Hello all, That's a rather long post to show my criteria on how to I define the weights for my 3 ETF portfolio: IWDA, EIMI and WSML. In this post I give some background information by briefly talking about the two main ETF portfolios for long-term passive investors: FTSE Global All-Cap ETF and IWDA+EIMI+WSML (I don't incorporate bonds, I only talk about the stock part). Highlighting that both are great, that the first aims at simplicity, while the second covers a bigger number of constituents and gives more flexibility but involves setting weights and rebalancing. To, then, talk about how allocation weights for the latter case can be calculated. And, finally, showing how I determine the allocation of the different ETFs in my personal case (portfolio composed of IWDA+EIMI+WSML) and how I automated the computation of those weights.
Common ETF Portfolios
In multiple posts I have seen people with a long-term portfolio composed of three ETFs: IWDA, EIMI and WSML. This is a great long-term portfolio following Bogle's strategy, which can be summarized by the following quote: "Don't look for the needle in the haystack. Just buy the haystack!". However, Bogle also advocated for simplicity and, for this reason, other simpler portfolios such as a portfolio with only FTSE Global All-Cap ETF may be preferred for long-term passive investors.
Ideas behind those portfolios
The goal of strategies such as the above is to invest in all areas and sectors of the world economy: large companies, medium-sized companies, small companies (in all developed and undeveloped countries) while at the same time having low expenses. As an analogy to his idea, you can think of an investor doing this as betting that the world economy is going to grow over the next few years.
However, to follow such a strategy, there is one important concept: the weight in your portfolio of each of these different areas and sectors of the economy should replicate the percentages that those sectors represent of the economy, thus having a portfolio that is an accurate representation of the overall economy. Adding a significantly higher percentage to one of these sectors would imply betting on that specific sector of the economy. For example, if you strongly believe that emerging markets will outperform all others, you might allocate a higher percentage to them, adding more risk to your portfolio but also potentially greater reward in the future.
Problems and advantages of simple vs more complex portfolios
Those percentages may change across time, and thus, your portfolio may need to be rebalanced. ETFs such as FTSE Global All-Cap have become so popular between passive investors because you do not need to rebalance the weights of different sectors by yourself, that's automatically done. However, those investing in IWDA, EIMI and WSML, will need to manually rebalance the weights of developed, non-developed economies and small caps.
Even so, a portfolio such as the latter also has advantages such as the fact of covering a greater number of companies (bigger number of constituents), giving greater diversification, and the fact of giving certain flexibility when establishing the weight of small caps, developed economies and emerging economies in our portfolio.
Defining weights
Those investing in long term portfolios composed of multiple ETFs representing the world economy will need to decide how much to allocate to each portion of the market and, also according to the criteria used, they will need to rebalance the portfolio once these weights change.
There exist different criteria to define those weights, for example one could calculate the contribution of each area or sector to the world GDP and use that as a weighting criterion. Or one could take the aggregate market capitalization of each area or sector and use that as a weighting criterion.
For those interested in different weighting criteria, I recommend reading this short article.
My allocation criterion
As I said, there are different criteria to define weights. In my case for example I prefer IWDA, EIMI and WSML. Mainly because from my point of view the allocation in small caps from FTSE Global All-Cap was a bit low and this allocation allows for a bigger level of flexibility on defining weights (furthermore having a higher number of constituents is a plus from my point of view). Furthermore, I also have a preference for a slightly different criterion than the ones previously mentioned. This criterion involves assuming that IWDA, EIMI and WSML are good representations of their respective areas of the economy and that each ETF’s market capitalization could be used as a weighting measure.
For this reason, I extracted the market capitalization of each of those ETFs from their information PDF (IWDA, EIMI and WSML) and calculated their allocations based on those market capitalizations.
Automatizing my allocation criterion
The calculation of those percentages is a pretty simple and repetitive task. The previously indicated URLs always lead to the most updated version of the information PDF and the information PDF always follows the same structure. Hence, I created some code to systematically extract this information as well as the ETF price in order to know how to allocate a portfolio composed of those three ETFs using the market cap criteria. Below you can find R and Python code to automatically extract this information.
R version
Note: The weighting criterion provided in this post by no means is the best, it is just my personal criterion. As I explain there are several criteria that can be used and each has its reasons and assumptions to be used.
Common ETF Portfolios
In multiple posts I have seen people with a long-term portfolio composed of three ETFs: IWDA, EIMI and WSML. This is a great long-term portfolio following Bogle's strategy, which can be summarized by the following quote: "Don't look for the needle in the haystack. Just buy the haystack!". However, Bogle also advocated for simplicity and, for this reason, other simpler portfolios such as a portfolio with only FTSE Global All-Cap ETF may be preferred for long-term passive investors.
Ideas behind those portfolios
The goal of strategies such as the above is to invest in all areas and sectors of the world economy: large companies, medium-sized companies, small companies (in all developed and undeveloped countries) while at the same time having low expenses. As an analogy to his idea, you can think of an investor doing this as betting that the world economy is going to grow over the next few years.
However, to follow such a strategy, there is one important concept: the weight in your portfolio of each of these different areas and sectors of the economy should replicate the percentages that those sectors represent of the economy, thus having a portfolio that is an accurate representation of the overall economy. Adding a significantly higher percentage to one of these sectors would imply betting on that specific sector of the economy. For example, if you strongly believe that emerging markets will outperform all others, you might allocate a higher percentage to them, adding more risk to your portfolio but also potentially greater reward in the future.
Problems and advantages of simple vs more complex portfolios
Those percentages may change across time, and thus, your portfolio may need to be rebalanced. ETFs such as FTSE Global All-Cap have become so popular between passive investors because you do not need to rebalance the weights of different sectors by yourself, that's automatically done. However, those investing in IWDA, EIMI and WSML, will need to manually rebalance the weights of developed, non-developed economies and small caps.
Even so, a portfolio such as the latter also has advantages such as the fact of covering a greater number of companies (bigger number of constituents), giving greater diversification, and the fact of giving certain flexibility when establishing the weight of small caps, developed economies and emerging economies in our portfolio.
Defining weights
Those investing in long term portfolios composed of multiple ETFs representing the world economy will need to decide how much to allocate to each portion of the market and, also according to the criteria used, they will need to rebalance the portfolio once these weights change.
There exist different criteria to define those weights, for example one could calculate the contribution of each area or sector to the world GDP and use that as a weighting criterion. Or one could take the aggregate market capitalization of each area or sector and use that as a weighting criterion.
For those interested in different weighting criteria, I recommend reading this short article.
My allocation criterion
As I said, there are different criteria to define weights. In my case for example I prefer IWDA, EIMI and WSML. Mainly because from my point of view the allocation in small caps from FTSE Global All-Cap was a bit low and this allocation allows for a bigger level of flexibility on defining weights (furthermore having a higher number of constituents is a plus from my point of view). Furthermore, I also have a preference for a slightly different criterion than the ones previously mentioned. This criterion involves assuming that IWDA, EIMI and WSML are good representations of their respective areas of the economy and that each ETF’s market capitalization could be used as a weighting measure.
For this reason, I extracted the market capitalization of each of those ETFs from their information PDF (IWDA, EIMI and WSML) and calculated their allocations based on those market capitalizations.
Automatizing my allocation criterion
The calculation of those percentages is a pretty simple and repetitive task. The previously indicated URLs always lead to the most updated version of the information PDF and the information PDF always follows the same structure. Hence, I created some code to systematically extract this information as well as the ETF price in order to know how to allocate a portfolio composed of those three ETFs using the market cap criteria. Below you can find R and Python code to automatically extract this information.
R version
Code:
library(pdftools)
library(rvest)
library(stringr)
#' Extracts iShares ETFs' number of constituents and market capitalization from
#' its info PDF
#'
#' @param pdfLocation url or path to the ishares ETF info PDF.
#' @return A data.frame containing the ETFs' Market cap and number of Constituents
#'
#' @example
#' extractPDFInfo('https://www.msci.com/documents/10199/178e6643-6ae6-47b9-82be-e1fc565ededb')
extractPDFInfo