Scholarlink

<p>This paper compares advanced nowcasting methods within the data context typical for small open CESEE economies, characterised by a limited set of mixed frequency and ragged edge domestic and international high frequency indicators. In nowcasting Slovenian real GDP growth, we evaluate bridge equation models, MIDAS models, MF-VAR models, the DFM, and the MF-3PRF. We also explore the benefits of model combinations, indicator preselection, and the role of news within the DFM. We find that factor models, particularly the DFM, perform best in this setting. Combining DFM and MF-3PRF nowcasts proves effective in periods of moderate volatility in real GDP growth. No model strongly favours a large scale information set, while using a smaller set shows greater variability. Within the DFM, news from both international environment related indicators and domestic hard data shows considerable importance for nowcasting in small open economies. Expanding the evaluation sample to include Covid-19 data further reinforces the DFM’s dominance, while the MF-3PRF’s performance declines. During this period, the relative importance of news from certain indicator categories increases compared to the original evaluation sample.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Nowcasting economic activity in a small open CESEE economy using mixed frequency data

  • 期刊论文

  • 发表日期：

    2025年8月7日

查看全文
  • Jan Radovan,
  • Igor Masten

摘要

This paper compares advanced nowcasting methods within the data context typical for small open CESEE economies, characterised by a limited set of mixed frequency and ragged edge domestic and international high frequency indicators. In nowcasting Slovenian real GDP growth, we evaluate bridge equation models, MIDAS models, MF-VAR models, the DFM, and the MF-3PRF. We also explore the benefits of model combinations, indicator preselection, and the role of news within the DFM. We find that factor models, particularly the DFM, perform best in this setting. Combining DFM and MF-3PRF nowcasts proves effective in periods of moderate volatility in real GDP growth. No model strongly favours a large scale information set, while using a smaller set shows greater variability. Within the DFM, news from both international environment related indicators and domestic hard data shows considerable importance for nowcasting in small open economies. Expanding the evaluation sample to include Covid-19 data further reinforces the DFM’s dominance, while the MF-3PRF’s performance declines. During this period, the relative importance of news from certain indicator categories increases compared to the original evaluation sample.