Yes, you need a PF Flyer magic decode ring to decode Powell’s latest message.

This week’s central-bank bonanza brings with it the usual set of ambiguous and often impenetrable statements and press conferences.

Traders and investors must spend precious time deciphering them. For those who wouldn’t mind a break from this parlor game, there are some markets that don’t march to the beat of the global policy cycle, and offer diversification benefits for portfolios.

The game was in full flow at Wednesday’s Federal Reserve meeting, as the central bank patiently unpicked the higher-for-longer stitching it had spent many months carefully inculcating in the market.

Anyone who had listened to it on financial conditions, keeping rates restrictive for an extended period, or who thought Powell meant it when he implied he was a Paul Volcker and not an Arthur Burns, is now left trying to figure out if the Fed’s reaction function has indeed changed.

Adding confusion to the game are the dots. They reinforced the dovish message, with the median dot implying three cuts next year versus two back in September.

But with one dot implying six cuts, two implying none and the rest all spread in between, this isn’t exactly sure-footed clarity.

It’ll be repeated again today with the ECB, BOE, the SNB and the Norges Bank all meeting to set rates (the SNB held rates steady and the Norges unexpectedly hiked). For those who can trade only G7 markets, there is not much choice but to play the game.

However, for those looking for markets less reliant on the Delphic utterances of central bankers, there are other options.

The problem is that most bond markets tend to be quite alike over the medium-to-longer term.

Interest rate cycles are typically quite synchronized (with the US the most influential), and capital can flow freely around most of the world.

To show the broad uniformity of bond markets we use a statistical tool called principal component analysis.

PCA is a way of making sense of large data sets. For instance, a retailer may have reams of data on how users use its website: how long they spend on it, which pages they visit, where they hover their mouse, etc. PCA will tell you which input – or combination of inputs – has the greatest explanatory power in determining the total time users spend on the website (something they would want to maximize).

Using PCA on yields from 40 countries, we can show that over half of global bond market moves are described by just one factor.

This not a trivial result. It means we could replace our data on 40 bond markets by this one component and it would capture more than half of the variance of the individual yields.

A bond market that was similar to this component would therefore be a good proxy for the global bond market. However, as the mathematician Carl Jacobi advised, we should always invert. A more interesting question is: which bond markets look least like this factor?

The chart below shows how each of the 40 bond markets are correlated to the first PCA factor. Most have a high correlation – more than 75% – but what stands out is the handful of markets with a low or negative correlation.

Japan and Turkey are among those with a negative correlation, while China’s correlation is close to zero.

This is intuitive.

Japan has been running a deliberately counter-cyclical monetary policy, while Turkey until recently was running a through-the-looking-glass one, cutting rates in the face of rampant inflation. And China’s economy is in deflation, at odds with every other major DM and EM country.

The above analysis provides analytical backing for the intuition that Japan et al have provided diversification for global bond portfolios in recent years, while most other markets, e.g. Europe, the US, Australia, New Zealand, etc are low-resolution facsimiles of each other and therefore have delivered few diversification benefits.

Of course, there will be other things to consider before adding a country’s bond index to a portfolio, such as how overbought or oversold it is, how liquid it is, the expected stance of the central bank, and so on.

But if the correlations persist, such a market should produce a more resilient and lower-volatility bond portfolio.

We can extend this analysis to look at stocks.

Global equity markets are even more similar to one another than bond markets, with the first component explaining 60% of the total move. Nonetheless, applying PCA offers up some portfolio diversification candidates (using as our data set MSCI country indices in USD).

Indeed, it can be shown that the first PCA component of equity markets is highly correlated to the US’s manufacturing ISM, elegantly demonstrating that in large part global stocks are driven by macro.

Indonesia, UAE and Chile’s stock markets are all negatively correlated to the global equity move, and therefore the global macro cycle, while European markets move almost in lockstep with it. As the chart below shows, Indonesia’s market (white line) remained supported in late 2022/23, as its market bottomed in July 2022 and rallied through, while other markets went on to make a new low in October of that year.

Indonesia et al also have had among the lowest returns of all global equity markets this year. It is relatively common for the markets that are among the worst performers one year to be the among the top the following year.

We can complete the analysis by looking at commodities.

The first PCA component here explains about half the move of all commodities, a similar proportion to global bond markets. PCA shows that precious metals, such as gold, silver and platinum are the most negatively correlated to commodities.

As almost all commodities are traded in dollars, the first component is quite similar to the DXY.

So to some extent we are asking which commodities are most negatively correlated to the dollar. Precious metals typically exhibit a more negative dollar correlation than most other commodities. Further, in the current environment they should offer some protection against inflation.

The lack of variation among macro-driven markets makes portfolio investment challenging as it is easier to add risk rather than uncorrelated factors.

The above PCA analysis shows that Japanese bonds, Indonesian equities and platinum are examples of assets that may confer some diversification advantages for global portfolios.

At the very least, you may not have to pay as much attention to central bankers.

The magic Fed Decoder ring!

No, that isn’t Mike Pence, RINO from Indiana.