Approaches to Liability-Relative Asset Allocation: Integrated Asset/Liability Approach

If you're the head asset manager of the world's largest bank, you've got an enormous responsibility. Not only are the bank's assets measured in the trillions, but each asset is directly tied to a liability, so managing the asset portfolio takes added responsibility. This pressure is compounded by the fact that typically liability-relative asset allocation approaches, like hedge portfolios and surplus optimization, can rely on a certain liability. When it comes to managing assets, what's the key difference between a bank liability and a pension liability?

No. Banks and pensions are both massive investment funds.

Not so. Both pension funds and banks serve their customers with a fiduciary focus.

Exactly! The duration and structure of a pension fund's liability won't evolve, but that of the bank will. Because the bank relies on depositors to fund assets, these depositors can make decisions that change the structure and duration of the liability. So the institution must manage the composition of the liability at the same time as its allocating assets. That's where the __integrated asset/liability approach__ comes in. The integrated asset/liability approach (also called asset/liability management in banking) uses multiperiod models to adjust assets to reflect a better correlation with the changing liability.

This approach applies to more than just banks. For example, insurance companies, long–short hedge funds, and reinsurance companies benefit from this integrated focus. Take a property and casualty company as an example. Within this book of business, decisions are made to assume risk for major natural disasters like earthquakes, floods, or hurricanes. As insurance policies are written, the liability of the insurer changes and so does the risk. So analyzing the probability of extreme events and the impact on the asset portfolio becomes critical. How might you evaluate extreme events?

Yes. Natural disasters are extreme tail events, so you need to accurately assess tail risk through a value at risk model or a conditional value at risk model. By using these models, you can pair the risk associated with an extreme event to the construction of the portfolio. This could look like a higher exposure to equities to offset potentially large losses every several years, or it could mean using reinsurance to reduce some of the overall risk.

That's not it. Natural disasters aren't market risks that beta measures.

That's not it. Extreme events don't occur with the usual three standard deviations of the normal distribution. You'll need to focus somewhere else.

As you can see, the portfolios of assets and liabilities are essentially linked in a circular flow. As the risks within the liability portfolio change, the asset portfolio must also adapt to address these risks so the business can have the necessary protection to take on additional customers. This continual loop naturally feeds into the primary goal of the integrated asset/liability approaches, which is to make the most informed decision that compromises between the assets and liabilities under its control so that asset growth is maximized subject to the worst-case outcome and other risk-measure constraints.

To sum it up: [[summary]]

Size

Fiduciary focus

Duration and structure evolution

Tail risk through 1% value at risk

Beta of the portfolio versus the market

Standard deviation of a normal distribution

Continue

Continue

Continue