240 Systematic Investor Series ft. Nick Baltas – April 22nd, 2023

[00:00:00] Intro

You’re about to join Niels Kaastrup-Larsen on a raw and honest journey into the world of systematic investing and learn about the most dependable and consistent yet often overlooked investment strategy. Welcome to the Systematic Investor Series.

[00:00:24] Niels

Welcome and Welcome back to this week’s edition of the Systematic Investor Series with Nick Baltas and I, Niels Kaastrup-Larsen, where we each week take the pulse of the global market through the lens of a rules-based investor. If you’re new to the show, I hope today’s episode will pique your interest and encourage you to explore our back catalogue, including my conversation with Andrew last week, where we discussed some of the recent challenges that his CTA replication strategy has faced and what surprised him about this recent period.

Also, I would encourage you to listen to the midweek episode where Stephen Roach, former Chief Economist at Morgan Stanley and Senior Fellow at Yale, discussed the China-US relationship and his new book Accidental Conflict, and where he offers a path towards healthier economies and insights on how to repair the current “rocky relationship”. Lastly, I am excited to invite you to enjoy the mini CTA series that I’ve done with Alan where we speak to the top decision-makers at the largest CTAs in the world so that you can gain unique access to these industry leaders who share valuable insights on pressing topics in the CTA space.

So head over and check them out after you’re done listening to myself and Nick today. Nick, it is wonderful to have you back this week. It’s been a little while. How are you doing this Friday in April?

[00:01:47] Nick

I’m doing well, Niels. It’s good to be back. I’m doing well. Post-Easter, it’s a less travel-heavy schedule on my side. Spending time doing some research projects. I think some of them we’re going to touch upon. Other than that, life is good. How about you?

[00:02:05] Niels

Fantastic. Yeah, absolutely. I think we both concluded recently trips to Asia and coming back from that, so our body clocks are getting back to normal. As wonderful as it is to be out there, it’s also quite nice to be back home for sure. We’ve got a pretty exciting lineup, because again this week, and because of your unique insights to what we do, we’re going to tackle some issues that we normally don’t (and also from an academic point of view). So I’m very much looking forward to that.

Now, normally at this stage, Nick, I would of course come up with some kind of brief market wrap, but I thought it’s Nick, let’s do something different. What I thought was just to quickly look at some of the news flashes that I’ve sort of noticed this week and just hear your thoughts on at least some of the topics. I actually think some of the topics don’t require much of a comment. But the first one might, actually. I hope you’re up for this, Nick. Is that okay with you?

[00:03:04] Nick

Why not? Let’s go for it.

[00:03:05] Niels

Why not? Let’s go for it. Exactly. So the first one comes to us from Bank of America. If I should summarize what I read, it would be something like investor allocation to equities relative to bonds has reached its lowest level since the global financial crisis driven by concerns about the possible recession according to Bank of America’s global fund manager survey. Bond allocation has risen to a net 10% overweight, the highest level since March 2009, with 63% of participants expecting a weaker economy. However, the bearish sentiments could be a contrarian signal for risk assets with some strategies suggesting that if recession fears are not realized, bond yields and bank stocks could rally.

So I’m kind of curious to hear when you speak, because you speak to a lot of institutional investors if this kind of overweight of bonds and these thoughts about asset allocation at this point in time is something that you sense that many of the people you speak to are also doing at the moment?

[00:04:10] Nick

I would possibly comment on this one from a slightly different angle, that being, what is the short-term forecast? Obviously, it’s very hard to foresee what’s going to happen, but there seems to be a number of indicators suggesting that a slowdown is more likely than not, possibly towards the later part of the year than in the very short-term. In this context, obviously, the question about being slightly more defensive and risk-off pops up. The question along the lines of is a fixed income allocation (my long-term defensive play setting aside last year) going to see another case or not?

But having said that, I think fixed income allocations have seen some form of increased allocation appetite. Conscious, however, of the substantially high 9and I think you have commented on this show\) interest rate volatility we have experienced in March, I think any move that we see from the investor side is, I would say, relatively more balanced. But I have to say that following March (I think last time we were here was before the interest rate spike in March), the beginning of the year was much more on the soft-landing kind of scenario being the more prevalent one. In this regard, it was more of a mild risk-on type of appetite. I would say post that event, it is kind of balancing out income-seeking strategies versus a bit more defensive, if you like.

[00:05:45] Niels

Yeah, I think the reaction, because I see it as well (I have it certainly in my conversations, for sure), that investors are suddenly becoming much more interested in fixed income again. But from my observations, it’s just this thing that yeah, this is a very rational thing to do. If suddenly you can get 5% on a one or two-year duration, that makes sense. But the only thing I would caveat that with and that is, I think a lot of these people are basing it on what’s been going on for the last 20, 30, 40 years where we’ve had this tailwind when it came to lower interest rates, meaning the interest rate cycle, the 40 years people talk about, was there.

But on the other hand, if the 40-year interest rate cycle flipped a couple of years ago and we’re now in the other part of the cycle which is interest rates heading higher for a very long time, then these normal reaction patterns are not going to play out necessarily as well as investors are hoping for. So to me, it is quite interesting because I think a lot of people who are doing this, their portfolios are not necessarily in great shape after last year.

Certainly, in the pension fund industry, we saw huge losses. So if they’re again jumping into something that looks like, oh, now is a great time to own a lot of fixed income, the next set of pain can be much worse than what we see now. Even though we don’t have to go down that route, I still think that the pension fund system is something that I worry a little bit about at this stage.

[00:07:20] Nick

I would not disagree with you. I wouldn’t disagree with you.

[00:07:24] Niels

Well, shifting to some other very exciting news, Nick, this week, we learned that the famed investor, also known as Taylor Swift, knows a lot about due diligence compared to many of the Wall Street pros. As it turns out, she asked a very simple question when FTX were offering her about $100 million dollars in sponsorship and she simply asked, “Can you tell me that these are not unregistered securities?” So good due diligence really does matter, and after all, having multi-talents can keep you out of trouble. So I’m not sure you have much of a comment to that, but I thought it was an interesting piece of news that came out this week.

[00:08:00] Nick

I didn’t see it. That’s the first time I’m hearing that.

[00:08:03] Niels

Okay. All right. Well, let’s go to more serious matters because just as I was preparing for our conversation we had in the US latest news from the Conference Board Leading Economic Index which measures the US business cycle that unexpectedly dropped sharply to its lowest level since 2020. It’s the 12th consecutive monthly decline, the longest run since the period 2007 to 2009. It’s not the most positive precedent, of course, yet we still have the CME FedWatch Tool showing a nearly 85% chance of a 25-basis point hike next month when they meet in May.

Equities are doing fine at the moment. So it’s kind of interesting, I think, the dynamics we see right now in the markets that the numbers are telling us one thing; people are still optimistic when you look at the equity markets; people are expecting, as we talked about before, recession and low interest rates and probably not the best economic environment on the bond side. So what are your big thoughts on these things?

[00:09:14] Nick

It’s fair to say that this year has kind of evolved month on month on month with different anticipation and opposing feedback in the equity market. I think the year started with lukewarm debate as to how the year is going play out. January was very strong, and then coming into February, sentiment turned more positive, but then equities underperformed. March obviously came with a significant rate spike. One of the indicators I was mentioning in the beginning is certainly the Leading Economic Indicator. It is one of the few that we’re using to monitor the business cycle.

By the way, we’re using some sort of a moving average to smooth out those changes month by month and therefore capture more like an evolving regime. But just to give you a few data points, since you mentioned this indicator, we started last year, from memory, from a rising growth, rising inflation standpoint. I think towards July, the moving average on the growth side drifted downwards. As you said, it’s the longest period that we have seen that fall. On a moving average basis, that was captured back in July.

So from July onwards, it was more of a falling growth, rising inflation, based on some sort of definition of indicators. But ever since November, we’re in a falling growth, falling inflation phase. Now, in anticipation of another hike, I think certainly before the rate spike in March, the anticipation was that the hike could be between 50 and 75 basis points. I think what happened during that week of banking sector turmoil (at least softly) can be seen as an implicit hike. That was a time, I believe, that there was consensus for this 25 basis points we saw, but I think the range was pretty wide between minus 50 and plus 50 or something along those lines.

So there’s a significant amount of debate and I think the asset class movements we’re seeing I think are partly driven by certain flows rather than strategic allocations. That’s my assessment. I don’t have obviously hard data to support that. But that’s how I perceive the short-term moves.

[00:09:24] Niels

Okay, that’s pretty interesting. The final thing I wanted to bring up with you just to round things off before we dive into all of our own topics, so to speak, is just also related to what’s going on in the bond markets. Because I read another article this morning saying the bond market is starting to signal concern about the debt ceiling as traders are swooping in on the one-month treasury bill because of growing concerns about Washington standoff over the government’s debt ceiling briefly pushing its rate down by more than half a percentage point on Thursday, which was yesterday, which I completely missed myself.

[00:11:27] Nick

Likewise, actually.

[00:12:03] Niels

Volatility in short-term fixed income is very interesting. You cover a lot of different strategies. Do you actually cover strategies that specifically operate in that area as well?

[00:12:18] Nick

You mean the short-term interest rate?

[00:12:20] Niels

Yeah. I mean just volatility or short-term interest rates or is it more broad?

[00:12:24] Nick

We do have some activity on the volatility side as well and we certainly look here now into more relative value type of dynamics in the present market environment. In other words, looking into which market. We’re not just talking about which market and which asset class but also which asset classes are priced expensively, which are priced relatively less expensive and therefore looking into some sort of relative value dynamics.

I think the equity rates interplay in the present moment is quite interesting, with your point, significantly elevated on the interest rate side, not as much on the equity side. If anything, fixed is now at pretty low levels. So yeah, this is something we’re looking at and this is something we’re trying to harvest systematically, while at the same time, highlight some sort of tactical entry point. I think this is probably an interesting one.

[00:13:19] Niels

All right. So before we do our little quick update on trend, I wanted to touch on the topic of trend and CTAs because, as we talked about a little bit last time, when we were just getting going with our initial publication of the episodes we have with these wonderful, wise people from the CTA industry. So I was curious now after a few weeks, and we’ve published most of the ones that we wanted to publish. We’ve got a few more. I’m sure we’ll do some on occasion, but not at the same rate.

I know you probably haven’t listened to all of them. But what you have heard so far, are there some things that have kind of stood out to you, maybe surprised you a little bit, or maybe something where you thought, well, that’s a different way or a new way to talk about certain things that we obviously talk about a lot?

[00:14:17] Nick

So I’ve actually heard most of them. I don’t think there was anything very surprising. But there are certainly a few things that I would easily highlight. I will start possibly with, I don’t know, more of the humorous part. When Roy Niederhoffer basically said I started investing in 1981, I’m like, dude, that’s when I was born. So it kind of brings a bit of perspective. On a more serious note, I think listening to people like Bruno or Roy or Doug, obviously the focus is on a more traditional, liquid trend following pivoting towards very short-term, possibly contrarian allocation and then towards alternative markets.

What becomes prevalent here is something I think we’ve discussed many times and I think you have discussed many times that it looks like trend following or some form of systematic investing in futures contracts is easy, but then you end up realizing how many nuances are in place specifically when you try to accommodate either more illiquid markets or alternative markets or short-term patterns. So as easy as it might be, I think there’s a lot of art and science going into it and a lot of research is going into it. I think it’s pretty much highlighted specifically once you start hearing people that have very different yet related mandates.

Their perspectives were so different, but at the same time, under the same hood. So that was the second thing that I was going to point out. In terms of philosophy and the way I’m kind of thinking about portfolio construction and running a strategy, I think Bruno’s points were pretty close to how I think about trend following. But I have to say, the whole discussion he was making about doing relative trends and looking into spread forecasts between market A and market B and therefore allowing them to largely multiply the number of assets they allocate to, that not being any more, just the markets, but any pair, it was an interesting idea. I’m still kind of running through my head, what’s the value out here? I’m not saying that in bad faith, I’m actually saying it in good faith.

I think the last point I would say, again, closing on another slightly less but possibly still humorous point is that when you asked Bruno what’s the one thing that you hear about trend that you disagree with, he basically said exactly the same thing that I said two months ago. He just disagrees with the fact that trend following is dead and every now and then people come along with excuses about its viability. So that’s a number of high-level points I was going to share.

[00:17:06] Niels

I appreciate that. When you hear the debate, for example, between funds that trade more liquid developed markets versus funds that have really embraced alternative markets, and as you say, off-exchange, difficult to trade, etc., how do you think about that in your shop? Do you embrace the idea that we should just trade as many markets, it doesn’t really matter whether they are on-exchange, off-exchange, and that there is some benefit to doing so? How do you think about that?

[00:17:44] Nick

Diversification is the key dynamic here in terms of motivation. So to the extent that we can believe there is some form of trend following, then by all means. To the extent we can have assets with enough liquidity, there’s a diversification element that comes along. But I just said the most important consideration, that being liquidity. Therefore, if inclusion requires a significantly reduced exposure precisely because liquidity is not high, then the marginal gain at scale is very small. So I think it’s always a balance between, can we scale it up and to the extent we can scale it up, I think diversification will play out.

Obviously, the rule of the nth asset that you’re adding has consecutively marginally lower contribution to your program is fair. I think that’s why I could possibly see the value of just focusing on those markets. I can possibly call Doug here and say, okay, if I focus on those markets and I don’t necessarily blend them with more liquid markets, then possibly, at least in that ecosystem, I can have value brought by every element in it and then I would let the allocators decide upon how much they would allocate into that bucket as opposed to just blending them together on a full look through base bottoms-up allocation.

So I can see the benefits on either side, or if you like, the disadvantages of just going into those markets. But we certainly look into those markets and we certainly look to deploy non-conventional markets (let me put it that way).

[00:18:37] Niels

Yeah. Final point before I jump to the next segment because we spoke before the SVB event. You’re obviously sitting in a huge shop each day. I was wondering how you guys think, I’m not going to ask you to talk for all of your firm but just you looking at all these markets and the reaction and so on and so forth. What I’m trying to say here is that we’ve talked, for a while on the podcast, about one of the concerns being liquidity, but I think often we think about liquidity from our own little niche saying, well, can we buy and sell the futures we want to buy and sell?

But this was, of course, another kind of liquidity event even though it did spill over to part of our portfolio, because clearly, the moves in the short-term interest rates and in bonds were big. So there were some slippages, I’m sure, somewhere in the system for managers, without a doubt.

But have you thought more about not so much whether it was SVB or Credit Suisse or whatever, that’s not what I’m trying to say, but just the event itself, what actually took place in the markets, this kind of drying up of “one form of liquidity” so quickly and what the loss of trust and confidence means to markets? Has this given you anything where you would go back and say, yeah, we need to revisit some of this stuff, because clearly, things can move much quicker than maybe anyone thought they could.

[00:21:04] Nick

So I think my answer would be a bit more generic here in the sense that we’re very, very focused on liquidity and execution and we’re very focused to making sure we deliver strategies that have no market impact precisely because our duty, if you like, from an index provider perspective, is to make sure that we can deliver to our end clients in the cleanest form whatever the premium is of question without that having any impact in the market.

Therefore, from the design phase, we want to make sure we constrain any trading in markets that are very liquid but to proportions of daily volume or daily activity that are in single digits or even smaller. Obviously, when liquidity goes through some phases and therefore we might feel that something is changing the market, we regularly review, and possibly we have to go back and rethink how we think about liquidity and utilization. But if anything, we started from the point of designing strategies that, at least on historical basis and also in stress environments, participate too little to even be meaningful.

With that being in mind, we try to evolve our product without having liquidity shocks being dominant on our day-to-day. We start from a place that we want to hedge against the possibility of having any impact in the first place. Should the market shift, we have to go back to the drawing board.

[00:22:39] Niels

Yeah. of course, a lot of people will be doing that as well. I think, for me what the surprising thing or what I think might be surprising to some people at least is that this event kind of happened in the most liquid markets that we trade. So it kind of gives you a little bit of a cause for real deep thought. Right. Exactly. Because if it can happen there, it can happen anywhere.

Anyways, let’s not go down that path anymore. Quick update on trend so far in April. We’re kind of a little bit past the halfway mark, and it’s looking pretty solid, I would say. I think this week was okay for trend followers.

So far in April, I think it’s also doing reasonably well. But interestingly enough, some of the markets that have actually done well for at least the trend following systems that I follow are things like sugar. They’re having a very strong trend at the moment. Even some of the meats and some of the grains are doing pretty well. Of course, we know that equities are still finding their way higher. So that’s been beneficial to trend following portfolios at the moment. For some, at least, depending on timeframe, the fact that interest rates have started to go back up again slightly for those who kept their short positions, that also has seemed to have worked out well.

My own trend barometer finished yesterday at 39, so it’s still pretty neutral. Nothing really to report there. Performance wise, BTOP 50 is up about 42 basis points this month, down about 3.75 for the year. SocGen CTA up about 1.14%, down 4% for the year. SocGen Trend up 1.7% for the month, down 5.7% for the year. The Short-Term Traders Index down about 10 basis points for the month and down 2.2% for the year. Equities up for the month a little bit, bonds down for the month a little bit, and that’s about it.

All right. As usual, I was about to say, Nick, whenever you come on, we get a question from Tim. So Tim is back with a question and so I will do my best to present it to you.

Tim writes, “Thanks for all the podcasts that you are churning out with an impressive speed and high quality. This week, I have a question relating to timing of trends. Here goes. Trend followers are known for being patient. They know from experience with their backtest that they need to be willing to work through environments where they incur small losses and see drawdowns take place. They do this because they know that, ultimately, strong trends will emerge and the profit from those trades more than make up for the losses and lead to phases of strong positive performance, as we saw in 2022.

Most, if not all, trend followers tend to firmly believe that there is no way to time trends. With Nick having a strong research background and perhaps not being 100% trend follower, he might think differently and would no doubt have some valuable insights into this. For example, some concepts that provide constructive results in backtests for timing trends are, one, using the trend strength metrics like Niels’ trend barometer, or two, a risk-on risk-off measure, or three, using some macro variables like inflation. All these approaches can be used to improve on overall results by adjusting the capital allocation depending on the results of the concept used, but they do lack in overall robustness. Of course, position sizing and whether this is done dynamically over time will also impact results. I apologize for the long-winded question, but very much looking forward to the response.”

No need to apologize, Tim. So yeah, your thoughts, Nick, on this particular topic.

[00:26:30] Nick

Thank you, Tim, first of all. Can we time trend? That’s I think the question here. Can we time trend? I would possibly even paraphrase the question and effectively say, should you deploy more risk or less risk in your trend following program? I think you and I, Niels, have discussed it or possibly have shared some comments back in the days on this topic. I think about two extremes here. The two extremes are if there is no trend in the market, what do you do? If all the signals are too low but you still have a particular volatility budget to deploy, do you optimize the noise?

The other extreme is that everything is trending so much, what should you do? Should just double the risk or not? I think these are the extreme questions. Trying to work backwards, what are the repercussions of us deciding upon reducing or increasing overall risk in those two extreme scenarios possibly can hint towards what could be a good indicator. So I’ll tell you some of the ideas we’re kind of brainstorming and looking at. I would basically say that when… I’ll start from the second extreme, that being that all markets have very strong trends.

If all markets have very strong trends, it is very likely that those trends are driven by the principal component, let’s put it this way. So if all rates are moving in tandem, it is more likely a core duration component driving everything, and therefore, the overall correlation is quite high but the dispersion across this trend is quite low. So what’s the risk here? Well, the risk here is that this dominant component kind of reverts and then the entire portfolio reverts.

This is very different to saying, hey, this one market has a very strong idiosyncratic move and now it becomes more of a risk concentration discussion in that market and mitigation potentially of that concentration if we want to taper that signal in the optimization. But if the entire ecosystem experiences very strong trends, it might be a signal and historical results kind of suggest that this could be the case that being a bit more nimble with the overall risk seems to pay off. Now, again, it’s a very extreme scenario.

But if I were to think about what is a warning signal here, that warning signal is that post very strong trends across a range of markets, we should be thinking about being more nimble, I would say. Conversely, if the signals are too little, is it worth reducing overall risk? Here’s a good debate because reducing risks, it would mean that it will take a bit of time to capture emerging trends. So that’s the counter. The for statement is that of course you’re optimizing on noise.

It’s unclear to me now whether what I’m suggesting is more of a tandem-shaped type of a vol profile or possibly like a decreasing profile with a single strand or possibly just acting on the extremes. I think that’s something that could or it is likely from a statistical perspective to be valid. I guess to the choices that Tim made, I think that’s closer to your barometer in a way, because that is almost like an estimation of the overall trendiness. I’m less convinced in the sense that risk-on risk-off is something that I think trend itself is acting upon obviously as a function of its own speed.

So if the market is turning into risk-off mode, that’s the time that you should possibly not cut your overall risk. We have discussions with regards to how we should be looking into trend following in a risk-on mode and whether we should be allowing all those equity moves, those being positive to be deployed, or let other pockets of the portfolio deal with that. But I think that’s more of a utility function discussion rather than a timing discussion. I think the last point I would say, because Tim is mentioning macro variables like inflation and so on, we’ve done some work at some point looking into how economic regimes and trend following interplay.

There’s a number of papers, I believe, in the last couple of years that have been circulated, mainly industry papers, kind of connecting trend following performance with rising or high levels of inflation. That is obviously contemporaneous. There’s no forecasting ability. I think the reason why I don’t necessarily think that macro variables can have a strong impact on timing trend following itself is because the historical performance or the signals themselves suggest that the asset is performing well in whatever regime we’re at or possibly not performing in the regime we’re at.

In other words, there’s a regime classification without even us giving it a tag, just purely by looking into the evolution of those signals. If equities are performing well, possibly we’re in a rising growth environment. If bonds are underperforming, possibly we’re in a high inflation environment. So somehow, there’s a subtle regime identification coming straight from the signal itself. That’s why I’m a bit more reluctant thinking about the micro regime itself, which by the way, can go against what Ooi said the other day, from AQR, talking about macro trend and so on and so forth. But anyway, I’ll leave that on the side.

[00:31:59] Niels

Yeah, it is interesting, because clearly, AQR told us about the fact that they use, I think, almost half their signal generation now is what they call economic trend, and not just pure price trend. So yeah, we shouldn’t rule any of these things out. I do agree with you. I think macro indicators can be very hard to put into a trend following perspective because once you are able to detect that inflation is high, it may be peaking and so you’re kind of adjusting it at possibly the wrong time.

I think from my perspective, and this is a general comment to Tim, is that, to some extent, I do understand the urge to try and improve things. I’m not saying that what you suggested here, Nick, that some things you do need to take into account. Of course, you can go old school and say, no, I’m not going to focus on changes to volatility, I’m not going to change focus on correlations, etc.

We know for a fact that that still works. The track records are there. People can go and look at those managers and it’s there. I’m not saying it’s better. I think they will say it’s better. I don’t think it is. That’s my view because the data doesn’t support that. But it could be better five years from now. Who knows? So we have to be open minded about it.

But in general, whether you use one methodology or the other, you’re still dealing with a strategy that tends to deliver about 150% to 200% of the S&P 500 returns with about the same drawdown, so to speak. Now, if you’re using an index of CTAs, the drawdowns are a lot lower. But even for a single manager, so to speak, obviously, we can have and have had in our long term track record, we have had similar drawdowns to the S&P, let’s just face that. So I am always curious that if you can get more returns with about the same drawdown risk, etc., why we are so focused or fixated about trying to improve upon that.

At some point, we should just say, this is fantastic. This is great. Because of course, the beauty of those returns that CTAs generate are completely uncorrelated to equities. So it’s actually a very useful return stream, let alone, it’s better. But anyways, I understand the question. I will say my trend barometer was not built in any way, shape, or form to be used as an indicator.

But I will say, even though this was done almost 20 years ago, believe it or not, it has an uncanny way of capturing the trendiness of our industry, meaning what I look for is where does it finish at the end of the month, and it is very, very highly correlated to the fact that if it’s outside its neutral zone, then it’s either going to be a positive month if it’s high or a negative month if it’s low. In the neutral zone, it’s obviously struggling a little bit to detect whether it’s an up or down month. But usually, that also means that the performance is pretty flat anyway.

So now we’re going to get to the meat of our conversation, Nick, and that is, of course, to some surprise, I think, a couple of papers that actually are somewhat critical about our industry. They are academic papers, so they are way above my paygrade. I’m glad that I have you as a guide because we’re going to be diving into Journal of Financial Markets and two papers, one from 2016, one from 2020. If we can get through these two attacks on our industry, if I can put it that way, we may have some positive news towards the end of that segment.

So I’m going to be just passing the mic to you, Nick, and then you’re going to set the stage for what we’re trying to do here, or rather, what you’re trying to do here, and talk us through some of these things, which are, of course, very relevant, because this is some of the critique that I think people, sitting on the allocator side, may be paying attention to as well. So that’s why it is super relevant even though we prefer to talk about things that support our case. But let’s go your way now for a little while.

[00:36:18] Nick

The idea that I had when you said let’s catch up again, what should we discuss? I was like, we talk about trend following, we talk about systematic investing. We have to be much more holistic and I guess radical sometimes in our thought process because unless we know the challenges, we live in a filter bubble. All of us agree among each other in the happy days. I think it’s always good to just take a step back and say, okay, listen, that’s what we do, that’s what we believe in, but certainly, there are challenges there.

I think challenges we have to face on a day-to-day basis, specifically when half of the people you have in the show, and that’s just a number, you might tell me if it’s more or less, suggested that some of the things they dislike hearing or disagree hearing is that in the trend following is dead. My idea was, hey, there’s like two or three academic papers out there that I have seen over the last like 5, 6, 7, 10 years probably, that make a case against trend following. So how about we discuss that and see what’s the other side and possibly just try to connect the dots or possibly try to defend the arguments?

Before even going to the papers, let’s just do a very brief summary of what trend following really is. What trend following really is, is a range of markets, you look back, you generate a signal direction, you follow that signal direction, and you typically scale the exposure by the volatility of the asset because you typically have assets with 2% volatility, that being a currency or interest rate or it can be 40% or 50% volatility like natural gas. That’s a very basic, and I don’t mean to play it easy, but a very quick description of how trend following can be designed.

So the two papers that I had in mind, one came out actually two years ago, three years ago- yeah, that was 2020. That’s the Journal of Financial Economics. It basically says, time series momentum, which is the academic way of saying trend following, and it kind of ask the question, is it there? The other one, that’s the Journal of Financial Markets, that’s in 2016, basically says time series momentum and volatility scaling. So those two papers try to showcase that two of the features we’re using in portfolio construction are either not statistically supported or they act like a cheat.

In other words, the first one that says is trend following there, suggests that past return does not predict future return. I’m going to come back to the point. The second point says, it’s not really predictability, you just happen to vol scale stuff and vote scaling helps so much, like basic long-only risk party portfolios have performed over the years precisely because they vol scale. So one paper goes against the predictability and the other one goes in favor of vol scaling, that being really the reason, not trend following itself.

So these are the two things that I want to put forward and possibly go through some of the findings and give my opinion, but I’ll make a pause here. Anything to say on your side or should I just go and dive deep into it?

[00:39:43] Niels

No, I think it is very interesting because clearly, it touches on some really key parts of what we do. Even though, of course, some people don’t vol scale; well, maybe at entry but not continuously. But of course, yes, people definitely have questioned this fact, whether there is really any predictability from using a time series momentum approach. So yeah, I’m all ears, Nick. I can’t wait for you to unpack these two papers.

[00:40:15] Nick

Okay. Let’s give it a go. Let’s have a predictability perspective. I’ll try to keep it at a level of conversation. I don’t want to go deep into the econometrics of it. By all means, I’m pretty sure you’re going to share the papers with the audience, right? So what this paper says is that if your past return over a number of months is predicting your future return, surely, you run a regression analysis of your predictive return versus your historical return and you expect some form of statistical relationship between the two, this kind of beta that we tend to talk about when we talk about regressions. They use 55 assets across commodities, equities, currency, and they run the regressions.

Lo and behold, about 45 out of the 55 assets have not only insignificant betas but possibly even negative, not significant, but a negative point estimate suggests that it’s a reversal signal. Then they contrast their work with some earlier work done actually by Ooi, Moskowitz and Pedersen and they effectively say that regression that they used back in the day, it’s a simple regression, it has a number of econometric issues. I’ll spare you the details.

They try to fix those issues and by fixing those issues, they suggest that their evidence is quite solid, that being that there is no time series predictability, past return does not predict future return. Above and beyond that, of course they say, okay, wait a second. If there’s no predictability, why has trend following performed? Which I think is a very fair question to ask. Like, hey, okay, I don’t believe that there’s any predictability, but something is going wrong here because hey, if I look at the line, it’s actually performing quite well.

So then they say, well, you know what? Possibly, what is happening here is that there is no beta in this predictability. But there’s also an alpha that we don’t necessarily talk about because the alpha itself is not necessarily this transmission of information from past to the future but it’s just a constant return, it keeps on ticking month on month on month on month. So their hypothesis is possibly assets that have delivered historically a very positive return, I’ll just benefit from holding them, and those that have historically underperformed, I’m benefiting from historically going short them.

So all that we’re actually doing is that we’re over biasing our long exposures by over-buying assets that have historically performed and over-selling assets have historically underperformed. So then they run a strategy whereby they said, okay, if I have an expanding window, I’m looking into the historical performance from the start of history, whatever “history” means here, and I’m just going in favor of the positive average historical return every single month I was expanding my information set, and I’m going short when this historical mean is negative.

They finish off the paper saying, hey, this form of doing some sort of time series historical return but not necessarily momentum return performs very similarly to a trend follower. So that’s the whole paper, and I’m going to try to defend or potentially suggest some of the points that I agree and disagree with. So to summarize, there’s no predictability but because trend following is performing, it might be that assets have statistically strong positive or negative returns over the history so that is an enough data point to follow, and therefore, kind of perform. It’s a blend between, if you like, econometrics and then financial economics.

So that’s the whole paper. That’s why the title is, is it there? Is it really there or we just happen to be biasing our exposures as a function of the historical performance of the assets, not a short-term unpredictability? So I think there are two points that I would want to make here. Point number one is that this serial correlation and this predictability, it’s a subtle concept.

I’ll make an extreme example. If I give you a dollar for the next thirty months, every month, the cash account would go up by a dollar. But statistically, the returns you’re experiencing which are a dollar per month have zero correlation among each other. Because the way correlations operate is by demeaning, if you like, the variables. I’m not going to go into the specifics of it, but really, what you have here is an average return which is $1, and no uncertainty around it. There’s no serial correlation whatsoever. This path of you growing from zero to 30 can be seen as a trend follower, but in reality, there’s no predictability per se.

So here’s an extreme example whereby you still perform, but then there’s no predictability whatsoever. So that is one point that is going to take us to the next point being that there are patterns out there that can be seen as trends, but statistically, can be seen as no predictability. But I think that’s the most important point here and that’s what it’s actually missing, and that is why I think there’s a bit of a disconnect between a very econometric approach and an investment approach is that expected returns are time varying.

This conditionality in expected returns is very important and I’m going to bring back the whole crisis alpha, the whole conditional analysis, and the whole performance of trend following strategies when the markets shift and when micro volatility rises. Because, yes, equities over the longer term have been performing. Yes, if I’m going long equities over the longer term, I will be performing over the longer term. But the ability of going short temporarily in anticipation of a negative trend continuing is going to give that variation of trend following the convexity element.

So I’m pretty sure if we now assess a time series implementation of trend following, as in the actual trend following strategy, versus just allocating on the basis of historical long-term returns, both might have over the longer term the same Sharpe, but when it actually matters and you get the convexity, you’re going to have a significantly different story. I think the short-term predictability or correlation, if you like, of returns is the one that is possibly missing from this analysis, which by all means, is absolutely very valid and econometrically sound.

But I think it misses the point of, we don’t just build trend following strategies for a longer-term Sharpe ratio. There’s way more that goes into it and that has to do with conditional assessment. On a conditional assessment basis, I’m not expecting a long term long the assets with positive return and long term short the assets with negative return is going to deliver that. That’s one point that I want to make with this paper. By the way, I’d be happy to see how this portfolio would have performed last year, this long-term mean being positive I’m going to buy, because that would be full on equities and credit and bonds.

Again, to that point, it’s not any more a point-to-point Sharpe ratio comparison. I think there is much more that goes into it. So yes, possibly over the longer term, the serial correlation might not be statistically strong. But there is this alpha component that somehow we should have to incorporate into the thought process, and secondly, the conditional alpha or the conditional mean, which is more about how we go through phases and regimes that makes predictability stronger or less strong.

I think that is the point of debate that I have with the result. I’m not negating the result. I’m just saying that a different perspective is required to connect the academic thought into what we actually experience and what we’re going to deliver from an investment standpoint.

[00:48:20] Niels

Yeah. How about the vol scaling?

[00:48:23] Nick

So how about the vol scaling? The other paper says something different. It says it’s not really predictability. It basically says if you do trend following and you vol scale, you get the same performance if you don’t do trend following and you still vol scale. For the sake of argument, you have S&P, WTI, US Treasuries and JPY. I’m just making it up. You go long or short as time goes by based on your predictor of performance using trend following, and you put the four together by some form of vol scaling. Whether that is done at the inception, whether that is done consecutively to your point, that’s less important for the point of discussion.

They contrast that to holding the same four assets, just vol scaling them, all of them being long. Over the longer term, all they suggest is that, hey, the performance is pretty similar and actually, if you remove vol scaling, what you end up having is statistically insignificant returns between the two. So whether you see a long-only buy and hold portfolio with equal weighting scheme, for the sake of argument, or a trend following portfolio with an equal weighting or equal gross exposure in its every asset, they’re very, very similar. So that’s the point they were making.

In other words, you don’t really need your signals, you just vol scale and you happen to believe that your signals are good enough and you just do the vol scaling, but that is more of a cheat. That’s their point, right? So what’s my counter to it or my defense to it or I guess some thoughts around that? The first one is that from an empirical standpoint or from a statistical standpoint or possibly from a research standpoint, you can run equally weighted portfolios of assets. But from an investment standpoint, I cannot just see us putting a dollar on a two-year treasury note and a dollar in natural gas, because what we’re basically doing is a dollar in natural gas and a fixed dollar next week.

So, from a portfolio perspective, diversification perspective, risk management perspective, no, we wouldn’t be able to deliver a balanced allocation had we not taken into account some form of volatility. So I think, yes, probably something is happening with vol and we’re going to get to it, but it’s not that it’s a fair comparison to say, hey, I’m having all those assets with different volatility profiles put together I guess on an equal dollar basis. So there’s a gap between just let’s do it in an Excel file or let’s actually try and run it and run money with it. So that’s the first point that I want to make.

There’s this disconnection here. It’s not really a reasonable portfolio to hold, right? Yes, I can run my analysis in Excel, but it’s not a portfolio I would ever hold because I’m concentrating my risk to equities and commodities. That’s basically it. That’s point number one. Point number two, and I think that’s way more subtle, they say that if you do vol scaling in a long-only portfolio, you bridge the gap between trend following. What they do not say is that volatility scaling embeds significant amount of momentum in itself when applied to assets that have enormous negative correlation between volatility innovations and returns.

In other words, if I look into equities, they typically experience drawdowns with contemporaneous spikes in volatility. In other words, if you volatility target an equity exposure, you’re going to start reducing the exposure to it when the volatility is going up. So even if you’re holding equities, by vol scaling them, what you’re acting upon is a subtle momentum signal. So what they’re kind of doing is they don’t translate momentum into a long-only portfolio. But possibly by vol scaling your long-only portfolio, they make it look more like trend following. I’m not sure if that makes sense.

It’s not that we simplify predictability and make it look like a long-only portfolio. It is rather making a long-only portfolio embedding some momentum fields in it. By the way, not all assets experience this behavior of like negative correlation between innovation in vol and performance, as in vol goes up, performance is negative. But most of the risk-on markets do. There are some subtle cases here, specifically rates back in the ‘70s, that interest rates were going up with very low volatility, and then vol targeting would be at your expense.

But setting that as an aside, volatility scaling has a number of econometric properties. I don’t want to dive too much into it. But one of the subtle points is that it specifically brings momentum into the portfolio. Possibly, you don’t go short, obviously. You just reduce the exposure or increase exposure following volatility reduction. But this dynamism in the allocation, even for a long-only portfolio, is very similar to trend following. What’s the difference? The shorting side.

So last year, a vol scale portfolio, a risk parity portfolio, a whatever portfolio that follows a risk parity rule, possibly starting to reduce exposure, but will never go short. So it will still have a very bad year, connecting my point to the first point I made for the previous paper that the conditional assessment is very important. So that’s kind of a bit of a defense here. It’s kind of a bit of a defense.

[00:54:01] Niels

Okay. So of the three papers that we had in total, do you want to touch on the third one?

[00:54:08] Nick

The third one is a much older paper. It’s from 2006. It was published in Management Science. The reason why I’m bringing it up, it’s not a trend following paper and I’m going to spend possibly like a couple of minutes on this one. It’s not a trend following paper but I think it’s extremely insightful in the sense that it is creating econometric conditions for trend following to perform even if there’s no serial correlation in the returns. So the point I made in the beginning, whereby one of the challenges here being that trend following is not statistically strong.

In other words, you cannot predict return using past return. That particular paper says even if return cannot be predicted, as long as volatility can be predicted, and I don’t think that anyone disagrees with the fact that volatility is persistent and it classes through time, and that’s the important bit now, you can predict the sign of return. But that is nothing more than trend following. Possibly, I cannot foresee what’s the magnitude of a return, but it’s enough for me to tell you, hey, return is going to be positive or return is going to be negative, go long, go short.

I think it’s very subtle because the whole paper just talks about the econometrics and the modeling implications as to how you can generate those dependencies. It never talks about trend following, but to me, it’s pretty obvious. It says, hey, you cannot predict return, but if you can predict volatility, then you can regenerate, as they call it, direction of change forecasting. I think that’s crucially important for trend following. They also make a few more points along the lines of even if there is sign predictability, because it’s very nonlinear, even basic econometric tests cannot deliver statistically strong results.

In other words, there is sign predictability purely driven by the fact that you have volatility predictability, but even basic empirical tests cannot easily document it. I think that’s a very strong defense to the argument being returns do not predict returns, there is no trend, because strength is about the direction of the move more than anything else, and obviously, vol scaling and signal scaling comes after that. So that’s the reason why I wanted to bring it up because I think it acts as a defense to the first paper I talked about.

[00:55:38] Niels

Yeah. Maybe I can’t speak for the research teams sitting out there in these different firms but maybe that is why so many of the managers, at least that we spoke to, have embraced dynamic position sizing, because actually, they do find that there is information and changes in volatility that needs to be used. So yeah, this is awesome, Nick, because although these papers are at a high level in terms of their academic nerdiness, if I can put it like that, it does actually lead me to something that came up this week.

So last week, I spoke with Andrew on the Systematic Investor Series. He was in New York, I was in Singapore. Then this week, we met halfway. We met in Switzerland.

[00:57:16] Nick

It’s convenient.

[00:57:33] Niels

Very convenient. Yeah. One of the things that he’s obsessed with and I’m obsessed with as well is this issue about narrative. We’re constantly thinking about how can we try and make the narrative more compelling for trend following because we have all this evidence, etc. Andrew made some excellent points that I wanted to put to you, but actually, I’m also going to try to weave in what you just talked about a little bit. I made some notes about it. So Andrew was saying that none of CTA, managed futures, or even trend following has a positive connotation. Value, growth, high-yield, quality, are all positive.

We all think trend following is good, but for the average investor, it sounds like the fools who rush into a trade after the smart guy has already made the money. Further, managed futures is, to many outsiders, a quantitative leveraged long-short derivatives-based black box. This is where I actually think it ties into what you said, because I think, in one of the papers, some of the people were from AQR. Correct? Yeah. Okay. So what Andrew has observed is something that… The papers obviously came out some of them mid-2010s and what have you.

I don’t know exactly when it was that there was this big surge into the AQR trend following portfolios, but I think a lot of it was that they did some really clever narrative and positioning of their strategy. Anyways, what Andrew observes is that, and I’m just paraphrasing what we talked about, what AQR and others did was to change the narrative that they were selling similar hard-to-understand strategies by framing the investment process as harvesting risk premium.

That sounds a lot safer because to many investors. It implies the excess returns are already there and AQR and others are just harvesting or mining it. The pitch also implied that the strategies would work over decades because according to backtests, they have worked. Describing the strategy as permanent features of the markets also sounds safer. So I was thinking about when you talked about it before, where you said, well, there’s no real beta but maybe there’s an alpha it’s always been there and so on and so forth.

I was thinking whether that’s actually what Andrew was trying to say that if you position it as something where, well, there’s this value, it’s there, whether we call it alpha, beta, it doesn’t really matter, but there’s something there, so our job is just to harvest it. That does sound a little bit easier than when we describe it as we have no idea what’s going to happen tomorrow. We can be short, we can be long. We’re just going to follow the price and it sounds very uncertain, so to speak, instead of saying, well, we’re just harvesting or we’re just mining this little premium that’s out there.

So am I on the right track in the sense that there is a little bit of a link here in these things? Because of course, one is a very academic approach to things, and we come from it as practitioners so we may not be the most, how should I put it, we may not phrase things as academically well-sounding as people who come purely from academia. So we’re more maybe direct saying, well, this is what we do, and so on and so forth. So anyways, over to you, Nick.

[01:01:29] Nick

I don’t necessarily agree with the fact that value and growth have a very positive connotation necessarily, or certainly value for the last like five, six years.

[01:01:44] Niels

This is important, Nick. You’re thinking about the performance. What Andrew is saying is value sounds valuable and growth sounds, oh, it’s going to go up. This is what he thinks. I think this is what he means.

[01:01:59] Nick

In this context, I think the primary difference and I think that’s what possibly AQR tried to do is that all his value, growth, quality, and so on and so forth, it is fundamentally linked to investing from the perspective of there is risk, there is some risk, and I’m happy to take on that risk, I’m expecting to be compensated for it. So if loss aversion is important, then we can give rise to a compensation for holding certain type of risks.

I think what trend following kind of goes against that premise is that it’s primarily based on the fact that we realize that human beings are not necessarily rational or there’s some sort of slower pace of consuming information, or possibly, we’re kind of constrained by some structural dynamics in the market.

This ideal, as we tend to call it in academia, a representative agent that is deciding upon all the investments very rationally and even takes into account all the risks, and there is no uncertainty around the estimation of risk and expectation or blah, blah, blah, blah, blah, it doesn’t necessarily exist. But now, it’s also against the human nature to accept the fact that we are irrational. I think that is possibly the leap of faith that has, in my view, potentially created this narrative, which again, I don’t necessarily agree with, but I’m happy to just discuss upon it.

Now, with that being there, if I were to go… Was it Bruno (I’m just trying to think), or Doug, one of the two? He basically said (I think it was Bruno), I cannot expect that the way that investors and the anchoring bias that has been in human nature for decades is going to stop. There’s no point for that to stop because we can see the reason why it shouldn’t stop. But back to the point of building a narrative, is it a premium that exists there and we’re harvesting it?

I would say there is a pattern which is purely based on the way investors act and interact with each other, and systematizing the way of harvesting that is certainly going hand in hand with the way we can deliver value and growth and high yield and quality and so on and so forth. So I don’t disagree with that. But equally, I don’t perceive too much of negative connotation with trend following. Possibly, if you just compare one with the other, yeah, probably value and growth have more of a positive connotation, but I don’t necessarily believe that trend following so badly associated.

[01:04:55] Niels

No, no.

[01:04:57] Nick

AQR and a range of industry practitioners have tried to approach the matter, that being via research, education, and evidence. I think it’s the right thing to do. But ultimately, we owe that to investors.

[01:05:13] Niels

Also, I think what has been really smart about taking an academic approach to this in the way they described it early on, and this is many years ago, of course, is the fact that I think a lot of the people they were essentially messaging with this, people sitting in pension funds, etc., they were kind of speaking that language. They were not speaking our language necessarily with breakouts and dungeon channels and what have you. I don’t think that is taught in many places leading to a CIO role. So yeah, I think that was very smart.

This is also why I do think narrative matters and I love this quest that we are on in terms of finding potentially a better narrative. We’ll see.

Anyways, let’s round off with just a couple of things. The Aspect paper also talks about the wonderfulness of trend following. Not surprisingly, of course. We all believe in that. But there’s one thing in particular that I took away from it, which is, I think they’ve done a nice job in visualizing it even though it’s something that I think we’ve certainly highlighted before. The more time that passes, the more convinced I think this is true.

This is the fact that the decade from around 2009 to 2019, what was described as the CTA winter or the lost decade, etc., I think a lot of people as we were heading into 2018, 2019, returns weren’t that strong, basically came to the conclusion that this is the new normal, markets have changed, trend following doesn’t work, at least not as well as it used to do, so we shouldn’t expect great profits from this. Of course, 2022 comes along and changes that to some extent.

But what I think is at least my conviction here is that we will look back at this period and realize that that decade was the anomaly. That actually, the decades before, whether it’s the ’70s, the ‘80s, the ’90s, the 2000s, and hopefully, the 2020s, will show normal return distribution, etc. But that period, where central banks were in control, to a large extent, was really the odd one out. That is my conviction. They show something visually in the paper that suggests the same even though we don’t have much future data, of course, and I thought that’s really interesting.

I think if people can internalize this, if people end up believing, as I do, that that’s exactly how we will find out that it is, then people should almost completely reverse the way they think about asset allocation and kind of start with the allocation to trend following saying, okay, this is really a core component because just look at how it’s done in all of these decades, maybe with the exception of one, and then let’s build something around that to build our portfolio from that point.

Of course, I’m taking an extreme point here. I’m aware of that. But anyway, that’s just my thoughts from that, which is hard for you to comment on if you haven’t read the paper.

[01:08:46] Nick

No, but I agree, because I think it goes back to the point that we had in the first time that we spoke together back in November following some work that I had done with regards to how that decade shifted and how this fundamental certainty back then gave rise if anything to reversals in asset prices across any asset class has significantly shifted, but also how my conviction was and possibly still is that this period of excess liquidity and issuance of cash is not going to return. It doesn’t feel that it gave as much as it has actually created at the end of it. So I completely agree with the statement.

[01:09:33] Niels

Yeah, cool. Well, Nick, this has been fun as usual and incredibly insightful so I really appreciate that. To all of you listening today, if you also appreciate all the time and preparation that Nick put into this, why don’t you go and leave a rating and review on any of the major podcast platforms. It only takes a few minutes but it does make it a lot easier for people to find the show and hopefully also join our little tribe here.

In terms of questions for next week where I’m joined by Alan, as usual, you can send them to info@toptraderunplugged.com. We’re recording a couple of days early so by all means, please send them over as soon as you have time. With that, I think we’re going to call it a day. So we look forward to being back next week, and until then, take care of yourself and take care of each other.

[01:10:26] Ending

Thanks for listening to the Systematic Investor podcast series. If you enjoyed this series, go on over to iTunes and leave an honest rating and review and be sure to listen to all the other episodes from Top Traders Unplugged. If you have questions about systematic investing, send us an email with the word “question” in the subject line to info@toptradersunplugged.com, and we’ll try to get it on the show.

Remember, all the discussion that we have about investment performance is about the past and past performance does not guarantee or even infer anything about future performance. Also understand that there’s a significant risk of financial loss with all investment strategies and you need to request and understand the specific risks from the investment manager about their products before you make investment decisions.

Thanks for spending some of your valuable time with us and we’ll see you on the next episode of the Systematic Investor.