Tagged: Alternative Data

The Paradox of Alternative Data

“My God, they’re purple and green. Do fish really take these lures?” And he said, “Mister, I don’t sell to fish.

Informational edge can drive fantastic alpha while it lasts. This explains the increasing investment industry focus on non-traditional data sources, aka alternative data. If you are the first to acquire a new alternative data set, you might be able to develop insights no one has.

Yet once a lot of people are using it, it is less likely to drive alpha. It might be table stakes to not get screwed, or it might be already be instantaneously reflected in the price of assets.

Furthermore, most opportunities really hinge on a couple factors- more info isn’t always useful. That won’t stop the alternative data industry from doubling to $400 million by 2021, as a widely cited Tabb Group report predicts This is worth considering while one is caught up in an alternative data arms race.

Yet some data sets genuinely will provide an edge.

Perhaps a data set that no one else is looking at provides the edge you need. If a data set isn’t established as useful, the provider of that data will probably offer it cheaper in the early days of their business. There are so many alternative data providers out there, that marketing strategy is important for startups.

Ironically, the provider can charge higher price once word about its value gets out among investors.. So later adopters might pay more for an edge that is already gone.

Of course eventually someone will put all the data online for free and meta data of how investors use that alternative data can also be useful.

Now can I interest you in an alternative data feed that will make all your dreams come true?

See also: https://ockhamsnotebook.com/2016/09/18/the-hard-thing-about-finding-easy-things/

Beyond headline numbers

Everybody has access to Bloomberg and Google. Every global macro investor closely follows macro data out of every country. To gain an an edge, one must look beyond headline numbers, and find underutilized datasets.

This applies when finding countries, industries, and individual companies in which to invest. Any time you want to combine top down and bottom up insights, you need to get creative with finding the right data.

Schumpeter and Perez

Joseph Schumpeter pointed out  that aggregate figures “conceal more than they reveal”.

Relations between aggregates are

“entirely inadequate to teach us anything about the nature of the processes which shape their variations, aggregative theories of the business cycle must be inadequate too…”

In Technological Revolutions and Financial Capital, ( Carlota Perez emphasizes that new technological paradigms can only be analyzed by looking closely at inner workings of an economy. Within the same country, or industry some subsectors will grow at astonishingly high rates, while others decline. Perez’s framework is valuable to analyzing times of great technological change, which is basically anytime. Examples she uses include the first British Industrial Revolution( the age of Steam and Railways, the Age of steel, electricity, and heavy engineering, age of oil, automobile and mass production.

Global Macro Cycles

Top line numbers such as GDP or earnings could deceive an analyst, especially when looking at a new market.

Valuation and pricing

“People living through the period of paradigm transition experience real uncertainty as to the ‘right’ price of things(including that of stocks, of course).”

Extreme jumps in productivity change relative price structures in the economy. “The change in relative price structure is radical and centrifugal. Money buying electronics and telecommunications today does not have the same value as money buying furniture or automobiles.” Therefore, looking at inflation or deflation in aggregate is deceptive. Many years after Perez’ book, this now exacerbated by the Amazon effect. To some effect this may impact valuation in some industries.

Research methods

Long term aggregate data, spanning multiple periods of technological change are senseless. This goes for GDP, corporate earnings etc. Yet disaggregated stats are rarely available(except during more stable phases), as Perez points out.

The internet has provided more opportunities to find disaggregated, unique, underutilized datasets. Often this means poking around on weird regulatory websites, and following up on footnotes to academic papers.

This process might be about to get a lot easier.

Google launched a new dataset search engine. I’m excited to see how its impact snowballs as more datasets are added. Although intended for journalists, it is likely to be a valuable tool for investors seeking differentiated alpha.

Of course that means today’s edge, will be tomorrow’s table stakes.

See also:  The hard thing about finding easy things

Cialdini’s Law of Data Smog

In the classic book Influence, Robert Cialdini outlines six principles(reciprocation, liking, social proof, authority, scarcity and consistency) that represent psychological universals in persuasion. In general, all these persuasive techniques exploit people’s heuristics. The proliferation of information in this digital age means people need to rely more on heuristics than ever before. This has broad and deep consequences.

Because technology can evolve much faster than we can, our natural capacity to process information is likely to be increasingly inadequate to handle the surfeit of change, choice, and challenge that is characteristic of modern life. More and more frequently we will find ourselves in the position of the lower animals, – with a mental apparatus that is unequipped to deal thoroughly with the intricacy and richness of the outside environment. “ ­ Influence

One of the thirteen laws of Data Smog outlined by David Shena is Cialdini’s Law: Though culture moves much more swiftly than evolution, it cannot change the pace of evolution. This of course leads to a dangerous situation, where the unwary can be tricked into making dumb decisions. Worse yet are the broader societal consequences.

In the electronic age, a good lie well-told can zip around th world and back in a matter of seconds while the truth is trapped, buried under a filing cabinet full of statistics.”Data Smog

Cialdini’s follow up book, Pre-Suasion discusses how persuasiveness can be enhanced by carefully crafting what is done and said before making a request. Information overload also makes people more susceptible to the “presuasive“ techniques:

“…(1)what is more accessible in the mind becomes more probable in action, and (2)accessibility is influenced by the informational cues around us, and our raw associations to them….

In addition to its time-challenged character, other aspects of modern life undermine our ability (and motivation) to think in a fully reasoned way about even important decisions. The sheer amount of information today can be overwhelming- its complexity befuddling, its relentlessness depleting, its range distracting, and its prospects agitating. Couple those culprits with with the concentration-disrupting alerts of devices nearly everyone now carries to deliver that input, and careful assessments role as a ready decision-making corrective becomes sorely diminished. Thus a communicator who channels attention to a particular concept in order to heighten audience receptivity to a forthcoming message- via the focus-based, automatic, crudely associative mechanisms of pre-suasion- won’t have to worry much about the tactics being defeated by deliberation. The calvary of deep analysis will rarely arrive to reverse the outcome because it will rarely be summoned.”Pre-Suasion

Summon the calvalry of deep analysis

What can one do about this?” Hueristics are necessary to function in the modern world, but they must be examined from time to time. The calvary of deep analysis must be summoned for big decisions. Cialdini also recommends forceful counters assault. Recognize the tricks being employed are often enough to blunt their force, but in other cases it may be necessary to aggressively fight against the tricks. These books are a great place to start.