Reading time: approx. 5min (fast), approx. 8min (slow, @leveragegrinding 😛), think about 3min (@DonkeyInvestor 🙉)
Over the last few days, I've been thinking about the weighting of some regions in my portfolio. There's always a heated debate about how heavily you should weight the USA, Europe, Asia and the rest of the world in your portfolio. Should you have "China shares" at all or how much of the USA is too high?
Especially when it comes to the weighting of US companies, the debate is heated: for some, an overweighting of US companies is perfectly fine, for others it is a cluster risk. In the MSCI World for example, the USA is weighted at 60%. This is why the classic 70/30 strategy MSCI World ETF, an emerging markets ETF is also emerging markets ETF is often added. It becomes even more complex when a Europe ETF on the STOXX Europe 600 is added.
Something that rarely plays a role in the debate, however, is the fact that this weighting only reflects the countries of origin of the stock corporations. Corporations like Apple $AAPL (+1,34 %) , Microsoft $MSFT (-1,05 %) or Google $GOOGL (+0,94 %) may be US companies on paper, but they generate their sales worldwide. One LVMH $MC (-0,25 %) may be from Europe, but it generates the majority of its sales in North America and Asia. Toyota $7203 (+2,16 %) is from Japan, but generates sales worldwide.
I think you recognize the problem: it is important to get an impression of your own regional exposure are not the countries of origin of your shares but rather the question where the sales are generated are generated. I wanted to investigate this question specifically for my portfolio.
Inventory
In my personal portfolio, the distribution of the individual position sizes (as of May 24, 2023) looks exactly like this:
19.3% Apple $AAPL (+1,34 %)
11.1% Costco $COST (-1,09 %)
9.9% Allianz $ALV (-0,66 %)
9.5% VICI $VICI (-0,32 %)
8.3% United Health $UNH (+1,64 %)
5.6% Microsoft $MSFT (-1,05 %)
5.4% Visa $V (-0,14 %)
4.4% KLA-Tencor $KLAC (+0,46 %)
4.1% BJ's Wholesale Club $BJ
4.0% Starbucks $SBUX (-1,55 %)
3.4% Paychex $PAYX (+0,65 %)
3.3% DEFAMA $DEF (-2,77 %)
3.1% Air Products & Chemicals $APD (+0,04 %)
2.9% Rio Tinto $RIO (-0,87 %)
2.7% McDonald's $MCD (-0,31 %)
1.4% Greencoat UK Wind💚 $UKW (+0,66 %)
0.5% Coca Cola $KO (-0,4 %)
This corresponds to a weighting of 81.4% USA and 18.6% Europe. It should be noted that this weighting corresponds to a weighting by countries of origin of the companies. So do I have a huge cluster risk as far as the USA is concerned? In addition, there are no Asian shares in my portfolio.
Now I have made the effort (in sweaty fashion) to all annual reports of the 17 companies on the corresponding investor relations pages. In some cases, sales are broken down by region. Unfortunately, it is clear that not every company provides the same breakdown. For example Apple breaks down its sales in [1] according to Americas, Europe, Greater China, Japan and Rest of Asia Pacific region. The situation is different for Visawhere in [2] sales are only broken down by U.S. and international is broken down.
After a lot of back and forth, I finally decided to split into the regions North America, Europe and Asia & Rest of the World region. North America comprises the USA, Canada and Mexico. Europe is mainland Europe and Great Britain. The last category then includes China, Japan, South Korea, Australia and everything else.
Virtually every company indicates the proportion of sales that are generated in the USA. As mentioned in the Visa example, this is not necessarily useful to find out the share of sales for Europe and Asia when reporting according to the principle "USA and the rest of the world", so in some cases I had to calculate values with a "good guess" estimate. A little online research helped me here. What is important here is not the last decimal place but rather the correct order of magnitude of the sales distribution.
Result
Before I present the results of my analysis, I would like to say a few words about the methodology. In my calculation, the sales shares of the individual companies are weighted by portfolio share.
ExampleApple has a 19.3% weighting in my portfolio and generates 18.8% of its sales in Europe. The 18.8% European share is then weighted at 19.3% in my calculation. Apple therefore accounts for around 3.6% of Europe in my portfolio.
Without further ado, here is the result of the weighted distribution of sales by region in my personal portfolio:
56% North America | 24% Europe | 20% Asia & Rest of the World
The difference to the country weightings in my portfolio are therefore
-25.4% North America | +5.4% Europe | +20% Asia & Rest of the World
Despite an 81.4% weighting of the USA in my portfolio, only 56% of sales are generated in North America, even though I have real USA-pure-plays such as VICI and BJ's Wholesale, which generate 100% of their sales in the USA, have a high weighting in the portfolio. The European share is about 5% larger than the countries of origin of my companies would suggest and despite not a single Asian company in my portfolio, 20% of the weighted sales are generated in Asia and the rest of the world.
Conclusion
Diversification by country is important. In my opinion, however, far too much attention is paid to the countries of origin instead of really looking at where the companies' sales are actually generated. Unfortunately, there is no quick way to find out. The route via the individual investor relations pages of the companies is laborious and still incomplete. I would like to see greater transparency from some companies here. The following were exemplary here Apple and KLA-Tencor.
In addition to the pure "sales risk", there is of course still the political- and regulatory risk. I can sleep peacefully with the distribution of turnover in my portfolio, even if it might be a nightmare for others.
What is your opinion on this?
Sources:
[1] Apple Financial Report 2022: https://annualreport.stocklight.com/nasdaq/aapl/221338448.pdf
[2] Visa Financial Report 2022: https://s29.q4cdn.com/385744025/files/doc_downloads/2022/Visa-Inc-Fiscal-2022-Annual-Report.pdf
