Business Cycles, Election Cycles and Potential Risks
Bill Hester, CFA
All rights reserved and actively enforced.
The group within the NBER that determines the length of recessions made news this week. The NBER's Business Cycle Dating Committee released a statement saying that it was unable to declare the end of the recession. “Although most indicators have turned up, the committee decided that the determination of the trough data on the basis of current data would be premature. “
This announcement prompted a spirited dissent from Robert Gordon, an economics professor at Northwestern University and a member of the committee. He released a letter saying that he strongly disagreed with the committee's decision not to announce the end of the recession because it was obvious that the recession is over. He argued three main points. First, he suggested that a double dip in the economy over the near term was “extremely implausible.” Second, he noted that based on economic forecasts, the level of GDP should reclaim its 2007 peak level sometime in the second or third quarter of this year. And third, he argued that the shortfall in GDP would need to be implausibly large to bring the level of real GDP from those forecasted levels below last June's trough in economic output.
Professor Gordon is not the only member of the committee who believes it's safe to consider the recession over. Following the announcement of the March employment data, where net gains in jobs were announced, Jeffrey Frankel announced on his blog that he believed that the recession is over. “The last piece has fallen into place”, he wrote. Also Robert Hall, who heads the committee, has said that it's pretty clear that the recession is over. Even so, the committee clearly felt that the data was not definitive enough to declare the end of the recession.
As I noted in January in A View from the NBER Recession Indicators , determining the end of this recession is going to be a difficult task for the committee for two reasons. The first one is that the metrics that the committee uses to date recessions have been - and continue to be - mixed. Metrics that follow production and sales are improving in a way that is typical for the average recovery (but contrary to the headlines, none of the broad measures that track the health of the economy are rebounding more strongly than during an average recovery). Measures of income and employment are still weak, and in the case of income the trends are weakening.
The second reason is that even though the level of GDP has been growing for two quarters, it remains far below its 2007 peak because the contraction in output was so severe. To call the end of the recession now would necessitate a strong degree of certainty that economic growth will continue on its current trajectory even if the factors that have created that growth (mainly government spending) will necessarily need to change.
Some investors may be willing to dismiss these differences of opinion between the members of the recession dating committee as nothing more than a spat between economic historians and unimportant to where the markets are heading. But the nuances within these arguments are important because they highlight what seems to have crept into investors expectations this year – the idea that the recovery will be V-shaped, and that there's little or no chance of it being derailed.
To get a better feel for the difficulties that the committee is facing, I've updated a set of charts from that earlier research piece which show the main measures that the group looks at – industrial production, manufacturing and trade sales, personal income aside from transfer payments, and employment. Each graph has the average change of the metric around the end of recessions since 1950, where data is available. The most recent change in the metric is also shown, making the assumption that the recession ended in June 2009.
It's clear that the recovery in industrial production is in line with prior recoveries. The index tracking real manufacturing and trades sales is lagging the typical recovery, but still expanding even if in a variable way. Personal income before transfer payments continues to weaken. The index declined the last two months and is near its low of September. And while payrolls increased in March, the change in net new jobs added still trails the average recovery by a large degree (and also still trails the “jobless recoveries” of 1991 and 2001).
Following the last few announcements of jobs data, some of the focus has moved from the change in the workforce as measured by the establishment survey to the changes in employment as measured by the household survey. It's easy to see why. In March, while 162,000 workers were added to payrolls, the household survey expanded by 264, 000. And during the first three months of the year, while payrolls expanded by 162,000 workers (after revisions, gains in January netted out job losses in February), the household survey reported that more than 1 million people were back to work.
Worth noting about this trend is that even with the large numbers of workers being added to the household survey, the gap still hasn't been closed when you compare the changes in the level of each of the surveys since June. The rate of job loss was so severe in the fourth quarter of last year in the household survey that the recent strength in the survey of late is now just bringing it back in line with the change in establishment survey payrolls. (For clarity, in the graph below I left out the average change in the household employment number around the end of recessions. For a frame of reference, during recoveries, the average changes in the level of employment measured by the household survey and those measured by the establishment survey are almost identical.)
It's clear from interviews with Harvard University's Martin Feldstein that he sits in the camp that believes the Business Cycle Dating Committee should postpone declaring the end of the recession. He told Bloomberg this week that there were two issues holding him back from endorsing that declaration. One is that since most of the data the group looks at will be revised over the next few months, it is currently difficult to identify a date that the economy turned up. The second point he made was that this time the recession dating process is “complicated by the fact that we are still far below the levels that the economy was at when it turned down, and in my judgment, there is a continuing risk that this economy could run out of steam and could turn down again”.
This is where the most interesting aspect of the debate emerges. From interviews, blogs and open letters, the members of the committee that want to declare the end of the recession seem to be convinced that there is enough internal momentum in the economy where re-attaining the prior peak levels of output is a near certainty. The group that would rather wait to date the recession believes that since the growth we've seen has relied significantly on government programs that are now being wound down, it's best to take a more measured approach to see if private demand picks up where the government leaves off. Only then will they be willing to have the confidence that this recovery will be self-sustaining.
Now that stocks are at rich valuations, bullish sentiment has moved higher, and the VIX index reflects low expectations for volatility - all suggesting that investors as a group are exhibiting a near-certainty that robust economic growth lies ahead – it may be a good the time for investors to consider what the effects would be if the recovery turns more uncertain.
Declines in Year 2 often Precede Year 3 Gains
The next six months will be interesting to watch. With investors showing confidence in the nascent recovery and above-average valuations now priced into stocks, the market is now heading toward a part of the calendar that historically has delivered weak returns to investors.
A few years back, I noted how the most distinct election cycle returns don't actually overlap with the calendar year. While the average return during the third calendar year of a presidential term is above average, the twelve-month period beginning with the fourth quarter of the second year of a presidential term is the true standout. Going back 60 years, the average annual total return for the S&P 500 during this period has been 28 percent.
Many causes have been offered up to explain the election cycle performance, usually focusing on fiscal or monetary policy decisions. But when you look through the economic data, strong and dependable trends don't jump out. One of the better explanations turns out to be the price action that is frequently seen during the second year of the presidential cycle, especially during the April to September period (which is the part of the election cycle that we just entered). Frequent declines during this period often set up the strong 12-month period beginning in the fourth quarter of year two.
The graph below shows the average return during each quarter of the election cycle. For example, the average return during the 1 st quarter of the second year of the cycle – the quarter that has just ended - is little more than 2 percent. The graph shows the strong average returns that have occurred during each of the four quarters beginning in the fourth quarter of year two. The first three quarters of this period are the strongest three-month periods in the cycle when measured this way. But look at the two quarters that precede these gains. These are the two worst quarters in the cycle, on average.
When you look at the individual quarters that make up this poor six-month period, an interesting pattern emerges. Not all of the periods are negative. And some of the periods have strong double digit returns. What explains the difference between the positive and negative outcomes? Part of the explanation can be attributed to the market's valuation at the beginning of the April-September span of year two.
The graph below shows the comparison of valuation and six-month returns. The P/E multiple is from the end of March for each 2 nd year of the election cycle. The six-month return reflects the total return of the S&P 5000 for the 6-month period ending on September 30 of each 2nd year.
In general, higher levels of valuation markets have resulted in weaker performance during this period, on average, while the market has done well from lower valuations. The data doesn't line up perfectly. Low valuations don't always help, as investors found in 1974, and high valuations don't always hurt, as investors found in 1994 and 2006 (although returns were still uninspiring). But it is worth noting that all of the strong returns during the April-September period of the second year of election cycle occurred when valuations were below average (typically below a 14.5 CAPE, the cyclically adjusted P/E ratio, where the denominator is the 10-year trailing average of S&P 500 earnings).
Big declines have tended to occur from higher levels of valuations. With a CAPE of 22 currently, there are only four periods that were more expensive than the current market, heading into this generally unfavorable six-month period of year 2. The average return that followed these periods was -11 percent, and it includes 2002 where the market fell by almost 30 percent.
Calendar cycles are an interesting backdrop in which to frame the analysis of other characteristics of the market, like valuation and market internals. We don't rely on calendar cycles in setting investment policy, because it is difficult to identify cause and effect mechanisms that would make these tendencies reliable (the links between valuations, market action, risk premiums and investment returns are clearer). Also, the average returns presented above can hide a lot of variation in the individual outcomes.
Even so, it might be worth keeping in mind over the next six months that this part of the election cycle has been mostly unkind to investors, and the more steeply valued the market, the less kind the market has tended to be.
|Home | The Funds | Open an Account | Account Access | Research & Insight | Site Map|
For more information about investing in the Hussman Funds, please call us at
513-326-3551 outside the United States
Site and site contents © copyright 2003 Hussman Funds. Brief quotations including attribution and a direct link to this site (www.hussmanfunds.com) are authorized. All other rights reserved and actively enforced. Extensive or unattributed reproduction of text or research findings are violations of copyright law.
Site design by 1WebsiteDesigners.