Who are you?
I am Gábor Komáromi, from Vienna, Austria, and I am passionate about financial markets and technology.
Can you tell us about your career in Fixed Income?
I have worked in investment banking, starting off in derivatives sales and then moving to the trading side. At Deutsche Bank I was doing derivatives sales, at Credit Suisse, I was a derivatives trader, which included market-making and proprietary trading. After the financial crisis, I changed to the buy-side, working as a portfolio manager for the largest Austrian institutional investor and most recently I worked for an asset manager as Head of Portfolio Management. I am now the Head of Fixed Income Analytics at RavenPack.
What do you do at RavenPack?
As the Head of Fixed Income Analytics, I work to improve RavenPack's capabilities in this very important asset class. Currently, we have challenging times in Fixed Income as interest rates are in many places at record low levels and it is a challenge for asset managers to generate returns. "Sliding down the credit curve", which means investing in issuers of lower credit quality, is one of the solutions with obvious risks - generating Alpha with alternative data is another. This is where RavenPack comes into play. Another exciting area that I am working on is ESG, not a new trend by any means, but in many cases, companies have only recently moved from "ticking the boxes" to really doing something.
"Interest rates are in many places at record low levels and it is a challenge for asset managers to generate returns. Sliding down the credit curve", which means investing in issuers of lower credit quality, is one of the solutions with obvious risks -, generating Alpha with alternative data is another.” Gábor Komáromi, Head of Fixed Income, RavenPack.
What made you think there was a value in news sentiment data?
It has been clear for a long time that sentiment is an important aspect of financial markets. I remember a cartoon that said something like “Today stock prices fell sharply on fears that stock prices might fall sharply.” Still, there wasn’t an indicator for sentiment except for the stock price itself. Many years ago, trading strategies that used the price as an indicator for future price moves worked quite well, but these momentum strategies have become overcrowded since An indicator measuring the sentiment directly from the news, therefore, would make a lot of sense.
When did you first hear about RavenPack?
In 2014. The year started with expectations that the “worst was behind us” with respect to the low interest rate environment but, as more often than not is the case in financial markets, it turned out to be different.
Due to a slowing of the Chinese economy and an increased oil production in the US, oil prices started to decline markedly and with it inflation expectations. This led to lower interest rates, 10-year US Treasury yields dropped from about 3 percent to below 2.5 percent by the end of the year.
Clearly, we had to start looking for yield somewhere. At that time I was working for the Vienna Insurance Group, which had a very conservative investment strategy, but the asset management department had a strong and perspicacious leadership. This encouraged me to look for innovative approaches to asset management.
What other data providers did you trial?
Over the last few years I was looking at other news sentiment providers: companies specializing in search engines for analysts, ESG focused analytics, and more. Most had nice websites and colorful factsheets but this is where the similarities ended, the value add differed immensely.
Why did you pick RavenPack over the competition?
It is not easy to evaluate alternative data properly during a trial period, especially if you are not full-time committed to the task. Getting a handle on the product and then actually building models to test takes a while and I always tried to do some real out-of-sample testing - it is just so easy to get caught by one of many biases that are out there.
The quality of the people was an important aspect for me. I started to make notes about the timing of the response to questions, evaluated the introductory calls – do they actually listen to what I need or is it the usual dull run-of-the-mill presentation? Here, RavenPack stood out, with very good sales and clearly the best client support.
“Do they actually listen to what I need or is it the usual dull run-of-the-mill presentation? Here, RavenPack stood out, with very good sales and clearly the best client support.” Says Komáromi.
Do you think alternative data lives up to the hype?
The real hype is to put AI onto everything, from refrigerators to brewing beer, and this is silly. Artificial intelligence is nothing new and we already had a boom in the 80s, after that, things quietened down and now marketing departments are hopping on the bandwagon. Alternative data is not hype and nothing new; it has been used by “smart money” already for years. Recently the use of alternative data has entered the mainstream and this has led to many articles in the financial press, for good reason.
This is not only true for applications in financial markets. Recently the Canadian Chief Statistician, Anil Arora said that traditional ways of gathering information, “are no longer sufficient to accurately measure Canada’s economy and societal changes.”
How is it useful to fixed-income investors?
Some alternative data has little value for investing, like Tweets, some have lost its alpha like the often-quoted satellite images of parking lots, other alternative data sets are very valuable.
News sentiment is one of the most useful types of alternative data. In fixed income, it is applied in a similar fashion to the way it is used for trading equities, which is to get insights into what is going on with a certain company.
Companies with a worsening sentiment tend to underperform companies with improving sentiment. But I have also found value in watching the sentiment of sovereigns to help me decide when to buy or sell its bonds, for example, the BTPS are often driven by sentiment around the political situation in Italy.
Sentiment data can also be useful for detecting signs of an upcoming change of direction in central bank rates, especially in 2019, when it was interesting to watch how the Fed changed direction given the pressure from politics. The platform is also useful for monitoring portfolios. Now I sound like one of our factsheets, hahaha.
What is your outlook for FI markets in 2020?
"Lower for longer”, sustainability continuing its path into the mainstream, and I expect not only hedge funds but also discretionary bond investors to use more and more alternative data to manage risks and generate alpha.
Alternative Data Research on Fixed Income
Research showing how to use Sentiment and Event Data for a variety of Fixed Income instruments is available here.