Cross-tested plans allow employers to demonstrate nondiscrimination by converting contributions into benefits and testing the benefits rather than contributions. Common cross-tested plan designs include age-weighted, service-weighted, and class allocation profit sharing plans as well as cash balance and defined benefit/defined contribution combinations. Cross-testing works best when using a profit sharing/cash balance or 401(k) safe harbor combination and can be applied to groups of related employers. Flexible plan designs and avoiding unfavorable demographics are keys to successful implementation of cross-tested plans.

1. Cross-Tested Plans
James A . van I waarden, F S A , E A
Con s u l t i ng A c t u a r y
Van I wa a rde n A s s o c i a tes
Minneapo l is, MN
ISCEBS Sympo s ium, B o s ton
Sep tember 2 5 , 2 013

2. Today’s Menu
y What are cross-tested plans?
y History
y Cross-testing “recipe”
y Plan designs
y Other uses
y What works, what to avoid
y How to use cross-tested plans with 401(k)’s

3. What are Cross-Tested Plans?
y Plans that don’t fit into
IRS pre-approved
designs
y Or...a group of non-identical
plans
y Show nondiscrimination
by converting
contributions into
benefits (cross) and
testing the benefits
(testing!)

4. History of Cross-Testing – Part I
§401(a)(4)
•...benefits or
contributions
do not
discriminate
in favor of
HCEs...
RR 81-
202
•comparability
of plans
410(b) &
401(a)(4)
regs
•200+ pages,
early 1990’s

5. History of Cross-Testing – Part II
New
Comparability
•a.k.a. class
allocation plans,
based on the new
regulations
Threats to
outlaw
• led to “gateways”
effective 2002
Stable for
10+ years
•now, even
prototype plans
include class
allocation option

6. Cross-Testing “Recipe” for DC Plans
Step 1: pass thru any of three “gateways”
Go!
Broadly available
allocation rates
Gradual age or
service schedule
Minimum alloc rate:
5% or 1/3 top HCE

7. Cross-Testing “Recipe” for DC Plans
y Step 2: prepare the ingredients
@ 8½%
interest
Contribution
for each ee
To age
65
÷ annuity factor (we’ll use 10) to get annual benefit
÷ pay = accrual rate

8. Cross-Testing “Recipe” for DC Plans
y Step 3: mix it all together
There’s a rate group for each HCE;
it includes the HCE & everyone
(HCE or NHCE) with a higher
accrual rate:
ees w/ rate
>=HCE#1
HCE#1
Test each rate group to see if it passes
the §410(b) coverage test:
• 70% ratio percentage test, or
• average benefit percentage test
Ratio % below is 75% ÷ 50% = 150%
%/HCEs
2/4=50%
%/NHCEs
3/4=75%

9. Cross-Tested Plan Designs
y Common designs include:
{ Age-weighted: guaranteed to pass
{ Service-weighted
{ Class allocation (aka “new comparability”)
{ DB/DC, usually cash balance & profit sharing
{ Examples for Dr. F & his faithful servants...

10. Cross-Tested Plan Designs
y Age-weighted profit sharing plan
{ Always passes §401(a)(4) test, but not very flexible
{ Gateway is gradual age or service schedule
{ One rate group covers HCE (Dr. F) & both NHCEs, ratio % = 100%
Age Service Pay PS % PS $
Benefit
Accrual
Rate
Dr. Frankenstein 60 40 $255,000 20% $51,000 3.01%
Igor 40 10 50,000 3.91% 1,955 3.01%
Inga 30 5 30,000 3%* 900 5.22%
* Age-weighted formula produces 1.73%, but 3% top-heavy minimum applies.

11. Cross-Tested Plan Designs
y Service-weighted profit sharing plan
{ Rewards loyalty, not very common
{ PS in this example is ½% of pay per year of service
{ Gateway is gradual age or service schedule
{ One rate group covers HCE (Dr. F) & both NHCEs, ratio % = 100%
Age Service Pay PS % PS $
Benefit
Accrual
Rate
Dr. Frankenstein 60 40 $255,000 20% $51,000 3.01%
Igor 40 10 50,000 5% 2,500 3.85%
Inga 30 5 30,000 3%* 900 5.22%
* Service-weighted formula produces 2.5%, but 3% top-heavy minimum applies.

12. Cross-Tested Plan Designs
y Class allocation (aka “new comparability”) profit sharing
{ Very common, even prototype documents include this option
{ Gateway is minimum allocation 5% or 1/3 of highest HCE rate
{ One rate group covers HCE (Dr. F) & both NHCEs, ratio % = 100%
Age Service Pay PS % PS $
Benefit
Accrual
Rate
Dr. Frankenstein 60 40 $255,000 20% $51,000 3.01%
Igor 40 10 50,000 5% 2,500 3.85%
Inga 30 5 30,000 5% 1,500 8.69%

13. Cross-Tested Plan Designs
y DB/DC combo: cash balance & profit sharing
{ Common for professional firms
{ Enables large deductible contributions for owners
{ Gateway here is 7½% DC; Igor & Inga get 3% cash balance credits
{ This example is for a PBGC-covered plan, Frankenstein Mfg. Inc.
Ù A PBGC-exempt professional firm may need to limit DC er contrs to 6%
Age Pay
Cash
Balance
Credit
PS &
3% Safe
Harbor
401(k) &
Catchup Total
Dr. F’stein 60 $255,000 $230,000 $33,500 $23,000 $286,500
Igor 40 50,000 1,500 3,750 whatever 5,250+401k
Inga 30 30,000 900 2,250 whatever 3,150+401k

14. Other Uses for Cross-Testing
y Can structure retirement plans by business unit
y Could be all DB, all DC or a combination
y It’s a lifesaver for mergers & acquisitions

15. Other Uses for Cross-Testing
y Case study: parent & acquisitions, 25,000 total ees
DB/DC DB only DC only
y Combine all the plans and test on a benefits basis
y Can apply this for any group of non-identical plans

16. Cross-Testing – What Works, What to Avoid
y What works great
{ Profit sharing / cash balance combination
Ù PS projected at 8½% interest
Ù Cash balance at interest crediting rate (usually much lower)
Ù Better results than pure DB or pure DC
Ù Effect is known as

17. Cross-Testing – What Works, What to Avoid
y What works great
{ 401(k) 3% safe harbor
{ You get a triple dip
Ù Free pass for ADP & ACP tests
Ù Covers DC top-heavy requirement
Ù Counts for cross-testing base & gateway
{ A few issues
Ù Notice required >30 days before plan year starts
Ù 100% vesting
Ù Can’t impute “permitted disparity” (FICA recognition) for testing
{ Add profit sharing, if needed to meet gateway

18. Cross-Testing – What Works, What to Avoid
y What to avoid
{ Unfavorable demographics
Ù HCEs younger than others
Ù Usually a solution, but it’s tricky
{ Inflexible plan designs
Ù Contribution rates specified in plan document
Ù Even prototypes now allow each participant in own class

19. Cross-Testing – What Works, What to Avoid
y What to avoid
{ Failing the §401(a)(4) test!
{ Causes
Ù Demographic changes, especially new young HCEs
Ù Inflexible plan designs
{ Prevention
Ù Test before contributing
Ù Flexible plan designs
{ Cure
Ù §1.401(a)(4)-11(g) corrective amendment
Ù Can provide extra contributions for a targeted group of NHCEs
Ù Must be done within 9½ months after end of year

20. Cross-Testing – What Works, What to Avoid
y What to avoid
{ 401(k) matching contributions (even safe harbor)
{ Really?
{ Deferrals & match don’t count for 1st stage of §401(a)(4) test
{ You can still match, but it’s an extra expense
{ What can you do?
Ù Convert match to 3% safe harbor, with advance notice
Ù Add profit sharing for gateway & cross-testing base

21. Using Cross-Tested Plans with 401(k)’s
y Get top people within deferral range of §415 DC maximum
{ In 2013, that’s $51,000 - $17,500 = $33,500 (13.14% of max pay)
y Cover the gateway for NHCEs
{ If HCEs are at 13.14%, NHCEs need to be at 1/3 of that: 4.38%
y Use 3% non-matching safe harbor
{ That leaves only 1.38% needed for NHCE profit sharing
{ Frees HCEs to maximize 401(k) deferrals
y Still need to run the test
{ But gateway is often enough to pass
20%
15%
10%
5%
0%
NHCE HCE
Def
PS
SH

22. Take-Out Menu
y Cross-testing is the most powerful testing tool there is!
y You can structure plans to fit business needs
y It makes some amazing plan designs possible