This study differentiates p-hacking from publication bias by examining biases resulting from selective reporting within studies versus selective publication of entire studies. Analyzing a dataset of 400 meta-studies, encompassing nearly 200,000 estimates from approximately 19,000 individual studies in economics and related social sciences, I observe a notably higher incidence of p-hacking as compared to selective publication. Employing various meta-regression methods, I find that selective reporting within studies is about 20% more prevalent than publication bias arising from selection among studies. This finding underscores the considerable influence of practices such as p-hacking and method-searching, suggesting that they contribute significantly to selection bias in the economic literature and could affect the perceived reliability of published findings.
The news-based Economic Policy Uncertainty indices (EPU) of Germany, France, and the United Kingdom have discernible trends that can be found neither in other European countries nor in other uncertainty indicators. Therefore, we replicate the EPU index of European countries and show that these trends are sensitive to a rather arbitrary choice of normalizing the raw counts of news related to economic policy uncertainty by a count of all newspaper articles. We show that an alternative normalization by news on economic policy leads to different long-term dynamics with less pronounced trends and markedly lower uncertainty during recent uncertainty periods such as Brexit or the COVID-19 pandemic. Consequently, our results suggest that the effects of uncertainty related to these events on economic activity could have been overestimated.
When people anticipate a change in the policy, they tend to adjust their behavior before the actual decision is made. The impact of anticipation on the results of VAR literature has been demonstrated particularly by Ramey (2009), who reexamines the differences between traditional Cholesky identification and the Ramey-Shapiro narrative approach with contradictory conclusions on the effect of government spending on consumption and wages. She discusses the importance of timing the identified shock (government spending) and argues that failing to count for the anticipation effect of the shock can cause these contradicting conclusions. So far, these aspects have not been discussed in the context of uncertainty shocks. However, some of the most prominent peaks in uncertainty, such as the Brexit referendum, could have been anticipated before the event actually happened. Daily data will allow studying the evolution of uncertainty related to specific events before and after the specific event so that an identification of the unexpected part of uncertainty will be possible.
Critical challenges in estimating the New Keynesian Phillips Curve (NKPC) equation lay in measuring the expectations of future inflation in the first part of the equation and choosing the driving variables in the second. Inflation expectations are not directly observable, and there is no agreement in the literature on the choice of driving variables. Hence the estimated NKPC is affected by various choices on inflation characteristics and the driving variable. Additionally, publication bias can be a salient factor affecting the variation of estimates. We use modern meta-analysis tools to study the impact of the estimated NKPC characteristics and publication bias in the literature. Finally, we conclude that heterogeneity in the chosen characteristics significantly affects the estimation of the NKPC equation and the resulting implications of the expected inflation and driving variable on the real economy. In conclusion, the research characteristics of the expected inflation and driving variable affect the real economy. Hence, our findings are important in understanding the role of heterogeneous characteristics in the implications of the estimated NKPC.