GCC and the Composite Leading Indicator

Today’s post looks at the Global Growth Cycle (GGC) strategy from Grzegorz Link. The strategy uses the OECD Composite Leading Indicator (CLI) as its TAA indicator.

Grzegorz Link

I don’t know anything about Grzegorz, but I found his strategy on Allocate Smartly (AS) when I was investigating Growth Trend Timing.

• They have written three posts about GGC, with the first dating back to December 2021.

AS is a great site, with back tests of many TAA strategies.

I’m keen to add more strategies to my TAA model that use economics rather than simply price indicators, so naturally GGC was of interest.

Rules and Results

What is the CLI?

OECD Composite Leading Indicator (CLI) data measures a wide range of “leading”(forward-looking) economic indicators for major world economies.

AS notes that since economic data is subject to later revisions, using historical data in backtests introduces lookahead bias.

• They plan to calculate the impact of this bias.

On the 15th calendar day of each month calculate a“Diffusion Index” based on CLI data at the previous month-end. [This is] the % of countries whose CLI value rose month-over-month. Absolute CLI values don’t matter.

If our Diffusion Index value is >= 50%, go long SPY (S&P 500) at the close, otherwise move to cash. Hold position until the 15th calendar day of the following month.

The strategy would have produced equity-like returns over the last 60+ years, while significantly reducing drawdowns. It’s rare for a strategy that’s only based on economic data to be so successful at limiting losses.

Revisions impact

We’ve attempted to model how the strategy might have performed if we traded based on the CLI data at that point in time rather than the CLI data as it looks today.

The most obvious difference between the backtests are those big drawdowns in 1987 and 2002/03. In both cases, the meat of the difference was a discrepancy in just one single month. Such is the risk of going “all in” on a single risk asset.

This is an imperfect analysis: (a) the OECD historical vintage data is very wonky and not even available prior to 2001, and (b) constituent countries have changed over time (survivorship bias).

Naturally, we plan to use GGC as just part of a composite TAA signal/system, and TAA as only one sleeve of our overall portfolio.

• So this problem would have limited real-world impact.

AS still likes the strategy:

The strategy successfully limited the worst market drawdowns and outperformed the market on a risk-adjusted basis. Further, the novel approach taken by the strategy is quite different than others we track.

Historically, it’s been most similar to Philosophical Economics’ Growth-Trend Timing, but even there, monthly correlation is a relatively low 70%.

That’s also the attraction for me, as a diversifying strategy different from all of the MA-driven systems.

Revisiting GGC

AS’ second article dates from August 2023. 

• The strategy had performed well in the interim and had received a lot of attention from AS members.

The CLI signal had also recently been streamlined from 43 countries down to 17, so AS felt a review was in order.

The old and new CLI indexes are very different in terms of constituent countries, and our tests of the two only agree on the direction of the market about 83% of the time. Despite that, the overall performance of the strategy would have been largely unaffected over the long-term.

The streamlined version is actually a bit better.

Adding Momentum

Article number three dates from December 2025, and looks at an enhanced version of GGC which uses momentum as well as CLI to determine its monthly allocation.

If the CLI “Diffusion Index” is > 50%, we are risk on.

• 12-month returns of US vs international stocks determine which asset to allocate to.

If CLI < 50%, we are risk off.

• 12-month returns of Bonds vs Bills determine which asset to allocate to.

Note that momentum here is used to choose within both regimes.

• With GGT, momentum/trend is a second test, used only as confirmation when a recession signal is present.

The enhanced version of the strategy is a significant improvement.

AS also prefers the enhanced version:

We would prefer the Enhanced version, because it takes into account long-term momentum, a fundamental market force that has worked for basically as long as financial markets have existed.

But with a caveat about portfolio/ensemble use:

If we were building a diversified portfolio of strategies, we might prefer the original version. Many of the strategies we track already have exposure to momentum. If the purpose of GGC is to smooth out performance, it may be more useful to have a “purer” signal with lower correlation to other strategies in the portfolio.

This makes sense and shows up in the AS portfolio optimiser:

Despite inferior performance on paper, the original GGC still appears in more optimized portfolios.

Without getting too Inception about this, this principle operates on multiple levels.

• The AS logic applies to a situation where your overall portfolio is a collection of TAA strategies.

In practice, I have a portfolio that includes DB pensions, property, venture capital, passive equities, equity factors, thematic allocations and leverage.

• You need a top-down approach to work out whether adding more trend/momentum gets you closer to your target allocations. 

Mid Month

The original GGC traded mid-month rather than at month end, so AS compared the two approaches.

On paper, trading at mid-month has outperformed trading at end-of-month, but most of that outperformance came during a brief period in the 1980s, and the two versions have tracked each other closely since.

Conclusions

GGC (and GGC enhanced) look very interesting to me and I plan to make use of them.

• I’m grateful to Grzegorz and AS for introducing me to them.

---

That’s it for today.

• Until next time.