By Rado Lipuš, on 21 September 2017
Estimated reading time: 4 minutes
I was recently on a data panel at a quant conference in New York. Two things surprised me: the large number of family offices approaching me for information about alternative data and, in particular, the same people wanting to know which fund managers are using alternative data.
Given that probably half the attendees were from quant investment groups, I was expecting more interest from quant fund managers themselves.
It has become increasingly clear to me that most of the family office and fund allocators' existing managers are not currently using esoteric sources. Put simply, alternative data use cases are still not well understood.
Alternative dataset types:
Too often, the allocators' externally managed investment strategies rely on traditional mainstream data sources and approaches. The primary alpha inputs are often price data, which can be the sole factor for certain strategies. Multi-factored alpha approaches employ widely utilised (and, in my opinion, both over-utilised and over-crowded) risk factors such as momentum, size, value, growth and other factor risk premia.
The same underlying and crowded risk factor serves as a model input factor or proxy, for the short-term (one to several days) to mid-term (up to four weeks) to long-term (several months to multiple years).
Time and time again we see a large disparity of opinions on the usefulness of the time horizon for these risk factors.
For example, one manager might be convinced of the mid-term (weeks) alpha potential of a factor whereas another might completely disagree with this application, finding alpha in a very short time-frame, perhaps a few days. Clearly these use-case beliefs will depend on the level of individual experience of the manager and her portfolio management research skills. In addition, model assumptions and correlations with other factors used in the portfolio construction process will determine different time horizon applications for alternative data.
It is generally true that most investment pitches received by allocators are all too similar with regard to both factors used, and the underlying sources. Rarely do fund manager pitches offer any truly distinctive insights into data sources.
A friend and CIO of a London-based family office states that the majority of investment managers' pitches are very similar and the most frequently stated USP refers to the unique quantitative research skills and the pedigree and track record of the portfolio manager. Big data expertise and know-how is rarely mentioned and does not make it into presentations, on the whole.
Allocators have always been looking for truly non-correlated and orthogonal sources of alpha, among many other important fund selection criteria. The supply of new data sources has been growing and more digital data on economic activity than ever before is making its way into the investment industry. It could be argued that the supply of digital data is infinite.
At Neudata we vet alternative datasets on behalf of investment groups and classify these novel sources in multiple dimensions. One of the most basic categorisations is by 19 dataset types (see chart) and we use another 60 tags and ratings to make sense of often unusual data, its quality and investment applicability.
Last week we met with a portfolio manager at a very large Swiss investment firm. He mentioned that, for the first time, their largest client (an allocator) had recently asked them what their plans were for alternative data. The portfolio manager was very excited as his CIO was present in the meeting and he has now finally convinced his senior management to back his alternative data research initiatives.
The supply of new alternative data is undoubtedly growing and so is its usage by the investment community - see a recent FT article entitled Hedge funds see a gold rush in data mining. We are still at an early adoption stage, however, I strongly believe there are many first-mover advantages of alpha harvesting.
Investment managers face many obstacles to tackling alternative data sources. One in particular is successfully integrating big data into their existing investment processes.
We will offer further insights into the challenges of, and misconceptions about, alternative data in future blogs.
Feel free to drop me a line at: [email protected]
Find out more about Neudata and keep up-to-date by signing up to our email newsletter.