Stockholm (Ekonamik) – Incorporating macroeconomic analysis, factors and news into portfolio management is not straight forward. Winton Capital illustrates this point rather well in a recent report about the reaction of the US Treasury market to macroeconomic variables.
Overall, Winton Capital’s report is in line with a previous perspective from BlackRock, but the quantitative contribution from Winton is important. Their research shows that on average, Treasuries display different sensitivity to different macroeconomic variables, but also that sensitivity to any given variable will change over time. However, one variable appears to stand out according to the research.
“Nonfarm payrolls stands out as moving markets more than any of the other statistics. Reported by the US Bureau of Labor Statistics on the first Friday of each month, nonfarm payrolls counts the number of US workers employed, excluding general government jobs, private household jobs, employees of non-profit organisations and farm employees. The greater emphasis that market participants seem to place on this statistic’s announcement should come as no surprise: employment data would appear to be a logical gauge of the health of the world’s largest economy,” the report argues.
Perhaps more importantly, unemployment data is rather useful to investors because contrarily to GDP data its initial estimates suffer from smaller subsequent revisions, according to research by Boragan Aruoba of the University of Michigan. Subsequent revisions are an important issue in the incorporation of macroeconomic insights into portfolio management. If the data is unreliable, it is natural for investors to ignore it.
The problem of revisions is not an issue in terms of FOMC announcements, which explains why this variable also has such a heavy weight on the market. More importantly, the overnight Fed Funds rate, which the FOMC basically controls, is the foundation upon which all other US$ rates are set, including the Treasuries yield curve.
Purchasing Manager Indexes are a popular industry measure and an important variable according to Winton Capital for good reason. Research from the Fed and the BoE has shown PMIs to be extremely useful to estimate present (nowcast) GDP, a fact not lost to IHS-Markit, the producer of these metrics. The opposite is, however, true of official output estimates which are heavily revised based on hindsight about the business cycle.
After unemployment, Aruoba argues that inflation is perhaps the least revised variable so, it seems appropriate that markets react to inflation information more strongly than they do to GDP.
Although the report does not explore this issue in more depth, it is not surprising to find that the market has become more sensitive to the announcements by the FOMC since the financial crisis. While the relative share of the Fed’s total holdings of Treasury securities has decreased from 46%, in June 2008, to 40% in the last quarter of 2017, its interventions during were much more targeted leading to reactions from other investors.
A noticeable absence from these models, perhaps to avoid endogeneity issues, is the term spread itself, which as a reliable predictor of recessions is also probably an important factor in the behaviour of rates itself. Given the recent inversion of the yield curve, this might soon become a more pertinent issue.
As this note shows, macroeconomic factors might, in some circumstances, allow models of portfolio returns to incorporate non-financial information to provide insights into the medium term. While this might seem a bit esoteric for some investors we should point out that there are services at the forefront of the market offering even longer-term insights. Templeton Global Macro uses factor analysis to incorporate Environmental, Social and Governance factors into their fixed income analysis. Finally, Ortec Finance has also started offering a model that incorporates the effects of pollution and climate change into economic forecasts in a manner previously unseen.
Picture from Pixabay