BY SOPHIE JIAYUAN WANG
Sophie Wang graduated from the International Development program in May 2020. She now works as an M&E professional on financial inclusion in China.
2020 is the year that China plans to comprehensively roll out its national social credit system. Since its inception in 2014[i], the system has undergone intense public scrutiny on the data ethics underlying its deployment. To which extent, though, is it really the Orwellian nightmare or the Black Mirror dystopia that it has been depicted to be?
Having a centralized national credit-scoring database is not in itself contentious policy. From a microfinance point of view, the lack of credit data centralization or harmonization between private credit-scoring providers can incur high transaction costs for lenders. A centralized solution offering a government-authenticated history of creditworthiness is a key enabling instrument for under-served sectors such as MSMEs to obtain financing.
What really lies in the ethical grey zone is behavioral data. One could argue that whereas transactional data such as payment data and history of loan defaults are within the acceptable realm of public audit, the behavior of individuals is not. China’s social credit system overtly targets the building of an “honest mentality”[ii]. It does this through bestowing a grading scheme on behavior that has included volunteering and donating blood[iii]. Such ratings produce not only documentation, but also behavioral nudges subject to arbitrarily defined parameters of civic virtue.
In light of this ethical debate, however, there are several myths to dispel. A codified system of civil conduct is definitely within the aim of China’s social credit scoring system. However, to say that it is equivalent to a system of moral coercion and ubiquitous control is deceiving on many levels.
Myth 1: Most social credit data collected is on individual behavior.
Reality: The vast majority of behavioral data collected is of a public or transactional nature, such as jaywalking, late payment of court bills and smoking in non-smoking areas. The system certainly aims to induce a degree of discipline and conformism of public behavior, but is largely unable to put constraints on individual liberty behind closed doors.
Myth 2: The central government carries out the collection of social credit data.
Reality: The primary agents for social credit data collection in China are technology companies and local governments. The biggest social credit data owners now are private tech giants such as the Alibaba-owned Sesame Credits and the mobility behemoth Didi[iv], while a handful of cities such as Hangzhou and Ningbo have run a pilot trial[v] on their own citywide algorithms. The national social credit database is formed by aggregating data submitted by private firms and local governments. There is no standardized algorithm on a national level.
Myth 3: The government has social credit data on everyone.
Reality: China has a higher share of the “credit-invisible” than most developed economies: only 320 million, or 23% of the population, had a traditional credit history in Central Bank records, according to 2017 sources[vi]. China’s social credit system was conceived as an effort to fix this trust deficit and catch up with more advanced credit systems such as the US[vii][viii].
The Bigger Worries
1. The dark world of data anarchy
The spotlight on China’s surveillance apparatus has distracted us from the bigger worries that underlie the collection of alternative data. The first of these is the unregulated or under-regulated collection of consumer personal data by private corporations. The opposite of centralizing behavioral data collection might not be ironclad consumer privacy, but could well be a race-to-the-bottom among private data brokers who seek gains from excessive collection and misappropriation of consumer behavioral data.
Both in and out of China, the most powerful collectors of personal data today are private fintechs, not the government. The Chinese government is neither a groundbreaker nor a leader in its behavioral data scoring capacity. The growth of the Digital Financial Services (DFS) market in Southeast Asia, for instance, has been powered by much more audacious exploitation of consumer data by private credit-scoring agencies. Lenddo, a Singapore-based credit scorer, collects over 12,000 behavioral data points from users’ phones including locations, call history, contacts, apps installed, browsing history and calendar events to generate a credit-worthiness score between 0 and 100[ix]. The scope of behavioral data collection by such fintechs is exponentially larger than that conducted by any public-sector body in China. Moreover, consumer data protection laws are staggeringly weak in some countries with a thriving private credit-scoring market. In the case of Indonesia for example, the country still has no specific personal data protection law in force[x].
In the fintechs’ defense, a functional alternative lending eco-system does run on behavioral data at its core. For small business owners who do not have a strong asset and cashflow profile, behavioral data such as social media footprint can be the saving straw to get them the funding they need.
On the other hand, in the absence of centralization or regulatory rigor, the risks of such personal data being illegally traded or hacked becomes much higher when they are owned by a plethora of private providers. A major credit-reporting agency was found to have sold consumer data to identity theft services in the US in 2013[xi]. Major data brokers have also been found to sell “sucker lists" which identify old or financially distressed individuals[xii]. Centralization would not eliminate these risks but would at least dampen them by consolidating the market of credit data.
2. The regressive nature of the social credit system
Another bigger worry than surveillance is the more far-reaching impact that the social credit system has on deepening social inequality. Despite its potential of becoming an enabler of financial inclusion, the current social credit system in China runs counter to this objective.
The most well-off percentiles of Chinese society are likely to register the most favorable social credit scores, which in turn position them well for receiving future financing. This could be first due to their superior “payment ability”, one of the five dimensions Sesame Credit uses to score an individual[xiii]. Those who regularly jet off to exotic locations on lavish holidays are usually high scorers. Secondly, it is to do with education and financial literacy. For instance, university graduates may have better financial management skills and knowledge about the social credit system, and are thus more apt at mitigating penalties-- from keeping track of loan payback dates to avoiding playing music too loudly on a train.
The Chinese social credit system practically translates the existing purchasing power disparities and education gap between the wealthy and poor into a social credit score, which becomes a regressive instrument in allocating opportunities. China has introduced the system as a policy experiment to police the credit market, but towards this end, the social credit system acts to penalize vulnerability rather than measure and respond to consumers’ economic needs.
These bigger worries do not just pertain to China. As alternative data collection and scoring capacities grow in both developed and emerging markets, private-sector regulation and potential impact on exacerbating inequality are issues of global concern. Data technologies should be harnessed to help the most vulnerable, not to further alienate them. This is the bottom line that any government engaged in a social big data experiment should abide with.
Endnotes
[i] State Council of the People’s Republic of China. “Notice about the planning outline for the construction of a Social Credit System (2014-2020)”. Gov.cn, June 14, 2014, http://www.gov.cn/zhengce/content/2014-06/27/content_8913.htm
[ii] Ming, Cheang. “FICO with Chinese characteristics: Nice rewards, but punishing penalties”. CNBC, March 16, 2017, https://www.cnbc.com/2017/03/16/china-social-credit-system-ant-financials-sesame-credit-and-others-give-scores-that-go-beyond-fico.html
[iii] Jiang, Suqian. “China’s “social credit” scheme involves cajolery and sanctions”. The Economist, May 28th, 2019, https://www.economist.com/china/2019/03/28/chinas-social-credit-scheme-involves-cajolery-and-sanctions
[iv] Hatton, Celia. “China 'social credit': Beijing sets up huge system”. BBC, October 26, 2015, https://www.bbc.com/news/world-asia-china-34592186
[v] Chen, Evelyn, and Tay, Shirley. “China wants to track and grade each citizen’s actions — it’s in the testing phase”. CNBC, July 25, 2019, https://www.cnbc.com/2019/07/26/china-social-credit-system-still-in-testing-phase-amid-trials.html
[vi] Hatton, Celia. “China 'social credit': Beijing sets up huge system”. BBC, October 26, 2015, https://www.bbc.com/news/world-asia-china-34592186
[v] Chen, Xinnian, and Wang, Yi. “The building of China's credit system from the perspective of the US credit crisis”. China.com.cn, January 21, 2003, http://www.china.com.cn/chinese/OP-c/265217.htm
[viii] State Council of the People’s Republic of China. “Notice about the planning outline for the construction of a Social Credit System (2014-2020)”. Gov.cn, June 14, 2014, http://www.gov.cn/zhengce/content/2014-06/27/content_8913.htm
[vi] “Fintech’s dirty little secret? Lenddo, Facebook and the challenge of identity”. Privacy International, October 23rd, 2018, https://privacyinternational.org/long-read/2323/fintechs-dirty-little-secret-lenddo-facebook-and-challenge-identity
[x] Davies, Ed, and Widianto, Stanley. “Indonesia needs to urgently establish data protection law: minister”. Reuters, November 14, 2019, https://www.reuters.com/article/us-indonesia-communications/indonesia-needs-to-urgently-establish-data-protection-law-minister-idUSKBN1XQ0B8
[xi] Krebs, Brian. “Experian Sold Consumer Data to ID Theft Service”. Krebs on Security, October 20th, 2013, https://krebsonsecurity.com/2013/10/experian-sold-consumer-data-to-id-theft-service/
[xii] Hurley, Mikella, and Julius Adebayo. "Credit scoring in the era of big data." Yale JL & Tech. 18 (2016): 148.
[xiii] Ming, Cheang. “FICO with Chinese characteristics: Nice rewards, but punishing penalties”. CNBC, March 16, 2017, https://www.cnbc.com/2017/03/16/china-social-credit-system-ant-financials-sesame-credit-and-others-give-scores-that-go-beyond-fico.html
Photo Credit: Free use image from Canva Pro.