• Provides appropriate integration of taxable and tax deferred investments, including taxes on distributions.
• Provides a "life balance" sheet approach to revising the investor’s risk tolerance through time to maximize the median of expected wealth accumulation.
• Includes a nearly exact solution to multi-period optimization without the need for complex dynamic programming.
• And includes an "expert: system approach" to consistent resolution of complex or conflicting preferences that are outside the simple expectations of return and risk.
  

The same consideration needs to be made when the client owns significant amounts of investment real estate.

Then there is the discussion of Use Assets versus Investment Assets. If an asset will be held and enjoyed until the estate is settled, I don't include it as an asset for allocation with retirement in mind. On the other hand, the client may plan on partially liquidating a Use Asset. Then, the present value of the liquidated portion should probably be included in the asset allocation as an illiquid asset.

For example, some clients plan on selling the family homestead in a high value area, and buy a smaller home in a more rustic locale. Both the reduction in size and change of locale may free up home equity in the future. Or, they may decide to keep the home intact but harvest some of the equity through a reverse mortgage or the like.

However, I believe many clients tend to "roll" their homes. If they want to sell the "8-bedroom white elephant" they often roll all of the equity into a non-leveraged "luxury condominium with concierge service." In essence, they intend to use and enjoy their home equity as home equity until they leave the site feet first.

Whether it's a home or business that will be held to the end, serious intra-family planning needs to take place before all parties have their senses and options. However, this is discussion is not in the scope of the Forum subject.

I am interested in your comment "most investments are taxable". Do you mean that most holdings managed by professional advisers are in regular taxable accounts? There must be some published data but I can't think of any off-hand.

In my former practice this was not the case, and so that's why I asked. A "typical" client had thieir first million in the RP, the scond million in real estate (that they tried to escape taxation) and beyond that used trust accounts, holding companies, deferred compensation and other and tax-free benefit plans to handle even. I had few, if any, clients who had exhausted tax advantaged options and relied primarily fully taxable investments.

I just wondered what was the tax status of  "average" holdings?

Merrill Lynch puts out annual surveys of the number of people in the world in various net "investible" worth categories.   If you're talking about people with a $1M investible, you're absolutely right that retirement plans and houses are a big piece.  

Our clientele is mostly private banks and family offices where their average "client" is a lot larger.   A hundred million dollar portfolio is considered modest.   At that size, retirement plans and a residence don't count for much of the total, and trust structures will be taxed differently but still taxed.

The expert system technology I am referring to is something called Analytic Hierarchy Process. Its a mathematical technique used in large scale industrial decision making.  There have been a dozen or so academic research studies on how to use it for investor decisions such as asset allocation and fund selection.   Here is a link to a paper from 2000 by two colleagues of mine that I think does a good job explaining the concepts.

http://www.northinfo.com/documents/62.pdf

As far as I know, our firm is the only one that actually has AHP software commercially available that has been tailored to investment decisisons.

With regard to the multiperiod problem, I introduced a pretty simple short cut that I think gets us prettty far in practice.  The basic idea is that in a world with no transaction costs, the "single period" Markowitz assumption is wrong but not harmful because we can costlessly revise our portfolio to new beliefs or preferences as often as we want.

When costs are substantial, the single period breaks down because the expected improvements in investor utility when I rebalance a portfolio are bassed on population statistics (because conditions don't change in a single period world), whereas in the real world we will realize the utility increase based on the sample statistics experienced only during the finite period until conditions change.  Our way of addressing this is to assign a "probability of realization, conditional on time horizon" to any expected increase in investor utility.   The tradeoff between utility (return adjusted for risk) and trading costs is therefore done in a much more economically rational way. 

Here is a link to a presentation I gave at our seminars in Asia last year that summarizes most of this:

http://www.northinfo.com/documents/229.pdf