Oleg is the CEO of dxFeed, a market data and financial services firm. Over 20 years of experience in information technology and finance.
Indices are not generally considered as being particularly interesting to the general public. They’re utilized when they coincide with a specific requirement or area of interest. This could be something like wanting to track how well an investment manager is performing against the broader market or to identify how property values have fared historically before deciding to move to a new area.
There is, however, a lot for business leaders to get excited about, especially in a world where we collectively create so much data on a daily basis and where data is increasingly becoming a commodity in its own right.
How are indices used?
Some people like to point to wacky indices, such as the Men’s Underwear Index, which purports to signal broader economic health when men spend more on their underwear. Or the Japanese Haircut Index, in which economically lean periods are supposedly correlated with short hairstyles among Japanese women.
Alternatively, traders have often looked to various indices of luxury goods, such as fine art, boat purchases or champagne, to detect irrational exuberance, or bubble-like behavior, in the underlying markets they trade. The logic is that when these items are overbought, this could signal the peak in other, more conventional, asset classes. Perhaps even affording them the highly sought-after “top-tick” trade of selling the absolute peak of a bubble before it pops.
In recent years, the sheer quantity of discussion that takes place online has allowed natural language processing (NLP) algorithms to trawl ever-richer datasets of publicly available human interactions for use in a growing number of sentiment indices. These are being used to gauge the collective feel of the public on a host of different topics.
What unites the above is the unconventional use of data in the service of garnering novel insights, which is something everybody can get behind. We all have problems to solve, and accessing the right data from which to generate cogent insights is something that’s becoming ever more accessible.
Further, the world’s mass transition to the online realm, while fundamentally disrupting a wide variety of industries and disintermediating many incumbents, has also led to the emergence of new business models and avenues of monetization. Music and video are perfect cases in point.
These industries made the wholesale transition from regarding their products almost exclusively as physical objects to pure data. This has fundamentally reshuffled the power dynamics in production and distribution and has led to a situation where the streaming of music and video has allowed interested parties to gain a much more profound understanding of how audiences engage with and consume these products. These are insights the relevant industries were previously forced to gather in a much more imprecise and haphazard manner.
What are the components of an index?
At dxFeed, we’re particularly intrigued by the prospect of outside-the-box approaches to data utilization and index creation. This can be in the form of novel approaches to old problems, as in the almost holy-grail-like search among portfolio managers for uncorrelated assets or negatively correlated assets.
The construction of indices with this end in mind can offer all sorts of valuable diversification opportunities to investors or even the ability to hedge against unforeseen market events by having a benchmark that doesn’t tank when the S&P 500 crashes or even does the opposite. For trading venues, these types of indices promise the ability to offer markets a variety of new types of investment vehicles.
How can business leaders leverage indices?
To start with, determining the initial concept is all-important. What problem is the index solving for the business? Is it to be used for hedging purposes, as the source for derivatives to be built on top, or is it providing the business with a crucial information signal, such as the benchmarking of ESG performance?
With the idea down, make sure you are using a reliable source of data. Then comes management. What is the business’s strategy for maintaining/rebalancing the index? Again, it should aid in the business’s goals rather than unnecessarily draining resources.
Indices are defined by their constituent components and a set of parameters associated with those components. An index may measure the performance of the 42 largest IT companies in the U.S. One way to do this is to compose a portfolio of those companies’ shares and then compute its value in real time. In this example, the components are the 42 companies, and the parameters are the number of shares per company in the index plus a divisor term for scaling the index value.
Index parameters are computed at construction time to reflect a certain design goal. In our example, that would be a certain proportion of each company in the index, e.g., market capitalization-based weight. However, since the parameters are static, but the situation on the market is dynamic, the optimality property of the index parameters gradually deteriorates. To restore it, indices must be periodically rebalanced. The question for the index maintainers is how often to rebalance the index and what procedure to utilize.
The second aspect of index maintenance is the processing of corporate actions and other relevant events. Every asset class has a set of associated events that require adjusting the index configuration. Algorithmically, the processing of these events is simple, but implementing it may give you a headache. One of the most important components of this scheme is to catch the relevant market events.
Let’s imagine there are 123 AAPL shares in our index portfolio. If we compute the de-facto weight as the dollar value of the 123 shares in proportion to the dollar value of the other companies’ shares in our portfolio, it will fluctuate as the AAPL stock price changes. Suppose, according to our analysis, the optimal weight of AAPL is 30%, but the de-facto weight increases or decreases due to the stock price dynamics; then we’ll have to periodically change the amount of AAPL stocks in our portfolio to meet the desired goal—a procedure known as rebalancing.
The index should be fully trackable, so the audit should work perfectly; if something happens with the data center, another instance should come into play; if, for some reason, the constituent is stuck on one level, investigate the reason.