“What is investment risk” is both simple and complex and in order to assess the information needed to assess risk we first of all need to qualify the world view of risk and return and then the model we are using to manage the risks and their impact. All this is beyond the subject matter expertise of the individual investor and why point of sale documentation can only ever inform about the broad nature of risk and return but never be sufficient to allow the investor to fully own the decision. Ownership of the transaction decision resides with ownership of the process and the significance of risk assumed with respect to risk via structure, assumptions and costs.
At a very basic level, risk is what happens to a price between two points in time and the impact of that movement in price – it can be both positive and negative. But what gives rise to risk is more important than a mere list of risk factors, or indeed a narrow quantitative measure of risk.
In a recent e mail exchange one of the mail recipients asked what is investment risk? A good question, but the question was more complicated than it appeared because it was raised in the context of what type of risks should be communicated to investors in point of sale documents.
Point of sale documents often deal with single securities or products and depending on the advice provided and the standing regulatory regime, the consumer may be required to fully understand the inherent risks and apply them to the position of a product or security within their portfolio.
This can be extremely complex, because as I have noted before you need a valuation, allocation and management model to be able to incorporate securities and products within a portfolio.
It is the model’s decision rules along with valuation models that determine which risks are important, how important and how they impact asset allocation. Some risks are offset by compensating assumptions and neutralised in terms of impact via structure and planning. While the decision tree may be complex, a good process allows it to be managed simply and effectively. But how an individual does all this is beyond me? Especially when you consider the often very limited range of information in these documents.
At a very basic level, risk is what happens to a price between two points in time and the impact of that movement in price – it can be both positive and negative.
Standard deviation is an estimate of the average price movement (above and below an “average price change”) but the impact can vary, and the price movement and the impact can compound.
Whether standard deviation is an appropriate measure of risk is a moot point and is to a certain extent dependent on whether or not all risk factors have been fully reflected in the price, and whether the estimate of price movement risk around return reflects the dynamics of the universe it is assumed to represent. This is not immediately obvious or explained.
The problem with standard deviation arises not because of its measurement of price movement per se, but because of its sole use in structuring “optimal” portfolios. In a non general equilibrium, non efficient market, the path of risk changes and hence so does the optimal nature of a portfolio that assumes a degree of certainty of a distribution that conflicts with that of the actual universe. In such an imperfect universe standard deviation is not the only risk impacting return and price, and standard deviation itself may be impaired by price movements which over or under emphasise risk factors.
In a perfect world risk is a price reaction to new information from a perfect price, but in an imperfect world risk is both a price risk due to new information and a latent price risk due to imperfect valuation and other structural economic imbalances.
Standard deviation ignores valuation risks, because the paradigm from which it stems assumes efficient markets and a general equilibrium. If markets were efficient and prices perfect investors would not need to stray beyond standard deviation as a guide to risk and would only be selecting individual investments as components of the market portfolio.
Risk management is how you manage risk and varies according to the impact you are looking to control for and your worldview of risk and return. So, in a point of sale document, what may be more important is the true nature of a product, so that risk can be inferred from this true nature or indeed calculated. Mere statement of the risks involved do not allow you to quantify the risks as they might impact the product or security, since these can only be derived from its true structure, nature and disposition.
Standard deviation of a mutual fund relative to its benchmark and market should raise more questions than answers – the market cap, sector, valuation ratios (P/E, P/B etc) relative to its benchmark and the major market index, the fund’s turnover and historical asset allocation and style profiles are also critical in terms of analysis, as are the decision rules and parameters under which a fund is currently run.
Factors which impact risk are many, but factors which are impacted by risk are few and some like liabilities are more or less fixed and unyielding.
If the price is (a) efficient and (b) the universe is in equilibrium, then standard deviation should reflect the market demand for and supply of an asset given the weighted distribution of all risk factors and therefore in certain dimensions assuming certain structures you can ignore the multitude of risk factors and focus on price movement which is (assuming a and b) representative of the impact of all risk factors. In other words, the universe of risk devolves to the universe of potential “uncertain” price movements, or standard deviation.
Price is determined at the margin so, in a way, it is also dependent on the distribution of tastes and preferences, saving and consumption along the margin remaining stable and representative of the whole.
The problem occurs when price does not fully reflect all current information (when fundamentals are not fully discounted in the price), because then you need to independently adjust for the those factors which are not fully represented in the price or take an additional risk which may or may or may not be rewarded. Passive investing assumes these risks but adjusts for these risks via costs.
The trouble is, even if price is incorrect, determining those risk factors which would allow you to adjust back to an equilibrium price is a complex and at times subjective process. If the price was incorrect and everybody knew it and acted on this information the price would adjust.
But even here we are looking at significance, since not all risk is significant and worthwhile adjusting for – risk has a cost and if the cost of the risk management exceeds the impact of the risk there is no point in controlling for it, which is why many low cost portfolio options can accommodate and ignore a large number of risk factors. A 3% MER compounded over 20 years is 80.61%, yet a 0.2% MER compounded over a similar time frame is 4.08%.
The biggest risk to an investor is the risk of adverse price movements that impact their ability to meet their financial needs from their assets. These risks are accentuated by imperfect pricing (think ridiculously high P/Es of the 1990s), economic and market cycles and structural economic and financial risks (think the build up of debt and global structural and economic imbalances).
Some risks can be managed by discipline, others by structure and planning, and those that cannot be avoided managed via assumptions and/or adjustment to withdrawals.
Adjusting allocations according to consumption/savings profiles (i.e. greater allocations to low risk, higher yielding, more liquid more defensive allocations for large liability profiles) assumes that price movements are to a significant degree path dependent and not random and independent. For example if your structure always allows you to manage a recession and the associated bear market and to adjust structure as economic and market cycles mature then you can manage liability risks within this universe.
But, if the universe is in equilibrium and future shocks to the universe are random and independent of current relationships then there can be no benefit from what is effectively making a bet regarding an element of certainty regarding price movements. In a perfect market the trade off would be with regard to risk and return and not necessarily a trade off between greater or lower certainty of meeting liabilities.
In MVO constructs, meeting liabilities through mean variance efficiency (my view not the world view) is fudged by assuming certainty with regard to the distribution of uncertainty, certainty with regard to the current equilibrium price, return and correlation and working back to an asset allocation that can best meet liabilities given that distribution and that price.
So in a world where markets are not efficient, where prices (demand and supply) are often, and at time significantly so, out of equilibrium, and where price movements are to a certain degree path dependent, standard deviation is not the only risk measure/structural decision rule that should be incorporated to allow you to structure portfolios to manage liability risks, at a point in time and over time. Indeed, if prices are wrong, then so is a statistical measure of price movement around return, and the return itself.
In a world where cyclical market and economic risks have impact on the ability of assets to meet financial needs, risk management would focus on the significance of the valuation differential and the time frame of risk exposed by that differential. Cyclical market and economic risk to return and time frame of risk (the time frame in which the returns on volatile higher risk longer term assets are likely to have a greater chance of out performing the return on lower risk/shorter term assets) impact asset allocation where liability profiles (expressed as an annual yield requirement) exceed the annual yield of the portfolio.
It is this time frame, which expands and contracts according to valuation and cyclical market and economic risks. So, we have another key risk component input in structuring portfolios that have liabilities, risk and return as the key decision factors.
In a sense, as the recent Noble Prize awarded to Shiller and Fama suggests, imperfect valuation can be reasonably assessed in terms of its long term impact, but there is little certainty over the short term magnitude and timing of direction of prices. In other words, risk management should not be focussed on trying to time but to mitigate significant risks via structure and assumptions. Structure manages significant short term risks to return, and hence withdrawals, by being conservative over the time frame of the risk event and by conservatively qualifying its long term impact.
What is investment risk is both simple and complex and in order to assess the information needed to assess risk we first of all need to qualify the world view of risk and return and then the model we are using to manage the risks and their impact. All this is beyond the subject matter expertise of the individual investor and why point of sale documentation can only ever inform about the broad nature of risk and return but never be sufficient to allow the investor to fully own the decision. Ownership of the transaction decision resides with ownership of the process and the significance of risk assumed with respect to risk via structure, assumptions and costs.