We continue our quest for unique and orthogonal sources of information. In the New Year of 2018, we shift our attention to an interesting textual database – S&P Capital IQ’s Call Transcripts. We showcase how our suite of proprietary and sophisticated NLP (Natural Language Processing), machine learning (e.g., deep learning via Convolutional Neutral Networks), linguistic, and psychological research are used to extract salient information from conference calls.
Search Coverage List, Models & Reports
Search Results1-10 out of 39
Introducing a New Research Series – Portfolio Compass. The focus of this new series is to deliver actionable investment ideas that investors can implement. The targeted audience is primarily non-quantitative specialists and quantamental investors. The Portfolio Compass series will be published in the 2nd/3rd week of each month.
We continue to explore the use of Alternative Big Data on the global macro front. In this research, we introduce Quandl, a premier platform for unique data from over 500 providers. In particular, we study the Dodge US Construction Activity (UCA) database, probably the largest and most reputable provider for US construction industry data. We focus on a sub-dataset called Project Life Cycle (PLC), which tracks 14 stages (planning, bidding, permit stages, etc.) of construction projects.
We conduct a thorough study on asset return seasonality, especially the January effect. The January effect is particularly strong among equity styles. In particular, in almost all countries, growth outperforms value in terms of earnings, cash flow and dividend multiples, with the exception of price-to-book where lower P/B tends to outperform. Price momentum is likely to suffer significant losses, while mean-reversal strategies outperform. Small-cap and illiquid stocks also perform strongly at the beginning of the year. Sector-wise, information technology stocks rally in January, while the consumer staples sector underperforms.
Last month on November 29, the market exhibited a significant sector and style rotation event. Among sectors, information technology sold off broadly relative to telecom services which rallied the strongest. Similarly, there was a sharp sell-off of price momentum and a significant rally across value strategies. The one-day shocks on November 29 exceeded two- or three-standard deviation events for many sectors/styles. However, we also find that rare events are far from uncommon in finance.
Opportunities and Challenges in China A Shares. With more than 1,000 large-cap stocks, robust trading liquidity, and plentiful alpha, China’s domestic equity market (aka China A) offers great investment opportunities. However, there are also unique features and challenges in China, ranging across data and vendor selection, factor construction, to market microstructure issues.
Wolfe Research’s QES team held its inaugural global quantitative and macro investing conference on November 14th and 15th in New York City. Nearly 500 buy-side investment professionals, academics, investment researchers, traditional as well as alternative data vendors, data scientist, and technologists attended the event.
Our proprietary parking lot traffic indicator based on satellite imagery is signaling a potential negative sales surprise for DKS, due to negative growth in quarterly YoY traffic. However, the latest measures for HD and LOW show that quarterly YoY traffic growth has been positive, signaling that a negative sales surprise is unlikely. We find that satellite imagery data tends to be more useful for home improvement retail and specialty stores possibly due to their higher in-store purchasing conversion rates and higher customer traffic during the mid-morning/early-afternoon hours when satellite imagery is captured.
October witnessed strong decoupling of cross-asset correlations. Currently, pairwise stock correlation, factor correlation, multi-asset and currency correlation all stand at moderately low levels. Our systematic global macro and stock selection models perform well in such an environment. In fact, our TAA, sector/industry rotation, and style timing models all delivered superior returns last month. Our bullish views on crude oil, EM equity, and US large-cap equity (and negative views on global REITs, US small-cap equity, and US high yield bonds) also paid off. Similarly, our predictions on the info tech (the best performing sector in October) and telecom services (the worst performing sector last month) were fairly accurate. Lastly, our style forecast of momentum/growth over size (small cap), illiquidity, and book-tomarket was also spot-on.
Accessing unique and high quality information is a tremendous
competitive advantage in today’s active management. The challenge of using alternative data is information
overload – there are simply too many data vendors and content to digest in a timely manner. Helping our
clients to vet and select the most interesting data vendors and datasets is an important part of our services.
- 1 of 4
- next →