Confidential data: How to do analysis when data is proprietary

Photo by: Casey Marshall / CC BY

In 2013, Barclays announced it was closing operations in Somalia. The exit of the last major bank in a country where local banks did not have the capacity to take over the market meant that a business decision became a development issue. Ultimately, markets found other ways to move money. But the decision was indicative of a continuing global trend of international banks consolidating away from perceived high-risk markets.

Exporters in developing countries are particularly affected due to their reliance on bank-intermediated trade finance. Trade-driven growth is a common development goal, but it requires trade finance to mitigate risk and enable cross-border payments — especially for nontraditional exporters including small and medium-sized enterprises.

Solutions to trade finance gaps can only be bluntly targeted due to lack of data. Even the Bank for International Settlements offers only an “interpretive characterization” of the size of the global market for trade finance.

Multilateral development banks including the Asian Development Bank and African Development Bank have introduced surveys to shed light on the sources of impediments to regional trade finance. This is a good start, but more needs to be done.  

Why is the data unavailable?

Understanding why data is not available is the first step to identifying an alternative. In the case of trade finance, data is unavailable for two reasons. One is confidentiality. For example, default rates are only shared in a highly aggregated form. We can see a disconnect in some regions between very low default rates and high rejection rates. But without country-level data, it is impossible to properly target capacity building or guarantee programs.

Sources: transaction default rate from International Chamber of Commerce Trade Register Report, 2015; rejection rates from ADB trade finance gap survey, 2015. View a large version of the image.

Another reason that data is unavailable is that collection can be expensive and time consuming. Banks’ information systems may not be harmonized internally, so culling data is difficult. Under reporting requirements for MDB guarantee programs, banks share the blunt number of funded transactions that go to SMEs. But a more informative breakdown would show whether SME lending increases are the result of new SMEs opening credit lines, or a static group of SMEs receiving additional credit volumes.

Fig 2. Number of ADB guarantee transactions by firm size (2010-2014)

Source: ADB Trade Finance Program

Surveys to estimate confidential data

In order to identify the causes of trade finance shortfalls at the country level, we have turned to anonymized surveys.

Trade finance has several characteristics that make this a reasonable strategy. First, a small number of global banks are responsible for much of global lending. BIS estimates that the banks in the International Chamber of Commerce trade register cover 25-30 percent of global market share. This offers a sense of the total coverage in any sample since randomization is not possible.

Second, banks may be willing to estimate data that cannot be shared. A widely used survey strategy is to request percentages rather than an exact value. This is a strategy the World Bank relied on in their look into the withdrawal from correspondent banking relationships.  

There is a third characteristic of the sector that is particularly helpful to check the robustness of responses — trade finance goes to companies. And companies are often more forthcoming about data than financial institutions. A survey of providers of trade finance can be matched with a survey of users to check if the data goes in the same direction and in a similar magnitude.

A global push for quantification

The difficulty MDBs face in responding to trade finance shortfalls is illustrative of a broader class of development issues where data confidentiality makes policy targeting difficult. Development banks have patched together information from surveys. This is a good first step.

A critical next step is a concerted global effort to quantify trade finance gaps and their development impact. MDBs already collaborate via the ICC annual trade finance survey. This could flow naturally into a multiregional exercise to collect data on trade finance mechanics. MDBs could collect data from regional banks and firms once every two to three years.

Barclays, along with many other global and regional banks, is now introducing digital solutions for emerging markets. Although it is not yet clear how this will impact exporters in underserved markets, more granular data on country-level trade finance impediments and impacts would allow faster and better analysis of shifting trade finance trends — and more precise targeting of solutions.

With potential to change the trajectory of crises, such as famines or the spread of diseases, the innovative use of data will drive a new era for global development. Throughout this monthlong Data Driven discussion, Devex and partners — the Agence Française de Développement, BroadReach, Chemonics and Johnson & Johnson — will explore how the data revolution is changing our approach to achieving development outcomes and reshaping the future of our industry. Help us drive the conversation forward by tagging #DataDriven and @devex.

The views in this opinion piece do not necessarily reflect Devex's editorial views.

About the author

  • Alisa DiCaprio

    Alisa DiCaprio is a trade economist in Asian Development Bank’s Economic Research and Regional Cooperation Department. She is currently on secondment to ADB Institute in Tokyo. Her research explores ways to make trade a more effective development tool. In particular, she highlights nontraditional trade topics including trade finance gaps, e-commerce, and fragile and landlocked states.