Several months ago, we wrote about how attractive US equity valuations were.
Seven months into 2013, US equities continue to rise and are up 16% year-to-date. Many are understandably nervous about how fast prices have risen. At the same time, rushing to sell after a big move up can be a regretful decision if markets continue to climb afterward.
For our portfolio as well, a +30% year-to-date performance demands serious introspection on whether to continue holding or to begin trimming positions.
One of the ways we answer questions like this is by aggregating the components of the major indices (Dow 30 or DIA, S&P 500 or SPY) and looking at the performance of these aggregates as a single business representing most of the US economy.
To get a sense of where today's prices are relative to business performance, what follows below are some figures from an aggregate analysis of 19 of the 30 Dow components. As this is a manual analysis, we don't yet have all 30 aggregated, but 63% of the market cap should give fairly similar results to the entire group (the main missing sector is the financial companies). For full disclosure, the exact stocks used in this data are CVX, CSCO, XOM, GE, HPQ, HD, INTC, IBM, JNJ, MCD, MSFT, PFE, PG, KO, UNH, VZ, WMT, and DIS). We use 10-k based data over ~20 years, which makes the data very accurate. The only major caveat is that we are using today's Dow components which includes a survivor bias because the list 20 years ago wasn't the same as it is today. However, since we are using prices from that same group, comparisons to other business metrics are still fair.
The chart below shows the annualized growth in revenues & prices from a particular year to 2012. If you compare the revenue and price bars in the 1990s, you have to go back to 1994 for the revenue growth to equal the price growth. In 1994, the forward 3-year price return was 32% per year because stocks were undervalued, and the market was about to go from there to crazy valuations for the "new economy".
Jumping ahead to 1999, this graph tells us that prices would have to rise by about 2x for us to get valuations similar to the 1999 bubble (you can work out the 2x by converting from annualized % to absolute numbers).
Finally, looking at the more recent period of 2005-2007, the bars are closer, implying that valuations today are similar to the relatively high and risky valuations before the Great Recession. However, there is an important factor we have ignored so far in this analysis. Revenue growth, especially over shorter time periods like 5 years, is cyclical. Market prices being forward-looking, prices will factor out some of this cyclicality (meaning that prices will be higher when growth is relatively poor, and lower when growth is relatively good). This is why it is good to buy stocks coming out of recessions even if PE multiples may be high.
The chart below shows that 5-year revenue growth is at a near low in 2012 compared to past years. Also, past revenue growth in 2007 was abnormally high, meaning that prices should have outperformed revenue growth in the next 5 years rather than only slightly underperforming.
So all of this analysis says that the relative risk-reward is still tilted in favour of US equities. We haven't even mentioned the alternatives yet, which further make equities more attractive (bond yields are much lower right now than in the past - 10-year treasury yields were ~7% in 1994, ~6% in 1999, ~4.5% in 2002, ~5% in 2007, and 2.6% right now).
As an investor, there are times to hold stocks through the ups and downs, and the above data implies this is one of those times to not get nervous early and sell out of sustained multi-year gains.