Trust and Privacy
Welcome to the first issue of Constituent Voices
At Omidyar Network, we start from a fundamental belief: People are inherently good and capable, but they often lack opportunity. We believe if we invest in people, through opportunity, they will create positive returns for themselves, their families, and the world at large.
But too often the voices of those at the far end of our interventions — the people we hope to empower — are not heard by the actors driving capital, policy, and resources for their benefit. Conversations center instead around entrepreneurs, capital markets, cost-benefit, or other top-down considerations.
We believe it’s essential to listen directly to the perspectives of the people we are working to serve. This series will share insights from those who engage with our own portfolio companies as well as those in the world more broadly. The goal is to help ground the activities of investors, philanthropists, and social change actors in the views of the actual people whom we all aim to empower, and to generate dialogue that can uncover changing trends to drive more effective outcomes.
In this first issue, Constituent Voices: Trust and Privacy, we present the findings from our global survey of 60,000 individuals’ perspectives on whom they trust — government or private companies — with their personal data. The survey captures data from 60 countries, and reveals a surprising degree of mistrust around the world that companies and governments alike will need to address if the digital economy is to succeed in the long run.
A Global Crisis of Trust
Research commissioned by Omidyar Network reveals the extent of the global “Data-Trust gap”, showing how pervasive the lack of trust is for individuals who have to share their personal data with either governments or companies.
It echoes the Edelman Trust Barometer, which has stark findings showing that the general population’s trust in four key institutions — business, government, NGOs, and the media — has declined to its worst position since they started measuring in 2012. This massive “Trust Gap” is affecting everything from social cohesion to the way we construct our political identities and interpret shared reality. This matters a great deal for any organization working to engage constituents or serve consumers.
At Omidyar Network, we are deeply concerned by this trend and committed to battling it, given that both individual empowerment and responsive institutions — two pillars of our impact framework — are greatly impeded in the absence of trust. As we outline in our recent report on the digital divide in India, we know our investments in services and applications will not improve people’s lives if there aren’t compelling value propositions riding on the rails of trust.
Across our portfolio, there are several areas where constituents’ trust is a key indicator of progress towards the impact goals driving our work. For instance, companies we invest in are using aggregated personal data to offer what should be empowering and inclusive services such as digital financial products or identity technologies. At the same time, we also work to encourage open, transparent government that delivers responsive services and independent media that holds the powerful to account, with the aim of increasing citizens’ trust in their societies. All this work involves engagement with, and the use of, personal data that citizens provide — directly or indirectly — to business and government.
To explore the current state of trust with these perspectives in mind, we commissioned a survey that obtained over 60,000 responses from people in 60 nations around the world. The survey asked respondents to rank which types of personal information they considered the most and least private — e.g. job and income histories versus web-browsing histories — results of which are shown above. Overall, there was a fairly balanced view that all of the data types asked were more or less equally private. Then, each respondent was randomly shown one of three types of data — location, web/app activity, and content of online conversations — and asked to indicate whether (or not) they trusted business and/or government to handle that information.¹ Throughout the brief, we refer to responses that indicate trust with either government or private companies to handle personal data as “Data-Trust,” and share our findings first across the whole sample then cut by demographics and finally by region.
Our Survey Found an Astounding Data-Trust Gap Across the Globe
The Scale of Distrust is Enormous
The majority of respondents across 60 countries trust neither the government nor business with key pieces of data that, in many cases, are already being collected. Across the sample, nearly three in five respondents, for instance, did not trust either party with their location data — which underlies many essential services — and about two in three said the same for the content of their phone or online conversations.
Distrust is Extremely Pervasive
While there were regional variations in the relative levels of distrust toward government versus business, distrust remains high ranging from 43 to 69 percentage points across regions for location data, as an example.
Trust Appears Binary
Across the sample, most respondents do not distinguish strongly between the private or the public sector when it comes to trust. In 58 of 60 countries surveyed, the most popular response with web/app activity data is trust in “neither,” and the next most popular response is “both.”
Data-Trust Varies by Gender, Education, or National Income
Data-Trust is Consistently Lower Among Women
Across the globe, the women in almost every country indicated lower levels of Data-Trust relative to their compatriot men, regardless of whether the absolute level of trust in the country was high or low. Figure 1 shows the differential in trust between men and women across countries in the sample. In Poland, for example, 35% of men exhibit data trust, while only 18% of women do — this gives the differential of 17 percentage points shown in the first data point below. Across the sample of sixty countries, the average differential is 7 percentage points, with almost two-thirds of countries revealing a “trust gap” of 5 percentage points or greater. Only in Bangladesh and Ethiopia did women indicate more trust than men with any notable margin of difference.
Data-Trust Drops Post Primary Education
Across the global sample, trust drops sharply by 18 percentage points between those with primary education and those with secondary education or higher. As Figure 2 shows, an average of 58% of respondents with a primary school education reported Data-Trust, while only 40% of individuals with advanced degrees indicated data trust.
Individuals in Nations with Higher Income Exhibit Less Data-Trust
Finally, the data shows a trend that trust diminishes as national income increases. Readers should note this does not reflect the income level of individual respondents — rather the income level of the countries in which they reside.
Regional Dynamics Reveal Surprising Findings
Regional Map Raises Questions on Neighbors’ Influence
Mapped at the country level, the variation in Data-Trust is stark in unexpected ways. Firstly, the degree of variation is significant: only 21% of respondents in Slovakia exhibit Data-Trust, versus 70% in the Philippines. Secondly, there appear to be contrasting Data-Trust levels between regional powers and their neighbors. Russia, with strong central government and known data surveillance, has low Data-Trust at 35% (averaged across personal data types); but this is not as low as in surrounding countries. Ukraine, Slovakia, Serbia, Poland, Hungary, and Czech Republic have on average 11% points lower Data-Trust. These are 6 of the 7 least trusting countries in our sample. In China, another country with a strong sphere of influence, residents report relatively higher Data-Trust at 46% than neighboring Hong Kong (37%), and Taiwan (33%).
As a Nation’s Fragility Increases, So Does…Trust?
Our initial hypothesis in exploring this analysis was that the level of political fragility² and extent of Data-Trust would move in opposite directions; surprisingly, they seem to move together. For personal web & app online activity, for example, the average level of Data-Trust in most fragile nations is 20 percentage points higher than it is in nations considered stable.
Formerly Authoritarian Regimes Indicate Lowest Data-Trust³
Finally, we compare the indications of Data-Trust from respondents living under autocratic, democratic, and transitioning regimes. The average level of Data-Trust was highest in nations that had never been autocratic, but only 3% points higher than the average of countries in current authoritarian regimes. Countries with formerly authoritarian governments revealed the lowest Data-Trust, where respondents expressed, on average, 16% points lower Data-Trust than those in current autocracies.
Personal Data is Increasingly the Most Valuable Resource — We Must Better Understand How People Feel About Its Use
Such mistrust in how personal data is handled is worrying. Unlike five years ago, the world’s top five listed companies all monetize personal data, accruing concentrated power and profits dependent on a contract with consumers that has so far been asymmetric. This is even more the case in emerging economies, where regulations further lag the innovation in services that utilize personal data. Mobile money services such as M-PESA, pioneered in Kenya and now used in several developing nations, give a mobile phone operator and financial institutions an immediate and full view into an individual’s transactions. Similarly, companies such as our investees Cignifi and Lenddo assess creditworthiness using data analytics tools on calling records or online social networks, helping clients provide loans and bring underserved consumers into the formal financial system. As outlined in Big Data, Small Credit, the sustainability of these offers must be enhanced by building strong relationships of trust with consumers. Similarly, governments rely on a relationship of trust with their citizens for their authority (at least in democratic regimes). These survey findings reveal there is much work to be done to build and maintain that trust for both the private and public sectors.
More Than Conclusions, This Survey Raises Questions
Running a quick comparison between our findings here and those from other sources, we find some confirmation and some conflict. The Data-Trust differential among women in India revealed in our global survey — 15% points lower than men in India — resonates with findings we surfaced in “Currency of Trust,” an investigation into digital financial services in India. Many women interviewed for that research didn’t have smartphones to share personal data, and those that did were often “dark users” — not accessing the internet at all, perhaps due to this lack of Data-Trust. By contrast, our global survey on Data-Trust by education level moves in the opposite direction of Edelman Trust Barometer’s measure of trust in institutions. While both uncover a notable trust gap, Edelman finds what we might call a positive correlation: that those with less education exhibit less trust in institutions than those with (college or) higher education. We, however, found a negative correlation: those with less education exhibit more Data-Trust. It is clear there is more to uncover in further research about trust, and we hope this data brief inspires more constituent research to understand these trends more deeply. In particular, this survey cannot tell us why these trends have arisen, and the questions that surface for further research are grounded in the reasons for our findings above:
- What has caused this Data-Trust gap and what can be done to rebuild trust with constituents?
- Why are women consistently less trusting than men of government and companies with their personal data?
- How does Data-Trust differ from trust in institutions, and what would explain the opposing findings across the education ladder?
- Why does higher education or living in a nation with higher income — things that one might interpret as corresponding to more opportunity — coincide with lower Data-Trust? And what does this mean about aspirations for societal and technological progress?
- Why isn’t the level of trust lower for people in countries marked by political, economic and social instability, or for those living in authoritarian states?
- What is it about moving away from an autocracy to democracy that markedly decreases trust? What does it take for these societies to recover trust to the level of long-time democratic nations?
Thank you for reading — look out for issue 2 in early 2018.
 Given there was fairly balanced perception of which types of data were most private, we present the survey results throughout this brief with respect to different types of data or sometimes averaged across the different data types surveyed, to give the reader different perspectives. In general, there was not significant variation by type of personal data asked.
 Based on Polity’s 2015 State Fragility Index (SFI), which scores countries based on their Security, Political, Economic and Social systems, to estimate a country’s “capacity to manage conflict; make and implement public policy; and deliver essential services, and its systemic resilience… responding effectively to challenges and crises, and sustaining progressive development”.
 For this analysis, current regime type in each of the 60 nations in the survey was defined according to Polity’s classification in its Dec 2015 SFI regime coding to arrive at democratic (what Polity deems institutionalized democracy or un-institutionalized/weak democracy), or authoritarian (Polity’s autocratic regime, or state failure — collapse of central authority). Historic regimes were classified using data from Economist Intelligence Unit’s Democracy Index, CIA World Factbook, and other sources, examining the past 100 years (or, for former colonies, since independence). Note: Some categorizations may be subjective as some countries can be considered on the cusp of sliding in or out of authoritarian regimes.