Please use this identifier to cite or link to this item: https://hdl.handle.net/10419/324510 
Year of Publication: 
2019
Citation: 
[Journal:] Central European Economic Journal (CEEJ) [ISSN:] 2543-6821 [Volume:] 6 [Issue:] 53 [Year:] 2019 [Pages:] 286-303
Publisher: 
Sciendo, Warsaw
Abstract: 
This article primarily aims to estimate the impact of the Armenian revolution and test the hypothesis, that is, the benefits of revolution and establishment of democracy can be seen even in the first year after the political change. To calculate the short-term net surplus of the revolution, we estimated the difference between the projection of Armenian economic activity for the four quarters after the revolution, using only pre-revolutionary (assuming there was no revolution) and real data for the same period after the revolution. Using deep neural network models, such as recurrent neural networks and convolutional neural networks (CNN), we compared prediction accuracy with structural econometrics, such as autoregressive integrated moving average and error correction model, using pre-revolutionary data (2000Q1-2018Q1) for Armenia and combinations of models using an ensembling mechanism. As a result, CNN overperformed the rest of the models. The CNN simulation on post-revolutionary data indicates that during the period 2018-Q2-2019-Q1, Armenia gained approximately 850 million EUR in terms of GDP, thanks to the revolution and the new government. Moreover, out of seven models, the five best models in terms of accuracy indicated that the revolution had no negative impact on the Armenian economy, as the actual values were within or above the 95% confidence interval of the prediction.
Subjects: 
Armenia
revolution
GDP
neural networks
ensembling mechanism
JEL: 
C45
E02
P16
Persistent Identifier of the first edition: 
Creative Commons License: 
cc-by-nc-nd Logo
Document Type: 
Article

Files in This Item:
File
Size





Items in EconStor are protected by copyright, with all rights reserved, unless otherwise indicated.