How to allocate asset classes in a multi-asset portfolio?

An institutional client needs to meet his pension liabilities, and suggested a multi-asset-class strategy. I'm trying to find ideas to pitch.

My experience is mostly from equity, so the way to go would be to balance some kind of covariance based risk minimization with a 1/n-style diversification, both to avoid concentration risks, as well as dependency on the model. Then possibly buy some puts to reduce downside risks, etc.

I'm really not sure however how/how much to allocate to other classes. Would you use a covariance based approach, and optimize over those? Would you pick a class allocation and stick with it? Follow marco-trends? I guess you can do any of those, but is there a 'standard' approach? Can you even use covariances to diversify bonds against stocks against commodities?

Right now my idea is to start 95% equity/5% cash, and then shift to duration matched bonds to meet the liabilities. That way, I can handle the equity with the options, potentially have some base risks, and take care of the liabilities in advance.

  • $\begingroup$ If any of the answers were helpful to you it would be great if you could accept one of them - Thank you :-) $\endgroup$ – vonjd Jun 28 '15 at 7:07
  • $\begingroup$ @vonjd Thanks, while your answer has some helpful points to start with, none of the answers actually answers the question of how to generate a multi-asset portfolio. An LDI-structure is about increasing the interest rate sensitivity to match (interest-sensitive) stochastic liabilities, and factor based investing is concerned with individual stocks, and mostly used in equity. While both concepts are great, I am looking to create a dynamic portfolio structure that balances equity and bonds by playing covariance against duration (and against 1/N diversification). $\endgroup$ – phi Jun 28 '15 at 13:47
  • 1
    $\begingroup$ Therefore I'd like to keep this open a little longer to see if someone has fresh ideas or sources, and perhaps sprinkle some bounty. Depending on how the project goes, I'll post our approach. I appreciate your answer and the link to Angs book especially, I was stuck with chincarini and I'm not too happy with it. $\endgroup$ – phi Jun 28 '15 at 13:51
  • $\begingroup$ Fair enough, good luck with your project! $\endgroup$ – vonjd Jun 28 '15 at 13:54

In this case it is important to differentiate between a liability-driven investment strategy (LDI) and a (the classical) benchmark-driven investment strategy. The first one is what you need in this case.

LDI was first established by Martin Leibowitz in 1986 ("Liability returns: A new perspective on asset allocation"). So googling that might help you already.

To dive into the matter and for many more details on all of your questions above (which are quite broad!) I would recommend the following current book:

Ang, A.: "Asset Management. A systematic approach to factor investing", Oxford University Press, 2014.


This is a huge topic in itself. It is impossible to answer without sitting down with the client. Modern Portfolio Theory would argue against a 1/n-style diversification, with the CFA calling it a behavioural finance bias.

The CFA answer would be: You want to develop an IPS with the client. What are their risk/return objectives? Also look at: time horizon, tax concerns, liquidity needs, legal constaints, and unique concerns.

For example, the client may have specific requirements, such as "no derivatives" or "no negative return in any year more than X%".

Normally for a pension plan you need to figure out what the required return is for the client to meet its pension obligations and start with that. You could propose a portfolio following Asset Liability Management principles. Allocate assets to meet pension needs, any surplus can be actively managed.

Also, beware of correlations between investment assets and the client's business.

These really just scratch the surface. Look at CFA Level 3 material, it goes into more detail.


Short Update on the specific way we have chosen:

  1. Have a risky and a safe portfolio, and shift assets over time into the safe one to protect liabilities. The safe portfolio is duration matched and holds Bunds and cash.

  2. The risky portfolio is multi asset class. Specific allocation is based on a constrained MV optimization on index level. This part is dynamic and will be reallocated quarterly. Implementation using index funds and fund-of-funds, as well as absolute return funds. Put options on the related benchmarks are used for each fund individually to minimize base risk.

  3. The allocation between risky and safe portfolio somewhat resembles a CPPI concept. First make sure that liabilities are met, then generate some performance. Shifting like this allows more capital to be employed in performance generation, and thus keep risk in check.

I'll keep this open to see other input.

  • $\begingroup$ @ phi would be great to see more colour in how performance is generated once the "liability met" trigger is active $\endgroup$ – rrg Nov 30 '16 at 11:05
  • $\begingroup$ @rrg considering the interest rate levels and markets since your comments, that remains an open-ended question. Institutional requirements are a bitch... $\endgroup$ – phi Dec 12 '16 at 22:08
  • $\begingroup$ CPPI - constant proportion portfolio insurance $\endgroup$ – rrg Dec 15 '16 at 14:11

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.