Profitable Market Timing with the Unemployment Rate

UNEMP is a good market timer.
Enter hedge rule: Close(0,#UNEMP) > Close(3,#UNEMP)
Exit hedge rule: Close(0,#UNEMP) <= Close(3,#UNEMP)

I backtested this to 1974 in excel. You can read all about it here:
http://imarketsignals.com/2016/profitable-market-timing-with-the-unemployment-rate-backtested-to-1974/

Thank you. This looks promising and makes sense.

Jessie Livermore of Philosophical Trader also wrote quite a lengthy blogpost about the historical power of unemployment rate as a market timer, going back quite extensively.

FWIW, he also has a quite extensive post on other Market Timing here …

Good article, thanks.

#RTLSales seems to be an interesting indicator like #UNEMP.

Too bad we don’t have the FRED industrial production growth.

Cool, Thank you Georg, please give as a heads up, when the model triggers a sell, or please create or recomend
a R2G from you that you think is you best timing model, so we can subscribe it…

I remember 2008 when unemployment startet to rise the real trouble begann…

Thank you again!

Regards

Andreas

More FRED data would be nice.

Just saw this on another blog. Are we at the top of the expansion cycle? And what about the trend of lower highs?


Top of the expansion cycle = Start of Recession.
We are not there yet. It looks as if we are entering, or about to enter, the “boom” phase of the business cycle, after which a recession inevitably follows. Let’s hope the boom phase lasts for a long time.

In most Tops interest rates were rising and approaching 10% or more right now they are at historical lows and no country can afford to raise them. Have we ever had these low interest rates at a Top of the expansion cycle? The free money is impacting everything but iits so hard to predict when the bubble bursts.

MV

Interest rates are far below the rates in past expansions.
Companies would be stupid not to barrow at these rates. :slight_smile:

Here’s YoY time series for both. Retail Sales ex food seems noisy and whipsaws a bit, so not sure how the blog author is smoothing this. Industrial production looks pretty powerful on first eyeball glance, but not sure whether it’s a leading or trailing indicator until I get an opportunity to poke around more.

Interest rates are in a secular decline that started 35 years ago. The 10 year at 5% was once considered low. Now it’s half of that.

Having said that, and being slightly bullish on bonds, I do think the new US government may be able to provide the fiscal stimulus to spark inflation. Monetary policy can’t do anymore.

Walter

For RTLSales, I saw the whipsawing as well. I think the author smoothed that out with his algo to use it in combination with his moving averages as he described in the paper.
#UNEMP looks better.
I use the attached in a ranking system (with buy and sell rules) as my current timer. This seems to work ok in 00-02 as well as 07-09. Each node contributes differently in different time periods. For instance, the Fed model worked in 00-02 but not in 07-09 (since the Fed started playing directly with interest rates). it may not work in the future, also, have not tested in earlier time periods (pre-1998):


Capture.JPG

But, with more and more automation, what will “unemployed” mean? Will historical data be relevant?

Kiosks and phone apps are replacing order takers at places like McDonald’s. Autonomous vehicles will replace cab and truck drivers. Automated burger flipping machines, which can cook, assemble, and bag 360 burgers per hour, may replace the cooks at places like McDonald’s. Can the Subway sandwich “artists” be far behind?

Industrial Production declines coincides with recessions, but does not seem to be a good leading indicator of recession. There is a Chemical Activity Barometer that seems to proceed Industrial Production declines by a few months, but I don’t think we have access to it in our p123 economic indicators.

I pulled the S&P and unemployment numbers and found that just hedging with cash using this signal significantly underperforms buy and hold since 1948. Is that why you chose to hedge with bonds?

In the conclusion of the referenced article it states that “Timing provided the same return from 1973 to 2001 as buy&hold stocks”. Only after 2001 timing with the UER provided significantly better returns, because one would have avoided two bear markets.

BTW this model running on P123 will exit stocks either this week or next week, depending on when P123 updates the monthly data.

Did anyone else notice a lag in the Unemployment timer? It seems to be lagging by 2 weeks. Today I noticed that it flipped to green (buy stocks). But it seems to have turned green 2 weekends ago!
I check it every weekend, and only today I noticed this lag.