How to Respond to Rising Stock Markets
As stock markets rise to ever larger price-to-earnings (P/E) ratios, the odds of a market crash grow. However, we can’t know when such a crash might come, so I’m not interested in trying to time a sell-off of all my stocks. Stocks remain the best bet for future returns, but how much higher can P/E ratios go before this is no longer true?
When we examine the relationship between Robert Shiller’s Cyclically-Adjusted Price-Earnings (CAPE) ratio to the following decade of stock returns, the correlation is quite weak; the result is closer to a cloud than a straight line. The most we can say is that when the CAPE is high, future expected stock returns appear to be somewhat lower. There is logic to the idea that P/E ratios will likely return to some form of normalcy in the future, but this may take a very long time. In the interim, stocks remain the best bet for future returns.
But at what P/E level can we decide that stocks are no longer a good bet? Shiller’s U.S. CAPE is at 38 as I write this. The highest it’s been in the last 150 years is about 45 in the year 2000. What if the CAPE gets to 45 or higher? At some point, the future of stocks won’t look very bright.
A few months ago I adjusted my investment spreadsheet to assume that my portfolio’s CAPE (a blended figure based on my allocation across Canadian, U.S., and international stock markets) would drop to 20 by the time I reach age 100. I kept the assumption that corporate earnings would keep growing at an average rate of 4% above inflation each year. The effect of this assumed slow reduction of the CAPE is that I would get lower stock returns for the rest of my life, and the amount I can safely spend in retirement is lower. For more about the details of how I calculate my retirement spending level and portfolio allocation, see my glidepath article.
So, this change has me spending a little less money each month, but it didn’t change my asset allocation. A minor technicality is that because I use a fixed income allocation of 5 years worth of my safe retirement spending level, this change would have had me lower my fixed income allocation. I added some calculations to prevent this slight shift to stocks. It would have been ironic if spending less because I’m worried about high stock prices had led me to own more stocks.
The way I made this technical adjustment was to increase the 5-year figure to keep my fixed income allocation the same as it would have been without the CAPE-based reduction in expected future stock returns. This led to another idea. What if I were to increase this 5-year figure even more when the CAPE is very high? The idea is to make a gradual shift toward fixed income as the CAPE grows ever larger.
Previously, I arbitrarily chose 20 as the CAPE level where I’d start to adjust downward the percentage of my portfolio I’d spend each month. So, as stocks keep rising, my spending level rises as well, but not quite as fast as my portfolio goes up. This time, I’m setting another (higher) CAPE level where I’ll gradually increase my fixed income allocation.
I haven’t decided on this second CAPE threshold, but let’s use 30 as an example. If the CAPE is above 30, then I take my 5-year spending figure (adjusted as described earlier if the CAPE is above 20) and multiply it by CAPE/30. So, as the CAPE rises above 30, my fixed income percentage rises. As my stocks rise in value, the absolute dollar amount I have in stocks will keep rising, but slower than it would have risen before, and my fixed income investments would grow faster than they would have before.
Unless the stock market does something very sudden, all these adjustments will happen at their usual glacial pace. If the stock market doesn’t crash soon, I’ll make somewhat less money with my lower stock allocation, but that’s fine because I’ve had the enormous benefit of a very long bull run. (Why keep playing the game when you’ve already won?) If the stock market crashes from some CAPE level above 30, I will have protected more of my portfolio than I would have before making this change.
Sharp-eyed readers may wonder whether I’m opening myself up to a bond crash. I don’t buy long-term bonds. My fixed income is currently in high-interest savings accounts, a few GICs, and some short-term government bonds. So, a crash in the bond market wouldn’t affect me much.
I don’t claim that this response to nosebleed stock levels is optimal in any sense. But I do believe it will help me in some stock market crash scenarios without taking away too much of my potential upside. A further benefit is that I can automate it in a spreadsheet and keep my feelings out of any decisions.
One of the benefits of having cash (even though earning ~0%), is that it's there to deploy when opportunities are available to earn above average rates of return. During the 2020 freefall during the early days of the pandemic, I was able to put my cash holdings to work. Around 23 March, the markets for certain securities were in freefall. Effectively, there was no bid and the few remaining bidders were setting the effective price. It's a wonderful feeling to have ample cash and the mental fortitude to invest when others can not or will not. Putting that cash to work has netted triples (200% gain) in 12 months. I'm pointing this out not to brag, but to identify that the gains available during panics are tremendous and out-sized. Finding opportunities to invest at 20%+ on an annualized basis can far outweigh the opportunity cost of a few basis points. Yes, you can point out ways that this doesn't work, but most of my portfolio gains have come from good investments made at market lows as a combination of purely low price and discounted price for future compounding of reinvestments. It worked in generally in 2002, 2009, 2020 along with individual wins in other specific securities.
ReplyDeleteAt the right price, a normal 10-15% compounder can turn into a 20%+ compounder which leads to excellent portfolio returns.
https://www.valueinvestingworld.com/2012/07/seth-klarman-on-painful-decision-to.html
Anonymous,
DeleteThis is the prevailing wisdom of stock pickers, including Warren Buffett. However, I'm not interested in going back to trying t pick stocks. I have little doubt that if I were to go back to age 30 with a good income, I'd be better off investing 100% in stock indexes and not worrying about having a cash reserve beyond a sensible emergency fund. If stocks do crater sometime soon, I don't intend to buy back in with the cash I have on the sidelines beyond my normal rebalancing strategy.
Thanks James. It’s reassuring to read that despite the potential higher gains in stocks, maximum portfolio value is not the only objective of the “game”. I was reluctant to shift to a higher fixed income ratio, so like you, I surrendered to a formula, and I sleep better for it. I have a rolling 5 years in fixed income plus a small buffer. The buffer is enough to cover the cost of a new roof or replacement vehicle.
ReplyDeleteLast week I developed a ‘market crash’ plan, where if the drop hits -30% I will quickly sell a years worth of fixed income and buy the discounted equity funds (index ETFs). Having finished my plan, I then realized that this totally undermines my five years of fixed income plan - I may as well have just four years of fixed income, and sink the other year into equity right now! I’ve subsequently filed the ‘market crash’ plan just to remind me that I’ve already looked at this option so I don’t waste too much time on it again.
My apologies, Michael, my comment should not have started with James.
ReplyDeleteMichael, do you have a ‘market crash’ plan? In response to an earlier comment you mentioned rebalancing. How soon or deep into a crash would you do the rebalancing, and if you’ve optimized to only having five years of fixed income coverage, how much of this would you be prepared to give up to effect the rebalance?
Hi Bob,
DeleteJames is actually my middle name, and I have a a couple of friends who use it to refer to me, so don't worry.
I don't have a specific crash plan, but my rebalancing plan would look like a crash plan. When I refer to my retirement spending level, I mean the amount I believe I can safely spend during retirement. This figure necessarily depends on the size of my portfolio, expected portfolio returns, expected taxes, etc. (and is calculated by my spreadsheet using near real-time figures). So, if the stock market were to crash, this "safe" spending level would drop somewhat. The changes I've made recently to tie expected returns to the CAPE level and to reduce my stock allocation percentage when the CAPE is very high tend to smooth the ride, but a stock market crash would still lower my safe spending level. This would then lower the 5-year amount in fixed income. So, my portfolio spreadsheet would declare that I need to rebalance somewhat from bonds back to stocks. So, I don't have a specific crash plan, but the set of formulas I use would have me buying partially back into stocks.
Wow, your spreadsheet sounds like a finely tuned machine. Thanks for explaining.
DeleteMichael, another question for you. When you receive dividends in your various accounts, with probably the biggest lumps appearing quarterly, what do you do with the dividends? As an example, in July you receive $4,000 in dividends spread across various accounts (RRSP, LIF, TFSA, non-registered X 2 if you have a partner), and in December you need to withdraw the cash for your 2022 expenses. What do you do with the cash in the intervening period? Also of note is that there are fees associated with withdrawing cash from at least the RRSPs and LIF. Thanks
ReplyDeleteHi Bob,
DeleteI chose a threshold level in each account to trigger a stock purchase. When the cash level builds to this threshold amount, I buy some stock ETFs based on which ones are below their target allocation. In the interim, I just let the small amount of cash sit there. In the case of taxable accounts, if my bond allocation is low, I would transfer the cash to a savings account.
Hi Michael,
DeleteThe process you use is pretty much the same as I used during the accumulation phase, although, I don't have any bonds in our non-registered accounts. Perhaps I'll look at swapping the equity ETFs and stocks with the bond ETFs in our TFSAs, remembering I could be realising some capital gains.
This is our first decumulation year, so now those dividends will be used for income in the following year (withdrawn the December before). It's not a huge amount, so I won't lose much if I just leave the cash sitting in the accounts, although, I am tempted to just buy more bond ETFs right up until the day I need to make the withdrawal. I'll keep pondering a little while longer.
Hi Michael,
ReplyDeleteThought provoking article as usual.
What do you think about using "Excess CAPE" to judge whether stocks are expensive and adjustments are required, per http://www.econ.yale.edu/~shiller/data/ie_data.xls , Data tab, column Q. Based on this measure, stocks do not appear that expensive, given interest rates on bonds is so low presently.
Hi Jim,
DeleteThis is a comparison of the yields of stocks to long-term bonds. If this were the only choice, then I'd take stocks, as the latest figure of +3.09% indicates. However, I have another choice -- cash (or short-term bonds). To me, long-term bonds look like return-free risk. The choice I've made to lean more to cash when the CAPE is very high is more about limiting the damage that a stock crash could do to me than it is about seeking yield. I've had the benefit of a huge run up in stock prices, and so my need for taking risk has declined at the same time as the risk of a stock crash is elevated. This makes it an easy choice to slightly lower my stock allocation according to the rebalancing formula I described. If stocks continue to rise, my cash allocation percentage will climb proportionally.
Thanks for writing this. Just wondering instead of assuming that CAPE will be 20 at a certain point in time, can you not just temper future expectations based on current value of CAPE using a simple linear model. For e.g. Future expected return = 4% * 20/30 ; where 4% is the average expected return, 20 being the CAPE threshold and 30 being the current threshold
ReplyDeleteHi ispeuq,
DeleteYour linear method works out similarly to mine for some time horizon, and is more pessimistic for longer horizons. I like the idea of assuming the CAPE will return to "normal" at some distant point in the future, but your method is more reasonable for shorter horizons.