STATEMENT OF JOSEPH M. BELTH
ON JOHN HANCOCK'S PLAN OF REORGANIZATION
(November 9, 1999)
I am Joseph M. Belth, professor emeritus of insurance in the Kelley School of Business at Indiana University, and editor of The Insurance Forum. For further information, see the biographical sketch on page 11. I am submitting this statement in the public interest, and I am not being compensated for it. The views expressed are my own. Although I am not able to attend the hearing on November 17 to present the statement, I will answer questions by telephone at an agreed-upon time. I ask that the statement be included in the record of the hearing.
Summary
This statement relates to the Plan of Reorganization under which John Hancock Mutual Life Insurance Company ("John Hancock") proposes to convert itself into a shareholder-owned company. The purposes of the statement are three: to commend John Hancock for its decision to demutualize rather than form a mutual holding company, to express regret about John Hancock's decision to exclude certain policyholders from sharing in the distribution, and to express concern about the formula used by John Hancock to allocate variable shares among eligible policyholders.
A Commendation
I commend John Hancock for its decision to demutualize rather than form a mutual holding company. John Hancock said it this way on page 13 in Part 1 of the Policyholder Information Statement:
The Board also seriously considered a conversion under a mutual holding company option. Under that option, a mutual insurer could reorganize into a stock insurance company which would become a subsidiary of a newly-created mutual holding company. The Board decided that the mutual holding company option was, at best, only a partial solution to the capital access issue. In addition, under that option the policyholders' membership rights would be transferred into the mutual holding company, without compensation, rather than exchanged for compensation in the form of cash, stock or policy credits. The Board of Directors ultimately decided that demutualization was the better alternative from both the policyholders' and the Company's perspectives.
I would say it more strongly. The effect of a mutual holding company reorganization is to terminate or dilute the ownership interests of the mutual insurer's policyholders without compensation. Also, the mutual holding company concept is fundamentally flawed. If the implications were disclosed to and understood by policyholders, most would vote against the reorganization. On the other hand, if strong safeguards were added to protect the policyholders' ownership interests, prospective shareholders would be reluctant to invest in the reorganized enterprise.
The Eligibility Issue
As indicated on page 14 of Part 1 of the Policyholder Information Statement, only John Hancock policyholders are eligible for compensation. Thus policyholders of John Hancock Variable Life Insurance Company ("JHVL") are not eligible.
That is regrettable. JHVL policyholders should share in the distribution, for at least three reasons. First, they probably bought their policies from John Hancock agents. Second, the agents probably did not disclose the absence of rights that would have existed if the policies had been issued by John Hancock. Third, JHVL policyholders, through their premium payments, probably contributed--albeit indirectly--to the surplus of John Hancock. Thus John Hancock has at least a moral obligation to treat JHVL policyholders as though they are John Hancock policyholders.
I am aware of the Massachusetts demutualization statute (Section 19E), which is reproduced beginning on page 71 of Part 1 of the Policyholder Information Statement, and I realize that the language of the statute appears to exclude JHVL policyholders from eligibility. Nonetheless, John Hancock, had it desired to do so, could have found a way to make JHVL policyholders eligible. As will be discussed later, John Hancock found a way to use the "historic plus prospective" method of distribution in spite of the language in the statute apparently requiring the "historic only" method.
The Allocation Formula
The formula used by John Hancock to allocate the variable component of the distribution among eligible policyholders is a matter of concern, for two reasons. First, the formula uses the historic plus prospective method, which considers past contributions to surplus and anticipated future contributions to surplus. Evaluating anticipated future contributions to surplus requires the use of assumptions extending far into the future. The historic plus prospective method is in contrast to the historic only method, which considers only past contributions to surplus. The historic plus prospective method favors recent buyers at the expense of long-time policyholders.
Second, John Hancock's allocation formula produces some strange results. For example, buyers of some large policies receive zero variable shares in the distribution, while buyers of some small policies receive significant numbers of variable shares.
The Historic Plus Prospective Method
The Massachusetts demutualization statute appears to prohibit the use of the historic plus prospective method, and appears to require the use of the historic only method. The statute says:
In exchange for all membership interests in the company, such plan shall give each eligible policyholder appropriate consideration. Said consideration shall be determinable under a fair and reasonable formula approved by the commissioner, and shall be based upon the insurer's entire surplus as shown by the insurer's financial statement most recently filed with the commissioner pursuant to section twenty-five, including all voluntary reserves but excluding contingently repayable funds and outstanding guaranty capital shares at the redemption value thereof, and without taking into account the value of nonadmitted assets or insurance business in force.
On page 11 of his statement dated October 30, 1999, Godfrey Perrott of Milliman & Robertson ("Milliman") quoted the above statutory language. He then said the historic plus prospective method was "chosen" in the John Hancock case. He did not discuss the apparent inconsistency between the historic plus prospective method and the statutory language. He said that the historic plus prospective method was used in every U. S. demutualization since 1990, including the case of Massachusetts-domiciled State Mutual (now First Allmerica Financial), and that the method is consistent with the 1987 report of the Society of Actuaries.
The Attorneys' Turnabout
Attorneys working for State Mutual in its demutualization a few years ago initially interpreted the statutory language to mean that the historic only method had to be used in the allocation process. In a memorandum dated May 4, 1994, the law firms of Debevoise & Plimpton and Ropes & Gray said:
Section 19E states that policyholder consideration is to be based upon an insurer's surplus as shown in its most recent annual statutory statement and that the value of insurance business in force is not to be taken into account in the calculation of consideration. Therefore, since current surplus reflects historical contributions and the value of in force business reflects projected future contributions to surplus, Section 19E requires that the variable component be calculated based solely upon historical contributions to surplus. [Emphasis in original.]
Accordingly, Section 19E not only permits but requires an insurer to base consideration solely upon historical contributions to surplus.
The attorneys changed their minds a year later, and said the historic plus prospective method may be used. In a memorandum dated April 14, 1995, the attorneys used the following tortuous reasoning to support their turnabout:
In light of the statutory ambiguities discussed above, and, bearing in mind particularly the requirement that the allocation of consideration be based upon "a fair and reasonable formula" approved by the Commissioner and the fact that the actuarial literature suggests that the historical plus prospective allocation method is fair, we believe that the historical plus prospective method of allocation is permitted by the statute under such circumstances.
The Actuaries' Turnabout
Milliman worked for State Mutual in its demutualization. Early in the process, Milliman said the historic only method should be used. In a memorandum dated May 4, 1994, Milliman said:
We believe the Massachusetts Law requires a historic only allocation base and have taken this approach in the bulk of our work.
The actuary's primary concern is fairness to the policyholders. We are comfortable with State Mutual's use of the historic only basis, rather than the historic plus prospective basis. We believe the resulting allocation is fair on its own terms, independent of the provisions of the statute.
We believe the historic plus prospective method in the State Mutual situation may allocate too much to Group Annuities . . .
Therefore, both a sense of the orphan surplus and a sense about the credibility of different future gains causes us to believe the historic only basis produces a fair allocation in the State Mutual case.
We distinguish this [State Mutual] situation from the normal situation that forms the basis for the recommendations in the [Society of Actuaries 1987] Report.
Milliman changed its mind a year later, and said the historic plus prospective method may be used. In a memorandum dated May 15, 1995, Milliman said:
There are no special facts applicable to State Mutual that indicate that the historic plus prospective method should not be used. Accordingly we conclude that the historic plus prospective method is fair and reasonable in the case of State Mutual.
The [Society of Actuaries 1987] Report recognizes that the historic only method must be used when required by the applicable statute and may be appropriate in other special situations. (Neither of these is applicable in the case of State Mutual, as explained below.)
We have been advised by State Mutual's outside counsel that the Massachusetts Law (l9E) permits use of the historic plus prospective allocation method for determining policyholder consideration.
We conclude that the historic plus prospective allocation method is fair and reasonable in the case of State Mutual.
The Strange Results
John Hancock's allocation formula produced some strange results. Here are a few examples, with the data altered to prevent identification of the policyholders without changing the relationships: zero variable shares for a $10 million second-to-die policy issued in 1992, 92 variable shares for a $100,000 term policy issued in 1994, 62 variable shares for a $250,000 whole life policy issued in 1992, and 1,352 variable shares for a $250,000 whole life policy issued in 1992.
John Hancock's dividend formula is supposed to be fair. To achieve fairness, the formula is supposed to be based on the contribution principle. That means a policyholder is supposed to share in the divisible surplus in proportion to the policy's contribution to surplus.
John Hancock's allocation formula for determining the number of variable shares a policyholder receives in the demutualization is supposed to be fair. To achieve fairness, the formula is supposed to be based on the contribution principle. That means a policyholder is supposed to share in the distribution in proportion to the policy's contribution to surplus.
The strange results of John Hancock's allocation formula suggest that something is wrong, and that either the dividend formula or the allocation formula is unfair. If the allocation formula is fair, the dividend formula must be unfair. On the other hand, if the dividend formula is fair, the allocation formula must be unfair.
For example, consider a John Hancock policyholder with a large second-to-die policy. He receives zero variable shares because the policy's contribution to surplus is zero or negative. The dividend formula, the results of which are reflected in the calculation showing that the policy makes a zero or negative contribution to surplus, must be unfairly tilted in favor of such second-to-die policies and against other policies.
Similarly, consider a John Hancock policyholder with a small term policy. He receives a significant number of variable shares because the policy's contribution to surplus is substantial. The dividend formula must be unfairly tilted against such term policies and in favor of other policies.
The problem may be the dividend formula. Perhaps it is based only in part on the contribution principle, and largely on competitive considerations. Perhaps John Hancock issued some policies that are not self-supporting, and allowed those policies to be subsidized by other policies. The public may never know precisely what happened, because the company uses a confidential ("trust us, it's fair") dividend formula and a confidential ("trust us, it's fair") allocation formula.
The Canadian Approach
The allocation formulas used by the Canadian mutual insurers that are demutualizing differ significantly from the formula used by John Hancock. The formulas used by the Canadian companies do not produce strange results, do not use proprietary information and therefore can be disclosed to policyholders, are simple and therefore can be understood by average policyholders, are perceived by policyholders as fair, and have been certified by actuaries as fair.
For example, consider The Mutual Life Assurance Company of Canada (Waterloo, Ontario), whose new name is Clarica Life Insurance Company. The information package describing the demutualization plan was dated March 15, 1999.
In his report, Mutual of Canada's appointed actuary identified the "guiding principles" used in developing the allocation formula. He said there were three:
The dilution of voting rights affecting all voting policyholders will be recognized explicitly in the allocation process.
The ownership interests of the eligible participating policyholders will be recognized in a manner that recognizes the Company's history and philosophies of conducting business. Factors that reflect both the length of the relationship of a policyholder with the Company as well as the size of a policyholder's financial stake in the Company will be used.
The allocation method must also be simple, practical and reasonable and understandable to policyholders.
Mutual of Canada's allocation formula was disclosed in the company's information package sent to policyholders. Each eligible owner of a participating or nonparticipating policy will receive 22 shares reflecting the "transfer of voting control to shareholders." That is the fixed component of the distribution.
An eligible owner of a participating policy will also receive a number of shares reflecting that he or she "helped to build Mutual Life by contributing to the company's value and accepting some of the risk associated with being an owner of the company." That is the variable component of the distribution.
Consider the hypothetical case of Mr. Jones. He owns two $100,000 individual cash-value participating policies in Mutual of Canada. The first has been in force for 15 years and has a cash value of $20,000. The second has been in force for 10 years and has a cash value of $12,000. Each eligible owner of that type of policy will receive two variable shares for each year of his or her relationship with the company. Because Mr. Jones has had at least one policy in the company for 15 years, he will receive 30 variable shares reflecting his relationship with the company (15 years multiplied by 2).
Each eligible owner of that type of policy will also receive a number of variable shares equal to the "coverage amount" (as defined in the package) multiplied by .0004864896 and the cash value multiplied by .01236, with the result rounded upward to the next whole share. Mr. Jones will receive 493 variable shares reflecting the coverage amounts and cash values of his policies. Here is the calculation:
Coverage amount: 200,000 X .0004864896 = 97.3
Cash value: 32,000 X .01236 = 395.5
97.3 + 395.5 = 492.8 = 493
In summary, Mr. Jones will receive a total of 545 shares. He will receive 22 fixed shares reflecting the change in voting control, 30 variable shares reflecting the length of his relationship with the company, and 493 variable shares reflecting his coverage amounts and cash values.
Actuaries have certified that the Canadian approach is fair. In his report, Mutual of Canada's appointed actuary expressed the opinion that "the benefits and the method used to determine the allocation of the benefits" are "fair and equitable to the eligible policyholders." Also, an outside consulting actuary, in a letter to Mutual of Canada's board of directors, expressed the opinion that "The benefits and method . . . to be used to apportion the value of the Company are fair and equitable to those policyholders eligible to receive these benefits."
John Hancock could have used an allocation formula such as that used by Mutual of Canada, or some other formula with similar characteristics. I do not know why John Hancock used a complex, confidential formula that causes consternation among policyholders and agents, when it could have used a simple, disclosed formula that would have been perceived as fair.
JOSEPH M. BELTH
Joseph M. Belth, Ph.D., is professor emeritus of insurance in the Kelley School of Business at Indiana University (Bloomington), editor of The Insurance Forum, and author of Life Insurance: A Consumer's Handbook. He is the author of several other books and numerous journal articles.
For one of his books Belth received the 1966 Elizur Wright award for "outstanding original contribution to the literature of insurance." For journal articles he received awards in 1962, 1964, 1965, 1967, 1971, and 1979. He received the 1987 Financial Security Nest Egg award from the Life Communicators Association for "professional activities in communicating to the general public." For The Insurance Forum he received the 1990 George Polk award in the "special publications" category. He received a 1999 Huebner Gold Medal from The American College "in recognition of distinguished service to education and professionalism."
Belth has received degrees from Auburn (NY) Community College (now Cayuga Community College), Syracuse University, and the University of Pennsylvania. He was a life insurance agent in Syracuse for five years in the 1950s. He is a past president of the American Risk and Insurance Association, an organization of insurance professors and others interested in insurance education. He has been a member of the Indiana University faculty since 1962.
Belth was the subject of a page-one profile in The Wall Street Journal on January 5, 1978. He was also profiled in Barron's on June 8, 1981, and in The New York Times on April 17, 1990. He is listed in Who's Who in America.