Safe haven by Mark Spitznagel

Try imagining a place where it’s always safe and warm
“Come in,” she said, “I’ll give you shelter from the storm”


I bought Safe Haven by hedgie Mark Spitznagel from a recommendation in one of Monevator’s comments. I’d agree with the comment that the book doesn’t leave you with anything actionable, but perhaps as Dion Fortune said of the Cosmic Doctrine, the object is to train the mind, not inform it. This Spitznagel achieves IMO. It isn’t a long book, I read it in a couple of hours in one sitting, albeit punctuated by watching a movie with Mrs Ermine.

Reading has its systole and diastole, which is why cramming is tough, which is why doing something else midway lets you digest it better – Darwin was a fan of walking for this purpose. I only find that useful for when I originate something creative, but the movie improved the digestion of the book’s 240 pages, presumably by letting something in the background reflect.

Spiznagel is pretty full-on, a reasonable storyteller, and uses metaphor and analogy well. The main takeaway is that many of us  evaluate investment prospects by expected value. Despite the standard FSCS warning that past performance is not a guarantee of future results, that’s sort of what happens. The author disses macro investing, and goes on to make the assertion that managing (tail) risk can be cost-effective. In particular, that it can improve your compound annual growth rate (CAGR) without costing you performance

cost‐effective risk mitigation—or raising compound growth rates and thus wealth through lower risk—is really our comprehensive goal as investors.

Spitznagel, spends the rest of his book showing you how you can recognise an asset class that could do that.

Tragically for you and I, dear reader, that asset class isn’t something that you or I could go out and buy, or synthesise from something we can. It might be possibly in the hedgie world. I am somewhat glad that intuitively I found one of the few assets that sort of comes close-ish. The book also has value in showing that you can compute the optimal amount of that asset class.

Yes, there really is a buried treasure for investors, one that solves our monumental problem by showing that the great dilemma of risk—the ostensible tradeoff between higher returns and lower risk—is actually a false choice. […] We need a more holistic approach; we also need a treasure map to know where to dig.

But just because that buried treasure exists doesn’t mean we will ever find it. The greatest value—more than in the treasure itself—will be in what we gain from the hunt.

I was tempted to issue a refund request, having gotten to the end and being told that the ideal was a chimera, for civilians at least, and since less than 24 hours had elapsed between buying it I would have got away with it.  Amazon track how much of a Kindle book you have read, though I don’t have a habit to returning Kindle books so I’d probably be OK.

But after sleeping on it I came to the conclusion that I did learn something, but in a Dion Fortune like way. My mind was trained, not informed. Most non-fiction reading is to inform the mind. So I got my £15 worth, but it wasn’t the £15-worth I expected.

Spitznagel insights – training the mind, not informing it

Take the Saint Petersburg dice game, a single roll of the dice offers


Wotcha going to pay to play this game? The expected value is ($1+$2+$6+$22+$200+$100000)/6=166,705

but I am guessing most people wouldn’t pay that much, intuitively. It seems obvious that with five chances of being largely wiped out you wouldn’t pay the expected value. Bernoulli’s computation shows if you compute the geometric mean of what you end up with, you can estimate what a reasonable proportion of your total wealth you would pay for this wager. If you had £100,000 then paying about £37k or less to take part gives you a better than even chance of ending up better off. It quantifies the fact that you can take more risk if you have more capital that you don’t immediately need.

Reading the methodology gives an analytical solution to the gut feel approach, and is intriguing. However, the training not informing shows, because most risks you take give a return proportional to amount you put in. However, Spitz has only got started at this point, and he uses a sequence of returns that includes a catastrophic loss (to 0%) to show that where you have a sequence of returns that build on each other then risk mitigation can be worth while,

The arithmetic cost of its risk mitigation is more than offset by its geometric effect—such that its net portfolio effect is positive.

Most of us invest in a single lifetime of a specific sequence of returns. I still remember hearing my German great-grandmother describing sequence of returns risk – they lost their (financial) life savings twice. Fortunately most Anglosphere stock drawdowns aren’t that extreme, but Spitznagels view on central bank meddling suggests that this is not an immutable law of nature, particularly in a declining Imperium.

The Spitznagel edge

Spitznagel despises modern portfolio theory, which is the rough assumption that you buy a mix of less volatile but lower-returning assets like bonds and more volatile but higher returning assets like equities. Inherently in that mix is the takeaway that you will give up some return, and Spitz has no time for such milquetoast ambition.

However, to this mustelid reader he spends a lot of his book in search of something that you could replace bonds with, bonds being the most common MPT risk mitigator of choice1.

As one example, say at the beginning of the ISA year I could save that £20k in an ISA, less an amount that I could go to an insurance firm and say here is £x. If this time next year the market falls more than y%, pay me some lump sum proportional to x (but note NOT proportional to the fall, this is a cliff-edge function and therefore non-linear).

He spends a fair amount of time showing how you would compute the right amount to spend on this insurance, and in his examples it’s not very much. I haven’t given enough thought to whether you can do this with options and CFDs, but I don’t know of anywhere you can go to buy this sort of thing.

You can spreadbet against losses, but in general it is always cheaper to simply buy less of the asset and sit on cash. I have spreadbetted against my ISA in times of market turmoil, but that’s not the same as doing this steady state, which is an exercise in futility.

However, to return to the training the mind aspects, one of his key statements is

We experience profits and losses and all accounting ledgers arithmetically; we experience life arithmetically—one thing after another. This is linear thinking versus geometric thinking. It’s a big difference and essential to our understanding of risk and the disastrous impact of losses on wealth. But it is highly counterintuitive. Here you face an inconvenient, uncomfortable but crucial truth:

Your raw, linear returns are a lie; your true returns are crooked.

Bernoulli’s call to map returns through the logarithmic function was thus a normative one, not a positive one. In basing decisions on the geometric average of expected wealth or returns, not on the arithmetic average, Bernoulli was showing us how we should view risk—not how we necessarily do view risk. And this is precisely where economists got it so wrong.

I find this reasonably compelling. It’s not totally new to me but this exposition is good. I have no idea of if economists got this wrong, but we generally experience a particular sequence of risk. In both the housing market and in the dotcom bust I experienced that the crawl back from a double-digit loss is long and slow, and best made up by Saving More than trying to make it back in that market. If you lose 50% you have to make a 100% profit on what you have left to get back to where you were before.

Some of this you can lean against by not being 100% invested in equities – you reduce your arithmetic return natch, as you are less exposed to the equity market. But you improve your geometric return, because you live to fight another day. Spitz gives you the lowdown in the bit on the Kelly ratio, but again, what makes that less actionable for most is that having seen the value of your equity holdings go titsup in the markets you need to get right back on the horse and throw some of your cash into that now undervalued market. Easy to say, not so easy to do. That’s why people have bonds, and I have gold.  I don’t do bonds, because I estimate 25 times my net DB pension as a bondholding, and unless I get a fair bit older I can’t manage the right mix.

Theory would therefore point me in the direction of 100% equities. But I have had a pretty decent run, I don’t need to shoot for the lights, and sometimes comfort is more valuable than performance. So while Spitznagel wouldn’t approve, I take a lower expected return, because I can.

The big killer is there is no safe haven for little people

Spitznagel has turned the handle on all the things people typically regard as safe havens and qualified them against his specific criteria of cost-effective safe havens (ie they get your CAGR above the 95% confidence interval  of the S&P over a representative set of trial periods)


And the results are in. Little people, you are hosed. As it happens an Ermine does use gold (and there is a useful piece of the Spitz in how you qualify how much gold you should hold, about 20% is right for me)  But before you all rush out to buy SGLP, most of the trial periods where gold lifted itself into Spitznagel success territory happened to be in the 1970s, after Nixon repudiated the convertibility of the dollar into gold at a fixed rate. So gold may not be all that after all.

“Gold is pretty darn good. You just have to understand there’s been a lot of noise around it.” – underlining gold’s value as a safe haven, while noting that it performs best when inflation expectations are high, and historically it’s been inconsistent in mitigating portfolio risk.

Obviously if you can buy insurance on Spitznagel’s terms then you are off to the races. But those terms are tough –

Any punter can devise a trade that does well in a crash. The key is how do you do in a crash relative to the rest of time.

Yeah, quite. From his Yahoo Finance interview via Business Insider interview

“The Federal Reserve is manipulating the most important information parameter in the economy, and that’s the interest rates.”

“I have this expectation of destruction in the financial markets. That doesn’t necessarily mean that someone should just hide away, because that may not be the best strategy either.”

Where’s Clint when you need him, eh? Do you feel lucky, punk?

Spitznagel’s Universa Investments hedge fund returned 4,144% in the first quarter of 2020

An Ermine felt pleased to get out of the first quarter of 2020 with the black tip to my tail intact after selling some crap and shorting some of my ISA. DNFS – bollocks to that. Going for a 40-bagger, now  that’s ambition.

More Spitznagel

Spitznagel’s company Universa

Spitznagel on the FT (Oct 20 this year)

“It would be very hard for bonds going forward to provide cost effectiveness. Bonds really represent the canonical case of the mean-variance approach of lowering the volatility in a portfolio, but being poorer because of it.”


No book is ever gonna tell you what to do successfully as an investor.

Well, this one sure ain’t. There’s a lot of good stuff in there, and I am sure I have brutalised the principles from a mixture of a lack of comprehension, not being as smart as Spitzy-boy and the exigencies of making it into a post. Nevertheless, it will probably reward re-reading, though I am almost 100% sure that it won’t give me anything actionable. Training the mind, not informing it…

  1. TIPS is the archetypal risk-free asset class – risk-free, that is, if you believe the CPI inflation index used by the Fed, which is a different matter. 

20 thoughts on “Safe haven by Mark Spitznagel”

  1. Wow. You turned that round incredibly quickly! From consumed comment to post in 24hrs or so!

    It sounds very similar to anything and everything that Talebs ever written. Contrarian and clever but completely in-actionable.

    I’ll put it on the to-read list..


    1. It’s an intriguing read. There’s a bit more about his modus operandi in this Forbes article which seems to indicate he steadily places slightly out-of-the-money puts below the current market price. I’m sure it’s all a bit more complex than that and he will get better rates than you or I. As you said, Contrarian and clever but not actionable by proles

      We should also remember the bear case always sounds smarter


    1. > but is your State Pension and your DB pension not your safe(st) haven assets?

      That’s why I don’t do bonds, although for my age profile if my ISA holdings were the entirety of my retirement savings it would normally be more than half bonds according to MPT.

      However, I don’t have a 20-year old, or even my 50-year old self’s appetite for volatility (accepting volatility risk), and I hate current valuations. While I don’t share Spitznagel’s Austrian School rabidity about central banks, I do find current interest rates at historic lows for ten years since the GFC fishy and share his view that will be stoking some sort of pathology somewhere, that eventually won’t go on because it can’t go on. However, experience has shown me that macro calls don’t really work for an individual investor because the market can be irrational for longer than I can be solvent.

      Although I bought the book to see if there was some actionable way to protect myself against these high valuations what I did learn was more the philosophy, and a better handle on qualifying risk. I wasn’t able to do the detail of Spitznagel’s approach justice in this short piece, but for instance he is quite emphatic that if you protect the downside you should not do that dynamically (say at high valuations). I find him intellectually compelling though it goes against the grain

      assandras typically and ironically lose more in their safety from looming crashes than those crashes would have even harmed them. (Cassandras make very bad investors.) Markets scare us far more than they harm us. They are very, very good at making us feel safe when we shouldn’t and scared when we needn’t. This is the market’s inherent, constant deception to rid investors of their positions.

      Cost‐effective risk mitigation cannot be a specific act; it must be an ongoing policy. In the words of Aristotle: “As it is not one swallow or a fine day that makes a spring, so it is not one day or a short time that makes a man blessed and happy.” In the same vein, it is not any well‐timed trade or prediction that mitigates risk. Risk mitigation needs to be a sustained way of life or habit, not a transient state.

      The concavity of risk (in a geometric progression, which is how an investor progresses through time) due to the log functions in Bernoulli’s expected value means losses slaughter you quickly. I haven’t spent enough time with this to qualify if insurance (in the form of put options at roughly -15% in the examples) would help with that, although his examples show that. It would still be a tough sort of trade to execute – for many years you will make minor losses on the options and then explosively make returns in down years, which tend to be much shorter periods of time that rising years. Most of us would struggle to do that, and probably pack it in just before a bear market, on the priciple that it’s all different now.

      So the problems are that it may be difficult for civilians rather than hedgies to get the derivatives at a reasonable price to carry deadweight for years at a time, it may be emotionally difficult to live with the small drag on performance. Though people seem chilled with bonds, and I am OK with the deadbeatness of gold, these are an invisible drag on performance you never know is there, whereas Spitznagel’s insurance option involves buying a little bit of insurance for many periods watching it explicitly going down the toilet. Until it doesn’t. While it’s rational to combine that with a decent slug (~90%) of normal equity allocations which is expected to carry the CAGR appreciation in the good times, it may be a challenge for civilians to actually do, even if they can find the insurance at the right price.

      But as an exercise for the noodle, £15 quite well spent on reflection!


      1. As you noted yourself in your post – you are not necessarily after optimising CAGR. That is, IMO, the framing of the problem matters to the applicability of the solution(s) proposed – whether they are actionable or not.

        I may have mentioned this blog before, but Will Selden has written at length about his “conversion” from arithmetic returns to geometric returns – amongst many other ret fin things, see e.g.

        Also, Mike D at seven circles has an on-going series of posts (IIRC nine and counting) looking at options. These may be of some interest too. The first part is at:

        Personally, I am sure the silver bullet will remain elusive; but some people will, of course, get lucky along the way. Having said that, it is probably fun to play with their ideas.


      2. > it is probably fun to play with their ideas.

        I backtested this against the SPX, roughing out what I could do with IGindex puts. Over six months a -15% puts costs about 2% of the index to buy, over one month it’s 0.14%. I could only simulate evaluating against the strike price at the maturity date. I appreciate you can do more with options, and I have no means of capturing this or the greeks to compute that. This was Jan 2007 to mid-2009. I targeted a nominal capital of £100000 in the SPX.

        In the six month maturity I purchase £100000/6 each month spread across a ladder over six months, in the one month maturity I simply bought the entire £100k nominal put each month

        The takeaway is that Spitz clearly buys his insurance at a rather better rate, but in a bear market at IG prices returns can be pretty decent, about a five-bagger across 2 1/2 years in the monthly case, and they arrive very peakily and non-intuitively.

        The downside is you are paying a steady 1.7% insurance premium p.a. in the monthly case. This assumes the targeted put amount matches your equity amount. That’s not necessarily dreadful, after all there are active finds with that much loading. But it’s also pretty high touch.

        So I could believe Spitz and his PhDs could do much better, because there is much I couldn’t simulate and they have scale and connections.

        Put options could be a useful tool in situations like last March, though to actually use them in a downturn like that will still be fairly high-touch. But as a steady insurance policy across time, well, that’s way above my pay grade, and the gold loading in the interests of

        Besser gut schlafen, als gut essen.

        is a better match for my resources and temperament. But it’s good to have some feel of the road not travelled.

        From 7circles ERN seems to be doing something different, in selling at the money puts. He seems to be using an anomaly in pricing due to buyers’ collective loss aversion, and integrating over many trials using the CLT to bring the negative skew normal.

        There is an argument that I should spend more of my finite time near standing stones, bird reserves or even solving technical problems for clients than mulling over this, but keeping the brain ticking over in different areas has value in itself…


  2. Very interesting article and thanks for posting it.

    Recent history has shown bonds to be a solid hedge against equity declines, which may or may not continue. The scenario, which concerns me is a sudden, unanticipated and uncontrolled rise in interest rates in which central banks have lost ability to keep rates to the floor. Not that this is expected and there are plenty of very good reasons why I don’t think this will happen. That’s not the point though. I don’t think my house will burn down but I still buy insurance. Plus I continually say to myself I don’t know what will happen.

    There’s limited insurance available to the average punter for such circumstances of unexpected and quickly rising rates (I’m not talking about a very talked about 25 bps rise that didn’t happen!). If last Thursday the BOE had increased rates by 5% in day (note this happened on Black Wednesday albeit today’s circumstances are entirely different) then one would expect all risk assets gold, LT bonds, ILG, UK equities, property to head south very quickly. There would also presumably be an almighty recession give our sensitivity to rates. With the only obvious asset that’s accessible to the average person being cash being unchanged and relatively speaking suddenly more valuable as real interest rates turned positive. So cash / no borrowings will reduce your exposure / provide an optionality of sorts but won’t have the affect of increasing your net wealth and obviously has a net drag given it’s negative real return. In that respect buying portfolio insurance (e.g. buying puts) seems sensible and one I’d be interesting in pursuing. The trouble is one could do that for years or decades without benefitting. Mind you it’s the same as buying any other insurance I suppose. It’s also hard to access as you noted.

    FWIW, I continue to think on the balance of probabilities that interest rates will be negative on a real basis for a long time to come (the manipulation of interest rates is flagged by the author of the book). And I’ve actually just borrowed a healthy six figure sum at <1% interest only and fixed for five years as a hedge against financial repression, which I'll invest. Which generally is a signal that we're at the top of the market :). Anyway holding cash LT is costly and why portfolio insurance is a good idea if you are trying to hedge the downside. Trouble is it's hard in practice to access and even harder mentally to stick with it….year after year after year.

    Slightly contradicting myself above – Naeclue (commentator on monevator) has regularly talked about his asset allocation being a number of years expenditure in cash and the rest in equities which is a reasonable hedge against a market crash impacting your standards of living, albeit it doesn't hedge what would happen to your net wealth. Anyway there's a lot to like about his / her's approach imho.

    The other hedge is just to have a massive margin of safety. if my net wealth halved due to rising rates, then I'd certainly be walking less tall but it would have no impact on my day to day living standards. Again no hedge to net wealth though.

    Another other option here is an index linked annuity if you believe RPI / CPI is the real measure of inflation as you say below in you notes.

    I think in the tail risks of say climate change (I mean the known issues accelerating much quicker than expected) and unexpected conflict between countries, gold is likely to perform quite well but who knows for certain.

    Much to ponder but little to action probably. As Chuck Prince said "When the music stops, in terms of liquidity, things will be complicated. But as long as the music is playing, you've got to get up and dance"

    Liked by 1 person

    1. I’m with Naeclue that holding a significant (about three years worth) of expenditure in cash makes for eaqsier sleep. Intellectually it’s a bit bonkers in my case – it made sense when my sole income was from savings and investments but with annuity income to cover that expenditure it’s a performance drag that’s hard to justify.

      The Harry Browne Permanent Portfolio isn’t bad for those in decumulation of a less optimistic nature, though as MMM notes, not being an optimist is about the biggest drag on returns you can get. As well as being an old man’s portfolio, it demands you start with more. The answer to nearly all personal finance worries is start with more money, but that’s at variance with the retire early side of FIRE.


      1. There’s a fund by Mark Spitznagel that’s good to invest in…..

        A combination of real assets – nothing particularly exciting….


      2. > a fund by Mark Spitznagel
        Although I wouldn’t necessarily go that way, my Google-fu failed me in how you’d actually do it. Universa seems to be the umbrella org but do you need the right handshake to find the ticker/access it at all


  3. “However, Spitz has only got started at this point, …”

    You reading his book taught him to swim? That is unusual.


  4. I am shocked to be reminded Amazon tracks your reading in a Kindle. I did know that, I had a very early Kindle, but when it came time to upgrade I bought a Kobo. At the time, it was better that the best Kindle.

    Anyway, I still buy e books from Amazon, but download them onto my computer and into Calibre ( , which I can recommend, to manage my ebooks.

    With a little help from *cough* Apprentice Alf I can convert them to any format to read on any device. E-pub format books go nicely on the Kobo. There is no way Amazon is tracking my reading habits and because I always pay for my books and don’t share them I think this is totally harmless.

    In response to your prevoius blog post I tried out importing a PDF document into Calible and sending it to my Kobo. That worked fine.
    I remember once watching people check in at the Eurostar terminal with their mobiles and it really did look like a system failure waiting to happen. Perhaps I will use my Kobo, with copies of PDF tickets loaded into it as a back up to printing the tickets out.

    Are there any hints you could give on how to read library books on my eReader?

    I will add ‘Safe Haven’ onto my wishlist, but I agree with The Rhino, it sounds very like Talebs approach to managing risk.


    1. > Are there any hints you could give on how to read library books on my eReader?

      If calibre is already on the machine you d/l your library books with, and that machine uses Adobe digital editions to view the library books, then Calibre can open the books downloaded by ADE and convert them. I’m not sure if Apprentice Alf gets involved in that, though I have his work on my copy of Calibre (and read library e-books on Kindle)


      1. Thank you. I will have a go at getting it to work.

        One book I can really recommend is ‘The Intelligent Asset Allocator’ by William Bernstein. It has a fair few graphs, scatter diagrams and tables which never render well on an eReader, so go for a hard copy if you decide to follow this one up. It is the best explantion of MPT I have come across and explains the concept of the efficient frontier. It has been hugely influential on my own investing and I think you would really enjoy it. It is not a light read though and I had to dig my old statistics tables out of the attic back when I read it.


    2. It certainly does; if you are reading a downloaded book on more than one device, finish reading one time at (say) p.100, swap to another device and read to p150, if you then swap back to the original device it will open at p100 but query if you want to move to p150. (This is on iPad / MacBook, but probably across the board)


  5. Not quite finished with Spitznagel’s book, but it is on my bedside table, so never more than 15 pages per day … Still, entertaining and compelling reading. If you think it isn’t actionable, have a read of his previous one, The Dao of Capital:
    One option (pun intended) to achieve what he is suggesting by way of insurance is to buy calls on UVXY –
    Agreed, not every broker will give you the option (!) to buy them, but Interactive Brokers does.


  6. I would agree with your synopsis. It’s a very good book but not easily actionable for retail investors. Also leave the Austrian economics and hyperbole at the door.

    His criticisms of economists and MPT are, in my view, valid. I’ve had to teach young asset managers the basics of MPT and when you go through all the initial and boundary assumptions, it’s pretty hard not to see it as anything more that a nice toy model. The assumption of ergodicity, for example, just runs through economics, despite not being valid. They want their models tractable, not a necessarily a good representation of the data. I find people constantly confuse ensemble probabilities with path-dependent time distributions. Hence they underestimate the probability of ruin in their portfolios.

    I have very small % of my portfolio in his fund. It’s not really a fund at all, more a risk management device to hedge ruin. Yes, it did phenomenally well in Mar 2020 but the prior big payoff (above 1000%) was 2015. In the meantime it decays very rapidly. A few percent, even 1% of your portfolio, is probably as much as many could stand.

    As someone who specializes in building positive convexity portfolios, it’s not as simply as buying some downside puts. That strategy is almost certainly guaranteed to lose money over the long term. In fact, the reason why we can construct positive convexity portfolios at all, with modest downside and substantial upside, is only because the majority believe in MPT/efficient markets/CAPM, are addicted to positive carry, prefer indexing etc. We need you to believe that active strategies cannot work. So don’t read that book again!

    Liked by 1 person

    1. It seemed the thrust of his argument that you only need a very small mount of the ‘insurance’ element, but I was left with the feeling yes but how do I make it happen. I was reasonably sold that it could be done and it would be a good idea, and at current valuations and general #FlyingPigs360 it would be a good thing for it to be done, just that it was way above my pay grade 😉

      I’m not sure I ever believed in efficient markets, but I was fortunate enough to go for maximum burn out of the GFC. It’s a bit like the way they teach you about laminar flow because it’s tractable, with the aside ‘oh but it all goes titsup when turbulence happens’ Well yes, quite!

      I wonder what the FI/RE landscape will look like five to ten years hence


Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s

%d bloggers like this: