Our Company and Sites

iShares evolved sector ETFs

Whitepaper:
Evolved sectors for a changing economy

Read more

Whitepaper:
Evolved sectors for a changing economy

Read more

Is Amazon a retailer or a tech firm? Identifying a company’s main business drivers is no longer clear-cut. Martin Small and Jeff Shen discuss a new way to approach investing in sectors using data science.

  • View transcript

    Do you do anything else in your life like it's still 1999? Do you listen to the Backstreet Boys on your Discman? Do you go pick up a video at Blockbuster? When you open your e-mail, do you still hear, "You've got mail"? Of course you don't. You don't do anything like it's 1999 anymore, so why would you keep making your sector investing decisions that way?

    MARTIN SMALL: Hi. I'm Martin Small, the Head of U.S. iShares, and I'm here today with my great partner and terrific colleague, the co-head of our systematic active equities business at BlackRock, Jeff Shen. I'm here today to talk about an exciting innovation that we're working on at BlackRock and at iShares, which brings together the power of technology and great investment in human acumen to reinvent sector investing.

    JEFF SHEN: There are really two characteristics to this approach. I think number one is the recognition that a particular company can be involved in multiple businesses. They can be originally doing one thing but then gradually venture into other businesses. Number two is really also a recognition that the economy keeps on evolving. The world keeps on changing, and the companies keep on changing as well. So a very simple example here is to look at a company like Amazon which started out as a book seller, then ventured into a technology sector, launching Amazon web service, cloud computing, and with the latest announcement certainly seem to be that it's getting into grocery and also healthcare. So, as Amazon is starting to do many different things, it's very important for our methodology to adapt to it to provide much more robust and intuitive exposure for our end investors.

    MARTIN SMALL: I think that's the most interesting part of this is that, when investors think about how they want to make sector allocation decisions, they want to have things, I think, that are intuitive, so that, if they want to buy technology stocks, some technology stocks might be classified as consumer discretionary, and that would be a bit of a surprise to them, so this methodology, as I really understand it, what we're aspiring to do here is to render parts of the market in a way that are more intuitive to investors. Can you tell me a little bit about some of the technology that's used to do that?

    JEFF SHEN: So machine learning is at the core of the methodology, and how we look at this is certainly looking at the disclosure a company makes on their own and how they talk about their business, how they talk about their business tends to be descriptive of not only where they have been in terms of business but also importantly on a forward-looking basis where they're going to go. So we use natural-language processing to look at the descriptions and then importantly using machine learning to cluster the companies which are doing similar business together. And we think this way of clustering companies for sector allocation can very much be forward-looking and not only describing multiple businesses that the company's involved in, but also importing where they're going to go.

A new way to approach sectors

Technological advancements are constantly changing the way businesses develop and deliver products and services. Modern companies have not only evolved their business models, they have revolutionized behaviors—Apple changed communication, Amazon changed commerce and Alphabet changed the way we access information.

Yet traditional sector classifications such as GICS tend to group together companies using a backward-looking lens. A company’s and even an economy’s changing dynamics may not be captured.

Anticipating change and challenging the status quo have always been at the heart of iShares’ mission. That’s why we've implemented a new way to classify companies – one that looks where they’re going rather than where they’ve been.

Case study: Amazon

  • Amazon started as an online retailer selling books, but over time it has expanded to other business with the build-out of Amazon Web Services (AWS), the acquisition of Whole Foods and its recent announcement involving the healthcare market.
  • GICS classification: consumer discretionary
  • Evolved classification: discretionary spending and technology

BlackRock’s evolved sector approach

BlackRock’s evolved approach seeks to provide a more representative view of sectors within the U.S. economy by:

  • Using forward-looking inputs guided by big data analysis
  • Allowing companies to sit in more than one sector
  • Allowing sector constituents to change more frequently than traditional classifications

The evolved sector approach uses text analysis, guided by machine learning statistical techniques, to identify words and phrases companies use to describe themselves in publicly available materials such as regulatory filings and earnings reports. Companies are grouped into sectors based on similarities in the language each uses when describing their businesses (see Figure 1). In the example below, words in larger font sizes occurred with greater frequency among the listed companies and had greater relevance.

Figure 1: Text analysis of the words that “innovative healthcare” companies use in describing themselves.

Figure 1: Text analysis of the words that “innovative healthcare” companies use in describing themselves

Names of corresponding companies in the “innovative healthcare” evolved sector

Names of corresponding companies in the “innovative healthcare” evolved sector

For illustrative purposes only.