax efficiency over thelonger horizon
• From ages 70 through 78, TE
54 JOURNAL OF FINANCIAL PLANNING I September 2011 www.FPAnet.org/Journal
C O O P E R S M I T H | S U M U T K A Contributions
tax-deferred withdrawals ($53,400to $56,100, respectively) exceedRMDs ($26,200 to $24,400,respectively)
• From ages 79 through 90, TE tax-deferred withdrawals ($23,700 to$32,900, respectively) equal RMDs;withdrawals from the more tax-efficient taxable savings ($35,100to $38,900, respectively) constitutethe bulk of withdrawals
Table 3 displays 25 years of accountbalances for CR and TE. As expected,CR depletes all taxable savings byage 75; at age 90, $724,000 of tax-inefficient, tax-deferred savings remains.In contrast, TE produces increasingtaxable savings balances through age 78;subsequent taxable savings balances stayabove 50 percent of the total accountbalance. At age 90, the TE balance is$824,000 ($359,000 tax-deferred plus$465,000 taxable), which is a $100,000or 13.9 percent improvement over CR.The final high balance in taxable savingsprovides further tax benefits afterage 90, for example, higher after-taxproceeds if annuities are chosen, or areduced tcix burden for heirs.
An interesting observation is that taxefficiency does not necessarily mean taxminimization. The cumulative taxes over25 years (computed using Tables 1 and 2)are $146,000 for CR and $225,000 for TE(about $3,000 more annually). Plannersneed to carefully explain this effectto clients, pointing out that increasesin taxes are offset by greater wealthaccumulation of TE. For CR (in Table 1),taxes are low before age 70 when incomesubject to tax includes interest on taxablesavings and 85 percent of Social Security;these are largely offset by deductions.However, at age 70 RMDs are $35,500and increase annually to $66,900 at age90, causing annual taxes to increasefrom $4,100 to $9,000, respectively. Incontrast, for TE (in Table 2) from ages 66to 70, taxes are significantly higher thanCR ($9,400 to $10,100 annually) because
Table 2: TE Withdrawal Plan: Initial Taxable Savings = $200K, Initial
Withdrawal = $40K,Taxable ROR = 7%,Tax-Deferred ROR = 5%
IIrawals
Age66
67
68
69
7071
72
73
74
75
7677
7879
808T
8283
8485
8687
88
89
90
•Living
Expenses($000)
60.0
61.2
62.4
63.7
64.966.2
67.6
68.9
70.371.7
73.1
74.6
76.1
77.6
79.280.882.4
84.0
85.787.4
89.290.9
92.894.6
96.5
IncomeTaxes($000)
9.4
9.59.7
9.9
10.110.3
10.5
10.8
11.0
11.2
11.4
11.6
11.9
7.0
7.17.1
7.27.3
7.4
7.4
7.5
7.5
7.5
7.5
7.5
•Total($000)69.4
70.7
72.1
73.6
75.076.5
78.179.7
81.3
82.9
84.5
86.2
88.0
84.6
86.387.9
89.691.3
93.1
94.8
96.698.4
100.3102.1
104.0
•Social
Security($000)
20.020.4
20.8
21.2
21.622.1
22.5
23.0
23.4
23.9
24.4
24.9
25.4
25.8
26.426.9
27.528.0
28.629.129.7
30.3
30.9
31.5
32.2
•Taxable($000)
0.5
1.62.8
4.0
5.2
6.535.1
35.235.4
35.6
35.8
36.036.4
36.8
37.2
37.7
38.2
38.9
ÏTax-
Deferred($000)
49.4
50.3
51.3
52.4
53.454.4
55.656.2
56.3
56.2
56.1
56.1
56.123.7
24.725.6
26.527.5
28.5
29.330.1
30.931.7
32.4
32.9
AdditionalTax-
DeferredWithdrawals
forTaxEfficiency
($000)
7.4
6.4
5.3
4.2
3.01.9
0.7
RMD($000
26.226.2
26.226.1
25.925.725.4
24.9
24.423.7
24.725.626.527.5
28.5
29.3
30.130.9
31.732.4
32.9
TE draws only from tax-deferred savings.However, these high withdrawals in earlyyears result in RMDs consistently lowerthan CR and fall below half that of CR atage 78. Specifically, at age 90 the RMDfor TE is $32,900, which is significantlybelow the RMD of $66,900 for CR. Fromages 70 through 78, taxes increiise from$10,100 to $11,900; but from age 79 to90, when RMDs are at their lowest (andequal total tax-deferred withdrawals),taxes average only around $7,000 per yeai.
In this example, CR generates lowercumulative taxes during retirement byfirst depleting the more tax-efficienttaxable account. However, becausethe taxable account is higher yielding,lower wealth results at age 90 and leaveshigher taxes for heirs by bequeathing
the less tax-efficient tax-deferredsavings. TE trades off higher cumulativetaxes earlier in retirement by depletinglower-yielding tax-deferred accountsto provide greater wealth at age 90 andlower taxes for heirs. Simply, TE buildsup the higher-yielding, more tax-efficienttaxable savings, which creates bothgreater wealth and cumulative taxes.
How do differences in taxable RORand tax-deferred ROR affect endingwealth for TE versus CR withdrawalsequences? Figure 1 compares resultsfor different tax-deferred RORs: 5percent in Graph A, 6 percent in GraphB, and 7 percent in Graph C. Each graphshows the percent improvement of totalremaining account balance of TE overCR for taxable RORs of 5 percent, 6
www.FPAnet.org/Journal September 2011 | JOURNAL OF FINANCIAL PLANNING 55
C o n t r i b u t i o n s COOPERSMITH I SUMUTKA
Table 3: Comparison of Account Balances: Initial Total Savings = $1M,
iAge
66
67
68
69
70
71
72
73
74
75
7677
78
79
80
81
82
83
84
85
8687
88
89
§90
Z r • • • •»•ai •»«»•*•«=>
^ 1 Taxable ROR =
Taxable($000)
170
138
103
64
53
43
32
22
13
5
CR
Tax-Deferred
($000)
840
882
926
972
986
998
1,0081,0171,0241,0281,0271,0191,009
998
985
969
952
933
911
887
860
831
798
763
7%, Tax-Deferred ROR =
Total($000)
1,0101,020
1,0291,0371,0391,0401,0401,0391,0371,0331,0271,0191,009
998
985969
952
933
911
887
860
831
798
763
724
Taxable($000)
221
243
266
288
312
335
359
384
409
435
461
488
515
513
511509
507
504
501
497
492
487
481
474
465
5%
TE
Tax-Deferred
($000)
781
761
739
717
694
670
644
617
589
559
528
496
462
460
457
453
448
442
434
425
414
403
389
375
359
Total($000)
1,0021,004
1,0051,0061,0051,0051,0031,001
998
994
989
983
977
973
969
963
955
946
935
922
907
890
870
848
824
percent, or 7 percent for different initialannual withdrawals of $30,000 (3 per-cent), $40,000 (4 percent), or $50,000(5 percent). These RORs, whichrepresent a mix of stocks and bonds, arein line with reported historical RORs.Spitzer and Singh (2006) assume a7.2 percent return for stocks and 3.5percent return for bonds for the 40-yearperiod ending in 2003. More recently,Weigand and Irons (2008) use a stockROR of 10.34 percent and a bond RORof 4.13 percent for the 25-year periodending in 2006.
Graph A shows that for initial with-drawals of 3 percent or 4 percent, whentax-deferred ROR is 5 percent and taxableROR is 5 percent or 6 ¡percent, the percentimprovement in TE over CR is minimal.However, when taxable ROR increases
to 7 percent, the improvement is muchhigher, jumping to more than 8 percent fora $30,000 (3 percent) initial withdrawal.For an initial withdrawal of $40,000 (4percent), the TE benefit is magnifiedto 12 percent because TE can withdrawmore from the lower-yielding tax-deferredsavings earlier to satisfy expenses.
Graphs B and C illustrate that as thetax-deferred ROR increases to 6 percentand 7 percent, respectively, the yieldadvantage of taxable savings is reduced(note values shown for $40,000 (4percent)) resulting in greatly reducedbenefits of TE over CR for any amountof the initial withdrawal. However, eachgraph in Figure 1 shows that the benefitsof TE over CR grow with increasinginitial withdrawal rates.
How do changes in levels of initial
taxable savings affect the improvementin ending wealth of TE over CR? Figure2 illustrates the changes when the initialwealth in taxable savings increases from$0 to $250,000, in $50,000 increments.Here tax-deferred ROR is 5 percent, andthe taxable ROR is 6 percent (Graph A)and 7 percent (Graph B). First, we seethat as the amount of initial wealth intaxable savings increases, the TE advan-tage grows. For all initial withdrawalrates, the improvement in TE over CRpeaks at around $150,000/$200,000of taxable savings. As taxable savingsrise beyond $200,000, the percentageimprovement in TE over CR declines.Obviously, if the entire amount ofwealth is in taxable savings, TE and CRwould produce identical results. Second,when the taxable ROR increases (6percent to 7 percent), the benefit ofretaining the more tax-efficient taxablesavings and withdrawing tax-deferredsavings increases. The greatest improve-ment is around 13.9 percent whentaxable ROR is 7 percent and the initialwithdrawal is $40,000 (4 percent).
Is there a greater level of improve-ment in TE over CR if a taxpayer item-izes deductions to lower taxable incomeand possibly reduce the marginal taxrate? Figure 3 illustrates the improve-ment when the basic standard deduc-tion is $12,500, but itemized deductionsare $15,625, $18,750, $21,875, and$25,000 (or 125 percent, 150 percent,175 percent, and 200 percent of thestandard deduction, respectively), andthe initial wealth in taxable savingsvaries by $50,000 increments from$0 to $250,000. When taxable savingsare between $150,000 and $250,000,TE provides a total remaining accountbalance more than 16 percent higherthan CR when itemized deductions aredouble the standard deduction.
ConclusionRetirement savings frequently include
both taxable and tax-deferred components.
56 JOURNAL OF FINANCIAL PLANNING I September 2011 www.FPAnet.org/Journal
C O O P E R S M I T H I S U M U T K A Contributions
At retirement, a common rule (CR)for withdrawal planning is to spendtaxable savings first. Although thisis a good rule, it is not always best.Tax-efficient models (TE) can be usedto plan withdrawals so that retirementlifestyle goals and financial objectivesare met while providing high growth intotal savings. Results for TE models canbe determined within a minute and canbe presented in a spreadsheet formatidentical to that of CR or other with-drawed planning methods. Key aspectsof complex tax laws and governmentrestrictions on use of retirement savings(for example, RMDs) are built into TEmodels that can easily be revised as lawschange. The fast and automatic natureof TE planning allows financial plannersto focus on "what if" analyses that mightvary by investment strategy and alterna-tive streams of annual living expenses.
We compare TE to CR results fordifferent financial scenarios in whichdata for the two methods are identicalin every way. Only the amounts ofómnual withdrawals from tcixable andtax-deferred accounts vary and conse-quently change the amount of federaltaxes due. Our measure of comparisonis the total remaining account balance(taxable plus tax-deferred) afrer 25years. Because the TE model uses LPto determine withdrawal amounts, thetotal remaining account balance forTE can never be less than that for CR.However, a TE plan can be identical orclose to CR under many conditions.
Our results imply that TE models cansignificantly outperform CR for combi-nations of the following conditions:
• Taxable ROR is greater than tax-deferred ROR
• Initial taxable savings are greaterthan 10 percent of total retirementsavings
• Itemized deductions are greaterthan the standard deduction
The first two conditions favorsustaining initially high taxable savings
Figure 1 : Total Remaining Account Balance After 25 Years of— — — Withdrawals, Percent Improvement TE over CR, Initial
Total Wealth = $1M,Taxable Savings Portion = $100K
Tax-Deferred ROR = 5% *
5% 6%
Taxable ROR*Finat savings exhausted for initial wealth ofiSOK
Graph B
14%
12%
« 10%
I "£ 6%
^ 4%
2%
0%
Graph C14%
Tax-Deferred ROR =6%
0.42% 0.48% 0.74%
5% 6%
Taxable ROR
Bf Tax-Deferred ROR =7%
IaE
12%
10%
8%
6%
4%
2%
0% 0.25% tO.28% I 0.33%
5% 6%Taxable ROR
Initial Withdrawal (% Initial Wealth)-$50K(5%) - • -$40K(4%)
(values shown)
7%
•$30K(3%)
www.FPAnet.org/Journal September 2011 | JOURNAL OF FINANCIAL PLANNING 57
C o n t r i b u t i o n s C O O P E R S H I T H I SUMUTKA
Figure 2:
Graph A14%
12%
10%
8%
6%
4%
2%
0%
Graph B
I
Total Remaining Account Balance After 25 Years ofWithdrawals, Percent Improvement TE over CR, by InitiaiWealth in Taxable Savings, Initial Totai Wealth = $1M,Tax-Deferred ROR = 5% I
Taxable Savings ROR = 6%
$0K $50K $100K SISOK $200K $250K
Initial Wealth in Taxable Savings
Taxable Savings ROR = 7%13.8%
'12.0%
13.9% 13.4%
$0K $50K SIOOK $150K $200K
Initial Wealth in Taxable Savings
Initial Withdrawal (% Initial Wealth)- • - $40K (4%) - • - $30K (3%)
$250K
to provide higher growth. Increasing taxdeductions may lower the marginal taxrate to enhance the financial advantageof transferring tax-deferred savings totaxable savings early in retirement.
TE models appear to performsimilarly to CR models for:
• Relatively low initial taxable savings• Taxable ROR that is less than, or
very close to, tax-deferred ROR• Standard or low itemized deductionsUnder these conditions there is
little or no growth advantage in earlywithdrawal of tax-deferred savings ortransfer of funds from tax-deferred totaxable accounts. The initial low level oftaxable savings is depleted early for both
planning methods.Employers and financial institutions
have traditionally promoted tax-deferredinvestments with the result that manyretirees are mostly or entirely investedin tax-deferred retirement savings.However, recent studies such asBernachhi (2008) have highlighted thebenefits of commencing retirement withboth taxable and tax-deferred savings.This provides incentive to financialplanners to encourage clients toaccumulate some taxable savings priorto retirement. Thus, we may expectmore retirees whose retirement savingsand tax requirements satisfy many of theconditions favoring the use of TE. Our
results show a combination of condi-tions favoring TE models can providebetter than a 16 percent improvementin savings growth over CR. Thus, TEmodels have the potential to providehighly effective support in planningretirement withdrawals.
Future research may show how therelative benefits of TE models mightchange as more details of the tax laws areconsidered explicitly, such as reducedtaxes on capital gains and Roth IRAs.Another area for further study is howbest to allocate assets between stocksand bonds in light of the tendency ofTE withdrawal planning to prolong theavailability of taxable savings.
ReferencesBcinachhi, Ben T. 2008. "Determining the Proper
Starting Balance for Taxable and Tax-Deferred
Savings at Retirement." Journal 0/Financial
Planning 21, 7: 56-62.
Buetell, Amy E. 2010a. "Ed Slott on Roth IRAs,
Automatic IRAs, and Retirement Planning."
Journal of Financial Planning 23,10: 20-24.
Buttell, Amy E. 2010b. "William Reichenstein on
Retirement Investing, Withdrawals, and Tax
Efficiency." Journal 0/Financial Planning 23,11:
18-22.
Coopersmith L., A. R. Sumutka, and ]. Arvesen.
2009. "Optimal Tax-Efficient Planning of With-
drawals from Retirement Accounts." Academy of
Financia! Services Proceedings (October).
www.academyfinancial.org/09Conference/09F
roceedings/%283B%29%20Coopersmith,%20
Sumutka,%20Arvesen.pdf.
Davis, Brad. 2009. "Six Steps to Building an
Effective Retirement Income Strategy." Journal
0/Financial Planning 22,12: 18-19.
Guyton, J. T. 2010. "The Withdrawal Policy
Statement." Journal 0/Financial Planning 23, 6:
41-44.
Horan, S. M. 2006. "Optimal Withdrawal
Strategies for Retirees with Multiple Savings
Accounts." Journal of Financial Planning 19,11:
62-75.
58 JOURNAL OF FINANCIAL PLANNING | September 2011 www.FPAnet.org/Journal
C O O P E R S M I T H I S U M U T K A Contributions
Ragsdale, Cliff T., Andrew F. Seila, and Philip
L. Liltle, 1994. "An Optimization Model for
Scheduling Withdrawals from Tax-Deferred
Retirement Accounts." Financial Services
Review 3, 2: 93-109.
Salter, John R., and Harold Evensky. 2008.
"Calculating a Sustainable Withdrawal Rate: A
Comprehensive Literature Review." ¡oumal of
Personal Finance 6, 4: 118-137.
Seibert, Kevin S., and Betty Meredith. 2010.
"Managing Retirement Income for the
Mid-market." ¡ournal of Financial Planning 23,
8: 32-37.
Spitzer. John J. 2008. "Do Required Minimum
Distributions Endanger 'Safe' Portfolio
Withdrawal Rates?" Journal of Financial
Planning 21, 8:40-51.
Spitzer, John J., and Sandeep Singh. 2006.
"Extending Retirement Payouts by Optimizing
the Sequence of Withdrawals." Journal of
Financial Planning 19. 4: 52-61.
Van Harlow. W, and S. Fein.schreiber. 2006.
Beyond Conventional Wisdom: New Strategies
Figure 3: Total Remaining Account Balance After 25 Years ofWithdrawals, Percent Improvement TE over CR, by Sizeof Itemized Deduction, Initial Total Savings = $1M, InitialWithdrawal = $40K (4%),Taxable ROR = 7%,Tax-DeferredROR = 5%
S 100k S 150k 5200k
Initial Wealth in Taxable SavingsS2S0k
Increase of Itemized Deduction over Standard Deduction•100% - • - 7 5 % - * - 5 0 % -« -25% Standard Ded.
for Lifetime income. Fidelity Research Institute.
Weigand, Robert A., and Robert Irons. 2008.
"When Does a Bonds-First Withdrawal
Sequence Extend Portfolio Longevity?" Journal
of Financial Planninf> 21, 11: 66-77.
ror more than 28 years, Weitz Funds has successfullyapplied the same common sense approach to selectinginvestments for our mutual fund portfolios. Our primarygoal has always been to earn good absolute returns overtime without exposing our clients' and our own capital toundue risk.
Investors should consider carefully the investment objectives, risks and charges and expenses oí the WeitzFunds before investing. The Weitz Funds Prospectus or Summary Prospectus contains this and otherinformation about the Weitz Funds. The Prospectus or Summary Prospectus should be read carefullybefore investing. A Prospectus or Summary Prospectus and addiüonal information is available from WeitzFunds.
Weitz Securities, Inc. is the Distributor of the Weitz Funds.
www.FPAnet.org/Journal September 2011 | JOURNAL OF FINANCIAL PLANNING 59
Copyright of Journal of Financial Planning is the property of Financial Planning Association and its content
may not be copied or emailed to multiple sites or posted to a listserv without the copyright holder's express
written permission. However, users may print, download, or email articles for individual use.
Top Related