Sunday, 10 November 2019

How much cash should you hold? The case for being fully invested

Over the past decade, a number of highly-regarded value investors with attractive long term track records have lagged the major benchmarks, and not just over one or two years, which can reasonably be expected to occur to any investor from time to time, but over stretches as long as 5-10 years.* There have been many reasons for this, including the much-discussed outperformance of 'growth' over 'value'; the rise of index funds; and markets becoming more competitive. In some cases, it has also been due to previously celebrated investors having only assembled their prior robust track records through a combination of luck and concentration, and having been subsequently revealed to have been far less proficient investors than previously believed (e.g. Berkowitz; Tilson; Ackman). However, one important contributor to many quite excellent investors' underperformance has been the simple fact that global indices have risen more than expected (by many) over the past 10 years, and these investors have held high levels of cash throughout this period - often 20% or more.

Although I admire many of these investors, I have long questioned the wisdom of holding so much cash, and my typical position is to be fully invested at all times (provided I am able to find sensible things to hold, which I always have been). The way I think about it is, if the market's expected return is say 5-7% (driven off earnings yields), then the opportunity cost of holding zero-yielding cash for any extended period of time is very substantial, even if your stock picks are no better than average (I would rather have a 'fair' return than no return, although fortunately I've never had to settle for fair). If the market return is say 6.0%, and you hold 20% cash, then the market component of your portfolio's expected return will fall to 4.8% (in the low/zero interest rate world in which we live; if interest rates were higher, market expected returns would likely also be higher, so the same principle would apply). And if markets rise by meaningfully more than this as they have over the past decade, because multiples expand, your degree of lag will increase commensurately. If markets go up at 15.0% a year, you'll make only 12.0% a year even if your stock picks match the market.

Platinum is an example of a fund that has been hurt a lot by this. They have an exceptional long term track record, and are (deservedly) highly regarded, but over the past 5-10 years they have lagged the market (they have generated about 90% of the market's return) - something which has attracted much comment/scrutiny - most of it unfairly negative. However, throughout this period they have held about 20% cash, as well as various index/stock shorts/hedges. The invested portion of their FUM on the long side has therefore outperformed, despite the significant underperformance of value over the past decade. Their stock picking has therefore remained excellent, on average. But that stock picking prowess has been applied to only about 70-80% of their FUM, because it's been diluted by their hefty cash holdings.

Now I'm of course aware that we happen to have been in a bull-market for equities over the past decade (or for US equities at least), and I'm also more than aware that markets don't always go up. If markets had gone down during this period, cash would have aided rather than detracted from performance. But I'm not using hindsight/happenstance here, as I hope to argue below that a strong case can be made that on average, over the long term, through multiple bear and bull cycles, holding large amounts of cash will be a losing proposition. I also practice what I preach, as I have been fully invested for the entirety of my so-far 20 year investment life, through good markets and bad.

The key point is that so long as markets are trading at a reasonable earnings yield that is well above bond yields, which they still are, the longs will win over time, because those earnings will be distributed to shareholders over time (dividends and buybacks), driving sustainable returns. The 'carry cost' of cash is high - it is essentially the market earnings yield forgone (5-7% at present, and significantly higher earlier in the decade, prior to multiples experiencing a decade-long increase).

The rationale usually proffered for holding large amounts of cash is that it gives you the opportunity to take advantage of any market downturns to acquire cheap stocks (or buy more of your existing holdings on the cheap). The cash has 'option value', it is argued. However, that option is very expensive, and there are alternative and better ways of taking advantage of intermittent market volatility and the opportunities it throws up, in my view. Provided correlations fall short of 1 (I'll discuss this later), a much better way is to have exposure to a wide variety of good ideas (i.e. without compromising idea quality) across a number of different countries, industries, and individual companies, and then take advantage of less-than-perfectly-correlated volatility to actively rebalance. When stocks are doing well somewhere, but poorly somewhere else (or a doing very poorly somewhere, but only modestly poorly somewhere else), there is an opportunity to actively cycle more capital into the more oversold areas. You don't need a lot of cash in order to do this - just agility.

For instance, in the early part of 2018, a lot of value stocks in the US performed very well (alongside a strong USD and US economy), but several emerging markets got hit very hard. I reduced my exposure to my US value plays and reallocated more into EM, including some selective picks in Indonesia. In the second half of 2018, Indonesia rallied hard. Some of my value picks there rose 50%, while the US suffered a significant sell-off, particularly in December. I sold down some of my Indonesian positions and significantly increased my US exposure in December.

I also blogged about a few examples in real time (I blogged twice on Turkey bullishly, both coinciding exactly with the lows, where I upped my exposure; I have subsequently reduced), and also explained my long thesis on Eurobank after it got disproportionately beaten up last year. In a subsequent blog post, I discussed how I was reducing my Greek banks exposure in favour of Italian banks, after the former had surged and the latter had heavily sold off. Unicredit is up 30% since then and the Greek banks are up 10%-ish. I've recently trimmed, and have been buying into other names that have come way down - including Indonesian thermal coal stocks which are down 50% over the past 12 months, as many other new ideas which I'll omit to mention as I'm still buying. You can do this all day long in markets if you have a broad mandate, are nimble, and correlations are significantly less than 1, which they are most of the time. From a distance, it might appear like I'm trading or market timing, but I'm actually just responding to a repricing of the fundamental risk-reward odds by adjusting the size and location of my bets.

But what if correlations do go to 1, and everything goes down significantly at the same time? It happens sometimes. It happened in the GFC for instance - equities everywhere fell 50%, and there was nowhere to hide. The issue is that while these events do happen, they are - at least to this degree - extremely infrequent. Meanwhile, the cost of missing out on an expected return of say 7% a year while you wait for such opportunities to occur, over any sustained period of time, can become very substantial due to the effect of compounding. 7% a year compounded doubles in approximately a decade. You would need everything, everywhere, to simultaneously go down by 50% once every decade, just to offset these lost gains, but drawdowns of this nature and magnitude are much rarer than that. They have historically happened only about once every couple of generations (prior to the GFC, the Great Depression was really the last episode; lots of stuff did well in the 1970s, as oil and commodities rallied and oil-dependent economies did very well).

The above also assumes you would deploy all of your cash near the bottom - something that is fairly unlikely. Indeed, people that like to hold cash usually end up retaining far too much of it for too long during downturns. After all, if you haven't deployed it by the time markets are down 30%, why would 50% be the magic number? No doubt you would either deploy it too early or hold it too long, waiting for further falls that fail to materialise. Even in an idealised situation of deploying all your cash at the very bottom, it is unlikely the gains you will make will offset the opportunity cost of spending many many years missing out on the market's underlying earnings yield of 5-7%+.

Consequently, the only way holding a lot of cash will work to your advantage is if there is a major, synchronous, correlated global market decline, and it happens relatively soon after you elect to hold a lot of cash. But this means you are essentially making a major bet on a market fallout happening, where the base-rate probability of it occurring is low, and the odds overall aren't in your favour.

I have discussed in past blog posts how the GFC has had a profound impact on investor behaviour. It has caused many investors to go from underestimating the probability and frequency of crises, to radically overestimating both dimensions in the post-GFC years. Post GFC, everyone became a macro expert. Disquiet has prevailed about central bank actions, and other recurring sources of macro anxiety - from strains in the EU/Eurozone to the sustainability of China's growth - have imbued investors with a constant sense of foreboding. The fundamental problem is that when everyone expects another major bear market, and holds a lot of cash in anticipation of deploying it during one, the said downturn becomes self-defeating - market's can't and don't go down very much for very long when everyone is chomping at the bit with cash looking to buy. Large cash balances held in anticipation of a downturn is likely one of the key reasons why we haven't had a meaningful and sustained bear market in the US and other major indices so far.

Furthermore, even if such a decline does occur when one is fully invested, one can still meaningfully benefit in the long term. How? If you own a portfolio of stocks that pay decent dividends, and/or own companies that routinely buy back their own stock (or buy back shares opportunistically when their stocks get oversold), you can benefit from market declines by reinvesting dividends at higher future rates of return, and/or from companies buying more stock for you via buybacks. Indeed, the rational long term investor ought to vastly prefer lower prices to higher prices, because if prices are low enough, getting rich is easy. It's high prices that are the enemy of wealth accumulation. This is another advantage of being 100% invested - it hedges the risk of markets going up a lot and staying up, which for the long term investor, is by far the biggest downside risk (discussed in this post in more detail).

Imagine for instance that tomorrow, for no fundamental reason (e.g. nuclear war), stocks were to open 80% lower than they traded on Friday. Coke no longer traded at 25x earnings, but instead at 5x, and furthermore, it stayed at 5x thereafter. Let's say you were fully invested prior to this drawdown. You lack 'cash', so you can't take advantage of this decline by buying stock at bargain basement levels, right? You're saddled with an 80% 'loss'. Catastrophic right?

Not so fast. At the moment, Coke pays a 3% dividend, and also buys back something in the order of 1% of its shares a year (i.e. distributing its 4% earnings yield). At its new price, which now represents a 20% earnings yield, its dividend yield would have risen to 15%, while the company would now be buying back 5% of its shares each year. In the long run, you would significantly benefit from this reality, and if it stayed at 5x, you'd eventually become very rich simply by reinvesting the dividends (and enjoying the compounding 5% buybacks), because you would now be able to compound your money at 20% a year into perpetuity, instead of merely at 4%.

For every $1,000 invested, if the stock stayed at 25x and you made 4% a year, you'd end up with $3,200 by year 30. If by contrast, your $1,000 dropped to $200 (5x P/E), and stayed at 5x, with dividends reinvested and 5% annual buybacks continuing, you'd end up with $47,500 by year 30. You'd have $47,500 at a 5x P/E, instead of $3,200 at a 25x P/E. No P/E 're-rating'/recovery would be necessary to make a fortune. Furthermore, your annual dividends would now be $7,100 a year, instead of merely $100. While this might seem counter-intuitive, it simply reflects the power of compounding money over long periods of time at high rates of return instead of low rates of return. Compounding at 20% a year instead of 4% easily trounces a higher level of starting invested funds (in this case, $200 after the 80% decline, instead of $1,000). The biggest long term risk to investors is not markets going down. It is markets going up a lot and pricing stocks at more and more expensive levels which thwart the opportunity for meaningful future gains. Ask government bond investors at the moment - they will glumly tell you how higher prices in the short term portend very miserable long term outcomes, as yields approach or submerse zero.

But what if you can't find any even modestly-good ideas? Ought you then not hold cash? If it's your own money, sure. It's better to keep your money than do something dumb and suffer a permanent loss of capital. But if you're an investment manager, I believe the responsible thing to do is to simply return the capital to your investors. Can't find a sensible way to invest 30% of the funds you have? Then return that 30%. This would be functionally equivalent to a company like Apple which had (say) 30% of its market cap in cash, deciding to buy back 30% of its stock. It allows investors/shareholders to gain exposure to the good parts which they want exposure to (Apple's cash generative business, or the good ideas you can find as an investment manager), without investors having to tie up a significant amount of additional capital in low-yielding cash as part of the bargain. The latter is something most investors do not want to do, and in any case is something they can do for themselves, without paying the investment manager a fee for the privilege. That, incidentally, is exactly what Warren Buffett did in the late 1960s when he found himself short of ideas.**


*Indeed, there are times when you should actually be happy your investment manager is underperforming. Those are times when there is a wild speculative orgy going on in a meaningful portion of the relevant index. Underperformance during these sorts of periods is actually a sign of discipline, and the manager's unwillingness to do anything stupid that could put their investors capital at risk of a major, permanent loss, merely for the short term benefit of the manager's P&L and headline numbers. 

The wise investor considers not just the level of headline performance, but the nature of those returns, and how much risk is being taken to generate them. Investing is a little bit like driving down a foggy road. The road might have been straight for quite some time, but a wise driver knows that visibility is poor and there could be a sharp curve in the road at any moment. The reckless driver, by contrast, puts their 'pedal to the metal'. So long as the road remains straight (i.e. currently extrapolated growth trends in markets continue), the latter driver might appear to be the 'better' driver, because they are progressing towards their destination at a faster rate. But if an unanticipated swerve in the road happens, they will careen over the cliff, with their investment passenger in tow (without seat-belts).

A long term track record through multiple market environments (i.e. multiple 'swerves' in the road) is therefore far more impressive/instructive than any short term performance assessment throughout a singular stretch of straight road. And what is 'long term' should be defined by the number of swerves, not an arbitrary number of years. It is for this reason that a lot of performance commentary that focus on the recent underperformance of managers such as Platinum, in full disregard of their long term track record, and the multi-year, uni-track nature of markets we have seen in recent times, is very misguided. I suspect many underperforming value managers with excellent long term track records, such as Platinum, will quickly redeem themselves as soon as the next curve in the road emerges, while many current go-go tech growth heroes that currently look like geniuses will end up looking anything but. Platinum will not, however, ever recover the performance headwind suffered from their decision to hold large amounts of cash over the past decade.

**It is also, incidentally, what I believe Berkshire Hathaway should be doing today as well with its >US$100bn (and growing) cash hoard - albeit that unlike investment managers, Buffett is not charging fees on this balance. BH has started to slowly buy back stock, but should be doing a lot more in my view (and likely will in the future). It took a long time, but BH has finally gotten too big to find sensible ways to redeploy all of its capital.