What's Really Driving the Trump Bump
With the stock market at highs and amid positive economic news, the so-called Trump bump is salient. The left credits Obama administration policies, which Republicans are screwing up. Right-leaning commentators treat the bump as optimism about the Republican agenda, which Democrats and party infighting are undermining.
Of course, boosting the stock market isn't the primary job of a president, and the president isn't the primary driver of stocks. Nonetheless, the 80 percent of the American populace that makes up the political center likes a vibrant economy. But that means real 1 earnings growth, not high asset prices. Earnings drive business expansion, hence wage and job growth, hence tax revenues.
Therefore, rather than looking at stock market gains since Trump's election, I ask:
- What did investors think when Donald Trump was elected?
- What effect did the administration have on real earnings growth?
Before looking at that, however, we have to consider a counterintuitive relation. You would think that when P/E ratios are high, the market is expecting high real earnings growth rates. But the reverse is true. For example, average real earnings growth over three years is about 27.6 percent minus 0.9 percent times the P/E at the beginning of the period. That means buying at the current P/E of 30.29 on average leads to 27.6 percent - 0.9 percent x 30.29 = 0.4 percent real earnings growth over the next three years. 4
The chart above shows the effect of P/E ratio on average real earnings growth up to eight years in the future. Buying at a higher P/E leads on average to slightly higher earnings growth for the first eight months, but then things turn sharply negative. At the minimum in three years, each extra point of P/E ratio costs on average 0.9 percent real earnings growth.
Now look at the same chart with a red line added using only data from November of presidential election years. 5 For the first two years the lines are about the same, but they diverge sharply from years two to five. Now a lot of things other than elections happen in Novembers of leap years, and it's reasonable to assume that those other things are driving the real earnings growth in the first two years, because the pattern is the same as for every month. It's therefore reasonable to attribute the divergence from years two to five to the presidential administration. 6
So to evaluate the business impact of presidential administrations, I look at the change in P/E ratios that happens at the election, versus real earnings growth between two and five years later. 7 The chart below shows the P/E bump on the horizontal axis and the real earnings growth rate on the vertical axis. The red line is a linear fit.
Although high P/Es are associated with lower future earnings growth, increases in P/E upon a president's election are associated with higher future earnings growth. However, the slope is pretty shallow. It's pulled down by three major misjudgments. On the far left, President Barack Obama in 2008 was the candidate investors liked the least, and he delivered average growth in his first term. 8 Next most disliked was Harry Truman in 1948, and his growth was quite good. On the far right is poor Herbert Hoover, the candidate most favored by investors, but who owns the Great Depression in this analysis.
The next two charts separate the data by political party. For Democrats, investors' opinions are worthless; in fact there is a slight (but statistically insignificant) negative slope. 9 Democratic presidents seem to deliver about the same average earnings growth 10 whatever investors thought about their elections.
Republicans on average are better liked, and deliver somewhat higher real earnings growth. The more investors like a candidate at the time of his election, the higher growth he delivers. 11
Finally, here is a list of all the presidents who completed their terms since 1888 (Republicans in red, Democrats in blue) with the market reaction to their election and the earnings growth rate of their administration. 12 Donald Trump's election bump puts him exactly in the middle. Based on that, history suggests investors can look for an average Republican real earnings growth rate of 29 percent from 2018 to 2021.
Bloomberg Prophets Professionals offering actionable insights on markets, the economy and monetary policy. Contributors may have a stake in the areas they write about.
That is, adjusted for inflation.
I use Yale professor Robert Shiller's cyclically adjusted price-earnings ratio, or CAPE, as the best common indicator of stock market valuation.
This is not an absolute measure of what investors think the president will do for the economy; it's a verdict on the difference between the president and his opponent, and it encompasses the entire election result including changes in Congress. Moreover, it reflects the surprise in the outcome. Investors may love a president, but if his election was considered highly likely, we wouldn't expect much market movement on the outcome.
Buying at a high P/E is a problem for another reason, P/E ratios tend to revert to a mean of about 18. Over three years you get about one-third reversion on average, so in three years the expected P/E is about (2 x 30.29 + 18) / 3 = 26.19. If real earnings grow at 0.4 percent, then the real value of the Standard & Poor's 500 Index in three years would be 29.19 x 1.004 / 30.29 = 0.967 or 97 percent of current levels, a 3 percent inflation-adjusted capital loss. However, this is a highly simplified analysis and doesn't include dividends. Even at the current P/E ratios, I think stocks remain the best choice for most investors for long-term, inflation-adjusted performance; although of course it's easier to defend stocks at lower P/E ratios.
The red line is kinkier because it is based on only 1/48th of the data as the blue line.
This also fits with intuition. It takes some time to change things in Washington, and some time for those changes to work through to corporate earnings. So changes start in year three, but by year six the next administration's policies are starting to drive things.
For example, for President Obama's first term, the election was in November 2008, so the base period for earnings runs from November 2007 to October 2010, one year before to two years after the election, and the period used for growth is November 2010 to October 2013, two to five years after election.
President Obama moved up to only third most disliked in 2012, leapfrogging Harry Truman, but we won't know the real earnings growth rate from that term until November. But it will probably be a solid real growth between 35 percent and 40 percent.
When President Obama 2012 enters the data in November, the slope will turn more negative.
Twenty-five percent, versus 29 percent for Republican presidents.
Again, this relation would be much stronger without Herbert Hoover.
Most of the entries conform to expectation, but there are some surprises. Ronald Reagan was only moderately popular with investors in 1980, and only average in 1984. Franklin Roosevelt was not hated. Investors rated him above average in 1936 and 1940, and only a bit below average in 1932. Both Bushes were disliked by investors, except for George W. Bush in his second term, and Dwight Eisenhower got no investor love for his second term.
To contact the editor responsible for this story:
James Greiff at firstname.lastname@example.org