When looking at economic cycles, the amount of data to measure can be overwhelming. It can get confusing to look at hundreds of different data points which is why we try and simplify as much useful data as possible into composite indicators. Once we have a collection of important data series in one standalone composite indicator, we can then measure what the basket of economic data looks like as well as dig into the dispersions of the underlying composite index for more detail.
When constructing a composite index, or a leading index more specifically, we are not trying to jam as many data points into a single composite as possible but rather carefully select a handful of time-tested and reliable indicators that logically have to move first over the course of an economic cycle.
If we take employment as an example, a company will likely lay-off temporary workers before full-time workers. Also, the average workweek length or overtime hours will likely be reduced before the firing of full-time employees. At every turning point in the economy, a slightly different indicator picks up the inflection point but if we put together a collection of the most logical early turning points, we can have an increasing amount of confidence that the aggregate index is moving in the same direction.
Once we couple several baskets of indicators and all of those baskets are moving in unison, the confidence in a cyclical turning point goes up yet again.
Each Thursday, the Department of Labor updates the initial jobless claims series which is another logical and time-tested leading indicator of actual payrolls. In this particular data release, the DOL revised the time series for seasonal factors dating back to 2014. With economic data, these revisions are common.
Initial claims came in better than expected and the trend that was starting to alarm some hawks of the initial claims data has subsided.
This is where the composite index approach comes in. It is difficult to monitor dozens of data points for every industry or sector of the economy. If we are looking for a read on the employment market, we now have to look at hours worked, temporary employment, overtime workers, etc. Not only is it difficult to synthesize all of these data points into one consistent outlook, but it is also hard to form an objective measure on when all the indicators turn as a collective basket.
I use two main leading indicators for the employment market. Once these indicators both inflect in the same direction, what that means is that underneath each of these indicators, there are dozens of logical leading metrics all moving in the same direction. I can then dig into each metric to find out where the inflection point is originating from in terms of sector or metric.
The first leading indicator that I use comes from hard economic data including three different sources (to avoid sourcing bias) including hours worked, overtime hours, initial claims and a ratio of temporary work. This basket includes data that all logically should lead the overall employment market and when this basket turns as a whole, my attention to an inflection point in the employment market is elevated.
Below is the year over year growth rate of the leading employment index.
Tracking the smoothed average of the leading employment index shows that a slowdown in the growth rate of employment is underway and this is consistent with the recent payrolls data that has been reported. If the smoothed average of the leading employment index moves into negative territory, that would be an early sign that the employment market may start to deteriorate quickly. After this occurs, we only need one or two payrolls reports to confirm a trend when consensus will still be attributing the weakness to one-off factors. We know the inflection is not due to one-off factors based on the logical basket of indicators that have inflected lower in aggregate.
To further confirm an inflection point in a given sector, survey data can be less reliable but useful when studied as a confirming measure.
There are five regional Federal Reserve banks that publish surveys on manufacturing conditions. Each survey has a present conditions section as well as a future expectations section. Each regional survey has its flaws and one-off reasons for monthly moves up or down but if we study all five regions as a whole and look at the trend, we can gather import trending data points.
I go into each regional Fed survey and pull out what the respondents are saying about future employment trends, including the expected change in employment numbers and the expected change in the average employee workweek. If respondents are lowering their outlook on future employment, in conjunction with a negative inflection in the hard employment data in the leading index above, we can have increased confidence in a turning point.
The smoothed average of the Regional Fed Future Employment Gauge has also started to inflect lower.
Neither of the leading indicators are showing trends fully consistent with massive job losses but rather a declaration in the pace of employment gains. When the next employment report comes out, I will not be focused on the headline data but rather what the internals of the report does to the trending direction of the leading employment index.
This was just one example used for the broad employment market but we can break this composite index approach down into sectors of employment and see when manufacturing employment may inflect negatively as compared to services employment.
Also, we take this same approach from a broader sense and use a basket of indicators for the entire economy to spot inflection points in both growth as well as inflationary trends.
A data-driven composite approach allows us to make consistent and confident calls on the trending direction of the economic cycle over 12-36 month rolling periods of time.
For now, most of the data in the US, and globally for that matter, is still inflecting lower. If there is an upturn in growth, I will be the first person to report such a development and spot where in the economy the upturn is emanating from.
Until then, I remain positioned for decelerating economic growth and have been expressing that view with an overweight position in Treasury bonds, spread out across the curve with a focus on extended duration ETFs such as TLT and EDV.
I am not short stocks on a net basis and have an underweight exposure to equities on average with a concentration in a few long defensive sectors.
If/when growth inflects higher, I will shift the balance to a more risk-loving portfolio or get even more defensive if the leading indicators of growth, manufacturing, employment and more all start to move even lower.
Regardless of where the economy moves next, the asset allocation and strategic positioning will come from an objective analysis of the trending direction in a series of leading economic data points.