Part of the World Gold Council’s investment service is a detailed quantitative analysis of gold’s correlation with a wide range of asset classes and commodities. This is calculated using a series of rigorous analytical tests, carried out on a country-by-country basis. It is accepted wisdom that gold is a hedge against the dollar, equities, bonds and so forth, but nothing is ever that simple, and gold’s effectiveness against other asset classes varies according to the country in which portfolio components are based. In fact, as our table below demonstrates, some asset classes are most closely correlated with gold while in others they are the least closely correlated. 

Here, then, is a distillation of the latest work from the Council, which in this most-recent analysis looked at nine countries, and also at commodities as an asset class. The methodology applies local investment research best practice to each area and includes domestic equities, regional equities in the case of Europe, international equities (non-Europe) and international bonds (both of these two latter are in domestic currency terms). The Bond index used is the Lehman Brothers’ Global Aggregate Bond Index series, which is weighted by market capitalisation. The gold price (and that of other commodities) is determined in local currency units, so that each matrix generated is specific to a country or a market. Calculations are based on weekly data using underlying daily data and the results are published at the end of each calendar quarter. The work looks at three-year and five-year periods, as well as generating volatility comparison. 

The following table gives an encapsulation of some of the results of the three-year analysis. For each country, gold is listed against the asset class with which it has the highest positive correlation and the lowest (generally negative) correlation. What may not come as much of a surprise for those versed in investment strategies is that there is a very high number of assets with which the correlation coefficient is not significantly different from zero. 

In the US the exercise was carried out against commodities, equities, bonds and real estate investment trusts (REITS), while in the majority of other countries the assets considered tend to be equities, bonds and interest rates. 

 

Country

Lowest correlation

Coefficient

Highest correlation

Coefficient

US

Three-month T-Bill yields

S+P 500

-28%

-8%

DJ AIG Commodity Index

56%

South Africa

BESA All Bonds Index

-14%

90-day deposits

-1%

Australia

MSCI* World excluding Australia

-48%

LB Global Treasuries Index (A$)

49%

Belgium

LB Global High Yield TR** Index, €

-23%

EURIBOR 3-month

-1%

France

CAC-40 (equities)

-14%

LB Global High Yield TR Index, €

+12%

Germany

LB Global High Yield TR Index, €

-33%

TecDAX

0

India

BSE 200 index

+2%

3-month Deposit Rate

-22%

Italy

EURIBOR 3-month

-1%

LB Global High Yield TR Index, €

-33%

Japan

MSCI World Excluding Japan

17%

LB Global Treasuries TR Index, ¥

43%

Netherlands

LB Global High Yield TR Index, €

-33%

EURIBOR 3-month

-1%

Switzerland

LB Global Treasuries Index TR Index – CHF

-16%

MSCI World excluding Europe

-2%

UK

MSCI World excluding Europe

-24%

FTSE UK AIM TR Index

8%

Commodities

Goldman Sachs livestock TR Index

5%

Silver

82%

*Morgan Stanley Commodity Index                                  ** Total Return 

Some results are as expected, but some are more surprising (two results are given for the lowest US correlation as the asset with the lowest correlation is a yield, rather than a price, so the next lowest tangible asset has also been included – which turns out to be the S+P 500). The MSCI inflation index shows up three times as the asset with the highest positive correlation – which should come as no surprise – but it has the lowest correlation in Switzerland, while Switzerland houses the highest correlation with the Lehman Brothers Treasuries index; and this Index has the lowest correlation with gold in Japan. The high correlation with bonds in Switzerland probably reflects cautious portfolio allocation (gold and bonds) while the low correlation in Japan may reflect the search for yield in the face of very low domestic rates. Yields in Australia have been high, however, and here the LB Treasuries index also has the lowest correlation with gold; this may reflect the use of both gold and global Treasuries as a hedge against local possible currency weakness. The high correlation with AIM Stocks in the UK would reflect the heavy mining component in that index. 

The Council’s analysis is more detailed than the encapsulation shown here and, as noted above, one of the most important results is that gold has low correlations with a wide enough range of asset classes to make it a worthy weapon in the battle to expand your portfolio efficiency frontier. The devil is in the detail, however, and if this exercise shows us one thing it is that, while markets may be increasingly globally integrated, they are not homogeneous.