In the last post entitled Business Cycle Recovery Indicator, I examined fundamental aggregates for company sales, operating income, and inventory. I concluded that I could make a reasonable "Recovery Indicator", but it would come too late to capture the start of stock market bull runs.
Now I am going to review what I did in this analysis, make some adjustments for errors (I make lots of those), and see what other rocks I can look under, before moving on to the next task in my quest for a business cycle market timer.
Quantitative analysis tools provided by Portfolio123.
The first observation of note from my previous analysis was that I threw Inventory into the same boat as Sales and Earnings. It wasn't my original intention, but somewhere in the middle of the night it seemed like a good thing to do, probably due to the similarity in the custom series charts.
However, according to my edict (Fidelity business cycle chart from previous post), the recovery phase has a signature of low inventory, rising sales, and rising profits. Bottom line is that I don't see evidence of simultaneous falling inventory and rising sales. It just didn't happen after the last two recessions, at least not according to my aggregate data.
Also, my original intention was to analyze change in inventory in conjunction with change in sales. Since I don't see any evidence of falling inventory along with rising sales, I can keep the analysis separate, meaning that I don't need to use the "SP1500 Manufacturers" custom universe for sales and earnings, only for inventory.
I also found that using Equal Weight for the aggregate fundamentals provides smoother, more decisive series/charts than those generated using MktCap-Weighted aggregates. I have therefore redone the custom series using equal-weight, and also the generic S&P 1500 universe in the case of Sales and Earnings.
Lets start with Sales and Earnings, then deal with Inventory later. Below are the new custom series rules and charts.
Note the criteria that I use for UnivAvg: "Abs(OpIncTTM)<10*Abs(OpIncPTM)&Abs(OpIncTTM)>Abs(OpIncPTM)/10". This criteria filters out some wacky values that cause huge glitches in the custom series chart. I will be using this type of filtering a lot in the future as custom series tend to be quite full of glitches.
The last step for Sales and Earnings is to put them together into one custom series, mostly for convenience.
As you can see from the above chart, the Earnings-Sales recovery phase is quite apparent. That is the good news. The bad news is that it is a lagging indicator. It can be used as what is called a "confirming indicator" but the stock market, which is a leading indicator, already knows that the economy is in recovery.
The question is "am I splitting hairs?" I will still need to have a trigger signal for a bull market and does it really matter whether it is located in the recovery phase or the consolidation phase? The answer is "I don't know". But for certain, I have not followed through on the entire signature for a business cycle recovery. Lets review the Fidelity diagram.
I have examined sales improvement and profits growing rapidly. What I haven't explored yet is "Activity Rebounds", "Credit begins to grow", and "Policy still stimulative". And "Inventories low" may not be applicable, unless the recovery stage can be moved ahead by use of other indicators.
Before I sign off for today, I would like to show some additional observations relating to inventory. But first, a modified custom universe is required specifically for the inventory custom series. The Earnings custom series is no longer using the SP1500 Manufacturers Universe, so I have deleted the operating income rules.
The first thing relating to Inventory of note is that the criteria that was identified by Fidelity was for "inventory to be low". Examining Inventory year-over-year as I did, clouds the picture. An easier way to visualize low inventory is by simply charting InventoryTTM, as shown below.
From this chart, it can be seen that inventory levels are clearly dropping in 2002 and 2009, and flatlining in 2015-2016. What I don't know is what is meant by "low" but I think that since inventory levels grow continuously except for those three time periods, one can conclude that when inventory is dropping that it is "low".
Getting back to Inventory Year-over-Year, it can be seen that there are two types of slowing inventory buildup. Where I have circled the chart below, is the start of rapid decrease in inventory buildup. This may be a sign of new activity going on. But that will have to wait for the next post.