Essay - Why Hold Cash?

[racemize]

Hi All, I've finally finished writing and revising my essay on holding cash in a portfolio.  Several of you have asked about it over the last few months (which was flattering by the way), and I'm sorry it took me so long to get it out.  This year has been incredibly eventful for me, so I didn't get to devote the time needed to finish this until recently. 

 

In any event, I hope it is useful to board members.

 

https://www.dropbox.com/s/96hwafdp5egf460/2014-15%20Why%20Hold%20Cash.pdf?dl=0

[LC]

Fantastic! I look forward to it.

[ASTA]

cool and thanks :D

[cobafdek]

Regarding the Variable CAPE Model (page 7), what was cash-level percentage-range used?  From late 1950 to about 1974, and then from about 1991 to 2009, looks like nearly 100% cash?  Would the flat-line appearance disappear if cash fluctuated between 0% to no higher than 50%?

 

Also on the Positive 10 Year Spread (page 9), from about 1980 - 2004, also 100% cash?

 

I realize these are nit-picky points, and minor adjustments likely would not change your thesis.

 

Anyway, impressive work.  Thanks for posting this.

 

 

[Liberty]

Awesome, thank you! Looking forward to reading it  8)

[racemize]

Regarding the Variable CAPE Model (page 7), what was cash-level percentage-range used?  From late 1950 to about 1974, and then from about 1991 to 2009, looks like nearly 100% cash?  Would the flat-line appearance disappear if cash fluctuated between 0% to no higher than 50%?

 

I used the percentile that it was, which in those years I think was 100%.  I just re-ran it with a 50% limit, and got a CAGR of 6.53% before tax and 5.28% after, so still not better than fully invested, but better than 100% bailing.  Or re-running it compressing the results to between 0.5 and 1 gives 7.39% and 6.01%, respectively (also underperforming).

 

Also on the Positive 10 Year Spread (page 9), from about 1980 - 2004, also 100% cash?

 

These spread experiments came from an academic paper that did have outperformance from 1970-2001 (and also in my own tests from 1970-2014), so I was mostly testing them to see if they worked over a longer historical period, that didn't largely sample the huge bull run to 2000.  In any event, they did full investment or not based on that.  I didn't do any testing to see if it would work with partial numbers (e.g., 50% instead of 0).  I just ran 0.5 instead of 0, and it is better, but not better than 100% invested. 

 

I realize these are nit-picky points, and minor adjustments likely would not change your thesis.

 

No worries, I tested quite a bit more while I was messing with it, and the tests I showed aren't all the possible ones.  Certainly there are other possible models and tests, but essentially nothing I found would work, so it didn't seem like there are any easy answers, at least (which again, should not be surprising--easy answers should be arbitraged out).

 

Anyway, impressive work.  Thanks for posting this.

 

Thanks! 

[Mephistopheles]

Thanks race!

[Liberty]

Just finished it. Good job! Obviously a lot of work went into this, you should be proud of the result.

 

One question: Have you looked at leverage? The next logical step, IMO, is to look at what'st he optimal level of leverage (as long as cost of leverage is below expected returns by a big enough amount that you have a margin of safety, and that the level of leverage is low enough that you can go through most downturns without being a forced seller. ie. maybe 1.3-1.4xx gives the best results over the long term and you would only delever a few times when interest rates spike up..?).

 

Just curious to know if you've looked into this.

 

Thanks again for sharing your work!

 

Update: kind of thinking of this paper:

 

http://islandia.law.yale.edu/ayres/Life-Cycle%20Investing%20Working%20Paper.pdf

[racemize]

Just finished it. Good job! Obviously a lot of work went into this, you should be proud of the result.

 

One question: Have you looked at leverage? The next logical step, IMO, is to look at what'st he optimal level of leverage (as long as cost of leverage is below expected returns by a big enough amount that you have a margin of safety, and that the level of leverage is low enough that you can go through most downturns without being a forced seller. ie. maybe 1.3-1.4xx gives the best results over the long term and you would only delever a few times when interest rates spike up..?).

 

Just curious to know if you've looked into this.

 

Thanks again for sharing your work!

 

Update: kind of thinking of this paper:

 

http://islandia.law.yale.edu/ayres/Life-Cycle%20Investing%20Working%20Paper.pdf

 

Thanks Liberty.

 

Regarding leverage, I haven't done a lot of actual testing, but I've been thinking about how it might work, given the testing I've done.  And I think the answer is that leverage is generally good to use, but you can't ever get margin called, as that is highly destructive.  So, it will come down to how volatile your portfolio is (which is similar to what this essay comes down to).  Thus, if you are very volatile, no leverage is likely best (and perhaps holding cash routinely as well).  But if you are less volatile, then some leverage is good.  So I tend to think that some leverage is probably warranted for the best long term results, assuming a long-enough time period.  Again though, I haven't done a lot of testing to be sure about that line of thinking.  Perhaps I should make it a follow-on essay...

[treasurehunt]

Wow, this is really impressive work! Any plans to get your paper peer-reviewed and published? Or to share your calculations for independent verification?

 

I think your results are very relevant to investors -- I have to rethink my cash strategy after reading your paper, for example -- so it would be nice to make sure that there are no errors in calculation and such.

 

In any case, thanks for sharing your work.

[racemize]

Wow, this is really impressive work! Any plans to get your paper peer-reviewed and published? Or to share your calculations for independent verification?

 

I think your results are very relevant to investors -- I have to rethink my cash strategy after reading your paper, for example -- so it would be nice to make sure that there are no errors in calculation and such.

 

In any case, thanks for sharing your work.

 

Thanks for your kind words.  I don't care too much about publishing it officially--I just enjoy working on these essays and am happy to make them available to whoever is interested.  I worked on the calculations with some other people to make sure I was doing it right and spent a lot of time on them, but I don't mind further verification if people are interested in looking at it and/or doing more tests.  Most of the calculations are just done in big google spreadsheets.

[cobafdek]

Perhaps I should make it a follow-on essay...

 

BE CAREFUL WITH WHAT SUGGESTIONS YOU CASUALLY THROW OUT HERE!  In no time, you'll get scores of kibitzers asking about maximum drawdowns, lengths of drawdowns, comparative beta (pardon the 4-letter word) numbers for rolling 10-, 15-, 20-year periods, etc., etc., etc.  You'll never sleep . . .

[philly value]

I'm wondering if the answer changes materially for fund managers who may face large investor redemptions correlated with drops in their portfolio. The way I have always thought about cash is that holding a large amount of cash in an individual portfolio is usually not optimal, and probably a slight amount of leverage would be ideal, given that you have the ability to remain fully invested even if your portfolio declines significantly in value. But for a fund manager with clients who are not as patient, it seems that holding cash may be optimal so as to not be forced into selling positions far below their value, particularly if the manager is holding assets that are not highly liquid.

[Liberty]

Just finished it. Good job! Obviously a lot of work went into this, you should be proud of the result.

 

One question: Have you looked at leverage? The next logical step, IMO, is to look at what'st he optimal level of leverage (as long as cost of leverage is below expected returns by a big enough amount that you have a margin of safety, and that the level of leverage is low enough that you can go through most downturns without being a forced seller. ie. maybe 1.3-1.4xx gives the best results over the long term and you would only delever a few times when interest rates spike up..?).

 

Just curious to know if you've looked into this.

 

Thanks again for sharing your work!

 

Update: kind of thinking of this paper:

 

http://islandia.law.yale.edu/ayres/Life-Cycle%20Investing%20Working%20Paper.pdf

 

Thanks Liberty.

 

Regarding leverage, I haven't done a lot of actual testing, but I've been thinking about how it might work, given the testing I've done.  And I think the answer is that leverage is generally good to use, but you can't ever get margin called, as that is highly destructive.  So, it will come down to how volatile your portfolio is (which is similar to what this essay comes down to).  Thus, if you are very volatile, no leverage is likely best (and perhaps holding cash routinely as well).  But if you are less volatile, then some leverage is good.  So I tend to think that some leverage is probably warranted for the best long term results, assuming a long-enough time period.  Again though, I haven't done a lot of testing to be sure about that line of thinking.  Perhaps I should make it a follow-on essay...

 

That makes a lot of sense. If I was routinely sitting on uninvested cash, your essay would definitely make me rethink my approach. But when I read the paper I linked above last year, I did a lot of thinking about it and ended up starting to use some margin. I'm 32, and it makes a lot of sense to me that I'll benefit more from compounding (as long as I don't have to be a forced seller at exactly the wrong time, which is why I'm not using a lot of margin capacity) by having more dollars invested in the early years of my life than I would without margin.

 

The unused margin capacity can also acts like cash would in extraordinary situations. If there's a fat pitch coming my way, I can always jump on it with margin and then delever later back to my target multiple.

 

I know that in some ways it's not as conservative as not using any debt, but if you take the longer view, I think it's in some ways riskier to not use at least some small amount of margin (especially at the 1.5% that I'm paying at IB); I feel that I have a higher chance of being poorer in 30 years if I don't use it than if I do.

 

Sure there might be another 2008-2009 crisis where everything is correlated and goes down 50% and I become a forced seller. But what if there's a 20-30 year stretch without any crisis bigger than I can absorb without forced selling at my relatively conservative level of margin use? I think the benefits during the good times would likely outweigh the negative impact during the really bad times.

[merkhet]

The problem with this, of course, is that it all comes down to timing.

 

That paper was co-authored with a professor of mine, Barry Nalebuff, and there was a subsequent book -- if I remember correctly, in the book, they show that taking on leverage with bad timing will cause you to have a worse result than had you not taken leverage (an obvious result), but it would, on balance, decrease the standard deviation of your overall returns (a less obvious result).

 

Or something along those lines -- I can't quite recall as it was a year or so ago.

[Liberty]

The problem with this, of course, is that it all comes down to timing.

 

That paper was co-authored with a professor of mine, Barry Nalebuff, and there was a subsequent book -- if I remember correctly, in the book, they show that taking on leverage with bad timing will cause you to have a worse result than had you not taken leverage (an obvious result), but it would, on balance, decrease the standard deviation of your overall returns (a less obvious result).

 

Or something along those lines -- I can't quite recall as it was a year or so ago.

 

Wouldn't this be mostly a problem for people buying regardless of valuation? If you do the bottoms up work and only buy things that you find attractive enough, it seems like it should mitigate that risk.

 

But if you just lever up to lever up and end up buying indexes at the top of the dot com bubble or whatever, then that obviously could cut the wrong way pretty deeply...

 

If you lever up to, say, 1.3x, it would take something pretty fierce to make you a forced seller. And over 30-40 years, if you can pick good businesses and if history is any guide, there will be a lot more good years than bad, so having extra exposure to good businesses certainly seems like a good thing, especially in light of the premise of the paper about lifecycle balancing (people invest fewer dollars when young, at the time when compounding would be most powerful for them).

 

I didn't know there was a book. Is this the one?

 

http://www.amazon.com/Lifecycle-Investing-Audacious-Performance-Retirement/dp/0465018297/

 

Interesting, the top review is by Robert Schiller:

 

"I have been getting flak for endorsing the Ayres and Nalebuff book (see above on Amazon.com), just as the authors are for writing it. People are thinking it certainly sounds reckless for young people to leverage two to one into stocks. For some young people, it certainly is. Those who do this with their personal savings and are contemplating buying a house soon could lose their down payment, and thus be unable to buy. This factor is increasingly important after the subprime crisis since no-down-payment mortgages are hard to find now. But Ayres and Nalebuff are advocating such aggressive stock market investing only for retirement portfolios. Most young people could survive an annihilation of their retirement savings; they still have plenty of time to rebuild it later and it may generally be a good bet to take just such a risk. This is the basic point that Ayres and Nalebuff make, and it is right. I worry that this book in the wrong hands (of people with a gambling impulse or who are really more precarious in their financial situation) the book could encourage irresponsible investing. At the present time, the stock market looks rather pricey, and I am less optimistic about young people leveraging stock market investments right now. But, as a general, long-haul advice book, for savvy people who can judge their situation and not get themselves into a corner, the book is indeed valuable. This is not "Dow 36,000" again, as one reviewer says. This is a book about overcoming standard prejudices about investing, and as such it is an important book.

 

Robert Shiller"

 

For those who haven't read the paper that I linked above (I recommend it), another review has a summary of the general idea:

 

"I believe that this is a ground-breaking book.

 

Here's a summary:

1) Most folks tend to have the overwhelming majority of their exposure to the stock market (i.e., most dollar-years of stock market exposure) during a one-to-two decade period late in life (e.g., perhaps during their 50s). The reason for this is that folks tend to accumulate wealth exponentially -- they save exponentially and their investments tend to grow exponentially.

2) This generally works fine UNLESS they are unlucky and the stock market suffers through a bad decade when they have the most dollars exposed thereto.

3) The main idea here is that you would be better off, conceptually, if it were somehow possible to more evenly spread out your stock market exposure over your entire life. This idea of "time diversification" is quite sound. If you were somehow able to do that, and during the decade of your 50s the stock market goes to h*ll, so what! You've hot lots of other decades of stock market exposure to make up for it!

4) Of course, the devil is in the details. Is it possible to more evenly spread out your stock market exposure through your entire life? Yes."

[enoch01]

I may be dense, and I apologize in advance if this sounds like cold water, but isn't the conclusion trivially true?

 

"...this essay is not focused on remaining fully invested at all times, but instead advocates remaining fully invested when there are still compelling investments to be made."

 

If compelling investments can be made, then by definition they are better than cash, right?

 

[Liberty]

I may be dense, and I apologize in advance if this sounds like cold water, but isn't the conclusion trivially true?

 

"...this essay is not focused on remaining fully invested at all times, but instead advocates remaining fully invested when there are still compelling investments to be made."

 

If compelling investments can be made, then by definition they are better than cash, right?

 

I can't speak for Racemize, but the way I understood it, some people target a certain % of cash "just in case" some big huge fat pitch comes along, because they believe that having the cash to jump on these infrequent opportunities makes up for the drag that the cash creates the rest of the time. So they might have a hurdle of 15% and invest in whatever they can find that meets that, but they keep an extra 20% cash on hand waiting for "blood in the streets" scenarios, even if they could use that 20% to just buy more of what they have in the other 80% of their portfolio, or similar things.

[cobafdek]

I may be dense, and I apologize in advance if this sounds like cold water, but isn't the conclusion trivially true?

 

"...this essay is not focused on remaining fully invested at all times, but instead advocates remaining fully invested when there are still compelling investments to be made."

 

If compelling investments can be made, then by definition they are better than cash, right?

 

I bet Liberty is correct in what racemize means, but enoch01 is correct to examine the concluding sentence closely.

 

To my mind, nearly all the data presented prior to the 2 short conclusion paragraphs IS FOCUSED ON and DOES SUPPORT the idea of remaining fully invested at all times.  After all, the cash strategies are compared to the S&P index, and we all know this index contains a lot of non-compelling investments. So even if the portfolio contains a lot of non-value stocks, racemize does conclude that most people should hold this non-compelling index and no cash.

 

But then he goes on with the concluding sentence, which you highlighted.  As such it's a sensible opinion, but strictly speaking seems like a non sequitur to nearly all that came before.  The exceptions are Portfolios 2 and 3, where it seems like racemize is bending over backwards trying to find some scenario to support the cash option.

 

So the conclusion is actually a racemic mixture, and racemize is being true to his name!

 

Still, an impressive paper.

 

[investor-man]

Thanks for posting!

[racemize]

Hey guys, thanks for all the discussion and comments!  So with regard to the purpose of the essay, I tried to lay it out a bit in the first paragraph, but perhaps didn't do a great job of it.

 

In essence, I've had a lot of discussion with people on the board about when and how they hold cash, and I wanted to know which were supported by data and which were not in terms of superior long-term results, so I started testing.  Here are some of the strategies that I ran into:

 

1) Stay fully invested all the time (e.g., Packer, stock screening methods, etc.).

 

2) Ignore the macro, and focus on individual investments.  Accordingly, invest when opportunities meet your hurdle and don't when they do not (I believe this is what Buffett believes and says repeatedly).

 

3) Pay attention to the macro and hold cash when things seem "heated", because when large stock corrections occur, all of your value investments will as well.  This has a lot of different aspects, such as paying attention to general stock market P/Es, CAPE, corporate margins, Stock market/GDP, etc.  Gio and I have gone back and forth quite a bit on this, and I'm sure he can provide a lot of investors that are on board with this concept.  It was this investing style versus 2) above that I wanted to answer, primarily.  The way that made sense to me to do it was the following: a) under the assumption that one could accurately catch the bottoms, determine whether or not outperformance was possible, but having the opportunity cost of holding cash in your portfolio outside of the bottoms, which was the first section of the essay and b) determine if market valuation metrics, and particularly CAPE (since it is more correlated with future returns than P/E is), could be used to identify how much cash to hold.  Basically, my point of view is, if you rely on a metric (such as CAPE) to make decisions about cash, then it should be predictive, and if not, should basically be ignored for that purpose.  I was just trying to see what the data supported.  I also went and looked up as many "outperforming" papers that I could find, and then test the ones that seemed to make sense and weren't based on hindsight data analysis.  One of the best ones was based on the spread between bonds and stocks, which I included in the essay.  I did not find any outperformance models that worked outside of sample, in a large historical context.

 

4) Pabrai's method, where the hurdle for each subsequent level of cash increases in order to ensure that each marginal dollar is used for progressively better investments.  I asked him about this, particularly whether or not it worked and how often these big events had to happen in order to result in outperformance versus his previous method.  He was short of time as it was at the FFH meeting, but essentially told me "well, you'd need to do the math".  So I attempted to test that with the hypothetical models.

 

And finally, I thought it would be interesting to just take actual portfolios and back test to see if holding cash ever made sense (which is related to item 4).  i.e., if holding cash for downturns would yield higher results. 

 

So, I wasn't ever really trying to approach the problem of comparing 1) or 2), but addressing 3/4 versus 2).  Basically, just comparing if holding cash outside of investment criteria made sense.  I think it is more difficult to make the leap that 1) is better than 2), but it is certainly possible.  The problem with 2) versus 1) is that it depends on the skill of the investor and the opportunities that arise, so is hard to test in any real sense.  That same criticism can be applied to my testing of 4), since the answer depends on how reliably one can identify the large opportunities.  If they can be found consistently, then holding cash for them makes sense.

 

Hopefully that all made sense, I need to go eat breakfast!

[racemize]

That makes a lot of sense. If I was routinely sitting on uninvested cash, your essay would definitely make me rethink my approach. But when I read the paper I linked above last year, I did a lot of thinking about it and ended up starting to use some margin. I'm 32, and it makes a lot of sense to me that I'll benefit more from compounding (as long as I don't have to be a forced seller at exactly the wrong time, which is why I'm not using a lot of margin capacity) by having more dollars invested in the early years of my life than I would without margin.

 

The unused margin capacity can also acts like cash would in extraordinary situations. If there's a fat pitch coming my way, I can always jump on it with margin and then delever later back to my target multiple.

 

I know that in some ways it's not as conservative as not using any debt, but if you take the longer view, I think it's in some ways riskier to not use at least some small amount of margin (especially at the 1.5% that I'm paying at IB); I feel that I have a higher chance of being poorer in 30 years if I don't use it than if I do.

 

Sure there might be another 2008-2009 crisis where everything is correlated and goes down 50% and I become a forced seller. But what if there's a 20-30 year stretch without any crisis bigger than I can absorb without forced selling at my relatively conservative level of margin use? I think the benefits during the good times would likely outweigh the negative impact during the really bad times.

 

I guess I agree with the younger age bit, but I tend to not care what the age is--I'm just focused on maximizing returns, regardless of age.  So, in that sense, if leverage makes sense, then I'm for it.  But I think it crucially depends on how volatile your portfolio is/will be, and that is very hard to know.  For example, if you hit a 60% downturn like Pabrai did in 2009, then it could be potentially devastating.

 

I like the idea of using margin for the extraordinary opportunities.  e.g., staying mostly invested all the time, and if 2009 hits (or broad down-turns), using the margin at that time.  Or similar, but low levels of margin in ordinary times (e.g., say 10%).  I think I would hesitate to have normal levels of margin at 30-40%.  But I haven't tested this to be sure.  I also think this is very hard to test well to come to any solid conclusions about what levels of margin should be held though. 

 

Addendum: I think Packer has given this more thought than I have, so it might be good to ask him, or see if he feels like pitching in his thoughts.

[racemize]

I may be dense, and I apologize in advance if this sounds like cold water, but isn't the conclusion trivially true?

 

"...this essay is not focused on remaining fully invested at all times, but instead advocates remaining fully invested when there are still compelling investments to be made."

 

If compelling investments can be made, then by definition they are better than cash, right?

 

I can't speak for Racemize, but the way I understood it, some people target a certain % of cash "just in case" some big huge fat pitch comes along, because they believe that having the cash to jump on these infrequent opportunities makes up for the drag that the cash creates the rest of the time. So they might have a hurdle of 15% and invest in whatever they can find that meets that, but they keep an extra 20% cash on hand waiting for "blood in the streets" scenarios, even if they could use that 20% to just buy more of what they have in the other 80% of their portfolio, or similar things.

 

I'll just nominate Liberty as my spokesman. 

 

Yes, this was the purpose of testing the hypothetical portfolios, and then testing the real portfolios to confirm with live results.  The other test I was doing was regarding general market timing/market overvaluation chatter for holding more cash.

 

I lean more and more to Howard Marks philosophy--being offensive when there is disconfidence (is that a word?) and defensive when there is confidence.  In his case, he is required to deploy the capital (or at least he has said that), so it isn't a matter of holding cash, but how aggressive he is on investments that he makes. 

 

Further, since we do not have to deploy cash to poor ideas, we can let the ideas determine the cash levels rather than any other factors. 

 

[frommi]

Thanks for your studies.

 

From what i remember the return of a simple 60/40 stock/bond portfolio would have the same performance of a 100% stock portfolio but with a lot lower volatility, this doesn`t really fit to your results.

In the graphs it looks like the portfolio in cash would have earned nothing, it that the reason?

 

[racemize]

Thanks for your studies.

 

From what i remember the return of a simple 60/40 stock/bond portfolio would have the same performance of a 100% stock portfolio but with a lot lower volatility, this doesn`t really fit to your results.

In the graphs it looks like the portfolio in cash would have earned nothing, it that the reason?

 

I did not assume the cash was invested in bonds.  A primary reason is because the cash needed to be available for investing in equities and I didn't want to  have exposure to interest rate risk.  Additionally, I don't ever buy bonds with my cash, so I just wanted to see what would happen on the equity side only.

 

Do you have a paper for the 60/40 stock/bond portfolio?  Usually they report the same risk adjusted returns, but not the same actual returns.  I'm not a huge fan of "risk adjusted" returns, particularly over long term investments.  As Marks would say, you can't eat risk adjusted returns...