President, Lohmann International Associates
Tokyo, Japan

Arriving the first time in Japan in July 1990, the actuarial culture shock was unexpected. I was to provide primarily FAS 87 valuations to subsidiaries of American firms that then (1990) needed to comply. There were communication problems (other than those related to my lack of language skills). There were surprises (other than my first earthquake).

But after nine years of providing retirement plan actuarial services in Japan, I have one remaining surprise: that Japanese do not recognize the superior private retirement plan system that they have. I am surprised that they don't recognize the strengths and build upon them. I'm surprised that no one seems to have enumerated the weaknesses and they are not being specifically addressed, while maintaining the overall system design.

A general form of Defined Contribution (DC) plan for employees is an employer-sponsored plan where part of an employee's cash compensation is accumulated in an employer-sponsored investment account. Although known as "retirement plans," they only function as retirement plans insomuch as employee participants are not permitted to use the money they have accumulated until retirement. Often, part of the employee's pay is paid conditionally, requiring the employee to serve a specified number of years (there can be other requirements also) before the money, including investment income earned by that money, vests.

When a plan meets certain requirements, the employee will be permitted to exclude the amounts contributed to the account and the related investment income from current taxation. Employers typically get to deduct amounts paid regardless of whether the plan is "qualified" or not. Since the investment income belongs to the participants, it is not taxable income to the employer sponsor. Employees are permitted varying degrees of investment choice with their accumulations.

When an employee retires, the amount accumulated is paid. Any amounts not vested revert to the employer, perhaps creating a taxable event for the employer. Some plans limit disbursement choices but most provide a single lump sum to the employee which is taxable at the employee's top marginal rate unless the amount can be "rolled over" into another tax-favored vehicle.

Please see figure 1. Plan participants get the immediate value of the lump sum calculated at retirement , meaning that, like Japanese Defined Benefit (DB) plans, DC plans provide 100% of the lump sum benefit no matter how old the employee is at retirement and, thus, provide reasonably good portability of benefits.

DC plans shift risk from the employer to the employee when compared to DB plans. And, while the truth of the Risk/Return paradigm (footnote 1: Over the long term, higher investment risks lead to higher investment returns) holds even here, we believe that the employer is a position to properly evaluate those risks while most employees are not. Like a lottery, the risks are seldom known to the average employee until the results are in. Unlike a lottery, there are, over the long term, more winners than losers, but we feel the losers were not truly cognizant of the risks they were adopting when they made the decision.

Employer costs of DC plans are more certain than DB plans; neither uncertainty of investment performance nor the effects of future pay changes affect employer cost. While one might look to this certainty with a certain amount of affection, it should be realized that it is a certainty of higher costs over the life of the retirement plan. One way American employers reduce the impact of these costs is by shifting the costs of retirement to employees through the 401(k) DC plan that uses the matching principle.

DC plans provide benefits related to the entire career of an employee. They cannot provide benefits related to the employee's final standard of living. And, they never provide the amount considered optimum in the original retirement design. In the US, the problem has been dealt with by adding additional employer cost through a device known as a "floor plan." A floor plan is an additional DB plan that guarantees a minimum benefit in case the DC plan fails to deliver the originally designed benefits. Of course, this is at extra employer cost and administrative headache.

The fact that no plan participant ever gets the benefits designed or expected from the plan is a serious weakness of DC plans. The reason for this phenomenon is that the participant gets the market value of the investment account at retirement, nothing more, nothing less. If the "market" is doing well, he will get more than planned to the detriment of the owners of the entity. If the market is "down," he will get less than planned, to his own detriment. In the latter situation, it is important to remember that the sponsor paid more than needed for what turned out to be the employee's inadequate benefit. The probability of an exact match is literally zero!

Finally, DC plans encourage saving over consumption. In a nation such as the United States, where savings rates are notoriously low compared to Japan, DC plans have had an important influence on increasing the rate of saving among employees. In Japan, where savings rates are already recognized as being high, encouraging additional savings can only lead to reduced consumption and related economic activity.

We see the following advantages of Japan's unique approach to defined benefit (DB) plans:

(i) Good design and appropriate cost

(ii) Reasonable portability

(iii) Little participant investment risk

(iv) Little discrimination in favor of the highly paid

Benefit Design and Cost. When designing retirement plans, an actuarial consultant emphasizes certain principles to the plan sponsor. Two of them are that the plan should provide the right benefits to the right people at the right time and that the cost of those benefits should be appropriate.

Japan's basic private system actually automatically provides for both of the "basic" design principles. Plan design is typically quite simple; a lump sum paid at retirement based on final pay and service. "Insured" plans (both those provided by Insurance companies and those provided by certain Trust Banks) also provide annuities, usually nominally actuarially equivalent to the lump sum, but the Japanese tax system so favors lump sum retirement benefits that retiring employees usually take the lump sum, even when there is a chance for arbitrage and/or anti-selection favoring the selection of the annuity. "Book Reserve System" plans can provide only lump sums, but, like the lump sum delivered by the externally funded plan, it is tax-favored. Both funding methods are designed to deliver the "right" benefit at retirement, since it is the one intended by the plan; it is neither larger nor smaller.

The typical Japanese retirement plan is a true final pay plan. DC plans are based on career earnings. This has little impact on the average worker. It can have an enormous impact on the worker who experiences higher than average increases in earnings over the last few working years. A final average plan will reflect this change, a career average plan won't.

It has often been said that a retirement plan should provide a level of benefits that will permit a career employee to live in retirement at about the same standard of living enjoyed while working just prior to retirement. Final average plans are capable of achieving that goal for all career employees. Career average plans achieve the goal only for average employees. Career average plans actually provide too much for the employee that turns out to be a less than average performer in his last years. They provide too little for the employee that turns in star performance in the last years.

There are two serious differences between DB and DC plans that affect sponsor costs and participant benefits: (i) individuals can tolerate less risk and have shorter investment horizons than sponsors and (ii) no plan participant ever gets the exact amount anticipated from a DC plan. These two differences are the source of my most immediate concern about DC being introduced.

It is widely accepted that higher risks lead to greater returns in an investment portfolio. Longer investment horizons also lead to greater returns due to the ability of a portfolio to withstand longer periods of poor performance, and, thus being able to accept higher risks. Assuming that a DC contribution plan is designed to produce similar (theoretically) results to the DB plan it is replacing, costs will be dramatically higher. In the U.S., we have often seen 200 basis points used as an approximation of the difference. While we have not studied the situation in Japan, we suspect that the difference is greater and that the impact, given the historically low "safe" returns in Japan, magnified.

Sponsors will have to pay this difference, assuming that benefits are designed to be unchanged from current levels.

One might question the assertion that the price is "right." The investment climate in Japan has been so bad for the past several years that retirement plan costs have been escalating. But, unlike the US where regulatory burden is seen as a major part of the high cost of Defined Benefit (DB) plans, everyone in Japan is subject to the same economic environment. While much of the rest of Japan is over-regulated (hence the investment climate), Japanese retirement plans are largely left alone.

Therefore, a plan sponsor is paying only for the benefits promised plus a small incremental charge for administration (sometimes not even incremental with a Book Reserve plan) and another for asset taxes (not applicable at all to Book Reserve plans). The available investment earnings serve to lower out-of-pocket. While external retirement funds seldom attract the same rate of return as funds invested in the business, it is often felt that the additional cost represented by lost earnings or increased borrowing costs is appropriate for the additional perceived security of an external plan. Of course, the hidden investment rate of return for a Book Reserve plan is the company's ROI (Return On Investment).

These funding methods (Insured and Book Reserve System plans) combined with the typical design, keep investment transaction costs relatively low (ignoring the high fees charged for trading by brokers). Companies, especially the insurers and trust banks, have kept two important principles in investing the funds in their care; (i) "Dollar Cost Average," as opposed to "Market-timing," purchases and (ii) "Buy and Hold." Anyone who has been "buying Japan" for the past six years will find they have bought a bargain after the turn of the century. Another investment principle comes to mind; "Buy low and sell high."

Another area that reduces administrative costs in Japan is the usual payment of benefits in lump sum form. The North American practice of keeping track of small amounts for former employees and paying annuities to many retirees leads to higher administrative costs. Administrative costs for Japanese DB plans are nearly identical to those for a DC plan, assuming that the cost of individual allocation and record keeping demanded of DC plans is roughly equal to the costs of valuation for a DB plan.

We feel that, after removing the effect of the investment climate in Japan, the cost of Japanese retirement plans is appropriate.

What is a "portable benefit?"; In a perfect world, a new employer would permit a new employee to transfer all of his previously accumulated retirement assets into the new employer's pension system. At retirement, the employee would receive a benefit based upon total service from the first employer with no diminution for having worked elsewhere. Social plans achieve this result to a degree by insulating benefits from the source of employment. But they often fail to provide an appropriate benefit since the retirement benefit is typically based on the employee's career earnings rather than his final earnings (a "career average plan").

Since many private plans are "final average" plans, the new employer is going to experience additional costs representing the benefits earned for pre-employer service but not purchased by the transferred funds. Few "new" employers would agree to such an approach since they do not feel obligated for benefits based on their plan but for service with a prior employer.

Further, the "old" employer, the one being left, shouldn't like to have to provide additional assets to a departing employee just because the new employer's plan is more generous or because the employee has managed to get a significant increase in earnings.

So, realistic portability is represented by the employee receiving at retirement an amount of money that could buy the anticipated future benefit from the former (the"old") employer when retiring at the age limit of the new employer. Japan's current system as typically applied achieves this result. Maybe better.

Typically, the Japanese employee's lump sum benefit is calculated without regard to age at actual retirement. There are no "early retirement penalties." The entire value is available to invest or, if there were a market in Japan, buy an individual annuity to start immediately or to start at a later date. Japanese plans are mathematically "front-loaded" based on how the "deferred annuity" is accrued over the employee's service lifetime.

Unlike Japan, North American DB plans have failed miserably at providing a reasonable degree of portability. The failure arises from two aspects of calculating the terminating employee's benefit; the amount is calculated as the present value of an annuity whose first payment is deferred and the present value typically includes a discount for survivorship. These result in the increasing concave curve of termination values associated with North American plans as shown by the lower line in figure 1.

Figure 1 illustrates the portion of the lump sum value that a retiring participant can get at each possible age of retirement compared to 100% of the lump sum value available at retirement at the age limit with the same amount of service. There are two lines. The higher one shows the proportion that a Japanese employee would get from a typical plan - always 100%. The lower one shows what an American employee might get. We have arbitrarily set both plans to a normal retirement age of 65 to ease the comparison, but, as can easily be seen, it doesn't make any difference to the Japanese employee since 100% is always available.

While Japanese plans were not specifically designed to be portable, in fact, they are. It is a consequence of benefit designs that ignored early severance due to the Japanese practice of "lifetime employment." Since few employees ever retired before the age limit, it was an unnecessary additional administrative burden to provide different benefits based upon the age at retirement.

As is shown, the Japanese employee gets a significant multiple of what his American counterpart does, magnified in the case of retirement at an early age. Taxes multiply the effect, since, if the American does not deposit his pittance into a qualified rollover instrument, he is generally taxed at his top marginal tax rate on the entire amount in the year of retirement. It is virtually impossible for the American employee to successfully replace lost benefits with the amount he is provided at early retirement.

Japanese plans further enhance the real intent of portability by distinguishing between voluntary and involuntary "retirement." The voluntarily retiring employee has a chance to consider the benefits of his new opportunities. The involuntary one often does not. America treats all departing employees the same way, reducing benefits even when the company is responsible for the departure and the employee may be faced with a significant period of unemployment. These periods put future retirement benefits at risk, since the former employee may need the money for current living costs.

Investment Risk. Japanese employees are able to rely on getting the benefits they have been promised by the plan. They are able to plan. They can comfortably make decisions about retirement knowing that the amount they calculated will be available to them when they ask for it. Voluntary retirement is a big decision. Even Americans examine the pros and cons well in advance of making a decision to retire. A defined benefit assures the participant that the amount of money expected is the amount of money paid.

Like any retirement plan delivering lump sum benefits at retirement, the participant has an investment risk after receiving the benefit. This characteristic holds for all plans. The lack of an individual annuity market in Japan tends to increase risks for Japanese retirees over their American counterparts, but this holds true for all disbursements.

Discrimination in Favor of the Highly Paid. Both Japan and America embody the concept that socially favored (tax, security, etc.) retirement benefits should accrue equally to all employees. America has achieved this through a system of specific calculations that are expensive to calculate and administer. Japan has achieved similar results by prohibiting employed members of the board (officers) from having externally secured benefits and by requiring benefits to be the same for everyone. While there are weaknesses in the Japanese method that have been exploited, one of the major reasons for employers moving to DC plans in the US is a perception of burden brought on by the rules. Japan does not present such a burden.

We believe that Japan needs to make a fine-tuning adjustment to company insolvency laws; remaining assets should be allocated first to unpaid salaries and allowances of rank & file employees, then to accrued involuntary benefits promised by the retirement plan regardless of funding method. We believe that officer accrued retirement benefits should be the last item just before distribution of any remaining net worth to the owners.

WEAKNESSES OF THE JAPANESE PRIVATE RETIREMENT SYSTEM.

We have observed some weaknesses in the Japanese private retirement system. Among them:

(i) Inadequate access to employer assets in insolvency

(ii) Abuse of the definition of "base pay" in order to enhance benefits for the highly paid

(iii) Reliance on a "product" approach to external funding (insured plans)

DC plans do not mitigate these problems. Since DC plans typically try to look at total compensation, rather than monthly base pay, they increase problems of discriminating in favor of the highly paid.

Insolvency. As in America, there is not enough protection of regular employees' retirement benefits in the case of sponsor insolvency. Confining our discussion to private retirement plans, we find that the difference between the total of involuntary benefits and the amount funded or recognized is not a legal liability of the employer.

In an insured plan (TQPP) this means that the employees run a risk of not being paid their benefits before other creditors when their company becomes insolvent. Fortunately, developing disclosure rules will at least inform investors and employees of the actuarial shortfall. Unfortunately, these disclosure rules are based on an "ongoing entity" assumption. The shortfall between total involuntary benefits and the TQPP liquidation (a number similar to market value) value is not disclosed. More importantly, it is not a true liability of the company; it does not have to be paid in insolvency. In fact, we do not believe the employees even have standing as creditors in the insolvency.

Book Reserve system plans present a similar problem. While the company must disclose the typically 100% (Footnote 2: This percentage can be as low as 40%, but, in our experience, companies typically push it up to 100% fairly rapidly. There are rules to discourage reducing the percentage after it has been raised.) of the voluntary termination amount as a liability, the difference between that amount and the involuntary amount is not disclosed. Employees do have some rights to remaining assets in an insolvency, but they are usually not sufficient to cover all amounts owed them under the retirement plan.

We believe that the FAS 87 concept of the ABO should be a minimum liability recognized on the balance sheet with the total of involuntary benefits disclosed in a footnote. Solvency should be based on equity calculated after recognition of the minimum liability. Essentially, the ABO represents the ongoing present value of amounts currently owed employees based on their continued employment and the survival of the entity. It uses current service and current pay. This amount should be recognized regardless of the funding method (insured or book reserve). We believe it fairly recognizes costs among generations of owners and employees of an ongoing entity.

We believe that all benefits promised to rank and file employees should have a first right against remaining assets in a sponsor insolvency. Essentially, at insolvency, the total of involuntary benefits owed rank and file employees would become a legal liability of the employer having first claim against remaining assets. The ABO would be replaced by the total of the involuntary benefits on the balance sheet as a minimum liability. These amounts would be paid before any other creditors were apportioned assets.

Base Pay. We have seen an abuse of the definition of "base pay" for purposes of retirement benefits. Retirement benefits should be based on the same "base" pay that unemployment, health and other social taxes are based on. We do not believe that bonuses and other allowances should be permitted to be included in the formula. To have a definition of base pay that differs from that amount used to calculate social benefits sidesteps the Japanese requirements for non-discrimination and should be prohibited.

Insured Plans. We would like to see external retirement funds (those currently referred to as "insured" plans or TQPP-Tax Qualified Pension Plans) established more on a trusteed basis than the current "product" basis. The current basis severely distorts cash flows and reported income between generations of employees and owners.

WHY AMERICA IS IN LOVE WITH DC PLANS.

The quick answer, of course, is "GE, MSFT, APPL, INTC, NASDAQ" and, more recently, "Amazon.com." While Japan has been suffering a crushing recession or worse, the U.S. has seen the greatest stock run-up in history. People who had only a few shares of stock ten years ago are now worried about paying too much tax. The mutual fund industry may present more funds than there are stocks! The US government is running a surplus...under a democrat president! Everyone in America is getting rich on DC plans. Like a lottery, a certain amount of envy occurs as we watch the winners walk away with their money.

Also, typical American participants fail to notice that most of the money contributed to the "company" 401(k) plan was taken directly out of their own paychecks. They are incapable of recognizing how little of the final benefit is sponsor-related. While we support tax-favored vehicles for individuals to save for retirement, we don't think they should be related to employment and invested in funds supported by employers. As the Teamsters discovered in the US, there is just too much temptation.

They also believe they are getting a tax break. Contributions are made before tax, while benefits are fully taxable at the top marginal rate at the time withdrawn. There are significant penalties for withdrawals failing to meet accepted criteria. Typically, participants compare taxes ignoring the significantly reduced rate for long-term capital gains in the U.S. The penalties and limitations attached to "deferred tax" vehicles may not be worth the actual tax benefit popularly attached to them.

Meanwhile, U.S. government regulation of DB plans has been crushing. Flyspeck rules requiring significant administrative expenses to demonstrate compliance as well as the requirement for several different valuations to establish deductible contributions are maddening and probably unnecessary.

Add a lack of portability of benefits from DB plans in the US, actual and perceived, to the above. This is a problem Japan just doesn't have; Japanese DB plans provide significant tax-favored termination (retirement before the age limit) benefits in cash and Japanese employees still enjoy a much more secure employment relationship than Americans do, resulting in much higher benefits at retirement.

Finally, there is that uniquely American demand for "freedom of choice!" Americans want to feel in control of their "fortunes." Defined contribution plans help achieve this while preventing employee access to their own money. At a cost, of course. A cost that employees believe is borne by the employer, but probably isn't.

The current Bull market probably has a long way to go yet, but there will be a significant change in which kind of plan is favored when returns drop back to historical levels, especially if they correct for the astronomical returns of the past 7 years. Buying high and selling low has never been a successful market strategy, despite its popularity among non-professionals.

FUTURE GAINS?

There is a final question that needs to be answered. Who deserves the coming investment gains in Japan as Japan recovers? Should it be the owners of entities that have born the pain of investment losses of the past years, or the current employees who were protected during the entire time by DB plans? Would the alternative of making market adjustments be acceptable to employees or their government? We think not.

We believe that DC plans can be an important contributor to saving for retirement. We do not believe they are appropriate as the basic retirement plan for rank and file employees. In a perfect world, we would like to see individuals permitted to have private, tax-deferred investments as a retirement funding choice. We would like to see employers with adequate DB plans permitted to create an extra layer of benefits in the form of individual DC plans.

SOME COMMON MISCONCEPTIONS ABOUT DC PLANS

Supporters of DC plans typically point to at least three areas where employers and participants will benefit from using DC plans in place of DB plans:

(i) They believe that a higher investment return is available for the funds invested for DC plans

(ii) They believe that employer costs will be lower

(iii) They believe that benefits are more secure

In addition to these benefits for employers and participants, supporters add a belief that the Japanese economy as a whole will benefit.

Investment return. This is probably the most common misconception. People see only the anecdotal evidence showing the "winners" with their gigantic accumulations and calculated returns significantly exceeding available returns. They do not see the retirees who are forced to liquidate during contractions.

Established investment theory confirms that two characteristics of an investment portfolio lead to higher returns; longer periods of investment (the investment "horizon") and higher risk (the "efficient frontier").

We know that invested funds in DC portfolios have shorter horizons and generally lower risk. The long-term expected return of DC plan funds is significantly lower than the long-term expected return of DB plan portfolios.

Illustrations of future results which use the same rate of return or a higher rate of return for DC plan invested funds are misleading and incorrect.

Lower employer cost. It is important to remember that the employer cost of a retirement plan is the cost of employer-provided benefits plus the cost of the administration of the plan less the benefit of investment return.

Administratively, DC plans tend to have a larger number of smaller transactions than DB plans providing the same benefits. These are expensive. In addition, there is a requirement for individual account tracking and, as discovered in the U.S., a need for the ability to update account information at least daily. This additional administrative expense is not applicable to DB plans.

We have already demonstrated that the expected return of DC plans is significantly lower than that of DB plans.

For the same level of employer-provided benefits, DC plans cost more than DB plans.

So why is there a perception that they cost less? In the US we call this phenomenon "cost-shifting." The most popular form of DC plan is the 401(k). An analysis of 401(k) plans will reveal that employer benefits depend upon employee participation. Employees who refuse to participate are denied employer-provided benefits. Those employees who do participate are often expected to "buy" at least twice as much benefit as the employer buys for them ("50% match"). Funds are commingled leading to employees failing to notice that employer-provided benefits are quite small!

Interestingly enough, even those small benefits would be less expensive is they were provided by a DB plan.

Higher benefit security. The individual account nature of DC plans gives a perception of benefit security.

It is important to analyze, however, why DB plans currently appear to be underfunded. Investment returns have been horrible. DC plans are never underfunded because the individual participants bare the investment risk; they are only entitled to the balance of their accounts, regardless of how investments perform. If Japan had permitted DC plans for the past 20 years, most career employee balances would be in trouble, but, of course, there would be no underfunding! As mentioned, DC plan funds tend to be invested in lower risk vehicles than DB plans and thus the principle investment tends to be more secure over the short term, at a cost of reduced long-term investment return; the more secure the principle of an investment, the lower the long-term expected rate of investment return.

In the U.S., many 401(k) plans have significant portions of invested funds in the employer's own stock; the DC plan is an owner of the enterprise. At insolvency of the enterprise, the owners get the residual assets after all other creditors have been paid - usually a small amount approaching zero. To the extent that such an approach is permitted, employees will be subject to a higher risk of insolvency than they currently are. Unfortunately, this risk is characterized as an investment risk and remains hidden from appropriate scrutiny.

Because the current investment return problems in Japan are structural, DC plans will be subject to the same problems as DB plans currently are, except for the fact that most believe that Japan is now in a period of recovery. This will shift investment gains from the companies who have borne the risk for the past decade to the employees who did not. We feel this is a negative result for investors.

We believe that the single most important change to improve benefit security is the requirement to provide rank and file retirement benefits before any other creditor is paid in an insolvency, regardless of the funding vehicle of the plan. After all, the promised benefits are based upon past employment. They should have a higher status than current pay based on current employment during the final days. This focuses attention of investors and management in an appropriate way.

Benefit the economy. In the U.S., where savings rates have been notoriously low and the expanding economy needs capital, DC plans probably help the economy by permitting more capital formation.

Japan already has a high rate of saving. The rate of saving is so high, in fact, that negative interest rates (essentially one would be further penalized for keeping money in a Japanese bank) have been recommended. DC plans will tend to increase the rate of saving and, thus, will likely have a negative impact on consumption and will tend to prolong the current economic malaise.

This also has an impact on the perceived rate of return on DC plan assets. Assets invested in debt instruments will experience a decreasing rate of return due to oversupply. Assets invested in ownership assets (primarily publicly traded shares) will appear to increase in value rapidly as the oversupply of capital chases an undersupply of shares. We do not believe that this will represent true increase in the value of Japanese assets.

CONCLUSIONS

While we believe that there is a place for DC plans in a well-designed plan of employee benefits and we particularly support the concept of individual DC plans, we do not believe that they are appropriate as a first or basic retirement benefit for rank and file employees.

Summarizing, we find that Japan has a working private retirement system with some weaknesses. Are these weaknesses those that will be addressed by the introduction of DC plans similar to 401(k) plans found in the United States? We do not believe they are.

Will Japan's economy be improved by the introduction of these new retirement vehicles? Again, we believe that the investment broker community will benefit enormously by the increase in transaction fees that they will collect. We do not see any tangible increase in the real output of goods and services from this change. In fact, we see the reduction in consumption related to the increase in saving as direct hit against the economy.

Leslie John Lohmann, FSA, FCIA, FNZSA, EA, FCA, MAAA

Index to
|Lohmann International Associates |