SUMMARY Developed CountriesEmerging CountriesMacro and Market forecastsDisclaimer to our forecasts |
||||
|
September 2020 |
Septembre 2020
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
The uncertainty around the macro forecasts is very high, and it triggers frequent reassessments any time fresh high frequency data are available. Our macroeconomic forecasts at this point include a higher qualitative component, reducing the statistical accuracy and increasing the uncertainty through wider ranges around them. A global recession is our base case today 1. How deep?
2. How long?
3. The fiscal impact
Financial targets
|
|
The probabilities reflect the likelihood of financial regimes (central, downside and upside scenario) which are conditioned and defined by our macro-financial forecasts. We use the k-means clustering algorithm to our enlarged macroeconomic dataset, splitting the observations into the K cluster, where K represents most of the variability in the dataset. Observations belong to one cluster or another based on their similarities. The grouping of the observations into the k clusters is obtained by minimizing the sum of squared Euclidean distances between observations and clusters centroids i.e. the reference values for each cluster. The greater the distance, the lower the probability to belong to a given regime. The GIC qualitative overlay is finally applied.
The probabilities of risks are the outcome of an internal survey. Risks to monitor are clustered in three categories: Economic, Financial and (Geo)politics. While the three categories are interconnected, they have specific epicentres related to their three drivers. The weights (percentages) are the composition of highest impact scenarios derived by the quarterly survey run on the investment floor. |
|