In this episode I speak with Jim Masturzo, Head of Asset Allocation at Research Affiliates.
In his role, Jim oversees the research and publication of the firm’s capital market assumptions as well as the implementation of those views into a suite of tactical portfolios.
We begin our conversation discussing the foundational assumptions behind the capital market assumptions. Like most firms, Research Affiliates takes a long-term view on return and risk. In line with the firm’s guiding philosophy, they also introduce long-term mean reversionary effects.
Not surprisingly, these assumptions have been relatively bearish on U.S. equity returns for a large part of the last decade, and we discuss how to view the dispersion between these model forecasts and realized results.
We then shift our conversation to the application of tactical views. With capital market assumptions serving as the strategic backbone, Jim and his team develop a number of regime-based model portfolios that can be blended to express different tactical views.
But the team does not take a purely quantitative approach. Jim proactively acknowledges and seeks out model blindness. Rather than try to force idiosyncratic fixes into the models that might bias results, however, he and his team adopt qualitative trades to adapt the portfolios.
From strategic to tactical and quantitative to qualitative, this is a wide ranging conversation all about asset allocation. I hope you enjoy.
Transcript
Corey Hoffstein 00:00
All right 321 Let’s do it. Hello and welcome everyone. I’m Corey Hoffstein. And this is flirting with models the podcast that pulls back the curtain to discover the human factor behind the quantitative strategy.
Narrator 00:20
Corey Hoffstein Is the co founder and chief investment officer of new found research due to industry regulations, he will not discuss any of new found researches funds on this podcast all opinions expressed by podcast participants are solely their own opinion and do not reflect the opinion of new found research. This podcast is for informational purposes only and should not be relied upon as a basis for investment decisions. Clients of newfound research may maintain positions and securities discussed in this podcast for more information is it think newfound.com.
Corey Hoffstein 00:52
In this episode, I speak with Jim Mr. Rizzo head of asset allocation at Research Affiliates. In his role, Jim oversees the research and publication of the firm’s capital market assumptions, as well as the implementation of those views into a suite of tactical portfolios. We begin our conversation discussing the foundational assumptions behind the capital market assumptions. Like most firms Research Affiliates, takes a long term view on return and risk. In line with the firm’s guiding philosophy. They also introduce long term mean reversion airy effects. Not surprisingly, these assumptions have been relatively bearish on US equity returns for the larger part of the last decade, and we discuss how to view the dispersion between these model forecasts and realized results. We then shift our conversation to the application of tactical views. With capital market assumptions serving as the strategic backbone, Jim and his team develop a number of regime based model portfolios that can be blended to express different tactical views. But the team does not take a purely quantitative approach. Jim proactively acknowledges and seeks out model blindness. Rather than trying to force idiosyncratic fixes into the models that might result model bias. However, he and his team adopt qualitative trades to adapt the portfolios, from strategic to tactical, and quantitative to qualitative. This is a wide ranging conversation all about asset allocation. I hope you enjoy. Jim, thank you for joining me on the show today. excited to have you here.
Jim Masturzo 02:28
Thanks for having me. Glad to be here,
Corey Hoffstein 02:30
you had a bit of a roundabout path leading to your current role as head of asset allocation at Research Affiliates. So I was hoping you could start us off with your background.
Jim Masturzo 02:41
Sure. Absolutely. And Cory, thanks for having me. I’m glad to be here. So it’s true. My, my path has not been the most traditional path when you think about others in the industry. But I believe that we’re the combination of our experiences and a more broad background works just as well as something more straight and narrow. So stepping back a little bit, I got my undergrad degree at Cornell and electrical engineering decided very quickly, I didn’t want to be an electrical engineer sitting at a bench and building circuits all day sounded like a good idea when I started and then decided, although I enjoyed what I was learning, and I enjoyed the field of electrical engineering, that was not the path for me, this was the late 90s, right at the end of the 90s. And I did what a lot of other my classmates did. And I actually joined a consulting firm back when the big four, I joined PricewaterhouseCoopers back with the big forehead, consulting arms, by consulting arms at that time, it was really software development. So I became a software developer, working in the financial services industry for a lot of the big banks did that for a couple years and then moved over to a small startup company called Aqualand, which was in the securities lending industry, actually acting as an intermediary are really a you out as a hub and spoke model. So they were the pipes between borrowers and lenders for the purpose of borrowing and lending securities. So worked there for a couple of years architecting their system building that out that company still exists today, and actually facilitates by and large, most of the securities loan or stock loan transactions that happen globally, although it’s a company that most people have never heard of, unless you’re really into the weeds of that industry. I’m a big proponent of learning how to write code How to Architect code, especially in our industry, it’s just a skill that I encourage everyone to get just the the gift that keeps on giving. So 2005 I left Aqualand and went back to business school decided it was time to get an MBA, think about long term future went to Duke 2005 to 2007 came out actually went to school thinking about equity research as something to do after school. Not sure that was a good or a bad idea at a time when post tech bubble still most of the big banks were getting rid of their equity research arms or farming that out to specialist firms got out of busy It’s gone. 2007 actually went back into consulting for a while and spent what I described as the global financial crisis working at a consulting firm in particular, strangely enough, focusing on electrical utilities, I learned a lot about that industry, electricity pricing, commodity pricing, how to think about or how utilities think about commodities. It’s a really interesting something, again, that I never expected to do, but really interesting, moved on to Bloomberg in 2010, in a portfolio analytics group, that was really kind of a step back towards where I am now. And in that role, I was focused on their port application really out talking to portfolio managers about how do you manage your portfolios, the job is teaching someone how to use a Bloomberg function. From my perspective, you can’t really do that well, if you don’t understand their process. So I learned a lot about how different managers think about their process. And you figure out in an industry where we all think we’re right, and we’ll sit around and tell everyone why we’re right. And we’ll try to convince everyone, you’re right, you realize that there’s a lot of different ways to do what we do. And people think about things in a different ways. None of us know what the future is. So it’s really how do we create our call it probabilistic view of the future and how to different people do that it’s really interesting. And then I joined Research Affiliates in about seven years ago, started by building their asset allocation interactive site, which is where we publish our long term expected returns. And that’s a free site for anybody who wants to go to our website and access those also being involved in our asset allocation strategies. And eventually, as you mentioned, the head of that group, running a number of different multi asset strategies that we manage or sub advice for, as many of your listeners know, research affiliates were really around publishing research and sub advising for others, we don’t actually manage any dollars, specifically, as Rob Arnott likes to say, we’re an investment manager that doesn’t manage any assets. And that’s where we are today.
Corey Hoffstein 07:03
To that point, Research Affiliates is probably most well known for your fundamental indexing methodology, which is the backbone for a lot of the equity strategies that are sort of the indices that are published by Research Affiliates, but also fairly well known for some of the tactical asset allocation strategies that you manage. So I was hoping that’s where we could really focus on this conversation. Obviously, with you being head of that role. Can you describe some of the investment process behind and maybe the philosophy behind those asset allocation strategies?
Jim Masturzo 07:34
Sure. And as you mentioned, we’re probably more well known for our fundamental index, the Rafi strategies. However, asset allocation strategies actually are where the initial strategies when the firm was founded in 2002. So they actually have a longer history, which surprises many people. Asset allocation can mean a lot of different things to a lot of different people, global macro kind of falls in the asset allocation, some people think about it is just traditional stock bond correlation. For us, it means investing in global and domestic stocks, bonds, credit commodities, currencies, but at an asset class level, we don’t drill in, or you can think about as an ETF or a fun level, we don’t trade within asset classes, we’re not going to go along corn and short coffee or anything like that. So it is at a high level sort of fund ETF level. And the way we do it is that we start with our long term capital market expectations. And the way we think about that is tenure expectations, build a portfolio. And if you know nothing else, if you just say, Look, I have no tactical views, I don’t know anything in the short term. That’s the portfolio you should hold. Now, most of the time, we like to think we do know something about the shorter term, the tactical part of tactical asset allocation. And so from there, we overlay a number of different signals. Some of those are economic and fundamental. You can think about growth, inflation, volatility, how those change not levels, I think it’s very interesting or very important to when you think about things like inflation. While we haven’t had high inflation, since the 70s, looking at the level of inflation is not all that helpful, but how it’s changing can be quite interesting and quite informative. Yield curve slope is another thing that’s really interesting and got a lot of play about almost a year ago now. And we talked about the yield curve, inverting and what that means for subsequent economic activity. But we don’t just look at economic and fundamental signals, we look at what we call factor portfolios or Reversal Momentum, sort of your technical indicators. So we look at all of those things from a tactical perspective. And what we do is we both integrate signals together. And we also build many different portfolios and blend those portfolios together. So you can think about as integrating signals going into portfolio construction and then coming out with weights or sending a bunch of different signals into their own portfolio construction mechanism and getting different weights in our portfolio construction mechanism is mean variance optimization.
Corey Hoffstein 10:14
So I want to be a little careful about getting too far ahead of myself, because I know we’ll dive into some of these weeds in a little bit. So let’s maybe take a step back and talk about the foundation, that idea of if you don’t have any tactical views, first how that portfolio is being built. And you’re talking about your long term capital market assumptions, of which there’s a variety of published capital market assumptions out there. And if you look at yours versus other firms that get a lot of play in the news, they’re directionally similar but often different in the levels that they’re forecasting. Most right now, we’re pretty bearish on the returns on treasuries pretty bearish on equities, but the level of bearishness obviously varies per firm. And I think that goes into some of the nuances of how you guys think about building those capital market assumptions. So can you talk a little bit about that? What’s really sort of the foundation for how you guys go about creating those forecasts?
Jim Masturzo 11:05
Sure. So our forecasts are based on what we call a building blocks model. And that building blocks model is probably more commonly known as the Gordon growth model with an idea about mean reversion. So just to step back for a moment, our Investment Beliefs are centered around the idea that long term mean reversion is the most persistent way to extract excess returns. And our capital market expectations, again, in the long term, reflect that idea. So our building blocks and again, the Gordon growth model, for those that are more familiar with that idea starts with the idea of a yield, that yield can be a dividend yield for stocks, it can be a treasury yield, or a credit spreads plus a yield. And when you get into credit and things like that, then you have to start thinking about ideas like defaults and downgrades, and all those other things. But essentially, you come out with a fundamental idea of yield some growth in cash flows, will dividends grow in the future earnings, however you want to think about it. And then at the end, we add on a mean reversion component. So we don’t say that prices are going to mean revert, we say things like multiples will mean revert. We write a lot and talk a lot about the cape ratio price over average real tenure earnings as a model for business cycle smoothed earnings. And we look at reversions of that, to some fair value might be a long term history, it might be an exponentially weighted history, there’s a lot of different ways. And we could have a whole conversation on what’s the right way to think about a fair value estimate. But that’s essentially what we looked at in our capital market expectation.
Corey Hoffstein 12:42
You guys have been doing the capital market expectations for quite some time now, have there been any really interesting results that have come out of developing those?
Jim Masturzo 12:50
Yeah, so we started publishing our capital market expectations on the website in 2014. So it’s been, what, six years now, seven years now since we started doing that. And what you see from this traditional model, and I don’t know if this is a surprise anymore, is that domestic and us, I guess, I should say US stocks and bonds have been expensive for most of that time on a cyclically adjusted P E ratio perspective or a bond yield perspective. Now, we’ve been saying that for a long time, hasn’t always played out very well over the last 10 years. And we admit that but if you think about asset allocation, and one of the areas where I think people get hung up is that you need to think about consistency. So the model is very consistent. And again, it’s a Gordon growth model, at least from the capital market expectations. It’s a very consistent model, we use the same yields for all different equities with the same dividend yield for all equities, the same yields for across global bonds. And so as we think about relative differences, and again, when we think about capital market expectations, I always say, we published a scatterplot on our website that shows risk and return, I always say, put your hand over the y axis doesn’t really matter. What matters are the relative differences. So we can talk about level shifts up and down, maybe the returns aren’t centered around 2%, it’s 4%, or 6%, or whatever. If you use an optimizer, it’s not really going to care that much, as long as the relative differences stay the same at the risk of resulting a bit here and just looking back towards what we know, which is that sort of 2010 to 2020, at least prior to recently, you had one of the best long term annualized Sharpe ratios for just naked 6040 US equities, US bonds ever historically. And you’re not alone as forecasting that that should have been a portfolio that over the last decade seemed like it was overvalued and shouldn’t have done well. Lots of firms were forecasting that stocks and bonds were expensive and that we should not have had the returns we had. With that in mind. I would love to get your thought and how you look back historically, were the models blind to something did the world surprise to the upside in a way that the models just couldn’t see? Have we pulled growth from the future into the present? And we made the future outlook words, how
Corey Hoffstein 15:15
do you interpret what the model said versus what sort of occurred? Yeah, that’s a great point.
Jim Masturzo 15:21
And something that comes up all the time with clients and just users of our website. So just to put some context around it over the last 10 years, the Sharpe ratio over the 6040 portfolio was about one. So nine point I think it’s 9.2% per year, with a volatility of pretty close to 9%. That’s pretty phenomenal. The s&p returned about 12% a year for 10 years. That’s phenomenal. return some of that is driven by the fangs. We know what they’ve done. We know what the growth versus value breakdown, in essence, s&p has done. Other reasons we look at, you know, buybacks are a big deal. And I’m not wanting to get into the argument of our buybacks, socially beneficial, or the pariah of all investments, they just are what they are people are doing them. As far as buybacks go, there was an interesting stat from Ed Yardeni about a year ago where he said, if you look at all the buybacks only about 30% are actually going to reducing the number of shares outstanding, the other 70% Stay in the market. And essentially, our hypothesis and the data we have or the work we’ve done says most of those are going being paid out as stock based compensation. And stock based compensation plans, which used to be just around officers of the firm are now pretty broadly distributed, which is fine. And so I do think there’s a natural updraft from companies buying back stocks and issuing them out to employees who in a lot of cases, they’re just holding those for retirement or for whatever. So I think there’s a little bit of a tail when that happened there, to your point, is it that our process was wrong, or that the outcome just you can have a good process and a bad outcome. And that’s okay. I look at it as from a different lens, which is a risk management lens. And I say, look, as Kane said, you know, the market can stay irrational longer than you can stay solvent. And that’s fine. I looked at it and say that a cake multiple of 30 1am, I comfortable putting a client’s money to work at that level. As we saw that market can keep going up. But from a you know, sort of risk tolerance, risk aversion perspective, is that good alpha or bad alpha, something that’s expensive that keeps going up, I like to think about it like, returns or returns. But that’s kind of bad alpha, I would much rather look for cheaper things that have a longer runway or a higher probability of going up. So I’m the first one to say we missed it. The first one to say that it just is what it is. But going forward, the model still makes sense, right? returns are still based on yields plus changes in cash flows, plus some multiple expansion or contraction. And when I look at multiples and develop markets, or even emerging markets, even if they’re priced at fair value, if you just go all the way back to the dividend yield, well, the dividend yield in developed markets are the MSCI ephah, if we want to use that is to x what it is for the s&p see a four and a half percent versus 2.2, or 2.25. And emerging markets are in the middle and 3.7. So I still think there’s value in the model. And yet it’s a long term model. It’s not going to tell you what’s going to happen tactically, there’s still a solid foundation that we’ve been tested for
Corey Hoffstein 18:44
decades, keeping in mind that this is a long term model and one that I think you would sort of seek to be coarsely correct directionally and probably not expecting exact numerical precision on to your point, the exact return level. But are there any refinements that you’ve made over the last decade from lessons learned and putting this to work?
Jim Masturzo 19:04
Yeah, there’s definitely and you’re right, it is course. And we don’t look at expected returns as point estimates, we do look at them as distributions. And we include those distributions when we build portfolios, and again, we show them on the website. For simplicity, we talk about the mean and what the numerical expected return is. But it’s important to understand that that number is just the mean of a distribution. When we think back of what did we change currencies is an excellent thing to think about. So historically, what we’ve seen is a catch up, especially in emerging markets, as they become more productive, they should get a tailwind to their currencies the currencies should appreciate and we had built in a meaningful appreciation consistent with what we’ve seen historically in other countries. Japan is the perfect example of where we saw you know, a huge productivity catch up coming out. At the post war period and through the 60s and 70s, but we’ve seen in other countries as well. And we had that baked into our models, as we started to look at it again, and really drilling deeper and look at particular countries, what we realized is, you know, we were probably a little bit or in our opinions, maybe a little bit overly optimistic on that. And so we’ve started to pare that down a little bit. Even though these are tenure expectations, we still believe in that catch up. But I would say it’s probably longer term than 10 years and so deserved, even over a decade horizon deserve more of a haircut than we were giving it. And there are other things as well as we, even though the model is simple. There’s a lot of underlying assumptions that have to go into that model. And we’re constantly reviewing those things. default rates are another thing, default rates around bank loans and high yield, and what does that really mean, in different states of the economy?
Corey Hoffstein 20:53
And correct me if I’m wrong here, but I believe you guys very transparently publish your methodology behind the capital market assumption, tools, right? If you go into the tools, there’s actually like a resources section where you guys very explicitly list out how you think about building it for each asset class, right? That’s right,
Jim Masturzo 21:09
we do have documents that we’ve written that explain our methodology, we try to keep them short and sweet, 10 pages or so maybe 15 pages with lots of images, but we show you all the equations we use, and we get a lot of feedback on that from various users of the site.
Corey Hoffstein 21:24
So I want to now sort of shift our focus to some of the tactical decisions that you were discussing a little bit earlier, because capital market assumptions are going to be tend to be for the most part fairly slow moving, so I wouldn’t expect that the models are changing wildly day to day in terms of strategically what they’re asking you to do. So given that what does your day to day typically look like,
Jim Masturzo 21:47
these days, day to day is shifting all the time, as markets continue to be volatile. But you’re right, capital market expectations by the very nature of them shouldn’t move very wildly, they should be relatively stable. So I spend most of my time asking one simple question, and it’s what are we missing? What are the models missing? I have this view. And all of us when we develop models can have a particular affinity for those models. And in particular, you can be overconfident about your models, I try to the best of my ability, take the other approach, which is when I build a model, that model in some ways, becomes the enemy. And I spend my entire time trying to figure out why that model is garbage. And beating it up. Because the model is going to do it we ask it to do and it’s going to give out those results. So spending time trying to recreate those results are not helpful. What’s helpful is figuring out what we’re missing. And so I asked to spend a ton of time asking that question. We’re always doing research on what are we missing? How can we improve what we did? We, like many firms in our industry have lots of software, lots of models of very large asset allocation infrastructure, which I’ve always viewed as a living organic system that evolves, it adapts. And so any model we have, we’re always asking the question, how can we make that model better? Or what models? Are we missing that we think we need to deal with the challenges ahead? So that’s a lot of what we’re doing is research and writing, thinking about outside the models? What can we do that maybe we won’t put in the models, right? There’s a lot of fundamental things that happen that we just say, look, there’s no real way to model it doesn’t mean it’s not important doesn’t mean that we shouldn’t be aware of idiosyncratic events. We talk about them as I can give an example. That goes back a few years. But when you think about tensions in the Middle East, what is that going to do to oil prices or commodity prices as a simple example? Well, building a war factor is a pretty challenging thing to do. But that doesn’t mean that that global event doesn’t have an impact on pricing, and asset pricing and risk tolerance and all of those things. So we need to be thinking about all of that stuff, even if the models aren’t going to be thinking about.
Corey Hoffstein 24:13
So how do you think about balancing those sorts of long term forecasts with short term events and impacts like the corona crisis, for example, that I could imagine you could either view as sort of transient risks, or they could have a permanent impact and create significant asset repricing that you need to suddenly incorporate into your views.
Jim Masturzo 24:37
Yeah, so this is, you know, in my mind, when we think about crises, this is really the crux of managing other people’s money. This is what you get paid for is to deal with crises. And in a crisis clients and colleagues, sometimes they expect you to do something. Sometimes the best thing to do is to do nothing because what we know that when we act just to act, it usually turns out poorly for any of us. And there’s been tons of literature on this and books written about this. The book Thinking Fast and Slow is talks about from a cognitive perspective, why we act fast and how to tame that. But when we think about crises, we do it from the perspective of reprioritizing objectives is the way I term it. And what that really means is, as I mentioned earlier, our Investment Belief is in long horizon mean reversion, right? And so we’re not going to throw that away. We’re not all of a sudden, you’re going to become completely trend following momentum investors. It’s just antithetical to how we think. But we also realize that during crises, valuations are less important, on the way down, nobody really cares about valuations on the way up, they do on the snapback, they definitely do. So there’s still a place for those. And we’re still important to us. But what we do realize is that during a crisis, correlations go up, betas go up. And so we shift our thinking to think about it as targeting betas or durations volatilities, we become much more risk focused, understanding that that’s how asset markets are moving. Now, again, we’re not throwing out our valuation perspective, that’s still bled in there. But instead of being you know, if you want to think about it simply a primary driver, it becomes a secondary driver. And that way we can get both of those perspectives in there. But allow us to take advantage of what we know how about the differences in markets at
Corey Hoffstein 26:38
different times? So how does that ultimately boil down? What’s it look like for you in terms of actually generating trades for the portfolios? And is it okay, we need to plug these things into the capital market assumptions and see how that changes? What the strategic asset allocation looks like? Do you have other quantitative models that are being used for tactical trades on top? How does this actually play through the process? Yeah,
Jim Masturzo 27:03
so what we do is, even during all times, we actually create a whole bunch of different model portfolios. And I mentioned earlier, we use mean variance optimization for our signals. And that mean, variance optimization is not a single optimization, it’s actually a bunch of them, that you can think about a dozen or two dozen different model portfolios that we create, some of those are unconstrained, they’re going to be driven largely by valuation. Some of those are driven by different for lack of a better term benchmarks or reference portfolios, high risk, low risk, inflation centric, duration, heavy, whatever it happens to be, we build a whole bunch of these. And the value of that is when we see these particular trades that that come out of these, you can look across these different regimes or perspectives. If the system wants to overweight, a particular asset. In a bunch of different scenarios, we realized, hey, you know what, we’re going to buy this and it’s going to add risk to the portfolio, but it’s also going to protect us against inflation. And it’s going to do XYZ to duration or whatever it’s going to do. If you see a trade that maybe it’s only overweight in one of those model portfolios, you say, Okay, maybe we still want to do it. But let’s understand that we’re only getting one perspective out of that particular trade. So this is the information that we’re always thinking about. And we don’t have a pure straight through quantitative system, put in the inputs, generate the portfolio, trade, the portfolio done, we have human interaction in the middle, which is we run the models, we get model portfolios, like I just said, and this whole suite of them, and then we look at those to determine the trades that we want to do. And some of its based on the value of those trades across multiple regimes or the regimes that we’re interested in and at a particular time, what’s the other side of the trade? If you’re going to overweight, something? What are you going to underweight? I think that is much more akin to you know, or people talk about that. And from the perspective of long short, I’m going to be long in this and short that. And we think about it less sometimes, but it’s just overweighting and underweighting. But it’s the same trade. It’s just based on what your reference is an equal weight portfolio, or is it cash, it’s the exact same trade. So we think about both sides of that trade. Anytime we’re thinking about putting a trade on.
Corey Hoffstein 29:23
I personally love that way of thinking about portfolios, I always talk about it sort of portfolio construction is long, short portfolios all the way down. You start with your benchmark, and you just layer on all these long, short trades that ultimately net out to be your new portfolio. But I personally love that way of thinking. And it’s the way I think about tactical asset allocation. And for me, at least it informs one of the reasons why I think tactical asset allocation is really hard. So when you think about sort of a long short equity portfolio, let’s just say a value portfolio. Typically you have a couple 100 stocks on the long side, a couple of 100 stocks on the short side and ideally there are sort of netting each other out from an idiosyncratic risk perspective. When you start thinking about doing that, at the asset allocation perspective, say you’re long US equities, short international equities, it would seem like the trade is the same, hey, we’re long a basket of US stocks, short a basket of international stocks. But you then also incorporate all these unintended bets potentially geopolitical risks, currency risks are in there, you could have all these sort of sector and growth discrepancies that make it a really hard trade. How do you think about trying to control for some of these unintended bets in portfolio construction?
Jim Masturzo 30:34
Yeah, that’s a great question. Currency is by far one of the largest that often gets lost in the calling mainstream articles that you’ve written on this stuff. And someone’s like, look, I wrote this back test, and it did so well. And I to the trade, you’ve just said, I’m long, one and short, the other. And you look at it, and you go, Well, yeah, all you did was create a currency trade. And so hedging and not hedging is a huge part of this. But you know, when we think about it, you know, there are the quantitative aspects that you can just focus specifically on kind of the distributions of those assets. And what does that mean? I like to think about it. And again, this is back to the what are we missing from the economic perspective, a great example I like to talk about is long em short commodities, or, you know, do this trade, as I mentioned earlier, because we don’t drill into specific countries. But if let’s just say, long, Brazil, short coffee, that’s an interesting trade. If that’s something that comes out of the model, I want to ask why, as we know, Brazil is a huge coffee exporter. So that’s just a strange trade. What is the model, either finding or missing that makes you want to do that trade. And so we not trying to build models that capture all those unintended risks, because pretty quickly, you’ll end up with pretty black box type models that have so many different features, they’re not parsimonious anymore, you’re probably overfitting. This is why we think human intervention in the process is so important. Because there are just things that we know that we probably can’t teach the model to do in a way that’s not going to just completely bias the results or overfit the results. But we also back to the earlier point, we do think about these things from a parent’s perspective, because I think it’s just hugely important to add in context to what am I, what do you want to buy? It’s sometimes easier, what do you want to sell? But when you think about them together, that why is just so important.
Corey Hoffstein 32:41
I want to use a phrase you just said to tee off a question, which was you said you couldn’t put some of these concepts into the model without the risk of biasing the results. And one of the whole arguments for using quantitative models is to avoid the inherent human bias, the qualitative bias, how do you think about managing that qualitative bias and knowing that you are looking at these different trades, how they appear across different regimes or risk factors and actually making a qualitative decision with your team as to how to potentially apply them in the portfolio?
Jim Masturzo 33:12
Yeah, so that’s a great question. And the way we do it, is that it’s not really 5050. And then the way I think I’ve been talking about it may sound like we give equal weights to the model and to our own judgments. And we really don’t our models are think well thought out, consistent, based on a long term, relationships, all that stuff is hugely important. And the model in a lot of cases, you can think about it as getting more weight. What we do from a qualitative perspective, sort of building a quote unquote, quantum mental type strategy is, again, try to poke holes in the model. It’s what is it missing? And if we can’t come up with a good rationale for what it’s missing, then we’re usually just do what the model what we built the model to do what usually follow those traits. But in cases where again, that whole you know, long Brazil short coffee trade, that’s one that you just look at, and you’re like, Wait, that’s just a weird trait. I don’t understand that trait. And that doesn’t mean you kill it. That means you’ll learn more about it. You try to dig in and say, Okay, why why would you want to do this trade? big believer of if you don’t understand a trade, don’t do it. The time to learn about a trade is not after you put it on. I think sometimes people forget that. They hear about an idea and they put on the trade and then they go back later after it didn’t go well. And like wait a minute, why did I lose money when you didn’t really understand what you were doing to begin with?
Corey Hoffstein 34:41
Taking a step way back for a moment, something I probably should have asked much earlier on. Before we really dove into the weeds here. But how do you think about benchmarking these portfolios? How do you think about defining the objectives of what you’re actually trying to achieve through this conversation? It sounds to me almost as if it is just a risk adjusted total return that you are trying to be very risk aware. But you are looking for those opportunistic trades that have the higher probability of occurring and a low downside risk. But how would you actually think about benchmarking
Jim Masturzo 35:13
it? That’s right, we do have some strategies that have very explicit benchmarks. And then obviously, we’re following those in our other strategies that are trying to maximize called maximize real return or maximize the Sharpe ratio, we have a number of those strategies, there becomes much more back to the idea that I was talking about earlier, where we we create models based on a lot of different types of regimes, whether it be risk or inflation, or whatever, because ultimately trying to maximize real return, we need to take into account all these other things, because that’s ultimately what clients are going to look at you against. So when you have a benchmark, and you say, look, here’s my benchmark and how to do against that benchmark, it’s a little bit more straightforward. It’s more straightforward for clients who said, Look, you know, the portfolio underperform because the benchmark was down, and we were down, and maybe you got some slight excess return or whatever. So, but you know, if you’re down 10, and you’re down nine and a half, clients don’t really care about the 50 basis points about performance, they care about the nine and a half percent down. But it’s a little more straightforward to have that conversation, when you talk about over a cycle, just maximizing real return, you need to be thinking about different perspectives, because ultimately, whether we admit it or not, that’s how clients are thinking about it. In my experience, most clients whether there’s an explicit benchmark or not, they have a benchmark in mind, that benchmark might be the 6040 portfolio, it might be, if you think about a real return strategy, maybe it’s just something as simple as tips or commodities. It might be the s&p, it might be actually what it always is, whether we like it or not, it’s the highest returning asset over the previous however, many periods. But that’s life. That’s the industry that we’ve all chosen. So we need to be thinking about all of those things, too.
Corey Hoffstein 37:01
So I find this idea of building all these different, for lack of a better word, almost regime based models really interesting. And I want to put words in your mouth. So correct me if I’m wrong, but it sounds like you have these different regime characteristics that you either build constraints into your optimizer or certain different types of targets into your optimizer, you come up with all these portfolios that might be appropriate for a given regime. Walk me through then how you think about either combining them or finding the trades across them and applying them to your strategic asset allocation. drill into that a little bit for me.
Jim Masturzo 37:32
Sure. So like you said, we run mean, variance optimization, and we do resampling and shrinkage and all that sort of stuff to produce better results for lack of a better term. And some of those optimizations Have you can think about is risk aversion to particular benchmarks or explicit constraints, tracking error budgets, that sort of thing around particular benchmarks. So then we have all these models. And there’s a number of different ways that we put them together, we have a set of signals that we look at some of those are business cycle type signals, where are we in the business cycle will inform where we want to which of those models we might want to lean more heavily on or away from, we think about it from we also have, again, some human interaction human insight into what we think we want to do, although that’s, again, you can think about it as secondary to some of the signals that we’re thinking about. So often, it’s the business cycle, sometimes, again, during crisis periods, where the business cycle is secondary to, you know, again, a beta based approach or volatility or risk based approach, then we’ll move to other signals that we have that are much more focused on risk. So there is a little bit of the art and the science of all of this. And the art is in putting these things together and choosing you know, which signals we want to use at different times. We rarely ever, when I talk about switching between things, we rarely ever wait things at zero. Usually everything we create has some amount of value or some amount of impact on the final portfolio, because it’s a very big decision to say, Look, I’m just gonna use an example. Inflation doesn’t matter to us. So we’re going to weight that portfolio at zero, that’s a big step to take. Now, we may downgrade it from its normal weight or up way to wherever the case may be, but but rarely go to zero
Corey Hoffstein 39:25
when you’re thinking tactically, is the signal more in what the portfolio is holding those different regime based portfolios or how those portfolios change? Because I would have to imagine something like an inflation sensitive portfolio is naturally just going to have a lot of tips and commodities, potentially and emerging market economies. It doesn’t take an optimizer to sort of tell you that is the signal in when something interesting, unique different pops into that portfolio, or is it how those things are weighted for the sort of risk return opportunity that they’re generating you
Jim Masturzo 40:00
That’s more than the ladder. It’s more how they’re weighted the portfolio that you described, and you’re absolutely right, the major holdings in a portfolio like that don’t change all that much. Now, the weightings will change a little bit, again, based on valuations, because we still have a valuation impact in everything that we do. So the weights will change. But by and large, the portfolio is what it is. So what we’re looking at is our belief in the need for that particular regime over the coming whatever time period we’re talking about a horizon so our signals to wait those portfolios and you can think about we just think about the business cycle. So you know, the business cycle and correction periods or bear markets deflationary periods, like we’re in right now, the risk of inflation over the short term over the coming quarter or so it’s pretty low prices are falling commodity prices are falling oil has rebounded a little bit from its lows. But by and large, what we’re seeing is broadly deflationary and we can have an argument over does the BLS prep correctly measure inflation and, and all of those things, we’re putting all that aside on inflation type portfolio is not something that we need to be as concerned with, again, over the coming quarter as we normally do. So we’ll look at and waiting these different portfolios to down wait that type of a portfolio, or at least that’s what we have been doing now. We may switch that going forward. But that’s what we’ve done over the last couple of months, I always
Corey Hoffstein 41:26
feel like some of the best lessons learned come from when something doesn’t play out the way you would expect it to. Are there any examples of some of these qualitative overlays that you guys have applied historically, that maybe didn’t play out the way you expected?
Jim Masturzo 41:40
I mean, there’s a number of times where we’ve looked and said, We think a particular you know, metric is undervalued, or we should overweight or underweight a particular assets. Sticking with inflation, a couple of years ago, we looked at breakevens had been, we’re down at one and a half percent at the time, which is pretty low, given the last 10 years and kind of where they’ve been going again, breakevens are usually called one and a half to two and a half percent most of the time, or at least over the last decade or so. And we looked at the bottom of the range and said, Well, you know, this doesn’t make a lot of sense to us. And we think we should overweight, our inflation protection. I wouldn’t say that trade was terrible, it was pretty sideways is just kind of a waste of time, by and large. And, you know, when we measure those things, and this is an important point, too. We don’t look at what the just what the trade did, not what we bought going back to the pair’s ideas, or what did we sell? What would we have held if we didn’t do the particular trade, and that long short pair becomes how we value or validates the success of a trade. So the trade that you know, this trade in particular traded, it was down a little bit, I don’t remember the exact amount, but let’s just call it 20 basis points, or whatever, over the year and a half that we held it. But it was basically just a sideways trade, we spent a lot of time thinking about we were smart about a particular trade, and it just didn’t really do anything
Corey Hoffstein 43:09
putting a trade on, it’s just half the equation. What’s the catalyst for exiting one of these trades?
Jim Masturzo 43:14
That can be a number of different things, the first thing we do, and again, a lot of our trades that we’ll put on, especially when they’re outside the model are overrides, is to go back to the model and say, Well, why did they miss this? If this is a persistent type, or something that we think could be persistent in the future or a particular environment that could come up? Well, we want the model to have that embedded to the extent possible. In a perfect world, we want the models to tell us exactly what to do. We want to be able to build models to do all of this stuff. And as I mentioned earlier, that’s not always possible for parsimonious reasons we’re fitting in all of that. But to the extent that we can look the model and say, Okay, here’s why it wasn’t captured, do we want to change the model. And oftentimes, we do identify parts of the model and say, Look, we can actually change the model, and in which case, it becomes less of a no longer a qualitative trade, it becomes now a model based trade, and the model tells us when to take it off. That’s the best case scenario. In other cases, we look at the model and say, Look, you know what, we’re not going to build this in or we think it’s more idiosyncratic. We’ll define review points for the trade. This might be quarterly or semi annually annually. Usually after a year, we’ll want to know was this tray is it doing what we wanted? If it’s not, why not? We want to keep it on. We want to take it off. But we’ll review you can think about it as quarterly. Sometimes it’s more frequently sometimes it may be less frequently just depends on what the trade is. I have
Corey Hoffstein 44:45
to imagine a scenario like what we’re going through currently, which feels a bit open ended, lends itself a bit more to the second aspect, the ongoing review where it’s difficult to say whether we I mean, the market certainly seems like it’s current Aren’t we pricing in a V shaped recovery? It’s hard to say whether the market will continue to price in a V shaped recovery whether the economy will actually have one. How do you think about navigating that sort of environment in real time where there is so much uncertainty as to how things can unfold? This is
Jim Masturzo 45:16
a reflection call it unprecedented, but it feels unprecedented. Sometimes, you know, when you’re in the middle of all these things, they always feel unprecedented. Until you go through the next one. Remember, the global financial crisis we’re live, we felt like this is unprecedented. And now here we are a decade or so later. So we things like now, and especially if we think about middle of March to the middle of April, markets are moving all over the place. And we were, as most managers in this industry, you’re in your portfolio, like not to say that we’re not in the portfolio during regular times, but you sit there and you stare at those positions, and you really have to understand that the next day, the market may move 3% against you might, the day after that it might move, you know, 2% in your favor. So there is that trade off of, as I said earlier, there’s a motivation to act all the time. And you have to fight that. But you also have to be willing to move the portfolio when you believe that it’s the right time to do it. Everything we do is really group oriented. It’s very rare that one person unto themselves makes a trade or alters the portfolio. So we have a pretty rigorous process to review changes to the portfolio. And by rigorous process, you can think in debates. And if you want to propose a particular trade, and even if it’s trades that the model wants to do so I can step back and say once we get our portfolio, we put our models together, we have our overall model portfolio, I take those trades, the trades that the model is flying, and we take them to our Investment Committee group, and we discuss those, and we debate those, and we try to poke holes in them. And so it’s a pretty rigorous process. Now, a time like now, when markets are moving all over the place. There’s a lot more of those. And so we want to be if we need to review the model daily, if we need to be debating trades daily, we did that doesn’t mean we were always trading that every day. But we were debating trades every day. And I think that’s necessary for navigating these times. As far as the V shaped recovery goes, I tend to be more on the you know, I think like a lot of folks surprised that the market continues to go up as it has. But I’m also a believer that the market is right. If they’re the buyer, and there’s a seller at a price, then we can believe what we want one, whether that’s a smart trade or a bad trade, but the market level is what it is. So we need to be thinking about is, again, what are we missing? Why do we think that’s wrong? And creating that probabilistic view of the future? And trying to poke holes in it? Because the market is all of us. And for those of us that think that things are overvalued or undervalued, we’re definitely in the minority. So let’s figure out why that is,
Corey Hoffstein 48:20
how do you think about dealing with an event like this? That is, in some sense, an N equals one event, this is an idiosyncratic event, but one in which we are likely to see another global pandemic in our lifetime. That’s just the reality of having a global market environment where there’s a lot of population movement, it’s just likely to happen again. Certainly, it doesn’t seem like necessarily something you would build into the model. Is this something that you might say, well, perhaps this is another regime, we might consider a regime type? Or is this something that you would always expect to be more of a qualitative type overlay that sort of adjust in time action that you would think about taking?
Jim Masturzo 48:59
I look at this, and I agree with you, I doubt although I hope I’m wrong, that this is the last pandemic that we see. I’m not even sure it’s the last COVID-19 occurrence that we’re going to see. What I think is more interesting is really around the response. You can think about your monetary and fiscal response. And so how is that going to change we saw in the global financial crisis, there was a ton of pushback on the Fed and Treasury for the stimulus that they did. The first tarp bill got shot down. And then the second one passed, and immediately it was well, that’s wasn’t even close to big enough. It’s just getting people used to trillion dollars or whatever it was, I remember they kept it in like the 950 billion range because nobody wanted that T number out there. And now here we are, 10 years later, and we’re into you know, 2 trillion 4 trillion 6 trillion. Some people if you annualize that, it’ll say well, it’s 20 trillion, whatever the number is, these are big, big numbers. And so, but I think it’s more interesting is now the response has become something bad happens and we need stimulus. And every time we give stimulus, it seems like it becomes easier to give more stimulus. What is the ramifications of that? If you’re on the MMT side of the argument, you would say, of course, and we should have low unemployment, the government should ensure low unemployment and infrastructure projects and all of that, maybe that’s right. Again, I’m not in the business of trying to think about what’s right or wrong. from a social perspective, I have my opinions, and that’s a different conversation. But if that is the case, what does that mean for asset prices? On the other side, you have the dollar bulls, who might say, look, the dollar keeps getting stronger against all other currencies? What’s that going to do to non US dollar denominated debt? Does this turn into a solvency crisis? What does that mean for markets? So those are like extracting away from COVID? Or whatever, which the next was housing last time this COVID? This? The responses are the same. It’s the same playbook just bigger. And so what does that mean? If every crisis is going to be more stimulus? What does that mean for asset prices? You can very quickly create a story for a V shaped recovery and higher markets or higher s&p, I should say, with a lot of stimulus,
Corey Hoffstein 51:25
what does that mean for model construction? And so all of a sudden, you have to build monetary and fiscal response as a variable into the way you think about building tactical portfolios.
Jim Masturzo 51:37
That’s one way that you could do it. Another straightforward way is just to say the way we build models is to say, look, Shiller P E ratios of 28 might be the new norm, forget about the long term history 16 or 17, or even the 30 year history at 24, there’s going to be propping up of asset prices. And again, that’s fine, that just the level up and that allows us to invest appropriately. So even though we focus on what are the causes of these things. There’s also on the flip side, what is the outcome, and if the outcome is something that we can very specifically fit into our models, that’s a much better approach, because we’re not building in, again, sort of large, maybe not more parsimonious, maybe overfit models that have unintended side effects and consequences. So we try to identify very specific places where we can enhance our models consistent with what they’re already doing, because we’re comfortable that what they do is the right thing. I mean, I should step back, and I’d say the right thing. It’s what we want them to do. That’s what they were intended to do.
Corey Hoffstein 52:47
As you look forward. Well, what are you really excited about? Ah, that’s
Jim Masturzo 52:51
a great question. I’m excited for what comes next. I think we’ve all been through these last couple of months and sheltering in place and all of that stuff. And I think we’ve gotten to that point where it’s become a little bit draining, but we’re in a pretty exciting time right now. This is again, what we in this industry, I think get paid for it’s to deal with crises in the best way we know how for our clients. So we’ve got unprecedented central bank activity, we’ve got a policy led global shutdown, I don’t think it’s ever happened. Is it a depression? Is it a recession? I have thoughts on that. But we’re gonna have to go through that. And it’s not to when I say excited, there’s a lot of people out there who are struggling right now. So I don’t mean excited from I’m happy that this is happening, but from intellectual, how are we going to navigate the future? This is a pretty exciting time, and trades are gonna go against you. Sometimes markets going to react the way we don’t want and that can be demoralizing, or it can be difficult, right? You have to enjoy what comes next and really focus on sort of keep a forward view. Otherwise, it’s very easy to kind of capitulate at the wrong time, or those sorts of things.
Corey Hoffstein 54:12
Well, I can’t think of a better way to close out the episode than with that. Jim, I can’t thank you enough for joining me. This has been a an absolutely fantastic conversation. Thanks, Cory. I really appreciate it.