After making some revisions to the recovery indicator since my last post, the recovery indicator for my business cycle model is now complete, at least for a first revision. The indicator has six signals, three related to company fundamentals, and three bank stimulus indicators.
Quantitative analysis tools provided by Portfolio123.
The recovery indicator was modeled from the guidelines from Fidelity's Typical Business Cycle chart, specifically the early phase of the business cycle. The ultimate plan is to create indicators for a three phase cycle, with mid and late combined, unless I find a reason to separate them. Once the phase is determined, then market timing signals can be designed for each phase, and appropriate investment instruments selected specific for the phase.
As shown below, the indicator was asserted twice in the last fifteen years. The first assertion marked the recovery from the 2001 recession and was in sync with the rebounding stock market. The second assertion relates to the 2009 recession. In this instance the signal was significantly delayed from the 2009 stock market recovery. Note that this is not a failure in modeling, as the indicator was modeled after the fidelity guidelines as best as possible with the Portfolio123 financial database.
The bottom line is that the stock market rebounded in advance of the economic recovery, so my business cycle investment strategy will have to take that into account. The next steps will be to model the contraction and expansion indicators, then follow that up with market timers for each.
The ranking system implementation is shown below, split into two parts. If you are implementing this project on your own, then please study these figures carefully, not only for new items, but previous formulas that have been modified as I corrected a mistake or two, and shuffled things around that weren't performing well.
I actually had included several more signals but I had trouble keeping the output glitch-free, and therefore had to reduce the number to make the result more robust. If you are doing this project yourself, you may be happy with using only Sales Year-over-Year (YoY) by itself, or in combination with Earnings YoY, as a simple version. There is no need to get too sophisticated with this signal.
I assume that the reader is already familiar with the specifics of the structure and how the ranking system works. Please review the previous posts on the business cycle timer for details.
Stimulative Bank Policies
Three signals were included that I refer to as "stimulative bank policies". These signals include declining Fed Funds Rate, declining Bank Prime Rate, and normal TED Spread. The declining Fed Funds Rate and Bank Prime suggest that the Federal Reserve is providing stimulation to the US economy by reducing interest rates. The TED Spread is an indication of credit risk. When the spread is above 0.5, the risk of overnight bank loan default is elevated. Anything less than 0.5 is considered normal.
Three signals were included that are aggregate company fundamentals and I refer to as "business recovery" signals. These have been described in past posts. It should be noted that I split apart the combined Sales-Earnings YoY signal into the original Sales YoY and Earnings YoY due to difficulties in providing a glitch-free output. I also only allow the sales and earnings recovery signals to be asserted for up to one year after earnings YoY are greater than 100% i.e. fully recovered. The first year of recovery tends to be a strong upswing in sales and profits. After that there is the potential for slow down and whipsawing of the recovery signal which is something to avoid.
Below is a recap of the three custom series.
As a reminder (again), don't forget to call up the three custom series as buy rules so that they will be updated. Details are in the previous post.
As a final note, the voting threshold for the ranking system was set at 90%. I did some robustness testing. below is the results with a threshold of 90% (for comparison), and 80%. 70% is same as 80%. Therefore there is a reasonable amount of tolerance within the range of voting results, with only minor variation from 70% to 100%.
Next I will be tackling the business cycle contraction phase.
Good night all.