Assignment X2 X2.6 the question is "describe how the actuary will have used their asset-liability model to derive an investment strategy for the fund" initially, my thoughts were constrained into the model of ALM, but after refer to the solution, it seems mostly answer the question from a general model concern, there is few points related to the ALM, i.e. how the assets allocation / investment strategy should be selected according to the liabilities' characteristics, cashflow etc. how do I understand the point of view for the solution? or, if the exam context, if I write more related to how to select assets to match the liabilities rather than general points subject to modelling, i.e. parameters, assumptions, scenarios testing, sensitive testing etc., will it be credited well?
Hi - the question is asking about how an actuary would use a model to do something, so the solution needs to reflect that: a description of the modelling process. This description needs to be tailored to the specified purpose: here, setting the investment strategy for the pension fund. Bear in mind that 'ALM' stands for asset-liability management, not asset-liability matching. You seem to be thinking about it from the perspective of finding assets to match the liabilities. A model wouldn't be needed to identify the types of assets that would best match the different types of liability outgo. A model could be used for matching asset selection if it were possible to achieve close cashflow matching: such a model could help the company to select the appropriate terms etc of bonds that would best match the liability outgo. But this is only really feasible for predictable cashflows such as from large portfolios of pensions (or annuities) in payment. In most situations, perfect matching is unlikely to be achievable - and even if it were, it may not be desired. ALM is about understanding and managing the impacts of mismatching between assets and liabilities. So, here, setting the investment strategy for the pension fund will involve assessing the extent to which mismatching can be tolerated. This is done by first setting an objective, eg a specified maximum probability of the funding level falling below x% in the next year. The model is then used to test different asset allocations against this objective. As the question is asking how this is done, the answer needs to set out the process: setting the objective, modelling assets and liabilities stochastically, turning the outputs into a decision. Section 6 of Chapter 15 ('Using a model to determine investment strategy') is relevant here. Hope that helps.
