It is often noted in the media that the current bull market in global equities has been running for over nine years. This naturally leads to a lot of discussion about when it will end. One of the most common points of analysis is to look at aggregate market valuations, such as the price to earnings multiple (P/E). The Australian market as a whole does not look particularly expensive on this measure, with banks and resources trading at close to historical multiples. However, the industrials part of the market (the market ex resources and financials) is clearly trading at a high multiple relative to history, as shown below. Is this a strong signal that the market run is coming to an end? Below, a number of considerations are presented which suggest that this conclusion ignores at least half of the story.
Figure 1. ASX Industrials forward P/E and the S&P/ASX 200 Accumulation Index
Firstly, let’s look at the recent history. The Industrials component of the market is currently trading on a forward P/E of 20x, compared to a 20-year average of 17x. The last time the industrials moved higher through a 20x P/E was in February 1999. Was that a good indicator that the market had peaked? Three years later, at the end of February 2002, the ASX 200 Accumulation Index was 36% higher. It then fell 9% over the balance of 2002, dragged down by the bursting of the Dot-com bubble in the US, before returning 15% in 2003. The large fall in the blue line in the chart above, denoting the Industrials P/E, largely reflects earnings growth rather than a market decline. In 2007, the Industrials reached a P/E of 19x and this did prove to be a very good signal of the market top. However, this multiple was also reached in July 2016 and the market has returned 21% over the subsequent two years.
Another important consideration is that the P/E multiple does not tell us much outside of the context of earnings growth and risk. If earnings growth is expected to be higher than average, then the earnings multiple should be higher to reflect this and vice versa. The below chart shows earnings per share growth for the ASX Industrials over the past 15 years. Excluding the post-GFC bounce, earnings per share growth has been lacklustre for over a decade but returned to reasonably strong growth in 2017.
Figure 2. ASX industrials earnings per share growth
Thirdly, the P/E multiple is based on current earnings expectations, but these expectations change over time. The chart below shows the progression of earnings per share forecasts for the broader market for each of the past 20 financial years. If you cast your eye on the time scale at 2003, you can see where consensus forecasts were for the next few financial years at that point, represented by the individual coloured lines. As you move along the time scale, through 2004, 2005 and 2006 you can see that these lines moved higher, reflecting an earnings upgrade cycle. If you looked at the consensus market P/E at the start of the 2005 financial year, the number you would have been considering was actually approximately 10% too high given the earnings actually achieved over the next year. It is a sad fact that since 2010 we have seen negative earnings per share forecast revisions every year through to the start of 2017. However, revisions have been more stable or even marginally positive over the past 18 months. Mirroring this earnings environment, real global GDP growth has just come through a decade in which growth averaged just 2.5% per annum (for the ten years to 2016), while the IMF forecasts 3.8% per annum growth for the period from 2017 to 2019.
Figure 3. Earnings forecasts revisions history over the past 20 financial years
Source. RIMES, IBES, Morgan Stanley Research
In addition, there is the question of the composition of the increase in the industrials P/E over the past three years. One company, CSL, now accounts for 14% of the market capitalisation of the industrials and has seen its P/E expand from 24x to 35x over this time period, accounting for a full point increase in the industrials P/E. Further, there are now twice as many information technology companies in this part of the market compared to three years ago, now accounting for 6% of the aggregate industrials market capitalisation, up from 3% three years ago and with this cohort trading on an average P/E of 24x. Excluding these two factors, the Industrials are trading at closer to a 10% P/E premium to the historical average than the 20% headline.
Finally, while the S&P 500 has not experienced a 20% drawdown for over nine years, which would mark the technical end of a bull market, a number of other equity markets have recorded 20% falls. Both the MSCI World Accumulation Index and the ASX 200 Accumulation Index fell by just over 20% between April and September 2011, during the Sovereign Debt Crisis. These two indices also came close to the definition of a bear market between May 2015 and February 2016, with 18% falls.
- Current P/E levels have not proven a consistent indicator that the market will not trend higher.
- The absolute P/E is not a good indicator by itself since it needs to be considered in the context of the earnings growth outlook. Presently, higher growth expectations can justify a higher P/E.
- Earnings expectations have and will change, which can have a significant impact on the market P/E. This last significant upgrade cycle was in 2004-2006.
- The Industrials P/E has been inflated by a single large stock, CSL, and the high P/E, high growth IT sector.
- It is misleading to describe the current bull market as nine years old, given there have been significant declines in a variety of equity markets over this period.
It is true that the Industrials are trading at a high earnings multiple relative to history. This would make the market vulnerable to a larger correction in the event that the earnings outlook deteriorates. However, the current macroeconomic and earnings environments remain supportive, notwithstanding identifiable risks such as an escalating trade war. A continuation of reasonably strong earnings growth in line with current expectations, or even an earnings upgrade cycle, would allow the market to continue to trend upwards while the earnings multiple contracts towards the long-run average.