Pros and cons of using big data in economic signals

0
965

Big data is being used for a variety of purposes, including assessing economic development.

A new wave of covid infections caused by the Omicron strain of the Sars-CoV-2 virus has restricted travel in most parts of the country. People spend more time at home than at work, retail outlets, or parks, according to anonymized data provided by Google for India-based mobile phone locations.

During the pandemic, such new types of data proved extremely useful in tracking economic activity regularly. Rather than waiting for official statisticians to compile the typical quarterly assessments of activity from across the economy using structured surveys and administrative data, they give us an immediate sense of what is going on in the economy.

Both have in common that we have evolved as a civilization in response to the epidemic. The relationship between mobility metrics and economic development has shifted over time. In more technical terms, the regression results have changed as economic actors have essentially learned to live with the virus.

According to OECD data, a ten-point increase in mobility was associated with a 2.2-point increase in the third quarter of 2020, but only a 0.9-point increase in the fourth quarter. This is a significant decrease.

It also implies that an analyst using the coefficients in the first and second quarters of 2020 to forecast the impact of shifts in focus on advancing on quarterly economic activity would reach significantly different conclusions than another analyst using the coefficient in the fourth quarter. This is significant when attempting to quantify the impact of the fourth wave on the Indian market alone, especially when movement data is a key factor to consider.

Other types of large data have their own set of issues

Consider artificial lighting, which some economists are increasingly using as a proxy for business growth. The data on lights left after dusk is collected by several satellites that can detect the brightness of lights produced by people in a specific region.

According to Ayush Patnaik, Ajay Shah, Anshul Tayal, and Susan Thomas of research firm xKDR in a recent working paper, clouds influence how data on night lights can be gathered by satellites that hang kilometres above the ground. The four researchers established that measurements have a downward bias during foggy months and created an algorithm to partially correct this downward bias.

Massive data analysis, on the other hand, poses its own set of challenges. Economists, for example, use night lights after sunset as a proxy for business growth. According to several research articles, clouds appear to obstruct satellites’ ability to acquire data on night lighting.

As a result, during cloudy months, measurements are low. Second, information derived from e-way bills generated during product transportation is a critical predictor of future economic activity. E-way bills are not generated for services.

The number of e bills must be kept to a minimum due to the shift in demand from commodities to services. As a result, it does not point to a slowing of economic activity.

 Follow and connect with us on Facebook, LinkedIn & Twitter