Big data, or big brother?
24 July 2018
David Lewis of FIS suggests there is little doubt that big data and the analysis of it can help society prosper, but like any tool, it has to be used in the right way
Image: Shutterstock
A wise man once told me that profit is not where other people are. The comment was made about data highlighting where the best rates could be found, for what asset to be lent against what collateral and under what terms. The point was that the trade structure in question was becoming crowded and that margins would clearly contract as too many people climbed aboard that bandwagon. When looking at trading signals and data, it can often be seen that the real gains are made by those that are early to the trade, with those joining later collecting the crumbs, relatively speaking. Carillion PLC, the UK construction giant that began to implode one year ago this month is a prime example. Before the ‘shock’ profit warning hit on 7 July, more than 70 percent of the shares available in the market had already been borrowed, and at remarkably low fee rates. Those that joined the deal later still made substantial gains, but paid a heavier price for the privilege.
Big data and the advancing capabilities of data analytics have certainly brought benefits to financial markets and the wider society in general. From intraday transparency in our own securities finance markets to the huge power of social media and the way it affects our lives—the pace and impact of the change we are experiencing would have been unthinkable only 10 years ago, let alone looking back a generation. At FIS, we are putting a great deal of effort into predictive analytics. Our new software is going beyond informing users what trades have been undertaken in the last few minutes. Now, it can also indicate where rates and volumes are likely be later in the day and going forward. Based on hundreds of millions of trades and over 13 years of historical data online, using past behaviours and results, our systems are predicting—within degrees of statistical confidence—where the rates will go next.
Looking at our trading systems, we are monitoring usage patterns, in both the front and back office, identifying repeated patterns to shorten work flows, suggesting corrections and improving accuracy. This kind of predictive analytics is not too far from the ‘if you bought that, you will like this’ features familiar to users of online music streaming sites or internet retailers. Moore’s Law, or, in fact, extensions and variations of the original law, which stated that the number of transistors on a chip roughly doubles every two years, is often quoted as an indicator of the speed at which computing capability is accelerating. Without this, the ability to cheaply and quickly process the vast quantities of data that now shape our lives would not be possible. This processing is going on around us continually, deciding what our best route to our next meeting might be, what credit card deal would be best suited to our financial status, and even who our best life partner might be. But where might the benefit stop, and the gains begin to narrow? When does the data and the algorithms behind it cease to be our servant, and instead become the master?
There are many amusing anecdotes about products and services being offered to people before they even knew they needed them, and much has been written about apps ‘listening’ to users without their apparent knowledge. But what about real world impacts? For example, debt default is a common outcome of relationship failure. It has been suggested that credit limits have been reduced for clients that exhibit behavioural and spending patterns associated with relationship stress. Many observers have noted that payday loan-style products are targeted at those who are already under financial stress, which is logical assuming those who are more financially stable would not consider them. But is this an example of predictive analytics restricting opportunities, or really helping humans make better decisions?
There is a clear difference between the retail user of financial services and the professional, but both need to be taken care of with regard to their interactions with data and systems. Arguably, the retail investor deserves greater protection to balance out their real or perceived lack of understanding of the financial products that they may be buying.
However, it is reassuring to see that the chair of the Financial Conduct Authority (FCA) and Payment System Regulator, Charles Randall, has been recently quoted reminding the market that the rules under which the FCA operates include the general principle that consumers should take responsibility for their decisions.
Randall made this statement with the associated caveat that fair disclosure is required if this approach is to be supported. Extrapolating that into the world of predictive analytics and big data, we need to be sure that the suggestions made to us, whether it be the quickest route to your next meeting or your best mortgage offer, are done with not only our best interests at the forefront, but also with fair disclosure of the options available to us. Multiple agent lenders in our client base tell us that most of their trades are now completed without being touched by human hand. Indeed,
FIS technology for price indications, as well as best execution benchmarking, is contributing to that progression. Much of this is relatively simple mathematics and engineering, with little opportunity for negative outcomes, financial or otherwise—but when we start to look at the more bilateral space, what potential impacts are on the horizon there?
Capital requirements and regulations are pushing trades toward the more creditworthy organisations, allowing counterparts to make the best use of scarce capital. However, this has the negative effects of creating new single points of failure, or to use the regulatory term, affect globally systemically important banks, while narrowing credit diversity. There is little doubt that big data and the analysis of it for patterns can help society prosper, but like any tool, it has to be used in the right way.
It has the potential to widen horizons, improve efficiency and help us to make better decisions; but it also has the potential to narrow our focus and limit our flexibility, ultimately removing the ability to make responsible choices, which, just sometimes, may not be where everyone else is.
Big data and the advancing capabilities of data analytics have certainly brought benefits to financial markets and the wider society in general. From intraday transparency in our own securities finance markets to the huge power of social media and the way it affects our lives—the pace and impact of the change we are experiencing would have been unthinkable only 10 years ago, let alone looking back a generation. At FIS, we are putting a great deal of effort into predictive analytics. Our new software is going beyond informing users what trades have been undertaken in the last few minutes. Now, it can also indicate where rates and volumes are likely be later in the day and going forward. Based on hundreds of millions of trades and over 13 years of historical data online, using past behaviours and results, our systems are predicting—within degrees of statistical confidence—where the rates will go next.
Looking at our trading systems, we are monitoring usage patterns, in both the front and back office, identifying repeated patterns to shorten work flows, suggesting corrections and improving accuracy. This kind of predictive analytics is not too far from the ‘if you bought that, you will like this’ features familiar to users of online music streaming sites or internet retailers. Moore’s Law, or, in fact, extensions and variations of the original law, which stated that the number of transistors on a chip roughly doubles every two years, is often quoted as an indicator of the speed at which computing capability is accelerating. Without this, the ability to cheaply and quickly process the vast quantities of data that now shape our lives would not be possible. This processing is going on around us continually, deciding what our best route to our next meeting might be, what credit card deal would be best suited to our financial status, and even who our best life partner might be. But where might the benefit stop, and the gains begin to narrow? When does the data and the algorithms behind it cease to be our servant, and instead become the master?
There are many amusing anecdotes about products and services being offered to people before they even knew they needed them, and much has been written about apps ‘listening’ to users without their apparent knowledge. But what about real world impacts? For example, debt default is a common outcome of relationship failure. It has been suggested that credit limits have been reduced for clients that exhibit behavioural and spending patterns associated with relationship stress. Many observers have noted that payday loan-style products are targeted at those who are already under financial stress, which is logical assuming those who are more financially stable would not consider them. But is this an example of predictive analytics restricting opportunities, or really helping humans make better decisions?
There is a clear difference between the retail user of financial services and the professional, but both need to be taken care of with regard to their interactions with data and systems. Arguably, the retail investor deserves greater protection to balance out their real or perceived lack of understanding of the financial products that they may be buying.
However, it is reassuring to see that the chair of the Financial Conduct Authority (FCA) and Payment System Regulator, Charles Randall, has been recently quoted reminding the market that the rules under which the FCA operates include the general principle that consumers should take responsibility for their decisions.
Randall made this statement with the associated caveat that fair disclosure is required if this approach is to be supported. Extrapolating that into the world of predictive analytics and big data, we need to be sure that the suggestions made to us, whether it be the quickest route to your next meeting or your best mortgage offer, are done with not only our best interests at the forefront, but also with fair disclosure of the options available to us. Multiple agent lenders in our client base tell us that most of their trades are now completed without being touched by human hand. Indeed,
FIS technology for price indications, as well as best execution benchmarking, is contributing to that progression. Much of this is relatively simple mathematics and engineering, with little opportunity for negative outcomes, financial or otherwise—but when we start to look at the more bilateral space, what potential impacts are on the horizon there?
Capital requirements and regulations are pushing trades toward the more creditworthy organisations, allowing counterparts to make the best use of scarce capital. However, this has the negative effects of creating new single points of failure, or to use the regulatory term, affect globally systemically important banks, while narrowing credit diversity. There is little doubt that big data and the analysis of it for patterns can help society prosper, but like any tool, it has to be used in the right way.
It has the potential to widen horizons, improve efficiency and help us to make better decisions; but it also has the potential to narrow our focus and limit our flexibility, ultimately removing the ability to make responsible choices, which, just sometimes, may not be where everyone else is.
NO FEE, NO RISK
100% ON RETURNS If you invest in only one securities finance news source this year, make sure it is your free subscription to Securities Finance Times
100% ON RETURNS If you invest in only one securities finance news source this year, make sure it is your free subscription to Securities Finance Times
