Views From the Floor

Why we believe a fall in earnings, rather than a de-rating, is more likely to be a driver an index collapse in the next recession; and who gives a frack?

How Much Could the S&P 500 Index Fall?

If we were to have a US recession, how much could the S&P 500 Index fall?

That’s a question that has been making the rounds on the trading floor. To answer it, we examined the previous two US recessions (the TMT bubble and Global Financial Crisis).

From the peak of the TMT bubble in March 2000 to the trough of October 2002, the S&P 500 almost halved, tumbling from 1527 to 777. However, earnings per share only fell 13% from peak to trough. The TMT unwind was basically a price-to-earnings unwind: the S&P 500 de-rated from 25x earnings to 13.5x from peak to trough.

In contrast, during the GFC, the fall of the S&P 500 was driven by a much steeper fall in earnings: The peak of the S&P 500 in July 2007 was somewhat corrected by the end of 2007; its PE ratio fell from 15.7x to 12x during the period. What followed next was an earnings collapse, resulting in a further fall in the PE ratio that continued until 2009.

As of 3 September, the S&P 500 is trading at a PE of 17.2x, marginally higher than the GFC peak of 15.7x but much lower than the TMT peak of 25.3x. However, EPS is currently at 176.2, compared with 102.3 at the GFC peak and 64.8 at the TMT peak. It is also worth noting that the range of return on equity for the S&P 500 is between 9 -18%, compared with a current ROE of 17.6%. In this light, rapid de-rating is less likely than a fall in earnings to be the driver of an index collapse in the next recession: there is simply more risk to the downside for earnings.

Figure 1. Analysis of S&P 500 Index in the Last Two US Recessions
  TMT Bubble Global Financial Crisis Current
  Level Date Level Date Level
SPX peak 1527 Mar-00 1565 Oct-07 3026
SPX trough 777 Oct-02 666 Mar-09 ?
% change -49%   -57%   ?
EPS peak 64.8 Sep-00 102.3 Oct-07 176.2
EPS trough 56.1 Jan-02 62.7 May-09 130.3
% change -13%   -39%   -26%*
PE peak 25.3 Jan-00 15.7 Jun-07 17.2
PE trough 13.5 Oct-02 9.0 Nov-08 11.3*
% change -47%   -43%   -35%

Source: Bloomberg, Man Solutions; As of 3 September 2019.
*Taking the average decline of the TMT Bubble and GFC

Problems loading this infographic? - Please click here

Source: Bloomberg, Man Solutions; As of 3 September 2019.

Who Gives A Frack?

Does the probability of an earthquake vary over time?

An instinctive reaction might be ‘no’. After all, earthquakes are natural phenomena, which seem to occur at random intervals. However, that is far from the case – the rise of man-made earthquakes is very visible in the data. Between 1973 and 2008 (inclusive), the number of earthquakes (of magnitude 3 and above) in the US averaged about 27 (Figure 3). This average spiked to 350 between 2009 and 2018 (inclusive).

The increase can be explained by onshore oil production. While the actual fracturing process can induce some earthquakes, the primary cause is the disposal of the fracturing fluid or waste water (as part of oil extraction) in deep wells. Indeed, the latter is produced whether a well has been hydraulically fractured or not.

What’s important to note is that the number of earthquakes with a magnitude of 5 and above remains low, averaging less than one per year in both periods, and with none reaching M6.0.

This is especially relevant to catastrophe bond investors: a 2018 cat bond deal for the California Earthquake Authority had a mere 0.4% of its total exposure to magnitudes of below 6. Even though earthquake frequency has rocketed, man-made quakes are of negligible importance to a cat bond portfolio.

Figure 3. Number of Central US Quakes by Year

Number of Central US Quakes by Year

Soure: USGS, Man AHL; Between 1973 and 2018 (inclusive). Note: Earthquakes with a magnitude of 3 and above, covering the central US (230N to 510N and 630W to 1090W).

With contribution from: Ben Funnell (Man Solutions, Portfolio Manager), Teun Draaisma (Man Solutions, Portfolio Manager), Henry Neville (Man Solutions, Analyst) and Andre Rzym (Man AHL, Portfolio Manager).

User Country: United States (237)
User Language: en-gb
User Role: Public (Guest) (1)
User Access Groups:
Node Access Groups: 1