How to build your first ETF Screen

In the previous article, we learned the basics of ETF screening and how can we use it to find suitable ETFs. In this article, we will understand the process of building your first ETF screen by taking a relatable real-life example. 

Building Your First ETF Screen

Building your first screen is super simple.

You can start from a template screen and edit the filter/query or start building a brand-new screen by adding your first query. We have highlighted the results in the images in yellow.

Let’s work with an example. 

Suppose an investor wants to identify ETFs in the large-cap category with high liquidity, the best performance, and a low expense ratio.

Now we will run through the steps required to create this screen.

Some users prefer a granular approach where they can apply the query, evaluate the list, and iterate on the query. This offers more flexibility in query building. In this user guide, we follow this approach.

Step 0: The default list

The starting list will be all the ETFs available. 

As we can see, we have 164 different fund options. This includes all the fund categories like equity ETFs, debt ETFs, etc. 

Step 1: Filtering the universe

Our first step is to narrow down the universe to ETFs in the large-cap category. This can be done by selecting the ETF category from the basic filter section. From the dropdown, we will select Equity: Large-cap. As index funds that follow broader indices also invest in large-cap stocks, they will be included in this list.

As you can see below, after adding this rule, we have filtered down our selection universe to 49 ETFs (from available 164 funds). 

Step 2: Identify ETFs with high liquidity

Now that we have 49 funds on our list, we can add another filter that will help us identify highly liquid ETFs.

For that, we will use the trading volume (turnover) as a parameter. We will shortlist funds that are in the top 25% based on trading volume.

Our categorization system makes it easy for users to find the metric they want. Suppose our investor decides to use AUM instead of the trading volume. She can see a detailed description of the AUM including the benefits and risks involved in using the metric.

With trading volume as the selected metric, the user can build the query (filter) through an easy-to-use and intuitive query builder.

Users can create an “absolute” comparison query (e.g., turnover greater than 1 Cr.) or a relative comparison query (turnover is in the top 25 percentile). The latter will pick the top 25% of funds (from the original list) with the highest turnover. In our case, applying this query will filter down the list to 13 ETFs with the highest turnover.

Note 1: In our screening system, whenever we screen based on percentile, metric values are always sorted in ascending order internally (low to high). So, if you want lower values (as is the case with drawdowns) you should select “bottom x percentile” and if you want to screen higher values (as will be the case with turnover) you should select “top y percentile”.

Step 3: Identify ETFs with low tracking error

We have picked 13 funds (top 25 percentile) with the highest liquidity. We can now proceed to apply our next filter – funds with low tracking error. We will select those funds that are in the bottom 25% based on the tracking error. 

The fund count after this step is 4. 

Step 4: Identify ETFs with high AUM

We need to apply one final filter to our filtered list from Step 3. We need to pick funds with large AUM. We apply the same procedure by selecting the AUM from ‘Basic filters’ and applying the relative (percentage-based) filter. Here we have selected funds that are in the top 50% based on the AUM.

After applying this filter, we have only 2 funds.

That’s it. We have just built our first ETF screen.

Note 2: In our screening system, whenever we screen based on percentile, the sequencing of the rules matters. If you change the sequencing, the results will also change. Below are the results of 2 queries with the same rules but different sequencing.

As you can see, the rules are the same and the sequencing is different. But both screens show different results. This happens because of internal sorting that happens when we use relative (percentile-based) queries.

In the next articles, we will see some examples of creating different ETF screens.