To Örebro University

In this paper, we analyse whether two key macroeconomic relationships in Australia-Okun's law and the Phillips curve-have been stable over time. This is done by estimating hybrid time-varying parameter Bayesian VAR models using quarterly data from 1978 to 2024. Model comparison based on marginal likelihoods indicates that Okun's law has been stable, whereas the Phillips curve has not. Using the preferred specification of the BVAR for the unemployment rate and inflation, we also calculate trend values for both variables. The model's trend unemployment rate at the end of the sample is approximately 5%; estimated trend inflation at the same point in time is close to the Reserve Bank of Australia's inflation target.

In this paper, we investigate whether key relations in the crude oil market have been stable over time. This is done by estimating hybrid time-varying parameter structural Bayesian VAR models using monthly data ranging from February 1973 to May 2023. Model selection suggests that while stochastic volatility is preferred over homoscedasticity, the dynamics of the model are best described by constant parameters in all equations.

In this paper, we evaluate the directional interest-rate forecasts of UK households from the Bank of England’s Inflation Attitudes Survey. Employing a test for directional forecast accuracy and data on the survey balance ranging from 1999Q4 to 2023Q2, we find that the balance is not able to predict in which direction the interest rate will move over the coming year. In addition, regression models based on the balance are not able to generate forecasts for the quantitative change in the interest rate over the coming twelve months that have higher precision than a naïve forecast of no change. In order to provide information as to whether our findings are due to the inherent difficulty when it comes to forecasting interest rates or if households are not very insightful regarding interest rates, we investigate – again using data on the survey balance and testing for directional accuracy – whether households have been able to correctly assess the directional change of the interest rate over the previous twelve months; our results indicate some amount of “literacy” among the households regarding the interest rates that they face. Finally, analyses based on individual-response level data suggest that literacy regarding interest rates – proxied by the respondent having been correct regarding the directional change over the previous twelve months – does not appear helpful when forecasting.

Employing an unobserved components stochastic volatility model, we estimate the trend 5-year government bond rate in Canada, Norway, Sweden, Switzerland, the United Kingdom and the United States. Our results suggest that the estimated trend rate has decreased substantially between 2000 and 2020 in all six countries. In Norway, Switzerland and the United Kingdom, the trend rate has increased non-negligibly since 2020; in Canada, Sweden and the United States, the trend rate has not been affected much by the increase in the actual 5-year government bond rate. At the end of the sample, none of the countries has a trend rate higher than 3%. The model hence suggests that the 5-year government bond rate will be fairly low in the medium-to-long run.

This chapter gives a brief introduction to the Festschrift. A short overview of Sune Karlsson’s career and research is provided. In addition, the content of each chapter is presented.

The original contributions on Bayesian econometrics gathered in this book pay tribute to Sune Karlsson, celebrating his significant work in time series econometrics and its applications in macroeconomics and finance. The volume consists of both methodological and empirical studies by leading experts in the field, with particular attention paid to Bayesian vector autoregressive (VAR) models and forecasting. It addresses forecasting with Bayesian VARs as a research field, mixed-frequency and high-dimensional Bayesian VARs, various forms of Bayesian VARs with stochastic volatility, forecast combination, analysis of time-varying parameter models in the frequency domain, and portfolio analysis in a Bayesian framework. Presenting cutting-edge research and providing valuable insights into the field of Bayesian econometrics, the book will appeal to researchers, practitioners in the banking sector, and government authorities.

In this chapter, we extend the standard Gaussian stochastic volatility Bayesian VAR by employing the generalized hyperbolic skew Student’s t distribution for the innovations. Allowing the skewness parameter to vary over time, our specification permits flexible modelling of innovations in terms of both fat tails and—potentially dynamic—asymmetry. In an empirical application using US data on industrial production, consumer prices and economic policy uncertainty, we find support—although to a moderate extent—for time-varying skewness. In addition, we find that shocks to economic policy uncertainty have a negative effect on both industrial production growth and CPI inflation.

We analyse how financial market analysts' expectations in the Czech National Bank's Financial Market Inflation Expectations survey perform relative to the random-walk forecast when it comes to predicting five financial variables. Using data from 2001 to 2022, our results indicate that the analysts are able to significantly outperform the random-walk forecast in terms of forecast precision for the repo rate and Prague Interbank Offered Rate at the one-month forecasting horizon. For the five- and ten-year interest rate swap rates, the random walk significantly outperforms the analysts at both the one-month and one-year forecasting horizons. For the CZK/EUR exchange rate, the random-walk forecast has a lower root mean squared forecast error than that of the analysts' forecast at the one-month horizon whereas at the one-year horizon the opposite is found; however, none of these differences are statistically significant.

In this paper, the natural rate of interest in Denmark, Norway and Sweden is estimated. This is done by augmenting the Laubach and Williams (Rev Econ Stat 85:1063-1070, 2003) framework with a dynamic factor model linked to economic indicators--a modelling choice which allows us to better identify business cycle fluctuations. We estimate the model using Bayesian methods on data ranging from 1990Q1 to 2022Q4. The results indicate that the natural rate has declined substantially and in all countries is at a low level at the end of the sample.