‘Statistics Should Be Available and Comprehensible to Everyone’

Implementing a digital analytical platform, opportunities for Big Data, and other prospects for the development of Russian statistics were discussed by participants at a plenary session of the XX April International Academic Conference.

Leonid Gokhberg, First Vice Rector and Director of the HSE Institute for Statistical Studies and Economics of Knowledge, opened the session and reminded the audience that HSE has been discussing the challenges and outlook for the development of Russian statistics for several years already. ‘Quality, reliability and immediacy of statistical information predetermine the quality of managerial solutions,’ the Vice Rector emphasized. ‘Meanwhile, administrators are not always capable of interpreting statistics and their limitations correctly, or of exploiting the opportunities they present. This is the reason behind the many negative pronouncements about statistics, most of which are false in nature.’

The Primary Objective is High-quality and Timely Statistics

Pavel Malkov, Head of the Federal State Statistics Service (Rosstat), spoke about some of the key points of the agency’s Development Strategy up to 2024, which is currently being finalized.

The main problem of official statistics is the low quality of incoming data

Users lack numbers, perspectives, the speed of statistical information delivery and so on, while businesses complain about excessive reporting burdens. That’s why a crucial task is to ensure high-quality and timely statistics, with minimal burden on businesses.

‘Worldwide, statistics is facing challenges that seem to be incompatible. As we make statistics gathering faster, we inevitably lose quality, and as we increase statistics output, we increase the burden on businesses. These problems can be solved by using modern technology, namely digitalization of statistical data,’ Mr. Malkov believes.

The first steps in this direction have already been taken. A new Rosstat website will be launched in the near future, and in addition, the agency is partnering with other government agencies (tax service, pension fund, and others) to develop a set of road maps on the use of administrative data. By the end of the year, the main components of a digital platform to analyse statistical data will be introduced. By 2020, several pilot projects on big data will be launched. Rosstat is also planning to initiate discussion of a bill on transition to digital statistical reporting and the creation of an international statistics resource centre, which will facilitate participation by Rosstat experts in international projects.

Big Data is Not a Silver Bullet

According to Mr. Malkov, big data is one of the main challenges today. The advantages of using big data are widely known. They include the high speed of information updates (almost in real time), maximum detail, minimal influence of human factors, and the lack of burden on those providing data. But there is a fly in the ointment. These data are poorly structured, their sources are unstable, and the terms of their use can change. It’s also difficult to determine the coverage and, accordingly, the data representativeness. New types of errors are appearing that did not exist previously, and the data quality assessment becomes more complicated. According to Malkov, there are no ideal solutions. However, Rosstat responds to the imperatives of our times and takes advantage of international experience: it is preparing a series of pilot projects on big data use, including statistics on tourism (using data from mobile operators, banks, payment systems, internet resources, etc.), population, agriculture, transportation and other fields.

Digital Analytics Platform

Another significant problem in official statistics is the lack of a unified methodology for collecting and processing data to develop statistical values. ‘Today, we have over 60 statistical authorities, but the data are often duplicated repeatedly and contradict each other, which leads to a lack of trust in statistics among users in general,’ Mr. Malkov said.

A unified digital analytical platform could help solve this problem.

The digital model should become an important part of the national data management system

At the point of entry, the designed model will include a continuous collection of data through one point of contact and imported information from alternative sources. ‘At the output, statistical master data will form a national data management system, and all interested parties will have the ability to generate figures in format that they need. The model should also allow data reporting in international formats,’ Mr. Malkov said.

Six Components of the Digital Platform at the First Stage

1. An indicator form registry will help decrease the reporting burden and prevent duplication; it should also put into practice the principle of one-time information submittal.
2. A statistical observation object registry is required to solve the problem of contradictory information and to create a unified record of objects for all government agencies.
3. Automation of the process for developing a federal statistics plan.
4. The one-contact principle for uploading primary statistical data and data file storage will solve the problem of inefficient use of a huge volume of primary statistical data. Further refinement of primary statistical data will be performed with information from the observation object registry and the indicator form registry.
5. A personal profile for respondents with a targeted approach to information collecting, automated reporting and a service for the platform’s interaction with respondents’ accounting systems.
6. BI technologies will solve data generation tasks requested by businesses.

An expert discussion on creating the digital analytical platform can be found here.

How to Keep Quality High

After Mr. Malkov’s presentation, several statistics experts and statistical data users expressed their views on the matter.

Ekaterina Prokunina, Director of the Department of Statistics at the Central Bank of Russia, reminded the audience that statistics should be part of contemporary project management. ‘We need to clearly formulate data requirements at the stage of standards approval and legal decisions, otherwise we lose quality since it’s not planned from the very beginning,’ she concluded.

Agreeing with Ms. Prokunina was Anton Kosianenko, Director of the Department for Expert Analysis and Oversight in Strategic Audit at the Accounts Chamber of Russia. He also noted that digitalization will only partly solve the dilemma between information immediacy and its quality and suggested using organizational measures in interim decision-making. The difference between preliminary and final estimations must be clear both in terms of methodology and in the fact of re-evaluation itself.

Methodology can’t be simple, but it can be clearly explained

Mr. Kosianenko admitted that some long, wordy and very professional reports are written at the Accounts Chamber as well, but that these reports are completely incomprehensible for users.

Leonid Gokhberg turned the audience’s attention to the need for a professional approach by organizations and individuals charged with decision making. ‘When statistical values and requirements for statistics are not generated in a professional way, this leads to negative consequences not only for the Rosstat system and those it surveys, but also for real processes, since these processes are adjusted according to the wrong statistical values,’ the Vice Rector said.

How to Survive the Crisis in Confidence

According to Ivan Begtin, Director of the Informational Culture NGO, Rosstat’s key problem that has been discussed in media for several years, is the crisis of confidence in the service. ‘Of course, we can explain this partly by media hysteria, but much of it is objective,’ he believes. Begtin emphasized that Rosstat is a government agency that works for the public good, but that administrative solutions are made not only within government agencies. This means that economic actors rather than decision makers should be the key consumers of statistical data.

Besides translating the methodology into a comprehensible language, the source code used to calculate the data needs to be published openly so that everyone can verify the figures

Begtin also said that Rosstat must act as a global methodologist and set requirements for obligatory collection, disclosure and input methods for all ministry-specific statistics that executive authorities possess. ‘In addition to the digital analytical platform, which is a promising initiative, we would like Rosstat to have the authority to impact ministry-specific statistics and their accessibility, despite the opposition of some of the more closed ministries,’ he concluded.