Let me draw on a simple example to make this clear. Suppose there were just two factors, and that they returned either +2% or -1% every month with 50% probability. Clearly these would be good things to add to your portfolio, as over the long run they would be expected to each add around 1% of excess return (let’s ignore compounding). If you had the choice of having all the +2% returns and all the -1% returns coinciding (correlation of one) then you should probably still consider taking that, given that it would add return. However, it’s easy to see that you’d now either win 4% or lose -2% every month. If, at the opposite extreme, the excess returns were perfectly negatively correlated they would always add 1% to the portfolio (the +2% and the -1% netted off each month). The expected return over the long run would be the same in both cases, but the volatility of the excess turns would be far lower in the negative correlation world (zero in fact in this extreme example).
So the bottom line is that, while even very highly correlated excess factor returns might still be beneficial, the relatively uncorrelated nature of international factor premia is a very nice result for investors because it offers the potential to lower the volatility of excess returns, effectively smoothing the alpha profile of a multifactor fund.
For factors, Japan is the destination of choice
One final comment on international factor exposures is the extent to which the fund’s investment strategy – which tilts towards value, size, momentum, quality, and low volatility – goes overweight in Japan and, from a sector perspective, in Industrials. With one of the biggest equity markets in the world, Japan, of course, is already the largest component of the market cap developed country index. In 2017 the average weight to Japan in the FTSE Developed ex US index was 21.5%. For the multifactor tilted version of that index, the weighting was taken up to 25.6%, an active overweight of more than 4%.
Now, it’s important to emphasize that with a bottoms up factor methodology like ours, what happens at the sector and country levels is driven by what happens at the stock level. So when one sees an overweight like this in Japan it ultimately comes from Japanese stocks exhibiting the types of characteristics that we want to tilt towards. What is it about Japan that the factors like? Well the country’s stocks generally score well across the factor board so are fairly consistent performers in all five factors.
At the sector level, the largest overweight is to the Industrials which were taken last year from 13.9% in the market cap index to around 21.2% in the multifactor portfolio, an active overweight of around 7.3%. The international Industrials sector has historically tended to score reasonably well on Quality and Size.
So the bottom line is not to overlook factor based approaches when considering your international allocations. Factors seem to like to travel, and to get a little off the market-cap beaten path when they do by visiting different countries, sectors, and stocks.