Not Another Investment Podcast
Understand investing beyond the headlines with Edward Finley, sometime Professor of Finance at the University of Virginia and veteran Wall Street investor.
Not Another Investment Podcast
Navigating Dynamic Asset Allocation and Portfolio Rebalancing Strategies (S1 E21)
Have you ever wondered how a seemingly straightforward investment strategy can spiral into unexpected risk over time? Join Edward Finley, a veteran Wall Street investor and occasional professor at the University of Virginia, as he unpacks the complexities of dynamic asset allocation. We challenge the status quo by illustrating how a typical strategic asset allocation, like one with a 75% equity and 25% credit risk, can morph into a much riskier portfolio if left unadjusted, potentially ballooning to 94% equity risk after 30 years. Learn why adaptability and periodic adjustments are crucial for maintaining a consistent risk profile and ensuring long-term investment success.
But that's not all—we also dive deep into the mechanics and psychology of portfolio rebalancing. Discover the merits of counter-cyclical strategies that compel you to sell winners and buy losers, despite the psychological hurdles. We dissect periodic and contingent rebalancing techniques to show how these approaches can stabilize your portfolio's risk levels, enhance returns, and minimize volatility.
Then, we explore the realm of tactical tilting, where portfolio managers make calculated moves based on economic indicators or expecting returns to revert to their mean. But as we will see, that is easier said than done.
Finally, we discuss how portfolio managers can use active risk in portfolios to fine-tune their risk allocations. Bringing into focus asset classes like hedge funds, private equity, real assets, and others, we see that active risk can be an equally important decision in asset allocation.
Tune in to understand the potential rewards and pitfalls of these dynamic strategies, and join the debate on the efficiency of markets over shorter intervals. This episode is packed with insights that will elevate your investment game.
Episode notes: https://1drv.ms/p/s!AqjfuX3WVgp8un_-h7LOKuedrf1a
Thanks for listening! Please be sure to review the podcast or send your comments to me by email at info@not-another-investment-podcast.com. And tell your friends!
Hi, I'm Edward Finley, a sometime professor at the University of Virginia and a veteran Wall Street investor, and you're listening to Not Another Investment Podcast. Here we explore topics in markets and investing that every educated person should understand to be a good citizen. Welcome to the podcast. I'm Edward Finley.
Speaker 1:Well, last time we spent our time discussing strategic asset allocation and, for your recollection, the strategic asset allocation just describes the allocation to different asset classes and different risks in a portfolio that are designed to achieve a desired objective, and the key part there is the desired objective. At the end of the day, managing money in a portfolio is really primarily about that achieving your objectives. We talked about how we have to identify what those objectives are. We have to somehow quantify our risk of failing to achieve the objective, and then we talked about several different methods that have been used traditionally and not so traditionally to manage portfolios, none of which really does a terrific job of accurately predicting the success or failure of an allocation. It's a little more art than it is science. Today, we're going to move on from that discussion and we're going to just assume that we've decided on a method of our strategic asset allocation, whatever that might be. So we now know what our long-term allocation to risk should look like in order to accomplish our objectives. Today, we're going to talk about something called dynamic asset allocation. What's the concern here? What are we focused on? Well, regardless of the method we use to develop the strategic asset allocation, we know that different asset classes have returns that pretty significantly time vary. We also know that their volatility time varies, though less so. We also know that their correlations time vary, but less so, and that means that any allocation that we do in a strategic allocation is limited by the fact that it's what's called by economists a single period framework. So what do called by economists? A single-period framework. So what do we mean by a single-period framework? Well, what we mean is we're designing the strategic allocation taking into account the average return, the average volatility and the average correlation over a long period of time. But in any given year or any given shorter time period, we know that those returns, volatilities and correlations will be different, and so we have to take that into account when figuring out how to manage our portfolio.
Speaker 1:One way for us to observe the importance of dynamically allocating a portfolio is to just consider the opposite. Consider what would happen if we set a strategic allocation. Let's just say 75% equity risk and 25% interest rate or credit risk, and then we didn't do anything. We just sort of invested the money on day one and let the portfolio roll. If you have access to the slides that are linked on the show notes, feel free to take a look at those. If not, I'll just walk you through the graphic. There's nothing you're really missing, but you'll see if you do look, I have a bar chart that shows just such a portfolio that starts now in 2024 and goes on for about 30 years, with the initial allocation being 75% equity risk and 25% credit risk.
Speaker 1:I also, in this model, have built a kind of randomized model where the returns to equities and to bonds each year vary randomly, with consistency around its volatility, so that it's not just that same return every year. And so I map out this portfolio with that kind of random return that all centers around each asset class's particular average return. Same for correlation. What happens? Well, the initial allocation 75% equity risk and 25% credit risk has an expected return of 7.2%. And I picked that 75-25 because in my mind I imagined a scenario where this is a portfolio that needs to grow in real terms but for which there aren't any distributions, but nor are there any additions, and so I picked the level of risk that I thought made sense, which would be an expected return of about 7%, and that had expected volatility of about 12% for an expected information ratio of 0.6.
Speaker 1:Okay, well, what you see is that if that portfolio is left to run, earning the returns that each of those asset classes earns over time, by 2054, we have a portfolio that's no longer 75% equity risk and 25% credit risk. We instead have a portfolio in year 30 that's 94% equity risk and 6% credit risk. That's 94% equity risk and 6% credit risk. So the first thing to observe is, by having a strategic allocation in year one and not doing anything more than that means that over time, your portfolio gets riskier and riskier. How risky 94% equity risk and 6% credit risk. It has expected returns going forward from 2054 of 8.3%, so a full percent higher, and expected volatility of 15.5%. Fully 3% percentage points higher for a roughly equal expected information ratio of 0.54. Higher for a roughly equal expected information ratio of 0.54.
Speaker 1:So that's a really different portfolio than what we originally designed and the fact that the expected information ratios of the initial allocation and the final allocation are so similar tells us that the portfolio became riskier, but it didn't become more efficient. It's not that it earned more returns per unit of risk. In fact, it earned a little less return per unit of risk. It just took more risk. And so what this tells us is that we really need to do something more dynamic, year to year, quarter to quarter, in order to maintain the risk allocation that's consistent with our goals, and that's why we call it dynamic asset allocation.
Speaker 1:Now, this is not to suggest that day trading is the best way to manage a long-term portfolio. Quite the contrary. What I'm trying to say is that, in order to be successful in the long term, it means that we have to make dynamic choices at shorter time periods in order to maintain the risk profile that we've designed from the beginning. In order to get an even clearer picture of this idea, imagine what that might look like if we had perfect foresight. Let's imagine a world in which, ex ante, we know each asset's return, volatility and correlation for that coming year, and then we allocate across those assets in order to have the optimal portfolio to achieve our 7% return goal. So what would that look like On the slides?
Speaker 1:The second slide you'll see is an image showing the portfolio mix of three assets equity, credit and real estate. If we had that perfect foresight credit and real estate. If we had that perfect foresight, the first thing that we observe, which is really quite shocking, is how wildly varied the portfolio allocation is year to year. If we knew perfectly ex ante what the returns, volatilities and correlations were of our assets, we would own a really really different mix than we would otherwise own. We would own a really really different mix than we would otherwise own.
Speaker 1:Second thing that comes out here is the degree to which equity risk is absolutely de minimis. It's tiny relative to the allocation in the static model that I gave you of 75% equity risk and 25% interest rate risk. I gave you of 75% equity risk and 25% interest rate risk no-transcript, not a combination of the three. All of which is to say that what this portfolio would achieve is very much what our strategy is. The average annual return of this perfect foresight portfolio would be 7% a year. That's exactly our goal when we design the strategic allocation of 75% equity, 25% credit risk.
Speaker 1:But the volatility of this portfolio is 1%. That's right, 1%. Now contrast that to the strategic allocation's volatility of 12%. Remember, volatility is a proxy for uncertainty, that is, volatility is the dispersion of monthly observations around the mean and, as we said a long time ago, if there's vast dispersion of monthly observations around the mean, there's going to be very high volatility. Why would we have such vast dispersions around the mean? Because it's highly uncertain. Markets are very, very uncertain, busy trying to predict what the true value of the asset is, and so when we're designing a portfolio with perfect foresight, how much uncertainty is there? None, and so, as a result, our portfolio has very, very little volatility, precisely because we have no uncertainty.
Speaker 1:So it's just a pretend thought experiment, but what I think it illustrates really quite importantly are a couple of things. First, that dynamic allocation doesn't have to be. Little tweaks around the strategic allocation. It can be, but it doesn't have to be. It can instead be quite robust. However, because we don't have perfect foresight, there's such dispersion in returns over time that it's super difficult to predict what those returns are, and the effect on outcomes of getting it wrong on predicting returns is very, very large on its own, and it's particularly large relative to getting it wrong when predicting volatility or correlation. So we don't have perfect foresight. We can see that there is reason to think that pretty substantial changes in allocation can be warranted, but there's an attendant risk to that of getting it wrong and not achieving our goals.
Speaker 1:So how do we account for this time variation in asset class returns and correlations? And the answer is there are a couple of different ways of doing it. Let's start by rebalancing. If we assume that returns and volatility are not predictable, so we can't know ex ante what they're going to be or, as we sometimes hear it called, it's a random walk then the simplest method of making dynamic asset allocation choices is to rebalance the portfolio periodically back to its strategic allocation. So think about it. What does that mean? Well, it means that if, in a year, equity returns are far higher than the average and bond returns let's just say theoretically are lower than the average, it means that, to rebalance back to our strategic allocation, we would sell equities because they grew faster, both relatively to bonds but also relative to the long-term average, so we probably own too much of them. We'd want to sell equities and then we'd want to buy bonds because the bonds earned lower returns relative to equities and lower returns relative to its long-term average.
Speaker 1:Well, that's really a counter-cyclical strategy, isn't it? It's selling your winners and buying your losers, and that makes a lot of sense if you imagine, over a very long period of time, that these asset returns will revert to their mean and, as we've discussed in prior episodes, there is some evidence for that to their mean and, as we've discussed in prior episodes, there is some evidence for that. It's just that that reversion to the mean might take upwards to 20 or 30 years, and so it's not perfect in that respect, but it's definitely one way to think about how to keep the strategic allocations risk profile the same. It's also the case that, by being counter-cyclical, rebalancing is really difficult to do in practice. It's really difficult to do because psychologically, as we talked about and we talked about behaviors psychologically to sell our winners cuts against the grain. That's not something that we like to do, because that suggests that we value our gains in the same way that we value our losses, and we know that's not true. And it's also very, very difficult to buy our losers. Our psychology tells us that when we own losers, we probably ought to cut and run, when, in fact, normal volatility will produce losers, and normal volatility producing losers is no reason to cut and run. It is really a reason to then invest in what is essentially if you believe the risk is accurately described what is essentially that risk on sale and who wouldn't want to own that risk when it's on sale? So countercyclical, which is a good thing, but behaviorally very, very difficult to do in practice.
Speaker 1:So one practice that investors adopt in order to overcome how difficult countercyclical rebalancing is is to just adopt periodic rebalancing. Is to just adopt periodic rebalancing. Call it annually, call it quarterly, really depends a little on one's goals and also the risk profile of the portfolio. I think more frequent rebalancing than quarterly runs the risk of creating inefficiency in the portfolio because there are costs to trading and those costs, even if you're just buying and selling ETFs at a discount brokerage house, they could be very low or maybe close to zero, but they're not zero in terms of opportunity cost. And so we don't want to do it too frequently. We also don't want to leave it for too long because, as we saw in the static portfolio, the allocation of risk can really get out of balance rather quickly, even just year to year. So some people do it annually, some people do it quarterly.
Speaker 1:Well, as I mentioned, rebalancing whether it's annually or quarterly absolutely has the risk of there being unnecessary trades and costs. And so it may be, for example, that if I do it annually and I sell equities in what's about to be I can't know it in advance, but what's about to be a prolonged bull market in equities it may be that I'm really losing, as an opportunity cost, some of those potential gains. But it's also the case that I might incur costs by making those trades at period intervals where, really, what I'm just observing is the regular volatility of the asset. What we want to do is rebalance the portfolio according to our strategic allocation at moments when the portfolio's balance is out of whack by more than just typical volatility. And so another way that investors will rebalance a portfolio is called contingent rebalancing, and what contingent rebalancing means is you evaluate the portfolio for rebalancing at whatever period you choose quarterly or annually but then you compare how far out of your strategic allocation you are as a percentage average volatility of that asset class for the period. So, for example, let's say I have an allocation to US large cap equities and US large cap equities have annual volatility of around 12%, let's call it. And so I rebalance annually and I find that my equity allocation is too high by 10%, well then I might say that is a really good reason to rebalance, but it looks an awful lot like just typical volatility, and so I shouldn't try to time the market, incur costs, both opportunity and transactions costs. I should just let the allocation ride. Whereas, in contrast, if I find that I'm 15% over my strategic allocation and the average annual volatility is 12%, then that's the moment for me to rebalance back to my strategic allocation, and there are lots of ways of doing it. I gave one very simple example, but all of them have in common the idea that you look first at the calendar, but then you also look at the degree to which the portfolio is out of whack.
Speaker 1:Well, the rebalanced portfolio I have as a third slide for you, the rebalanced portfolio, does something that I think is rather interesting. Portfolio does something that I think is rather interesting. First, in the first year, when it's 75% equity and 25% credit, we have the same expected return 7.2% the same expected volatility. However, by rebalancing the portfolio annually and in my model I didn't do contingent rebalancing, I just rebalanced it annually back to our strategic allocation we see that there is no change in the risk allocation, so that's a pretty good thing. We are owning the risk that we said we wanted to own, and that's a very, very important fact. It's also the case that if we look at the average annual return of that rebalanced portfolio and we contrast it, say, to what the expected annual return was, we see that we've actually earned about 40 basis points more in return. We earned 7.6% as opposed to 7.2%. But, much more importantly, the volatility of the portfolio is now 7.3% instead of 12.2%.
Speaker 1:Now why is that? What is it about rebalancing like this that causes the return to perhaps be a little higher, but, importantly, that the volatility is so much lower? Partly it's because the expected returns when we're looking at expected returns, expected returns assume continuous rebalancing, and that continuous rebalancing is a good example of just doing it too much. There are certain frictions and not just costs, but frictions in terms of opportunity costs that cause the return to be slightly diminished but that give the volatility a whole lot more room to bounce. It's also partly because we're buying or selling the lower volatility asset class in this case, credit and we're doing the opposite. And we're doing the opposite If we're buying credit, we're selling the higher volatility asset class equities, and if we're selling credit, we're buying the higher volatility asset class, and the result is we have a much more balanced, stable portfolio. In terms of its volatility is. Our counter-cyclical behaviors actually mean that we are moving from higher volatility to low volatility and vice versa, depending on the relative and absolute returns of each asset class, and so rebalancing is a really, really potent and important way to make sure we're running our portfolio efficiently, but, importantly, also to make sure that our portfolio is tuned to exactly our goals and our purposes.
Speaker 1:Another method of dynamic asset allocation available to portfolio managers is called tactical tilting. What are we talking about here? Well, the manager starts from the same starting point, which is having a strategic asset allocation, and uses a form of dynamic asset allocation usually rebalancing to keep the risks at that strategic level, usually rebalancing to keep the risks at that strategic level. What tactical tilting does instead is the manager will make choices about which risks to overweight and which risks to underweight, and they'll make those choices because, as we know, time variation in returns and time variation in volatilities and correlations suggest that there may be reason to own our risks in different weights at different periods of time.
Speaker 1:There are a few different ways that managers can engage in tactical tilting. One is sort of a version of our perfect foresight model, which is managers can simply look at economic data and they can try to make prognostications about whether a particular kind of risk will have future returns that are higher than their long-term average or future returns that are lower, and based on that prognostication which is a fancy word to say guess based on that prognostication, they'll overweight or they're underweight. Based on that prognostication, they'll overweight or they're underweight. Well, you'll recall from earlier core episodes about equity risk, we looked at lots of different ways to try to understand how to forecast equity risk. One of them was the economic data, because of course, equity is exposure to the productive economy. Of course equity is exposure to the productive economy. But what we saw there was very weak evidence that equity returns are correlated to conventional economic data, and so that's probably not going to be a terribly useful way to tactically tilt.
Speaker 1:There are lots of other ways that managers can do tactical tilting. Lots of other ways that managers can do tactical tilting. They might tactical tilt not based on economic data, but based on price earnings ratios, which is just simply a ratio of how expensive a current stream of earnings is in the market. But all of them will have in common the very significant risk of getting your estimate wrong, of getting your estimate wrong. And if you get your estimate wrong and you're overweight or underweight, depending, as we saw many times in the original discussion of portfolio construction, getting it wrong is very, very costly in terms of ultimate outcomes. So that form of tactical tilting is available. There are lots of managers that do it. It is not for the faint at heart.
Speaker 1:Another method of tactical tilting is to take a less subjective view and to take a more objective view. The objective view would be to say I don't know if I can predict equity returns properly or predict bond returns properly, but I do know that over time returns tend to mean revert. That is, if recent returns are higher than average, we would expect in the future returns to be lower, and vice versa. So one form of tactical tilting is to just look at the prior three years of an asset class's returns. If they're higher than average, then underweight them, because if they're higher than average, we would expect, all things being equal, the future returns to be lower than that, and vice versa. If the last three years' returns are lower, overweight that because we have the expectation of higher future returns.
Speaker 1:The difficulty with this method of tactical tilting is similar to the first, which is it sounds great because it sounds objective, but there's really significant data out there to suggest that really which usually does not bode well for your keeping your job, or, if it's not your job and you're just managing your own portfolio, underperforming for a significant period of time can really undo your confidence and maybe make you make mistakes. So are there other ways? Yeah, there are a bunch of other ways, but at the end of the day, tactical tilting, while it's out there and you should be aware of it, is a method of dynamic allocation that's well-intended but very, very difficult to execute. Well, well, the last form of dynamic asset allocation that I want to bring up is something called active risk. Now remember back to earlier core episodes when we talked about some finance theory and the efficient markets hypothesis, and you will recall that back then we said that in market theory speak, we would expect over a long period of time for all accurate information to be fully incorporated into price. That's not a terribly controversial thing to say, because I've introduced the phrase over the long term, and that erases a whole lot of imprecision.
Speaker 1:The real debate around efficient markets hypothesis and what it means is over shorter time intervals, and you'll recall that over shorter time intervals there are really two schools of thought. There is one school of thought that suggests that at shorter time intervals there are really two schools of thought. There is one school of thought that suggests that at shorter time intervals the market is simply random. And if the market is simply random, there's no way to predict what future prices will be in the next period, and so you shouldn't try. You should just own the market risk. That view is what animates the strategic asset allocation and rebalancing approach. It's that reality.
Speaker 1:The other view is to say that at shorter time intervals it's not random, that there are behaviors at work, preferences at work, and those preferences cause prices to move more predictably in shorter time intervals. And if you believe that version of how markets operate and how they are efficient at incorporating information, then that might lead you to believe that there are managers who can do the kind of tactical tilting I described a minute ago. They can own a set of risks that's different than the systematic risks that we use to design our portfolio. They can alter their mix of the ownership of those risks so that in the short run they earn excess return. And if the long run is a period of short runs, then managers who successfully trade like that are going to be managers who, in turn, will earn, over the long run, excess return over their risks.
Speaker 1:Likewise, going back to our discussion of asset classes like various hedge fund strategies, like some of the things like private equity and real estate and so on, we will remember that there are also certain asset classes where the risks that you own are non-normal, that is to say, they either have very significant skew, or they have lots and lots of outliers, or they may not even be linear. The relationship between risk and return might be non-linear. So it could be non-normal, could be non-linear, could be both. And those risks make volatility and beta very unreliable measures for the overall risk of a portfolio and it's very difficult, as we saw, to measure what the actual risk is. But what we do know is that investors earn some kind of return for those non-normal, non-linear risks, and so managers can also not choose to overweight different risks and own it in a different proportion to the market.
Speaker 1:Managers might own exactly the amount of risk in the market that they want to own, but they might own it with additional non-normal or non-linear risks, earning return for those, and that will not turn up in typical summary data when we look at an investment. And so when we have these two things in hand, it means that as a portfolio manager, we not only want to design a strategic allocation likely to achieve our goals, we not only want to be sure that we rebalance the portfolio back to the strategic allocation in order to have dynamic management of the risk allocation. Dynamic management of the risk allocation. We might tactically make decisions about which risks to own more or less of, with all the attendant dangers in doing that. But we also might make decisions about whether to own certain risks in their pure systematic form, which is to say that's the strong view of the efficient markets.
Speaker 1:Hypothesis prices are a random walk, utterly unpredictable. Don't try. Or we could choose to allocate that portion of risk to managers that pursue strategies with active risk, that is, owning the same aggregate market risk but in different proportions, or owning exposure to that risk along with non-normal and non-linear risks, and that means that the manager, that the portfolio manager making those choices, is still ostensibly earning returns for the risks that they've designed the portfolio to own, but they're doing so with some additional return, either because they've chosen managers who are good at what they do think venture capital in that example where we showed that the best managers earn far more return than their systematic risk and they do so persistently or because we're choosing owning that risk in a form that includes non-normal, non-linear risks Think hedge fund strategies. So we can also do that in a portfolio which, again, doesn't change the amount of risk we own as much as it does offer the possibility that these managers will earn returns in excess of the systematic risk. Well, that's it for dynamic asset allocation.
Speaker 1:When we come back in our next episode, we're going to take active risk allocation. How do we evaluate the portfolio and the way in which we would expect it to achieve our goals? Thanks for listening and we'll talk to you next time. You've been listening to Not Another Investment Podcast hosted by me, edward Finlay. You can find research links and charts at notanotherinvestmentpodcastcom, and don't forget to follow us on your favorite platform and leave comments. Thanks for listening.
.jpeg)