This post is a continuation of a series where we will be providing some of our own thoughts and commentary on the conversations we had in the first season of our new podcast.

This post covers our conversation with Tobias Carlisle, which you can listen to here.


2:09 – Toby starts at the beginning: with school classes that included sheering sheep in Australia.

Corey Hoffstein (“CH”): I was so taken aback by this introduction that I was totally caught off-guard.  I knew Toby had grown up in a fairly remote town in Australia (he likes to joke he’s the only Australian living in LA that doesn’t surf), but I had no idea it was that remote.

Nathan Faber (“NF”):  I wish I had practical skills like that.

Justin Sibears (“JS”):  If anyone knows Nathan even a little bit, they know that last statement is completely false.

4:03 – The peak of the dot-com bubble, mergers & acquisitions, and, “you can’t do a DCF on something that’s losing money.”

CH: I liked this line a lot.  A good reminder that blindly applying rules and equations without any thought can mean you miss the bigger picture.  Yes, a company losing value should have negative value: but only if you expect it to lose money forever.  Trying to look at the problem from another angle allowed investors to step in and buy these companies at a steep discount.

4:40 – How Toby became an “activist” and the influence of the latter chapters of Benjamin Graham’s Security Analysis.

NF: There was a paper published a while ago that showed that you could earn returns above the market return by riding Warren Buffett’s coattails and investing in the holdings Berkshire Hathaway disclosed in their 13F filings.

Following activists can lead to similar results. However, a few specific examples do not justify the rule.

JS: To play devil’s advocate, another paper suggests that Buffett’s alpha is not in fact attributable to being better able to run the companies that he buys.

7:16 – Corey asks Toby to discuss the differences between the Australian and U.S. equity markets.

CH: My thought process in asking this question is that an investor’s approach may be shaped by the landscape they live in.  In other words, U.S. investors may be so equity centric because U.S. equities have performed so well.  I was hoping to get a sense, in asking this question, if Toby’s approach to investing was subtly shaped by the investment opportunities that are actually available in Australia.

NF: Having sector weights so skewed based on market cap leaves a lot of room for smarter sector weighting using factors like value, momentum, and quality. On the flip side, any large deviation from market cap weighting has the potential to exhibit very high tracking error.

JS: A sector-based view is very helpful to just in domestic, but also in international, equity investing.

8:51 – The catalyst for starting Toby’s blog, Greenbackd and the net-net opportunities available in the credit crisis.

CH: Blogging in the financial space is very popular today, but back in 2008, it was far less so.  What I should have asked Toby was, bluntly, “what in the world drove you to start blogging in the first place?”

10:14 – Toby reflects on his 250% return in 2009 and learns a significant lesson.

CH: I loved this story.  250% is an astounding result that almost anybody would be pleased with.  But here’s Toby, asking himself the question, “how did I mess up so badly?”

But for me, the follow up question I asked is an important one to consider.  How did Toby manage to miss investing in so many opportunities?  Did he pro-actively screen them down, and therefore detract value?  Did he simply miss them because his process to identify them was not scalable?  Or was he simply capital constrained from investing in everything and just got unlucky?

JS: I agree that this is a great story.  I think it highlights the importance being cautious when evaluating success/failure on a single realization of history.  Toby’s picks delivered 250% return.  The whole cohort delivered 750%.  One interpretation is that Toby did a great job identifying the theme/factor (net-net investing), but could have done a better job executing on that theme.  This seems to be Toby’s conclusion.  However, as Toby mentions, there was another possibility: that the world was actually ending (at least financially).  Had the financial world ended, it’s at least possible that Toby could have outperformed the cohort substantially if say, for example, his process was able to identify firms that could survive the apocalypse while the other firms in the cohort failed.  Herein lies the difficulty in evaluating managers based on one draw of history.

NF: Cherry picking can look good when taken out of context. The results of the broader picture could have been much better or much worse. Most investment strategies are benchmarked against an alternative investment (e.g. benchmark a value manager to a small-cap value index ETF) to give the returns context. These relative performance differences can be much more informative than absolutes for determining expectations ex-ante and if a strategy is working as expected ex-post.

12:14 – The common occurrence in deep value investing: it’s scary because it’s often the companies that are losing the most money.

CH: This is one of the reasons that value is often considered a risk premium: investors who buy these stocks are insuring investors who are selling them against the business failure risk. For this service, they demand a premium (by buying at a significant discount).

NF: Catching a falling knife is a risk with value investing. Our natural tendency is to shy away from companies that seem like they may be in this situation, but as Toby said, those are the ones that turn out to be true deep-value plays. Momentum overlays can help to some extent, but value investing works because no one else wants to stomach the ride to the bottom before the turnaround happens.

13:42 – The influence of James Montier’s Painting by Numbers: An Ode to Quant and evidence suggesting that simple models can beat experts.

JS: Quoting from Montier: “Grove et al consider an impressive 136 studies of simple quant models versus human judgements. The range of studies covered areas as diverse as criminal recidivism to occupational choice, diagnosis of heart attacks to academic performance. Across these studies 64 clearly favoured the model, 64 showed approximately the same result between the model and human judgement, and a mere 8 studies found in favour of human judgements. All of these eight shared one trait in common; the humans had more information than the quant models. If the quant models had the same information it is highly likely they would have outperformed.”

CH: One of my favorite papers on this topic is Andrew Haldane’s The Dog and the Frisbee.

NF: We have written about how bad “experts” can be at forecasting in the context of professional forecasters predicting interest rate moves.

16:28 – The broken leg theory and asking, “when is it okay to override my model?”

CH: I think there are several subtle points worth acknowledging here.  First, overriding the model is a slippery slope.  If we override the model for a broken leg, will we override the model for another ailment or sickness?  Second, with idiosyncratic situations it can be difficult to discern what is signal and what is noise because we do not have any prior data to work from.  And if we think the new data we have learned is highly predictive, we should probably ask ourselves why it was not part of the model in the first place.  Finally, we should be reluctant to over-parameterize our model, lest we find ourselves with so many input variables that we lose any forecasting strength.  The goal is to find the balance between parsimony and accuracy.

NF: We may disagree with our models at times, but any modification we make basically creates a new, and untested, model.

17:35 – Toby discusses the origins of The Acquirer’s Multiple.

JS: For an activist investor, a capital structure agnostic measure, like the Acquirer’s Multiple, makes a lot of sense because it may be feasible for the activist to change the capital structure post purchase.  For such a person, using something like P/E would ignore this flexibility and potentially leaded to missed opportunity.

19:48 – Discussing Joel Greenblatt’s The Little Book That Beats the Market and the mean-reversionary nature of return on invested capital.

CH: This is a common thread I find among quantitative value investors.  Most have read Joel Greenblatt’s book, but most disagree with the use of ROIC.  The narrative sounds good – cheap stocks that are compounding at a high rate – but it’s worth asking, “why is nobody else buying these?”

I was not wholly surprise that removing ROIC increased return.  After all, you would expect that focusing on higher ROIC companies means you’re buying “safer” stocks and should therefore harvest a lower risk premium.  But I was rather surprised it reduced risk-adjusted return.

NF: Return on invested capital (ROIC) can be a good measure of companies that compounding at a high rate high at a cheap price, but there can be false positives when combining this with value.

23:38 – Re-reading Buffett’s letters and why ROIC may be a poor measure of a “moat.”

CH: As a quant, I’m allergic to qualitative concepts.  But the evidence that Toby discusses by Michael Mauboussin highlights the risk: companies with moats will have high ROIC, but not all high ROIC companies have moats.  Trying to identify the 4% may be a fool’s errand, as there may simply be too few examples over time to try to find the stable, predictive measures.  Furthermore, we cannot discount the role of luck and cumulative advantage may play in establishing moats.  The frustrating answer to, “why did high ROIC persist for this company when it did not for another?” may simply be “randomness.”

NF: “What does the market know that I don’t?” is always a good question to ask.

Toby brings up a good point with mean reversion. With value investing, you are hoping for mean reversion, but if you incorporate a fundamental measure (ROIC, in this case) that mean reverts because there is no economic moat, then you value strategy may be diluted.

27:06 – Corey asks Toby to provide his perspective on the value investing landscape, where he falls within it, and why the Acquirer’s Multiple is the right metric for the investing problem he is trying to solve.

CH: Again, my view is often that an investor’s approach can be shaped by their environment.  In this case, the environment may not be physical, but could simply be based upon how they perceive the world.  Asking Toby to break down the world of value investing was, for me, simply a way to try to get an insight into how he views the environment in which he operates.

NF: While many value strategies are hard to systematize because of special situations or qualitative input, Toby operates at the cusp of quantitative and more traditional value investing.

JS: Ultimately, value investing is about buying companies who are trading at a discount to fair value.  Toby does a nice job of segmenting the value landscaping as a function of how different investors come up with that fair value.  He argues that the Buffett group is willing to give more credit to future growth, allowing for the idea that you buy “expensive” companies that you view will get “more expensive.”  Toby’s deep value side of things is more about balance sheet value, what the company could be potentially liquidated for today.

31:34 – Toby discusses why Apple can screen well when using the Acquirer’s Multiple.

CH: I was excited to ask this question because it strikes me as being so counter-intuitive.  Yet it highlights the fact that how you approach measuring value can dramatically change your perspective on what is cheap and what is expensive.  The counter-point to this line of thinking, however, is that it can be all to easy to make numbers say whatever you want them to.

NF: Different measures of value can lead to very different conclusions. Even buying a well-established company at the right time (e.g. based on cash on the balance sheet with no meaningful debt) can lead to a strong value play.

34:15 – Corey asks Toby pivot to discussing special situations and whether it is possible to employ a systematic investment approach.

NF: There may be some aspects that you can systematize, but this is a really tough area.

JS: It doesn’t have to be systematic or not systematic.  Toby talks about looking for special situations where the stock is cheap regardless or whether or not the catalyst happens.  It would be easy to systematize the value portion of the process, even if the special situations portion remains less systematized.

40:41 – Toby provides an example of why special situations is more of a “Swiss army knife” approach to investing.

CH: My head started spinning during this conversation when I started thinking about all the ways in which in which the same special situation could be played.  This makes it a “high dimensionality” problem, making it very hard to implement in a pure quantitative manager.

NF: There is more than one way to invest in a special situation.

43:01 – Toby discusses how he thinks about sourcing ideas and building a book of special situations investments.

CH: I wasn’t ready to give up on finding some way in which quant could be applied to special situations.

NF: I am normally skeptical of machine learning when broadly applied to investing, but there are some good applications of natural language processing and machine learning in this area.

When sizing positions, Kelly betting can help to balance the risk of ruin and the benefit of the payout.

JS: Kelly betting is a rule used to determine the optimal size of bets where the goal is the maximize log wealth.  For an investment that has a p probability of returning b percent and a q = (1-p) probability of losing a percent, the optimal bet will be equal to p/a – q/b.  All else being equal, the bet size will be larger when the probability of making money is higher, the size of the potential loss is lower, or the size of the potential gain is higher.

44:54 – Looking for a “lazy” balance sheet.

CH: One of the things I found most appealing about Toby’s approach was that he was looking for value plays that had characteristics that would invite a catalyst to bring them out of the realm of value.  In other words, instead of buying a deep value cohort and simply waiting for mean reversion, he is screening for securities that he believes have a higher probability of being acquisition targets or targets of activists. These sorts of catalysts can force the valuation reversion prevent his capital from being locked up in a perpetually undervalued security.

NF: “Too much” cash, buying back stock, paying down debt, paying out dividends can be good signals of a healthy company when looking at valuations.

47:07 – Why using the inverse of your long signals may not be sufficient for creating an effective short book.

CH: The asymmetry of shorting returns (+100% max return versus theoretically infinite loss) plus explicit and implicit cost of carry (borrowing costs, dividends, and the equity risk premium) means that timing in shorting is probably very important. I was surprised to hear Toby mention using momentum as a screen in the short book.

NF: “The market can stay irrational longer than you can stay solvent.” – John Maynard Keynes.  When you are shorting, the downside is theoretically infinite.

JS: There are two potential motivations for adding a short side to your factor portfolio: (1) increase exposure to the factor in question (a long/short portfolio that buys the cheapest companies and sells the most expensive will have more value exposure than a long-only portfolio that just buys the cheapest companies and/or (2) remove market exposure from the portfolio (i.e. I want to make money if my value stocks outperform expensive stocks, regardless of what the market does).  If you are shorting mainly due to reason #2, based on this discussion it may make sense to just hedge out market risk with something like a broad market futures contract.

51:03 – Moving beyond the Acquirer’s Multiple and thinking about how to build a portfolio of deep value stocks.

CH: Toby’s comments on sector neutrality particularly resonated with me.  I put sectors in the category of risk factor: knowing how the sector is behaving, as a whole, likely provides information about how the individual securities within the sector or behavior.

However, sectors are not a risk premium: naively overweighting a sector should not provide an investor with an excess return.  So in most market environments, it probably makes sense to remain sector neutral, but there may be environments where large sector tilts are justified (e.g. the dot-com era).  Evidence seems to suggest that sector neutral implementations offer a higher information ratio, while unconstrained implementations offer a higher excess return.

Trying to choose when to be neutral and when to be unconstrained implies some sort of market timing, which adds a whole extra layer of complexity and risk to the investing process.  Therefore, it likely makes sense to simply choose one and stick to it.

NF: When most value (or any other factor) research constructs short and long portfolios in the same way, it is interesting to think about how you might construct the different sides in different ways.

53:29 – Corey asks Toby about the results of the quant value “horse race” that was performed in Toby’s book Quantitative Value (co-authored with Wesley Gray).  Toby discusses using compound versus individual measures of value.

CH: After my conversation with Adam Butler, which revolved around the benefits of process diversification, I was interested in hearing an opposite opinion.  Toby’s answer caught me totally off guard, since I was under the impression his process was based upon a single metric.  By and large, it is driven by the Acquirer’s Multiple, but he employs other valuation metrics as a check-and-balance process, acknowledging that his core metric can be “tricked” in certain situations.

I thought this was a very thoughtful approach to leveraging the intuition and transparency he has with using single metric, but still diversifying his model risk.

JS: Value strategies tend to be very reliant on accounting measures.  In an ideal world, we should be careful to always make sure we are comparing apples to apples.  One way to do this would be to laboriously clean all of the financial statements of the companies in the universe to make sure everything is comparable.  While doable, this is very time intensive.  Using multiple measures helps to address this risk with a whole lot more efficiency.

NF: Rather than using a compound measure, another option would be to use all of the measures separately and blend the portfolio.

56:18 – If you were an investment strategy, what would you be and why?

NF: Like Adam Butler, Toby has some great diversification!

Corey is co-founder and Chief Investment Officer of Newfound Research, a quantitative asset manager offering a suite of separately managed accounts and mutual funds. At Newfound, Corey is responsible for portfolio management, investment research, strategy development, and communication of the firm's views to clients. Prior to offering asset management services, Newfound licensed research from the quantitative investment models developed by Corey. At peak, this research helped steer the tactical allocation decisions for upwards of $10bn. Corey holds a Master of Science in Computational Finance from Carnegie Mellon University and a Bachelor of Science in Computer Science, cum laude, from Cornell University. You can connect with Corey on LinkedIn or Twitter.