Corey Hoffstein is a quantitative researcher, portfolio manager, and financial innovator.
In this episode, Corey shares his insights on return stacking, the overlooked impact of rebalancing timing, and how to find alpha in increasingly efficient markets. We also explore his views on trend following, the 13-month momentum anomaly, and how his innovative ETF strategies are reshaping portfolio construction.
Please enjoy our conversation with Corey Hoffstein.
The Investipal Podcast is produced by www.investipal.co. Past guests include Meb Faber, Brent Beshore, Peter Lazaroff, Douglas Boneparth, Jamie Hopkins, Tyrone Ross and many more.
Follow us on LinkedIn: www.linkedin.com/company/investipal | www.linkedin.com/in/cameronhowe/; Twitter: www.twitter.com/camhowe16 | www.twitter.com/investipal; Tiktok: www.tiktok.com/@camhowe16 | www.tiktok.com/@investipal; or Instagram: www.instagram.com/investipal/
Find Corey Hoffstein at:
https://www.newfoundresearch.com/
https://www.returnstacking.com/
https://flirtingwithmodels.com/
https://twitter.com/choffstein
https://www.linkedin.com/in/coreyhoffstein/
https://www.amazon.com/Corey-Hoffstein/e/B0BLYBQ4XR
https://www.youtube.com/c/FlirtingwithModels
https://papers.ssrn.com/sol3/papers.cfm?abstract_id=3981826
https://blog.thinknewfound.com/
00:00 The Unsustainability of the Private Equity Bubble
02:17 Comparing the Private Equity Craze to the SPAC Craze
09:41 The Potential for a New Super Cycle in Commodities
12:06 Understanding the Relationship Between Time and Money
17:45 Managing Finances: Keeping Money Separate to Avoid Conflicts
20:05 Investment Strategies for Retail Investors: Active Management and Diversification
21:57 The 'Awesome Portfolio': A Balanced and Low-Volatility Investment Strategy
23:38 Consistency in Asset Allocation Throughout an Investor's Career
Cameron Howe:
Hi everyone. Welcome back to the Invest Ball podcast. Today's guest is Corey Hoffstein. He is the co-founder and the CIO of Newfound Research. Corey's very well known for his work in quantitative investing, particularly around innovative strategies like return stacking. He's also the host of Flirting with Models. It's not that type of podcast everyone. He dives deep into the world of quantitative finance with top experts. And he's also a social media tycoon, very prolific on Twitter. Corey, it's great to have you on.
Corey Hoffstein:
Thank you for having me. Not sure I qualify as a social media tycoon, but I will say having almost 70,000 followers for a guy who only talks about quant finance, I assume I attract a certain type of bot with that behavior.
Cameron Howe:
To the top 1% of algo traders doing sentiment scans on Twitter, I'm sure.
Corey Hoffstein:
That's right, and it's negative alpha to follow me on that one.
Cameron Howe:
You save all your real alpha strategies for the ETFs.
Corey Hoffstein:
I just assume what I'm trying to do is put out there is like fake sentiment stuff to throw off everyone else's model so I can counter trade them.
Cameron Howe:
I remember like when I was a practicing quant at one of the banks, we did an NLP study for predicting the 2016 election, or 2018 election, was it? And it predicted Trump was going to win, but I imagine that was a lot of bot traffic. So who knows if you could actually trust Twitter with sentiment analysis.
Corey Hoffstein:
I find that sentiment is probably one of the areas I'm most skeptical of because what people say they do and what they really do in private are very different things. And so I've heard of some very clever strategies from folks, especially on like these fundamental quant crossover desks where they'll use, they won't use sentiment data, but they'll do things like they'll run ads for consumer products to figure out click-through rates of people's actual behavior to say, "Oh, I think this company has a hit product on their hands" or "How much do these ads cost to run" and do some competitive spend analysis to try to back out profitability based on costs of client acquisition and stuff like that. And I think there's some really clever internet-based quant analysis you can do if you want to spend the money. But again, like something like sentiment is hard to drill into because not only the bots, but you know what people say is often incredibly different than the way they behave.
Cameron Howe:
That's actually very fascinating. I remember looking at like Mastercard's data sets to try to predict like consumer buying trends, but this is like, I mean, you could actually probably boost the stock price because if people are engaging with those ads, you're driving more conversion, like real conversion over into that product and then bullying the stock price. And that company's probably quite happy that they're not spending any money on ads, but seeing the benefit of it.
Corey Hoffstein:
Yeah. Well, I mean, maybe that's right. You could front run them and then do their ad spend for them but by the stock and say, "Hey, I actually think your ads suck. Let me do better ads for you and I'll help support your growth." I'm not sure I'm moving the needle for any large, you know, mega cap companies, but the point being, there's some really interesting ways with the internet now you can track people's actual clicks and traffic and behavior. If you're clever about it, that I think then tells that I'm not sure I'd call that sentiment, but yeah that's an alternative type of data that can be quite powerful.
Cameron Howe:
Yeah. I wonder like with the shutdown of, I know like Quandl, a lot of Quandl's data sets I think got shut down on like tracking flights and all that. Like what's, do you actually have any opinion on like what the new age of alt data in the stock market is?
Corey Hoffstein:
I don't, and I'll tell you why, because I've never spoken to anyone that's used a public alt data set that found it useful. Because once it's public like that, it's usually priced in. And I hate to be like the, everything's priced in meme. But the reality is most of the alternative data sets that I hear people using are not publicly available. They're hard to come by or they're homegrown in some fashion. Like the example I gave of running ads, right? On different things, running analysis on different communities or whatever. If you have 15 hedge funds trading the same data, it's likely that the alpha is extracted, squeezed out, and competed towards zero quite quickly.
Cameron Howe:
Yeah. It seemed like that happened with like analyzing Wall Street bets, buying behavior back in 2021. I imagine no one's paying attention to that anymore.
Corey Hoffstein:
No. Surprise, surprise, markets are rather efficient.
Cameron Howe:
I think that's a very good segue. So as a trend following guy yourself, what is your view on the market? Like do you view it's maybe like the US market is perfectly efficient or where do you see areas of opportunity on your strategies?
Corey Hoffstein:
I don't think anything's perfectly efficient, but I think things are pretty darn efficient. And I think it's worth asking, where should you shoot your shot, right? The problem most people have is that the average allocator, I'm just gonna talk equities here, but the average allocator has 60% of their portfolio in US equities. And we know that because that's, if you look at passive markets, passive markets on a global scale, 60% of that exposure is US equities, and therefore by definition, that is how much active investors have on average, because all that has to average out. That is historically the hardest place to find alpha. And there's different ways you can look at that, but let's just, I mean, you can go look at the Morningstar active passive barometer, you can look at the SPIVA reports, the survivability of funds in those US large cap categories is very, very low and their ability to produce consistent outperformances negligible.
And so you have this contradiction of like the part of your portfolio that's taking up the most real estate that you have the most capital in is the hardest place to fish. And I think that's somewhat tautological, which is everyone has capital there, so everyone's trying to harvest alpha there, and therefore it's very competitive and efficient. Doesn't mean it's entirely squeezed out. I think there's evidence about index additions and subtractions, that there's some areas of event-driven opportunities, merger arbitrage, though I would guess I would call that a risk premium. Like there's certain events that I think there's still opportunity, as well as, I think there's a growing field of academic literature around the impact of passive flows.
I want to be careful when I talk about passive. Like you know the role of 401k and target date funds, for example, there's some evidence that companies that are included or excluded within certain indices and having indices that are more likely to be traded within a target date fund versus not changes the in and out group behavior among those stocks. So I do think there are places where the market isn't perfectly efficient. But I would personally say, look, if I am an investor and I have a certain risk budget for how active I want to be, I'm not personally spending that risk budget in the large cap U.S. equities. I will tell you, across my PA, I have basically gone passive in large cap U.S. equity because there's other places I would rather shoot my shot.
Cameron Howe:
Yeah. I've, since I left the flow of information have prescribed to an index investing strategy myself. Cause whenever I make bets made a bet on snowflake, I think two months ago when I was looking at the stock price, what am I doing cam? Just stick with the S and P.
Corey Hoffstein:
Well, what's hard, I think what's very hard too is there can be opportunities that are temporal. And so I like, I look back and I say, okay, 2020, 2021, there were very obvious bubble characteristics to me that I took certain trades on, but that to me, isn't something you can always trade. So I placed certain trades that were meant to capitalize on the environment, but that doesn't mean I'm gonna persistently be active in large cap equities because that environment disappeared. And I no longer think I have an edge. Another example is I traded a whole lot of crypto stuff just for fun in 2020, because I couldn't do anything else besides my day job. And I think my edge in that market has largely disappeared because a lot of the retail traders have left, a lot of the platforms I used to trade, like I hate to say this, I use a trade on FTX. Oh, that's clearly gone, right? But a lot of the stuff that was happening that was bubblish behavior is no longer there. And a lot of the people I know who have come into that space are much more sophisticated operators and my edge has disappeared. And so I can't play there anymore.
So I think there's these concepts of like, is a market on average pretty efficient? Is it always efficient? Are there times it's less efficient? I think that's certainly true. And you can take advantage of that, but I think you also have to be aware of, all right, I think you know most of the time large gap US equities are pretty efficient. And unless I'm certain there is bubble behavior going on, I probably just want to passively allocate and spend my time elsewhere.
Cameron Howe:
Yeah, I know we spoke about this last time, but when I was running quant strategies, we'd apply the same strategy to the Canadian market where I am relative to the US market. And these were like multi-factor models. We'd splice them out, run the quality strategy, run the momentum strategy, run a value growth strategy, et cetera. And the US market had no alpha on like a fundamental based quant strategy. And then in Canada, eked out like consistent returns, mostly on the quality side, but I think it plays into that, like where are the algos, all the algos from the US market, way more efficient, everything's priced in a lot more quickly versus like a more underdeveloped market like here, where you can apply more of those, I'll call them like basic strategies or strategies of yesteryear.
Corey Hoffstein:
Canada is an interesting market to me because somehow when you look at a US allocator portfolio, you're almost left out most of the time. In the US, we have your US slice, you have foreign developed and you have emerging and you would think Canada would be in foreign developed, but that's always benchmarked to the MSCI EFI of which you're not a part of. So American investors somehow totally just forget Canada and exposure in our portfolios, which is wild because I actually, I have a good friend who's Canadian who once mentioned to me, he thought the best portfolio was he called the North America, Canada barbell, which was just by 50% of your portfolio and in, excuse me, the US large cap by half your portfolio in Canada. And you had this nice, like natural resources, tech barbell going on. And it was a pretty good looking portfolio.
Cameron Howe:
It's interesting because I would run an oligopoly index that was like long Canadian oligopolies, like grocers banks, telecoms, railway companies, and then you'd go short, natural resources, junior mining companies, and it would like eke out better alpha or better consistent returns than like the S&P would. It was just like, where's all the, yeah, where, like that's why it's a quality theme. Cause the quality theme here is like big five banks and like major enterprises that own pretty much the entire market. And you go short the small caps.
Corey Hoffstein:
And you want to write again, who's competing in Canada for that alpha? I don't often talk to a lot of managers who go, "Oh, yeah, we're setting up shop to trade Canadian markets as our source of alpha." It's like, you don't hear about it. And that's probably why the opportunity persists. And that might simply because there's not a lot of appetite for that mandate. I don't see a lot of institutions going like, "I want alpha from Canadian markets." And then it's also the infrastructure, dealing with regulators. Is the capacity there? Is the demand there? And I think when you answer all those questions, you start to go, all right, maybe that market should be less efficient and that's an opportunity for someone.
Cameron Howe:
Yep, absolutely. So I guess, yeah, good segue. I'm curious to pick your brain a little bit about your overarching strategy, the return stacking approach and how you apply that to markets.
Corey Hoffstein:
Yeah. So if I can maybe back up a bit, to just describe what return stacking is. So there's this concept in institutional investing called portable alpha. And the idea here is to, and it goes back to the 1980s. Like this is not a novel or new concept. But the idea here is to separate alpha from beta. So I talked about this idea of the average allocator equity allocator has 60% of their portfolio in US equities. And so they're kind of stuck hunting for alpha there. What portable alpha does is it says, well, let's separate alpha from beta and say, let's just get our beta where we think beta is efficient, but do it with a capital efficient investment vehicle. And then what that will do, and I'll explain that in a minute, but what that will do is allow us to then take our capital that's left over and allocated to a place or a strategy where we think we can harvest some excess return.
So as an example, let's say I think the S&P 500 or US large cap stocks are very, very efficient. So if I have $100, instead of buying $100 of the S&P 500, I might put, I don't know, $20 in T-bills, use that as cash collateral to buy $100 of S&P 500 futures. Effectively using leverage here, right and then I'll take that $80 left over and I'll go invest it in some sexy hedge fund because I think that sexy hedge fund can go generate alpha trading. I don't know anywhere else in the world too. Another example of doing something like this might be well, I really believe I can generate alpha and small cap stocks. But I've got a you know, everyone wants a large cap mandate. Well, what I could do is I could buy all the small cap stocks I think I have alpha in short the Russell 2000 futures, go long the S&P 500 futures, and now I have effectively stacked small cap alpha on S&P 500 beta, and I've ported where I'm getting the alpha from onto the beta I want.
So that's the idea of portable alpha. Return stacking, which is a strategy for which I've launched a suite of ETFs, I guess, starting 18 months ago, we now have five ETFs, 750 million in assets under management here. Return stacking is taking the concept of portable alpha and putting it into tickerized products. And the name return stacking comes from my colleague, Rodrigo Gordillo, who, when we were sort of talking about this concept of capital efficiency and portable alpha, we didn't think they were very approachable terms. And he came up with the phrase return stacking, which I think more clearly explains the general concept, which is you're taking one return stream and stacking it on top of each another. It might be alpha. It might be a risk premium. It might just be a diversifying beta. But you're effectively stacking it on top of another.
And so for us, one of the things we've been talking about as we bring these products to market as sort of these Lego building blocks is saying, look, there might be other strategies that we have a high degree of belief in their efficacy that are hard to put into a portfolio otherwise for a variety of behavioral reasons that you can now introduce through this returns checking or portable alpha structure. And so that's things like managed futures trend following or managed futures carry merger arbitrage. You know there's a whole list of these types of strategies that that you can look to now introduce as an overlay rather than having to sell stocks and bonds to make room in your portfolio.
Cameron Howe:
Interesting. So is that, is the idea behind return stacking that the core part of your portfolio is in whatever set strategy you're mandated and this is like a 10% sleeve or is it in replace of like your core strategy? It's an all in one solution.
Corey Hoffstein:
Yeah, so let's say you have, let's say you're your the average 60, 40 investor, right? And you wanted to introduce managed futures into your portfolio. You know manage futures trend following which historically low correlation of stocks and bonds positive excess return historically did well during stock and bond crises right did really really well in 2022 it's a great like empirically historically great diversifier the problem is that diversification has historically been this process of addition through subtraction to add managed futures into the portfolio I need to so make room by selling stocks and bonds. And when you're talking about a lot of these publicly available liquid strategies, you're talking about strategies that have a Sharpe ratio of 0.3, 0.4, right? These aren't too sharp strategies. And so 0.3, 0.4 means for every 10 vol, I expect 300 basis points of excess returns, 400 basis points of excess returns.
The probability of going through five-year, 10-year drawdowns is it's pretty high. And we saw that with something like managed futures trend following in the 2010s where post-2008, they did really well. There was a lot of attention. People started to sell stocks and bonds, make room, allocate to the strategy. And then the strategy basically went nowhere for a decade. No different than a lost decade in stocks, but infinitely more frustrating for allocators who had given up exposure to stocks and bonds that had then gone on to do quite well. And they were now allocated into this trading strategy that was more opaque, higher cost, less tax efficient. And so huge behavioral frictions, huge performance hurdles that had to get over, especially on a relative basis to what had been sold to make room. And a lot of people abandoned the strategy before it then went on to do quite well in 2022.
And so what return stacking is trying to say is, what if we got rid of this either or decision? So if you're a 60/40 investor that sold some stocks and sold some bonds to make room for managed futures, what if we said, well, what if we had a product that when you invest with us, we're going to give you, if you give us a dollar, we'll give you a dollar of large cap equity plus a dollar of managed futures. So you can sell 10% of your equity exposure by our product. You get the equity exposure back, but now the managed futures is layered on top of your portfolio. So you're not sacrificing any of that underlying strategic policy allocation. You're just getting the alternatives as an overlay. And we think that is long-term, not only beneficial from a return enhancement, as well as the added diversification, but behaviorally, it's much easier to stick with.
Now, full caveat here for full fairness, like we are employing leverage. That's what capital efficiency is, that's what portable alpha is, that's what return stacking is, like there is a leverage here. And leverage cuts both ways. It enhances the good and accentuates the bad. And so what we need to be really careful of when we use leverage is to make sure that whatever we're stacking on top, we truly believe is diversifying to the underlying portfolio we're stacking it on, as well as making sure that we're never using so much leverage that if correlations crash to one during some sort of liquidity crunch or crisis, that the portfolio would be permanently impaired.
Cameron Howe:
That's very interesting stuff. And this is based on the work you've been doing the past 15 years with Newfound.
Corey Hoffstein:
Okay, so Newfound got its start back in August 2008 as a research firm. Hence the name Newfound Research. And we sort of pivoted over time. The emphasis of the work was always quant, very trend following driven actually, but we went from providing research to being a sub advisor, index provider to eventually launching our own suite of ETFs. And it was that suite of ETFs was born out of the frustration of trying to get people to adopt alternative quantitative investment strategies. And again, in the 2010s, this is, I mean, maybe I'm bright, maybe I'm smart as a quant, but I clearly don't have that much wisdom because I kept running into the wall again and again and trying to tell people, hey, sell your stocks and bonds to buy these diversifiers. And that's just not a great way of going about it. And so return stacking was really a way for us to try to repackage the benefits of diversification in a way that's far more palatable for investors to hold for the long run.
Cameron Howe:
Very interesting. And so I know you touched on it previously. I did a lot of never manage futures, but trend following on pure equities. I'm curious, and you don't have to divulge your secret sauce, but what are some like key indicators you look for on the trend following side?
Corey Hoffstein:
Yeah so this goes There's always an interesting balance I found when you're a pure ivory tower quant versus having to put something into a product. And one of the decisions we made in putting something into a product was we wanted to manage future strategy that hit the ball as sort of smooth down the middle of the fairway as we could. What I mean by that is Trend following, managed futures trend following is a category that is notorious for dispersion among managers. So 2022 was ostensibly a very good year for the category. On average, if you look at sort of category benchmarks, whether it's the SocGen trend index or the BTOP 50, they were up 20, 30%. You can find individual managers that were down 20% in 2022. So there's huge manager selection risk because it is a strategy in which there's a large number, a large degrees of freedom in how you can design the strategy.
Including what markets are you trading? What types of indicators are you using? What trends speeds are you following? How are you allocating risk among markets? Are you commodity heavy? Are you balanced across the different asset classes? Are you trading this fat tail of a hundred markets or are you focused really just on the liquid instruments? And so when we were putting together this return stacking concept, we said, well, if what we're giving in the underlying component is just meant to be beta, right? Just like, for example, large cap US equity beta, and we're not making active stock picking decisions there, how can we stack sort of this alternative strategy beta on top? In this case, maybe a futures trend following. And for us, what we wanted to do then was to say, can we look at sort of a category average, whether it's a benchmark or find a pool of five, six, seven, eight, 10 liquid managers, average their returns together and get what we think represent is representative of you know the industry average and then build a model that can track that average.
And so that's ultimately what we did. And there's two ways we went about that. There's one approach is that we call top down. And the top-down approach is just a pure regression-based approach where we say, what are the prior returns of that sort of category average, and how would we have best fit them given the markets we trade, of which we trade 27 stock, bond, currency, and commodity markets, and then find the replicating portfolio that would have fit it best, and assume that's the portfolio we're going to hold for the next day. Because we don't think these managers, while they're making changes, on average, those positions are changing pretty slowly. And so we think if we look at, hey, let's look at the last 20, 30 days of returns, find the best fit portfolio, we think that's a decent approximation for what they're holding today and will continue to hold tomorrow. That's called top down. The benefits there are that it is totally agnostic to how the managers are making decisions, but the con is it is definitively looking in the rearview mirror, right?
The other approach is what we call bottom-up, which is we know these are all trend followers. So can we, with the benefit of hindsight, choose, to your original question, the indicators and the speed and the combination of markets from a risk-weighting perspective that would create a system that would have historically looked like the average? And no surprise, what you get is an average of a whole bunch of signals and speeds and markets. And then you say, OK, I think that process is representative of the average. I'm going to continue to trade that process on an ongoing basis. So you're no longer looking in the rear view mirror trying to track what the index did. So you can adapt in real time, which is the pro. But the con is if those managers start changing the way they behave over time, you might miss that.
And so given these two different approaches to replication, no surprise about, given all our chats about diversification, turns out there's benefits in process diversification. We use both approaches, get target weights from both approaches and average them together. And so we have a system that's ultimately designed to try to capture the average behavior of the trend following space, but using two very different approaches at doing it.
Cameron Howe:
Very interesting. Have you ever looked at applying ETF for index flows as a proxy or a signal within trend following?
Corey Hoffstein:
Explain what you mean by that.
Cameron Howe:
If you see positive inflows into certain strategies or certain indices is using that based on maybe like market cap weighted funds that are like, you know, we're going to go like the Magnificent Seven, if that's now an indicator of positive momentum for the stock individual stocks that are highly concentrated within it.
Corey Hoffstein:
Yeah, so we don't trade the individual stocks. We're only trading index level futures. So we're not trading single stock futures. So we're trading S&P 500, NASDAQ, FTSE 100, German DAX type exposures. And we're the strategy itself is focused solely on price momentum. That said, what's interesting, right so our bottom-up approach is purely price signals on those types of indices. But not just stocks. We trade oil. We trade the Japanese yen dollar cross. We you know we trade all these different asset classes. But you do see certain managers start to make what I would call proxy trades. So AQR, they take some basis risk, basically. So AQR, for example, has a model they've introduced where they look at fundamental trends, trends in fundamental data, and they say, we think that's predictive of future price movement. And so they've incorporated fundamental trend following in their trend following models. That sort of thing may or may not be picked up. It might just be a different angle at the same price trend following concept, but our bottom up may miss that. The top down approach though, which is totally agnostic, might pick that up if AQR is in our basket. And so that's all to say, no, we don't look at that stuff. I do know there are managers who look at, they might use price trends as a, precursor excuse me, volume trends as a precursor to price trends or fundamental trends as a precursor to price trends. That in a way gets picked up in our top down approach, not in our bottom up approach though.
Cameron Howe:
Yeah, I remember looking into applying technical indicators on fundamentals, but it's just such a slow signal. Like you would need probably eight quarters worth of data, but like, is that really, you know, that one news event on an earnings release or some sort of macro event that happens kind of blows that signal out of the water and it seems tough unless you're looking for like a very durable momentum strategy or trend following strategy to apply on the fundamental side.
Corey Hoffstein:
Yeah, I think you either need to take a lot of breath, right? In the sense that it's a very low information signal. So you're taking a lot of breath across what you're trading to make, to increase the power of what you're doing, or you're trying to blend a multitude of different correlated signals. Right? So you might only get one of these signals every quarter, but it's a signal that is highly correlated to another economic type of reading. And so you're using a mixture of these signals, monthly and quarterly signals to create a smoother time series. But I generally agree. I haven't found particular application in it. With that said, I know the folks at AQR are quite bright, and so they went from having none of it in their trend strategy to having a pretty significant proportion of their trend strategy being driven by these types of models. I will give them the benefit of the doubt that they have found some something there.
Cameron Howe:
Very interesting. And so changing tune a little bit, but down the fund manager side, you put out some really interesting research on rebalancing luck and timing luck. Could you elaborate a little bit on that?
Corey Hoffstein:
Yeah, this is where I'm going to put anyone listening to this podcast right to sleep. If we haven't already, actually, if you are, if you're not listening, you know, before bed and you, you need some something to just tune you out for the night, you're having trouble sleeping. This is save this for later. Look, I'm Don Quijote tilting at windmills on this topic. What is rebalance timing? Look, I believe there's three axes of diversification when it comes to running a portfolio. There's the traditional type of diversification everyone knows about, which I call your what. What are you investing in? And that's sort of your, Hey, how many stocks do you need for your variance to sort of stop having a marginal decline? And okay. You need 20, 30 stocks. That's the, what are you investing in?
The next axis is how are you making those investment decisions? And here I would give the example of like a value investor. A value investor could look at, and I'll take a quant value investor. They could look at sorting stocks based on book to price. They could look at sorting stocks based on earnings to price. They could look at sorting stocks on enterprise value to EBITDA. All sorts of different ways of capturing the same sort of quote unquote fundamental measure of value, all correlated over the long run, but all have pretty significant dispersion in the short run. And the idea here being if you don't have any statistical or fundamental belief why one is better than the other, you should use all the approaches, right? Because you're gonna get diversification benefits. So that's your how.
The last one is when. And this is, I find, is totally ignored. And you can go and look at every smart beta ETF that's come out over the last decade to see that most of them have totally ignored this concept, which is, as it turns out, when you rebalance can have a massive impact on your portfolio. So if you were say a value investor who, and you were creating an index for an ETF to track and that portfolio was going to rebalance once a quarter. If it rebalanced at the middle of the quarter or the end of the quarter, as it turns out, you can end up with wildly different performance based on what screens is a value stock at that point in time.
So I will give you a nice little story here to drive this home. So folks listening may have heard of a firm called Research Affiliates. It's a pretty famous quant equity firm here in the States out of Southern California. And back in 2005 or 6, I believe, Rob Arnott, their principal, wrote a paper called Fundamental Indexing. And his core idea was equities should be weighted on their fundamental footprint. It's just a value screen, but he's a brilliant marketer. Anyway, he writes this paper and he develops this index methodology and the index rebalances every March. Ends up getting an ETF launched on the concept. I think there were several, but there's there's one that comes to mind. That was gonna rebalance once a year in March. And lo and behold, 2008 comes around, bottom of the financial crisis, March 2009 from an equity market perspective. His index rebalances right into it and ends up and absolutely blowing large cap equities out of the water from an alpha perspective in 2009.
Fast forward to 2010 and a couple of researchers from Robeco say, well, that's sort of interesting. What would have happened if we took the exact same rules, but instead of rebalancing in March, we rebalanced in June? Or September or December or really just any other time of the year. And what they found was that the exact same rules applied at a different point in the year led to thousands of basis points of performance dispersion within a given year. So same strategy simply applied a different year. In fact, his 2009 performance, which I would argue catapulted research affiliates to become the monstrosity it is today, and I don't use that in a bad way. Monstrosity is like a massive firm. He would have had negative alpha had he rebalanced in September instead of March. Nothing to do with the quality of the strategy, just simply the luck of years earlier saying I'm gonna rebalance in March and turns out the crisis bottomed right when he was rebalancing.
So the solve for this is the same way you can diversify across assets, the same way you can diversify across process, you can diversify across time. You could say, okay, I think this portfolio should be rebalanced once a year, because that's how long I want my holding period to be to match sort of the forecast period of my alpha versus the transaction costs I'm going to incur, maybe a holding period of a year is optimal. How do I do that while diversifying over time? Well, what I can do in my head is think of having multiple managers. So maybe there's a manager that rebalances every March and holds for a year. There's another manager who does the same thing in June and holds for a year. One in September it holds for a year, one in December that holds for a year. And what I'm gonna do is I'm gonna give a quarter of my capital to each of those managers. In effect, they create these staggered portfolios. And as long as I keep rebalancing among those managers, so they all have equal amount of my capital, I have more or less helped eliminate some of this rebalance timing lock.
And so this has been an area of huge focus for me. I've looked at it from a strategic asset allocation perspective. I've looked at it from an equity factor perspective. I've looked at it from an options strategy perspective. It has profound impacts on returns and can be the difference between managers being hired or fired based upon their track record, but also on a flip side, many benchmarks that managers are held to, those benchmarks rebalance just once a year. But if you're a value manager being held accountable to the Russell 1000 benchmark, well, that benchmark has a huge amount of rebalance timing luck embedded that may hurt or help its performance relative to you. And so my rallying cry has sort of been, as an industry, we're doing this all wrong, but no one cares. So that, thank you for listening. I appreciate it if you're asleep, sleep well. But it to me is so easy to fix, so massive in its implication, and yet no one talks about it.
Cameron Howe:
Have you ever looked at like a quadwitching event and how that impacts? Like if you're not allocating to managers, but allocating to individual securities, you know, is it like better timing to front run that day to rebalance on that day to like wait a few days afterwards?
Corey Hoffstein:
So I haven't looked at quad-witching that, you know, there's a lot of conversations around, all right, Corey, you want to, you want to do this rebalancing, timing lock, tronching, or staggered portfolio approach, but aren't there seasonality effects that you can benefit from, right? It's sort of what you're talking about with quad-witching and turn a month effects seem to come up with the biggest. Like, you know, I talk about, do you rebalance at the end of the month or mid month? A lot of people say, well, actually I think there's alpha to be taken advantage of end of month. I'm of the view that one, I think the evidence for a lot of these seasonality effects, especially that sort of seasonality effect applied, one, once at a time, sort of at the whole portfolio level are very, very small. Like the statistical significance is not great. The poor potential portfolio impact is not high and you're taking on a huge amount of timing luck to implement that, so it's not worth you know going all in on that seasonality factor for what is effectively a market timing call, because you're not applying it across a breadth of bets. You're just applying it at the whole portfolio level.
That said, doesn't mean that you can't slightly tweak the way you're doing things. So let's say I do believe that there's a beneficial turn of month effect in my rebalancing. One of the things I can do is I can just, if I'm doing these staggered portfolios that I'm rebalancing a little bit of my portfolio every single day or part of my portfolio once a month, you know, I can overweight certain rebalance periods to adjust for the fact that I think those rebalance periods have some alpha associated with them. So I talked about hey, there's four managers I'm allocating to. One's March, one's June, one's September, one's December. What if I actually thought there was March alpha, right? 2008 crisis ended in March 2009. We had March 2020. Maybe I'm just sort of a believer once a decade we get a big March event. But in the meantime, I wanna manage that lock. Well, maybe I don't give it a quarter to each. Maybe I give 33% of my money to that March person or whatever optimal combination it is based on my view of alpha.
So you can account for some of these seasonality patterns if you believe in them. But again, I don't think the optimization math would ever suggest, yes unless the alpha was so extreme, that you would ever go all the way into one because the dispersion, the potential for dispersion it creates from this rebalance timing lock effect is so massive. And by the way, the last thing I'll add here is that I think the reason no one ever talks about rebalance timing lock is because it's hard to see the counterfactual, right?
Cameron Howe:
Yeah.
Corey Hoffstein:
The only reason the research affiliates was sort of someone finally pointed it out was because they went through and replicated the public approach and calculated the counterfactual of, well, what happened if you rebalance on another date? And so it's not like the math is the same where we can clearly see a 20 stock portfolio has lower vol on average than a five stock portfolio. This is like, I need to go do the work and show you from a counterfactual basis, actually you could have performed better if you rebalance on another date or you just got super lucky and happened to rebalance on the best date of the year. And it's not trivial necessarily to take someone's historical track record and figure out what the counterfactual would have been.
Cameron Howe:
Yeah, it'd be very tough. I feel like to run a back test and decide based on my strategy, like, yeah, it's, it is timing luck. You can't, it's an uncontrolled variable. I don't know if you could ever really solve for that besides your approach on staggering your rebalances.
Corey Hoffstein:
So what I find is that a lot of quant firms actually do this. Like if you talk to a lot of the bigger quant shops, they tend to have some sort of staggered rebalancing, two approaches maybe, staggered rebalancing or what they're explicitly doing is they're creating a forward curve of their alphas. They're and right they take all their different alpha signals and they're forecasting these alphas forward in a curve. And then they're running a portfolio optimization that tries to maximize their exposure, but minimize the transaction costs that it takes to get the portfolio in the right place. And that's in effect somewhat doing the same thing.
Cameron Howe:
Yeah.
Corey Hoffstein:
But that's more like a hedge fund idea. You don't see that happening in an ETF, right? In an ETF where we have this structure of index providers creating an index that's rules based that then the ETF passively tracks and we get these super low cost tax efficient ETFs, those tend to be rebalancing twice a year or last day of the quarter or five days before the last day of the quarter struck on data calculated three weeks before that.
Cameron Howe:
Yeah
Corey Hoffstein:
And it's done because you have this totally independent group who's running this index that needs to be running it for massive scale and transparency and simplicity. But I actually think it's at the cost of a much superior outcome for investors, which would be, hey, if you're running a momentum strategy, rebalancing twice a year is kind of insane. And yet, the MTUM ETF from iShares, which is based on an MSCI index, did that for a decade plus.
Cameron Howe:
Yeah.
Corey Hoffstein:
I don't know, maybe if you think six months is the right holding period, which you're a guy who did equity momentum's holding for six months kind of sounds insane to me, but even if you want to do that, you know, maybe you do it with six staggered portfolios. And you get rid of the timing lock, because if you go back and look at MTUM over the last five years, it got like perfectly out of sync with all the major market movements that were happening from 2020 through 2022.
Cameron Howe:
Yeah, I would think I would only apply a six month rebal on a momentum strategy if it was rooted in fundamentals. If you're running a like a technical technical chart based on ROI or ROE or some sort of like fundamental factor, that I think I could stomach a six month rebal on, but like we used to run a strategy month to month and we'd rebalance our ETF on a monthly basis just because those stocks come so aggressively in and out of favor that like a six month rebal would, to your point, you'd have a lot of momentum crashes going on in between them.
Corey Hoffstein:
So every once in a while you come across a paper that sort of shakes your foundational views of everything you know or think you know. And I read this paper I'll find it to you after this podcast and send it to you because this is one that as a guy who's traded momentum equity, you'll typically you think to yourself, momentum is a short lived anomaly. I'm looking at maybe the past nine, 12 months of returns to help me forecast the next two, three, four weeks out of relative performance among securities. And what this paper found was actually there wasn't an optimal forecast period, but the optimal number was the length of time in which you were forming the portfolio. So that look back period plus the holding period. And they found that the optimal number there was 13.
So for example, if you did a 12 month look back, you want a one month holding period. If you had a six month look back, you wanted a seven month holding period. If you did a one month look back, you wanted a 12 month holding period. All of those adding up to 13 and all of those strategies performed equally well.
Cameron Howe:
Huh, that's very...
Corey Hoffstein:
And it just sort of blew my, it changed it. I mean, and who knows that could have been somewhat weirdly data mined that that was the result. But so every once in a while, you know, you see all this literature that looks one way and this paper said, well, actually there's this other weird aspect to it, which is it's the holding period for momentum may actually be, you know, there just might be this magic seasonality constant that is 13 months that this all sort of aligns nicely with.
Cameron Howe:
Magic check number 13. Huh. Yeah, send me that. I'm very interested. You're making me miss my research days where I could actually spend the time to dive in on this stuff.
Corey Hoffstein:
Well, as you know, all too well, 99.9% of the time you spend diving in, it ends up in a research graveyard. So I'm sure if you if you went back and spent another two weeks doing it, you go, well, maybe actually having a productive gig isn't such a bad thing.
Cameron Howe:
Corey, I really appreciate you coming on today. Hopefully there's still some listeners after our Nerd Out session. If anyone's interested in learning more about the work you're doing, what's the best place they can find you on?
Corey Hoffstein:
Yeah. So you can find me on Twitter, at C Hofstein. You can find my podcast flirting with models on Spotify, Apple, YouTube, pretty much every major podcast platform. If you liked this conversation, it's probably this dialed to 11 in terms of nerdiness.
Cameron Howe:
Okay
Corey Hoffstein:
So whether it was good for helping and put you to sleep or you enjoyed it, check that out. And then, if you're interested in some of the work we're doing on return stacking and portable alpha, you can go to returnstack.com.
Cameron Howe:
Wonderful. All right, sir. Really appreciate you coming on today.
Corey Hoffstein:
Thanks so much for the opportunity, appreciate it.
.png)
Curious how Investipal can help accelerate your firm's growth? Chat with one of our solution experts.
Stay connected for the latest updates, features, and curated content.
.png)
