As an example, in the spring of 2018 it was very negative due to a significant spread widening -on the back of political uncertainty in Italy- and a decline in German yields. The variability of the beta raises the question of which period to use when assessing whether the observed spread is warranted given the level of Bund yields. The table shows in row 1 that, using the most recent 20 weeks of data, the observed spread on 17 June was 20 basis points below the estimated spread. Row 2 uses 20 weeks of data starting in August 2021. The beta is significantly lower and the observed spread on 17 June was 65 basis points above the estimated one. Using the entire sample (row 3), with data going back to 2013, the observed spread on 17 June was 17 basis points below the estimate. Different conclusions can be drawn from these calculations.
One, based on a 20-week moving window, the relationship (beta) between the BTP-Bund spread and Bund yields fluctuates over time. After being in negative territory since early 2021, the beta has recently seen a significant increase and is now positive again. Two, considering that Bund yields have moved higher, this has created a feeling of a ‘double whammy’ for the spread on the back of a rise in German yields and an increased beta. Three, the variability of the short-term beta calls for taking a longer perspective. Using data since 2013, the current spread is more or less in line with an estimate based on current Bund yields. Four, clearly there is a need to incorporate additional economic variables in the analysis because they will condition the reaction of Italian yields to changes in German yields. One can think of the stance of monetary policy -the role of net asset purchases, the outlook for the deposit rate-, the outlook for growth and public finances, the risk appetite of investors, etc. It shows the complexity of the task when committing to address unwarranted spread widening, all the more so given the number of spreads that need to be monitored.