Essays on the payment system in Indonesia: macroeconomics, behaviour analysis and policy implication

Badrawani, Wishnu ORCID: 0000-0001-8563-5408 (2024). Essays on the payment system in Indonesia: macroeconomics, behaviour analysis and policy implication. University of Birmingham. Ph.D.

Preview

Badrawani2024PhD.pdf
Text - Accepted Version
Available under License All rights reserved.
Download (3MB) | Preview

Abstract

The use of non-cash payment instruments in developing nations, particularly Indonesia, has changed significantly compared to conventional payment instruments such as cash and coins. This rapid increase of non-cash payment instruments was made possible not only by the rapid development of the financial and payment system and telecommunications technology but also by government and central bank support, as well as unprecedented events such as pandemics and economic crises. This thesis comprises four articles with intertwined general purposes to present empirical evidence of economic analysis of payment system data, analysis of people’s behaviour in adopting a new payment instrument and policy impact analysis, as well as an examination of contemporary methodology to forecast inflation using payment system data. The first essay examines the determinants of money demand in Indonesia. This study examines the stability of money demand by incorporating the payment system innovation variable and the possibility of a structural break related to central bank policies. The second and third essays analyse the behaviour of consumers and merchants in adopting a new payment platform in the presence of a new central bank policy that was hindered by the COVID-19 pandemic. These studies use an online survey with self-administered questionnaires to obtain data from 31 Indonesian provinces. We introduced additional exogenous latent variables embedded inside the unified technology adoption theory; our research successfully unravels and distinguishes the effects of central bank policies and the pandemic. Finally, the fourth study explores payment system data to forecast inflation using various machine learning (ML) techniques and compares it to that of the univariate time series ARIMA and SARIMA models. We also perform various out-of-bag sample periods in each ML model to determine the appropriate data-splitting ratios for the regression case study. The study reveals that the ML models outperformed the ARIMA benchmark in terms of prediction accuracy and discovered that numerous payment system factors were highly predictive of inflation. We demonstrate the interpretation of ML forecast results that complement the causal inference of the more established econometric method.

Type of Work:

Thesis (Doctorates > Ph.D.)

Award Type:

Doctorates > Ph.D.

Supervisor(s):