Maastricht University research thesis project aiming to quantify differences in the banking competition publications, between central bank and university economists.

Current dataset consists of 69 economic articles, of which:

• 15 are affiliated to various central banks (authors are employed by CB)
• 54 are written by university or public organisations researchers

• Natural Language Processing

  • Cosine similarity
  • Hierarchical clustering
  • Topic models: Latent Semantic Indexing
  • K-means clustering on LSI components

• Reported statistics analysis

  • reviewing reported p-values and t-statistics to determine if Central Bank researcher show tendency to commit Type II Error more frequently

• Panzar-Rosse: H-static

• ∇ correspond to central bank paper
• Colours show corresponding class