27/09/2021

On the demand side, AI to learn needs to ‘drink’ from vast, diverse data pools collected from multiple sources. On the supply side, many companies also are looking at how to ‘monetise’ data for uses outside their own business. We are heading into a world of ‘data as a commodity’, but what would a data trading environment look like?

The World Economic Forum has been working up a model of ‘data exchanges’ operated by data marketplace service providers (DMSPs).

WEF has observed that, relative to the vast amount of data that exists in the world, the amount that is exchanged between the parties that control it remains small. The market for processing and analysing data lacks depth for a variety of complex reasons, but in major part because the public remains concerned about and solely focused on data privacy. Simply put, people feel that they lack sufficient control over their data.

The WEF sees data exchanges as addressing two hurdles to a trading data: they replace personal trust in bilateral relationships with systemic trust through rules, enforcement mechanisms, transparency procedures, etc; and given the problems of valuing data, they create a ‘price discovery’ mechanism by minimizing information asymmetry and expanding the pool of buyers and sellers.

A data exchange is depicted below:

The functions of the DMSP would be:

  • Providing a settlement function between buyers and sellers of data: as the balance between convenience and price matters, reducing the cost of payments (transaction costs) is essential.
  • Screening of market participants: DMSPs should audit sellers and buyers based on specific eligibility requirements e.g. DMSPs should assess whether a data seller has a system to properly obtain the consent of every original data provider.
  • Ensuring quality of data: through uniform formats and vocabulary, and the use of quality assessment criteria for traded data.
  • Ensuring an informed market: as with stock exchanges, the DMSP should ‘help make the market’ by disclosing final transaction prices and other information about traded data.
  • Providing means of dispute resolution and enforcement, such as through trading suspensions.

What constitutes a data exchange in in the eye of the beholder

The term data exchange is one of those ‘spray-on’ terms that is applied to all manner of data initiatives.

Way back in 2014, the Federal Department of Social Services set up a ‘data exchange’ for organisations which receive grants from the DSS. The data exchange has over 3,000 registered organisations and 15,000 individual users. However, at its heart, this is mainly a standardised monitoring and reporting tool for grant organisations to the DSS. However, there is functionality which allows individual organisations compare their performance against like organisations and for use in managing their own organisation.

The combination of a traditional trading platform and data does not a data exchange make. The London Stock Exchange recently acquired the financial information provider Refinitiv for $27bn. LSE has built itself into a major information provider, not only financial information but also economic data, shipping trackers, satellite-based imagery. But there is no sign yet that the LSE will apply its trading expertise to build a trading market in data.

Hong Kong’s and Japan’s regulators are implementing ‘open banking’ through use of a data exchange rather than by regulatory mechanisms such as the consumer data right used here in Australia. The Hong Kong Monetary Authority will build a Commercial Data Exchange to which all banks will connect and through which they can exchange information. The HKMA says that the CDI means that:

“with customer consent, banks will gain access to a substantial body of data, including merchant point-of-sale information so banks can forecast a businesses’ future cash flows, identify cashflow patterns, understand counterparty risks, and make loans without having to ask for collateral. Merchants that outperform projected sales can then begin to improve their credit standing.”

While the European and Australian CDR initiatives are primarily driven by consumer rights, the HKMA initiative, while emphasising consumer portability between banks, also seems to be as much driven by facilitating the banks building new use cases for data exchanged between each other.

The Australian AgriFood Data Exchange has recently been announced by a consortium of leading agrifood stakeholders including government, industry and research bodies, backed by $4 million in government and industry funding. Farmers are concerned that data about their farms and farming operations (which they regard as ‘our data’) is being ‘harvested’ by digitally enabled equipment such as tractors and harvesters for the benefit of the equipment vendors, not the farmer. The digital platform will enable:

  • The permissioned exchange of data between AgriFood industry participants;
  • Timely access to information that supports decision making for the AgriFood value chain;
  • Release management capacity;
  • standardisation and consistency of industry data assets;
  • the capacity to adapt, incorporating new use cases for data exchange that deliver value and support resilience of AgriFood value chain participants’
  • increased transparency of AgriFood industry data to support multiple use cases (e.g. regulatory compliance, collaboration between public & private data sets).

The AADE depicts its scope and role as follows:

Given the diversity of agricultural sectors, the number of participants and the planned reach of the project, hopefully it will not be a case of ‘boy swallows universe’.

Closer to the WEF’s vision of a data exchange is One Creation. It says that:

“we're a Digital Rights Exchange that is location agnostic. This means that your data remains where it is. We enforce the Digital Rights of data so that you know who the data owner is, who is using the data, and when they can see it. Most importantly, you are able to maintain full visibility of your data lineage”

Issues to be tackled in a data exchange

Experience shows that a trading system can be developed for just about anything, although not always sustainably (ask Eron’s shareholders). But as the WEF points out, there are unique challenges to configuring data as a tradeable commodity.

Data is infinitely copyable and shareable, making it a poor fit with traditional ownership rights that assume exclusivity. The thing that is being sold on an exchange is likely to be a more intangible thing called “access”, rather than outright ownership, raising the question of how access rights are defined and secured.

The ability to resell is an important part of price discovery and building critical market mass. If data can be on sold, how do you create ‘title’ that can be passed on? Do you need to prevent value dilution by the first buyer replicating the data for sale to multiple parties (i.e. A form of speculation)? If data is purchased and combined with other data, how is the ‘new’ property recognised for on-sale? Encryption technology provides part of the answer, but a legal framework is required to express the ‘rights’ which the encryption protects and which can be sold on to others.

How will price function? The WEF discusses pricing in terms of stabilisation measures, such as price floors or caps, to create stability during a start-up phase. But the more fundamental issue is how a ‘common currency’ or measure of value can be established in the first palce: each megabyte of data is not worth the same because the data itself is different.

So, may be your data won’t be hitting the trading room floor any time soon, but the pressure to develop a more systemic model to unlock more value from data will continue to build.

 

Read more: Developing a Responsible and Well-designed Governance Structure for Data Marketplaces

""