Get Your 401(k) Design Right

I happened to read a few things today about 401(k) matching contributions. One in particular talked about stretching the match. Apparently that means that if you are willing to spend 3% of pay on your workforce, consider making your match 50 cents on the dollar on the first 6% of pay deferred instead of dollar for dollar on the first 3% of pay deferred. This will encourage employees to save more.

That might be a really good idea ... for some companies. For other companies, it might not be.

First and foremost a 401(k) plan is, and should be, an employee benefit plan. Taken quite literally, that means that it should be for the benefit of employees.

Plan sponsors may look at the plan and say that they get a tax deduction. That's true, but they also get a tax deduction for reasonable compensation. And, there is probably less of a compliance burden with paying cash than there is with maintaining a 401(k) plan.

Where does the typical 401(k) design come from? Usually, it's the brainchild, or lack thereof, of someone internal to the plan sponsor or of an external adviser. Either way, that could be a good thing or a bad thing. Most plan sponsors have plenty of smart and thoughtful employees and many external advisers are really good.

On the other hand, when it comes to designing a 401(k) plan, some people just don't ask the right questions. And, just as important, they don't answer the right questions. We often see this in marketing pieces or other similar propaganda that talk about designing the best plan. We might see that the best plan has all of these features:

• Safe harbor design (to avoid ADP and ACP testing)
• Auto-enrollment (to get higher participation rates)
• Auto-escalation (so that people will save more)
• Target date fund as a QDIA (because virtually every recordkeeper wants you in their target date funds)

All of these could be great features for your 401(k) plan, but on the other hand, they might not be. Let's consider why.

Safe harbor designs are really nice. They eliminate the need for ADP and ACP nondiscrimination testing. They also provide for immediate vesting of matching contributions. Suppose your goal, as plan sponsor, is to use your 401(k) plan at least in part as a retention device. Suppose further that every year, you pass your ADP and ACP tests with ease. Then, one would wonder why you are adopting a plan with immediate vesting whose sole benefit is the elimination of ADP and ACP testing. Perhaps someone told you that safe harbor plans were the best and you listened. Perhaps nobody bothered to find out why you were sponsoring a 401(k) plan and what you expected to gain from having that plan.

Auto-enrollment is another feature that is considered a best practice. (Oh I despise that term and would prefer to call it something other than best, but best practice is a consulting buzzword.) Most surveys that I have read indicate that where auto-enrollment is in place, the most common auto-enrollment level is 3% of pay. Your adviser who just knows that he has to tell you about auto-enrollment tells you that it is a best practice. Perhaps he didn't consider that prior to auto-enrollment, you had 93% participation and that 87% of those 93% already deferred more than 3% of pay. Since he heard it was the thing to do, he advised you to re-enroll everyone and now, you are up to 95% participation, but only 45% of them defer more than 3% of pay. Perhaps nobody bothered to ask you how your current plan was doing.

In the words of a generation younger than me, this is an epic fail.

I could go on and on about other highly recommended features, but the moral of the story is largely the same. Your plan design should fit with your company, your employees, your recruiting and retention needs, and your budget. That your largest competitor has a safe harbor plan doesn't make it right for you. It may not even be right for them. That the company whose headquarters are across the hall from yours has auto-escalation doesn't make it right for you. It may not be right for them either.

If you are designing or redesigning a plan for your company, ask some basic questions before you go there.

• What do you want to accomplish with the plan?
• Enough wealth accumulation so that your employees can retire based solely on that plan?
• Enough so that the plan is competitive?
• Something else?
• Will eliminating nondiscrimination testing be important?
• What is your budget? Will it change from year to year? As a dollar amount? As a percentage of payroll?
• What do you want your employees to think of the plan?
• It's a primary retirement vehicle.
• My employer has a 401(k) plan; that's all I need to know.
• My employer has a great 401(k) plan.
• My 401(k) is a great place to save, but I need additional savings as well.
• Will any complexity that I add to the plan help my company to meet its goals or my employees to meet their goals? If not, why did I add that complexity?

These are the types of questions that your adviser asked you when you designed or last redesigned your plan, aren't they?

They're not?

Perhaps it's time to rethink your plan.

The 401(k) Design Quandary

It's a question that is really in vogue these days: what is the optimal 401(k) design? Should you auto-enroll? Should you match? How big should the match be? How quickly should the match fully accrue?

I read an article yesterday that was written as if it had the authoritative answers. I was amused, to say the least. It said that if you are currently matching 100% on the first 3% of pay deferrals, you should switch to 50% on the first 6% of deferrals. People would still get the same match, but they would defer more in order to get it and the plan sponsor's cost wouldn't change.

Whoa! Hold on.

What are your goals? Can your typical participant afford to defer 6% of pay? If they are currently deferring 3% of pay to get the full match, how much will they have to change their use of their take-home pay in order to be comfortable deferring 6% of pay? Will this affect your ADP and ACP nondiscrimination testing?

Many of the fund houses that are recordkeepers write (and speak) as if they have all the answers. Don't they have an ulterior motive though? They want to get more assets under management. They make more money that way.

Don't get me wrong. This is not intended to say that participants shouldn't be encouraged to save more. They should save as much as possible. But, it's not fair to make blanket statements like the one that was made, and purport that they apply to a general situation.

Plan design studies should not be done based on articles found in internet searches. Just like any other study, they should start with an enunciation of parameters -- goals, constraints, and desired outcomes. And, don't forget risks. No new design should leave a company exposed to risks that are too extreme.

I'll give you a real-life example. I was working with a large company last year that was considering a change to their 401(k) plan. Fortunately, this is an extremely cash-rich company, so if they had decided to make the change they were contemplating, even though the cost increase would have been measured in 10s of millions of dollars, this would not have broken the bank.

Currently, the company offers a match of something like (I don't recall the details exactly) 100% on the first 5% of pay deferred either pre-tax or Roth. They also allow after-tax contributions. They were considering a change to match both (not either, but both) pre-tax (orRoth) and after-tax contributions dollar for dollar on the first 5% of pay. It didn't occur to them that their highly educated workforce would find a way to defer at least 5% of pay pre-tax (or Roth) AND 5% after-tax, to double their match. I'm not sure why.

In any event, do your work properly. Understand your goals, risks, constraints, and desired outcomes. And, just because a design is popular or written up in an article, that doesn't mean it's the right one for you.

Employer Matches Being Restored and Even Increased?

A recent study by the Profit Sharing/401(k) Council of America (PSCA) suggests that among companies who reduced or eliminated their 401(k) matching contributions, most are restoring them. The data also suggests that fewer companies than the media would have led us to believe actually reduced or eliminated those contributions and that a fair number are actually increasing those matching contributions.

A summary of the study results is here: http://www.plansponsor.com/Suspended_401(k)_Matches_are_Returning.aspx

Key finding include these:

• 14.8% of employers suspended their matching contributions in recent years
• Of those, 39.3% have restored those matching contributions
• An additional 37.8% of those were planning to restore those matching contributions by April 2011
• Roughly 10% of companies have increased their matching contributions
• More than 75% of companies that suspended their matching contributions reported lower employee participation
• More than 90% of companies in the survey have Committees that review fund performance
• More than 70% of companies (as compared to slightly more than 50% in 2009) reported making changes to their fund lineups with the predominant reason being poor performance

I'd like to take a moment to look at these findings. One in particular looks misleading to me, that being the roughly 10% that have increased their matching contributions. I don't have the data to support my statement, but I would certainly hazard a guess that of those increasing their matching contributions, most were done at the expense of some other benefit or compensation plan. There just aren't that many companies out there that are willing to increase their allocation of dollars to employee benefits, especially retirement benefits.

To me, this points out that when you are reading any survey results, unless you have all the raw data to support it, remember the cogent words of Benjamin Disraeli (Mark Twain only popularized them): "There are three kinds of lies: lies, damned lies, and statistics." 

Just as good magician can deceive the eyes, so can one who has the data ... and sometimes has their own agenda.

Does 401(k) Match Not Promote Savings?

A study conducted by James J. Choi (Yale), David Laibson (Harvard), and Brigitte C. Madrian (Harvard) suggests that employer matching contributions to 401(k) plans are not a significant motivator in increasing participation. While we don't have all of the details of the study, we do know that the researchers focused on participants at least age 59 1/2; that is, they focused on participants who, assuming their plans allowed it could make a deferral, get their match and withdraw the amount immediately, creating a taxable event, but not an excise taxable event.

I wish I had access to the actual questions and data. The arbitrage technique that the researchers note is probably not well known. I have not done the research myself, but I would hazard a guess that more benefits professionals than not would neither know of this technique, nor would think to avail themselves of it if they could.

Other factors that could easily be contributing to the data depend upon the sample population. Additional data that might be helpful in understanding this analysis and drawing a conclusion include (but are not limited to) these:

• Did the researchers eliminate people from their study who had recently taken a hardship withdrawal?
• Were any or all funds eligible for immediate in-service withdrawal in all of the plans?
• If not, would factors such as children in college, parents in long-term care, or uncertainty with regard to plan investments have been deterrents?
• Were plan communications to participants clear enough that participants would know AND UNDERSTAND that the technique the researchers describe was available?
• The researchers appear to suggest that even with education, participants would not change their behaviors. Is there reason to believe that participants have become an untrusting group, especially when it comes to employer-provided benefits?

The researchers have drawn surprising conclusions, and while their data may support it, my personal experience in consulting suggests that their conclusions either are flawed or not clearly enough reported in the following article : http://fiduciarynews.com/2010/11/new-study-explains-why-the-401k-match-fails/

In any event, it's interesting food for thought.