Friday, January 13, 2017

January 2017 Data Update 2: The Resilience of US Equities!

If asked to list the biggest threats to US equities at the start of 2016, most people would have pointed to the Federal Reserve’s imminent retreat from quantitative easing and the possibility of a slowdown in China spilling into lower global growth. Those fears contributed to a very bad start to 2016 for US stock markets, and as stocks dropped by about 5% in January, those who have warned us about a bubble looked prescient. But the stock market, as is its wont, surprised us again. Not only did US equities come back from those setbacks but it weathered other crises during the year, including the decision by UK voters to exit the EU in June and by US voters to elect Donald Trump as president in November to end the year with healthy gains. As we enter a year with potentially big changes to the US tax code and trade policy looming, it is time to take stock of where we are and where we might be going in the next year.

Stocks and Bonds: Looking Back
The best place to see  how the year unfolded for stocks is to trace out how the S&P 500 (large cap stocks), the S&P 600 (for small cap stocks) and US ten-year treasury bond rate did on a month by month basis through 2016.
Monthly returns, using month-end values
To convert the index values into returns each month, I first computed price changes for the indices each month (and cumulatively over the year) and added the dividends for the year to estimate annual returns of 11.74% for the S&P 500 and 26.46% for the S&P 600; it was a very good year for small cap stocks and a good one for large cap stocks.  I converted the treasury bond rates into bond price changes each month and cumulatively (for a 10-year constant maturity bond) over the year and added the coupon at the start of the year to get a return of 0.58% for the year; the rise in interest rates cause bond prices to drop by 1.68% during the year.

To put these returns in perspective, I added the S&P 500 and treasury bond return for 2016 to my historical data series which goes back to 1928 and computed both simple and compounded (geometric) annual averages in both for the entire period and compared them to a annualized 3-month treasury bill return (which you can think of as the return for holding cash).
Download spreadsheet with historical data

This table (or some variant of it) is used by practitioners to get the equity risk premium for US markets, by subtracting the average return on treasuries (bills or bonds) from the average return on stocks over a historical time period. Using my estimates, I get the following values for the historical equity risk premium for the US market.
Download spreadsheet with historical data
Note that the equity risk premium varies widely, from 2.3% to 7.96%,  depending on how long a time period you use, how you  compute averages (simple or compounded) and whether you use treasury bills or bonds as your measure of a risk free investment. Adding a statistical note of caution, each of these estimated premiums comes with a standard error, reported in red numbers below the estimated number. Thus, if you decide to use 6.24%, the difference between the arithmetic average returns on stocks and bonds from 1928-2016, as your historical risk premium, that number comes with a standard error of 2.26%. That would mean that your true equity risk premium, with 95% confidence, could be anywhere from 1.72% to 10.76% (plus and minus two standard errors).

Stocks: Looking forward
Looking at the past may give us comfort but investing is always about the future. I have been a long-time skeptic of historical risk premiums for two reasons.  First, as noted in the table above, they are noisy (have high standard errors). Second, they assume mean reversion, i.e., that US equity markets will revert back to what they have historically delivered as returns and that is an increasingly tenuous assumption. It is for this reason that I compute a forward-looking estimate of the equity risk premium for the US, using the S&P 500 Index as my measure of US stocks. Specifically, I estimate expected cash flows from dividends and buybacks from holding the S&P 500 for the next five years, using the trailing 12-month cash flow as my starting point and an expected growth rate in earnings as my proxy for cash flow growth and use these estimates, in conjunction with the index level on January 1, 2017, to compute an internal rate of return (a discount rate that will make the present value of the expected cash flows on the index equal to the traded level of the index).
Given the level of the index (2238.83 on January 1, 2017) and expected cash flows, I estimate an expected return on 8.14% for stocks and netting out the T.Bond rate of 2.45% on January 1, 2017, yields an implied ERP for the index of 5.69%. That number is down from the 6.12% that I estimated at the start of 2016 but is still well above the historical average (from 1960-2016) for this implied ERP of about 4.11%.

There is one troubling feature to the trailing 12 month cash flows on the S&P 500 that gives me pause. As was the case last year, the cash flows returned by S&P 500 companies represented more than 100% of earnings during the trailing 12 months, an unsustainable pace even in a mature market. I recomputed the ERP on the assumption that the cash payout ratio will decrease over time to sustainable levels, i.e., levels that would allow for enough reinvestment given the growth rate. The results are shown below:
The implied ERP for the index, with payout adjusting to about 82.3% of earnings in year 5, is 4.50%, still higher than historic norms but with a much slimmer buffer for safety. Looking at the next year, though, the potential for tax law changes will roil estimates. Not only are many analysts expecting significant increases in earnings next year of 12-15%, as they expect corporate tax rates to get lowered (at least in the aggregate) but there may also be a return of some of the trapped cash ($2 trillion or higher) back to the US, if that portion of the law is modified. Either change will relieve the pressure on cash flows and make it less likely that you will see dramatic cuts in stock buybacks or dividends.

Interest Rates: What lies ahead?
With bonds, I will take a different tack. I believe that, rather than waiting on the Fed, the path for interest rates this year will be determined by the path of the economy, with higher real growth and/or higher inflation pushing up rates. Updating a figure that I have used before, where I compare the T.Bond rate to an intrinsic interest rate (computed by adding expected inflation to expected real growth), you do see the beginning of a gap between the two at the end of 2016:
Entering 2017, the ten-year treasury bond at 2.45% is well below the intrinsic risk free of 3.60%, obtained by adding the inflation rate to real GDP growth through much of 2016. It is entirely possible that the economy will revert back to its post-2008 sluggishness or that there will be other shocks to the global economic system that will cause inflation and real growth to recede and interest rates to stay low, but for the moment at least, it looks like interest rates are their journey back to a new normal. If I were advising the Fed, my suggestion is for them is to act quickly on rates (perhaps as early as the next meeting) in order to preserve the fiction that it is they who are setting rates, rather than following them.

PE, CAPE and Bond PE Ratios
I am not a fan of PE crystal ball gazing but I know that there are many who make their market judgments based on PE ratios. Updating a graph that I last used when I posted on CAPE last year to reflect the numbers at the start of the 2017, here is what the updated PE ratios look like for the S&P 500:
Spreadsheet with data
While current PE ratios, in all their variants, are not at 1999 levels, they have clearly climbed back to 2007 levels and are well above historical averages. Scary, right? This will inevitably lead to the warnings about markets overheating and a coming crash, just as it has for much of the last five years. While one of these years, that predicted crash will come, you may want to look at stock PE ratios relative to the PE ratio on a treasury bond today, another comparison that I made in my CAPE post;
Spreadsheet with data
It is true that stocks look expensive today (at 27 times earnings) but they start to look much better when you compare them to bonds (at 40 times earnings). If you are concerned that bond rates will climb this year to reflect higher inflation/real growth, you may be forced to take another look at how you are pricing stocks at that time. There is one final divergence that needs explaining. In the last section, I noted that implied equity risk premiums on the US market look reasonable or even high relative to historical norms (a sign that the market is not over valued) but in this section, I have pointed to PE ratios being higher than historical norms (a sign of stock prices overheating). How do you reconcile the two findings? The answer lies in this final graph:
Spreadsheet with data
While PE ratios have risen over the last five or six years by almost 35-40%, the ratio of price to cash returned to stockholders (in the form of dividends and buybacks) has barely budged for the last five years. Here again, you should heed the warnings in the last section, where I noted that US companies are returning almost 107% of their earnings as cash to stockholders, unsustainable in the long term. If companies abruptly pull back on stock buybacks, the delicate balance that has allowed for the long bull market will be threatened.

The Closing
In summary,  the primary threats to stocks at the start of 2017, whether you look at implied equity risk premiums or PE ratios, come from two sources. The first is that interest rates will rise quickly, without a concurrent increase in earnings, and the second is that companies will  scale back the cash they return to stockholders to get back to a sustainable payout. Is there a reasonable probability that these events could occur? Of course, and if they both do, it will be a bad year for stocks. However, there is almost equal likelihood that as interest rates rise, earnings will rise even more (partly because of higher inflation/growth and partly because of cuts in corporate taxes) and that companies are able to sustain or even augment cash returned to stockholders. If this scenario unfolds, it will be a very good year for stocks. I will predict that you will be hearing from absolutists on both sides of this argument, one side preaching gloom and doom and the other predicting a market surge. I am in awe of the conviction that each side has in its market-timing judgment, but I am afraid that my market crystal ball is much too cloudy for me to make strong market predictions. So, I will do what I have always done, invest in individual stocks that I find to be priced right and accept that I have little or no control over the market.

YouTube Video


Datasets
  1. Historical Returns on Stocks, T.Bond and T.Bills from 1928 to 2016
  2. Implied Equity Risk Premium - January 2017 (Calculation Spreadsheet)
  3. Historical Implied Equity Risk Premiums - 1960 to 2016 
  4. T.Bond Rate - Actual versus Implied from 1954-2016
  5. PE, CAPE, Shiller PE and Bond PE from 1954-2016

2 comments:

Dajo9 said...

Higher interest rates driven by higher growth / higher inflation sounds like it is a given whether you read mainstream financial media or Professor Damodaran's blog. I don't think there is any reason to expect higher growth, higher inflation, or higher interest rates. There is no policy motion in place that would drive up middle class incomes enough to generate any of the above. In fact, the only policy motions in place are the Fed increasing short term rates and the elimination of Obamacare both of which will have the opposite impact on middle class incomes / balance sheets.

The one Trump policy that could have a positive effect in this regard is his infrastructure spending plan (depending on how it would be implemented). Infrastructure spending was in doubt due to Congress while Trump was in the glow of victory. As Trump stumbles out of the gate, Congress becomes stronger and infrastructure spending becomes less likely. I expect interest rates to go down from here.

Tim said...

Dear Professor Damodaran,

I have had a look at your latest implied ERP calculation for the month of February and it's puzzling to me why trailing earnings, dividends, and buybacks (in index units) are computed using seemingly outdated 'unit adjuster' as of September-end, instead of December-end, given quarterly updates (that data is provided on the worksheet named "Buyback & Dividend Computation").

On a side note, it is also not clear why there is a difference between market cap values on worksheets "Buyback & Dividend Computation" and "S&P500 monthly data (CapIQ)" for the year, say, 2015.

I would greatly appreciate if you could clarify these points.

Kind regards,

Tim