By Stephen Jen (via Itau Global Connections).
Bottom line: Global equity prices have historically exhibited a strange but powerful seasonality. The trading performance from May to October has been consistently, persistently, and significantly inferior to the equity returns from the November to April period. Using a sample of six developed markets and data from the past 32 years, we find un-weighted equity returns to be -1.8% for the May-October period and 14.3% for the November to April period. This disparity in equity performance over the two parts of the year holds across different markets and over different time frames. Further, in terms of the economic fundamentals, in the last three years, US and Chinese macro data have exhibited weak patches during the middle of the calendar year. While we do not have a strong opinion on whether global equities are about to tank, we are increasingly worried about such a risk, particularly in light of the weaker-than-expected data around the world lately. Having said this, the tail risks are substantially lower now than those seen last summer: when China ran a serious risk of a hard landing, with European Crisis reaching the zenith of its intensity, and with the US economy turning south without the ‘protection’ of QE3. Therefore, if we see a mid-year sell-off in global equities, it might not turn out to be as violent as last year. Further, we suspect JPY will no longer be treated as the safe haven currency of choice, and the Nikkei may still manage to rally, against the prospective global trend in equities.
Some observations about the US economic seasonality. The April NFPRs were worrisomely weak. Other indicators in the establishment and household surveys also confirm that the US labor market was weak in March. Further, the ISM, jobless claims, retail sales, and consumer confidence all disappointed expectations in recent weeks. US GDP growth is generally expected to decelerate from around 3.0% in the first quarter to only 1.0% or so in the second quarter.
The charts below (see attached pdf) show that, in the last three or four years, there has been a soft patch in the middle of the year for both the ISM and the NFPRs.
These powerful seasonal patterns have led the overall macro outlook of the US economy to exhibit a soft patch, or negative surprise, in the middle of the year. The chart on the left below (see attached pdf) shows the Citigroup Economic Surprise Index (CESI) and its seasonality over the past four years.
There are two important caveats about this seasonal pattern. First, the chart on the right shows that the opposite pattern was observed in the 2003-09 period and that the middle of the year actually showed more strength. Second, the same seasonal pattern (economic weakness in the middle of the year) is not seen in most other countries and looks like a US-specific issue. China also had a soft patch in the middle of last year and in 2010, but not in the other years.
Seasonality in global equity prices. Regardless of the stability of the seasonality of the US economy, there has been a rather powerful seasonal pattern in global equity prices. The first table below reports the average monthly returns across six major MSCI country equity indices (US, Germany, Japan, UK, Canada and Switzerland) over the past 32 years. The second table shows the paired t-statistics for the statistical significance of the returns in each month. Specifically, the test compares the volatility-adjusted returns in each month against the returns in the rest of the sample in the past 32 years.
The well-documented September (negative) and December (positive) effects are evident in this table. September is especially negative because several large market sell-offs happened in September, e.g., the 1987 crash and the September 2008 crash. December is usually positive, and there is a lot of literature trying to justify this.
While returns during the month of May are on average dismal (around 0.1% across six countries on average), contrary to popular assumption, over the last 32 years the ‘Sell-in-May-Go-Away’ effect does not appear to be statistically significant or consistent across countries.
Using the same data, the table below highlights the stark difference in equity returns during the May-October period (really bad) and the November-April period (much better). This dichotomy was consistent across the countries in our sample. Correspondingly, risk-adjusted returns (return/vol ratios below) for the latter period are much better than those for the former period.
Even removing the September (negative) and the December (positive) effects mentioned above, the disparities between these two periods persist. (Broadening these calculations to include some southern hemisphere countries, such as Australia, we find very similar effects. And this is most likely related to the global common factor effect on the rest of world equity markets).
April is not bad for equities. While our calculation results don’t guarantee equity returns in the various months, April is historically a decent month for global equities. The un-weighted average across the six countries in our sample is 1.6, and the equity returns for April were positive for all six countries.
– Is the yen still seen as a safe-haven currency? We are no longer convinced that the JPY will behave like a safe-haven currency as it has in the past. The BOJ’s announcement on April 4, 2013, was so powerful that we suspect the JPY may not rally much in risk-off scenarios. The market’s reactions to last Friday’s NFPRs release are a good illustration of our point: that the divestment from JGBs by Japanese institutional investors could be so powerful that USDJPY could remain supported in a risk-off environment.
– The summer of 2012 was particularly bad. Last summer was marked by extreme risks in all three time zones: rising risk of a hard economic crash in China, intense market pressures in Europe, and a sharp economic slowdown in the US. While China’s growth seems to be flattening out, it does not seem to be at risk of a hard landing. However problematic still, Europe does not seem to be close to the crisis threshold it found itself in last summer. Also, the US private sector seems to have better growth support now (an improving housing market and labor market). Finally, last summer, the Fed had not committed to QE3 and the ECB was forced to announce the OMT. Both of these monetary operations, plus the latest announcement by the BOJ, should provide a bit more support than last year. In short, the left tails just don’t seem to be that scary this time around.
– On seasonality itself. The seasonality concept has always been subject to debate in the academic literature and among market participants. In particular, the justification and robustness of the underlying causes of seasonality have been the major sources of debate. We admit that we have no clear explanations for the disparities and the seasonality presented above in equity returns in the two periods. However, at the same time, we think the results are far too strong and consistent to dismiss. Why such strong effects remain in returns after several decades is strange. One possibility is that there might be related seasonality in volatility dynamics (level and range of volatility) during same periods, and in turn this seasonality in volatility might be linked to unexpected news and data hitting the markets during those periods (we are investigating this effect ourselves).
Bottom line. Without making an explicit call on whether global equities go higher or lower from here, we cite some interesting historical facts. Most important is the fact that developed-market equity returns have been consistently, persistently, and significantly lower for the May-October period compared with the November-April period. This fact worries us. However, the left tails just don’t seem to be as severe now as they looked last summer. Finally, in contrast to last summer, financial markets have the partial protection of the Fed’s QE3, the ECB’s OMT, and the BOJ’s ‘Q-squared’ (quantitative and qualitative) monetary easing. If we have to guess, we suspect any sell-off this summer will likely be less severe than those in the previous summers.
Stephen Jen is the managing partner at SLJ Macro Partners.
Prior to establishing SLJ Macro Partners in April 2011, Stephen was a Managing Director at BlueGold Capital (since May 2009), working as the key risk-taker in currencies and as its macro strategist. Before BlueGold, Stephen was a Managing Director at Morgan Stanley and, from October 1996 to April 2009, held various roles, including the Global Head of Currency Research and the Chief Global Foreign Exchange and Emerging Markets Strategist. Prior to Morgan Stanley, Stephen spent four years as an economist with the International Monetary Fund (IMF) in Washington, D.C., covering economies in Eastern Europe and Asia. In addition, Stephen was actively involved in the design of the IMF’s framework for providing debt relief to highly indebted countries. Stephen holds a PhD in Economics from the Massachusetts Institute of Technology, with concentrations in International Economics and Monetary Economics. He also earned a BSc in Electrical Engineering (summa cum laude) from the University of California at Irvine. Stephen was born in Taipei, Taiwan, and now lives in London with his wife and two children.