The 'Sound Asset Allocation Strategy': The Missing Piece

Includes: SPY
by: The Vice Investor


This article is a follow up of an article written by ''US Investor'' which had me intrigued.

The shape of the yield curve is a good advanced indicator of the real economy.

With the stock market as an advanced indicator of the real economy as well, a timely exit of the market is challenging as we observe patterns in the yield curve.

An inverted yield curve is a signal of troubles ahead; outperforming the market is possible by using the indicator and following a few rules.

In a recent article, ''US Investor'' provided general guidelines on when to buy, hold or sell the market.

The idea was to use the steepness of the yield curve (the difference between the yield of 10 year T-Notes and 3 Month T-Bills) in order to buy or sell equities.

He based his theory on three economic arguments:

1) When short-term yield goes up more than long-term, investors have no incentive to buy long-term debt. As corporations issue long-term debt, the funding dries up.

2) Banks borrow at short-term and lend at long-term. Bank issues fewer loans, and economic activity cools.

3) Default rates start going up as Libor and Prime rates go up with an increase in short-term rates. The profitability of companies go down as interest expenses go up. The stock market decline ensues.

These are sound arguments, except for a lacking piece of puzzle: Since both the shape of the yield curve AND the S&P500 are advanced indicators of the real economy, we need to discover which one predicts the other. They might as well be coincident, in which case there would be no edge available from trading one using the other. Moreover, if the bond market sends us a ''sell signal'', should we panic and buy T-Bills instantly? When shall we return to the market?

This is my attempt at an answer:


1. Setting the ''signal sine'' indicator (10 year T-Note yield minus 3 month T-Bill yield)

I chose a frequency that is sound for our ''advanced yield curve indicator''.

Legend: In order, Monthly, Quarterly and annual reading of the Yield Curve spread. In orange, recessions in the US. Monthly graph looks fine. Credit to the St-Louis FED for the data.

It is important that our selling indicator does not generate too much noise, i.e. does not cross its monitored threshold too often. Every time I exit the market for T-Bills, I incur transaction costs and taxes.

The monthly yield curve spread (Graph 1) does not give many signals, follows trends and is the most accurate (by definition) of the 3 graphs above, I therefore chose it.

2. Setting rules

Timing our equity selling is our next challenge. In some cases, market returns turned negative shortly after the yield curve reversed (meaning the market consensus started to anticipate the recession to come), in some other cases it took a while.

Our only historical postulate for our test is that when the yield curve reversed, a recession followed. The graph below shows why we feel that way:

Legend: Signal line Dips below zero, economy gets punished. (The indicator dipped below zero in June 1989; the zoom makes it indiscernible.)

I therefore set the following rules for our testing.

· Upon obtaining our selling signal, we will sell all equities after 6 months, which is the calculated average historical time before the stock market (proxied by the Wilshire 5000 Index) turns sour.

· We will remain out of the market until the indicator returns above our selling zone and will remain out for at least 6 months, since the definition of a recession, as per the NBER, is at least two consecutive quarters of negative GDP growth.

· Upon exit, we invest 100% of our cash in 3 month T-Bills, and roll it over as long as we remain out.

3. Result

Hypothetical Compound Annual Growth Rate for various strategies (1982-Today)

1: Buy and Hold (Wilshire 5000 Total return)


2: Instantly Switch to T-BILLS if Signal Line below ''0 ''


3: Switch to T-BILLS if Signal Line dips below ''1'', 6 months from now


4: Switch to T-BILLS if Signal Line dips below ''0,3'', 6 months from now


5: Switch to T-BILLS if Signal Line dips below ''0'', 6 months from now


6: Switch to T-BILLS if Signal Line dips below ''-0,3'', 6 months from now


7: Switch to CASH if Signal Line dips below ''0'', 6 months from now


Legend and comments:

1: Base Scenario

2: Panic sell if signal line dips below zero

3: A flat yield curve is not enough to justify switching to T-BILLS

4: 0.3 is not yet a STRONG SELL, contrary to ''US Investor's'' initial conclusions

5: BEST OPTION. Wait for the market to ''understand'' a recession is coming, then sell!

6: Overly conservative option: Sub-optimal.

7: The strategy outperforms even if investing in 3 month T-Bills is unavailable.


Green line: Binary (''1'' = 100% in Wilshire 5000, ''0'' = 100% in T-BILLS)

Red dotted: Signal Line

Blue: Buy and hold Wilshire 5000 (Hypothetical growth of 1$)

Red: Our strategy, scenario ''5'' in the table above (hypothetical growth of 1$).

Credit: Myself, FRED Excel Add-on for the Data.

The looming recession acts as a catalyst that makes the market underperform. As ''US Investor'' mentioned, it takes a while to stop a moving train, especially when that train is the US economy.


Special thanks to ''US Investor'' for outlining a smart way to predict recessions. It appears that when the bond market, as per our Signal Line indicator, predicts a recession, the stock market does not fully assimilate the information. It is therefore an investable mispricing.

As much as I could ''data snoop'' the very best past historical exit and re-entry points, it seems like the ''6 months wait then 6 months in T-Bills'' rule worked very well. It outperformed a buy and hold strategy by a wide margin, while being ''reasonable'' in laymen terms.

Finally, I will open the discussion by proposing a related strategy, which I intend to use for my personal portfolio: How about buying puts on market indexes ((NYSEARCA:SPY), for example), shortly after the yield curve gets inverted? Knowing implied volatility is expected to rise and the market's future returns are expected to be below average, could that be superior to our binary ''Wilshire5000 or T-Bills'' strategy outlined in the article?

Disclosure: The author has no positions in any stocks mentioned, and no plans to initiate any positions within the next 72 hours.

The author wrote this article themselves, and it expresses their own opinions. The author is not receiving compensation for it (other than from Seeking Alpha). The author has no business relationship with any company whose stock is mentioned in this article.