Central banks are turning to Big Data to help them craft policy
CENTRAL bankers around the world have set up or are creating departments to embrace Big Data in the quest for deeper insight into the economies they manage.
David Hardoon, chief data officer at the Monetary Authority of Singapore, in a recent speech said: “Isaac Asimov once said, ‘I do not fear computers. I fear the lack of them.'”
“We are now starting to put in place the necessary tools, infrastructure and skillsets to harness the power of data science to unlock insights, sharpen surveillance of risks, enhance regulatory compliance and transform the way we do work.”
Authorities like Mr Hardoon are tapping publicly available sources such as Google Trends and jobs websites to help “nowcast” their economies, and confidential data like credit registers that can help identify a stressed bank.
Collection of micro data increased after the financial crisis, when policy makers realised they lacked the depth of information to make appropriate decisions.
“Central banks are considering or already collecting data that is transaction by transaction, trade by trade, asset by asset, mortgage by mortgage, loan by loan,” said Maciej Piechocki, Frankfurt-based partner at BearingPoint, which has worked on regulatory technology issues with officials around the world.
“That is allowing them the flexibility necessary to answer policy questions that could not be answered before.”
The Bank for International Settlements, known as the central banks’ bank, is also in on the act. In March, it hustled statisticians and economists away from their computers to a Bali beach resort to explore the potential of big data for policy and supervisory purposes.
Here’s a snapshot of how central banks around the world are using or plan to use Big Data:
The Bank of Japan (BOJ) has been using Big Data since 2013 to analyse economic statistics, starting off by beating private forecasts on the accuracy of its GDP (gross domestic product) predictions and evolving its own experimental index that has pushed the government to assess if it’s understating growth.
Yoshimasa Hayashi, a member of parliament from the ruling party, wants to use big data to capture economic numbers directly rather than rely on surveys. Hiroshi Nakaso, deputy governor at the BOJ, in a speech in November cited the movie Moneyball to say that “open-minded and thorough analysis of data, free of prejudice and convention” can drastically change traditional tactics.
The People’s Bank of China in May said that it will increase the use of Big Data, artificial intelligence and cloud computing to boost its ability to recognise, prevent and reduce cross-sector and cross-market financial risks.
China’s regulators have generally steered clear of financial technology innovations, which have spawned success stories such as Jack Ma’s Ant Financial, but have also led to problems such as online lending fraud. In its latest step, China last month halted approvals for new online micro-lenders.
The central bank last year set up a separate technology unit to look at cyber security as well as research and innovation. Its board includes Viral Acharya, deputy governor in charge of monetary policy at the Reserve Bank of India. In a country that lacks privacy laws, an RBI panel in August called for a rights-based privacy framework in household finance rather than the common consent-based method.
“Technological advances such as machine learning and Big Data have changed the ways in which we process data and as a result, have made consent a less-than-effective tool to protect personal privacy,” it said.
The Monetary Authority of Singapore this year started a Data Analytics Group and hired Mr Hardoon to lead it. “To us, ‘data science’ is the business of algorithmically deriving insight from data. Now, this is not the same as AI, or artificial intelligence – the insights and data themselves are inherently not ‘intelligent’,” Mr Hardoon said.
He laid out a hub-and-spokes partnership plan, where the DAG will work with the various departments of the monetary authority and the broader financial sector to identify issues of study, such as anti-money laundering and countering funding of terrorism.
For a fortnight before a policy announcement, Bank Indonesia’s statistics department scours social media, news sites and other content to monitor public perception and rate expectations. The growth in online shopping means the bank is also now receiving information from big players in the e-commerce market, said Yati Kurniati, who heads the department.
“The machines can screen very large amounts of information,” she said, although “we still have to use our own eyes to make sure the context is appropriate”.
Tired of relying on existing data that shows Thailand having one of the lowest unemployment rates in the world even amid tepid economic growth, the central bank is building its own employment index based on data from online job-search portals and is creating a property indicator to give it a better sense of supply and demand in the housing market.
“Existing indicators that we are familiar with may not be enough to answer the questions, not in time, and sometimes even lead to wrong answers,” said Jaturong Jantarangs, an assistant governor at the Bank of Thailand.
“We want to do evidence-based policy so Big Data is useful.”
Bangko Sentral ng Pilipinas last month issued new regulations to boost information technology security in the local banking sector. Governor Nestor Espenilla had said in June that while there is a need to address threats brought about by social media, cloud computing, Big Data and AI, the real challenge is in determining how to strike a balance between encouraging the use of technology and ensuring security amid these innovations.
Earlier this year, Federal Reserve chair Janet Yellen professed that officials at the US central bank are “excited about and want to find ways to use Big Data”, but she also made it clear that policy making continues to, and will for some time, rely on traditional data series.
Big Data has been put to good use by Fed staff economists studying specific issues, like post-hurricane spending activity. But the Fed sees a lot of shortcomings in Big Data, especially the limited periods of time that super-rich data series cover. That significantly reduces its predictive power. Moreover, the data sets are often produced by private companies focused on something other than economic analysis. That can make it less reliable and has made the Fed wary of applying it to policymaking.
The European Central Bank (ECB) has been exploring Big Data since 2013. Information on some 40,000 daily money market transactions will form the basis of an alternative reference rate as traditional benchmarks become unreliable. It has also bought a large set of prices from actual consumer purchases and is exploring ways to scrape the Internet to measure inflation in real time.
Researchers at the Frankfurt-based institution monitor Google Trends to gauge changes in unemployment and use algorithms to scan media reports to assess whether its communication is seen as “hawkish” or “dovish”. Yet, caution remains. “Just as there are concerns about ‘fake news’ dominating social media, there is a risk of ‘fake,’ or at least poor quality, statistics driving out better quality ones in public discourse,” executive board member Benoit Coeure said in November.
In his first year as governor, Mark Carney created a data council – now called the Data Governance Group – as well as a data lab and an advanced analytics unit. Its leader, Paul Robinson, reports to Andy Haldane, the Bank of England’s chief economist. It also contains the bank’s data lab, and boasts physicists and computer scientists.
BOE researchers have recently used Big Data to examine the pass-through of large exchange-rate changes and are building a platform for trade repository data as part of their work on financial stability. One staff member also built a Twitter tool around the time of the 2014 Scottish referendum on independence from the United Kingdom for signs of a run on Scottish banks.
Researchers at Sveriges Riksbank showed this year that using data from online retailers could improve the precision of inflation forecasts by providing up-to-date information on goods with volatile prices such as fruit and vegetables.
“I’d like to build a whole line of indexes using Big Data,” said Alexander Morozov, a former HSBC economist who’s headed the Bank of Russia’s research and forecasting department since 2015. So far he’s developed a gauge of economic activity by scouring a news site and believes the central bank can rely on Big Data to pry information that is even more forward-looking than that contained in purchasing managers’ indexes, which are usually published around the first of each month.
Researchers at Norges Bank fed 26 years of news from local business daily Dagens Naringsliv into a macroeconomic model to create an index of business cycles. The “hypothesis is quite simple,” said Leif-Anders Thorsrud, a senior researcher who started the project.