Last update on March 23, 2021
** Based on the data available as of March 21, 2021. Click to enlarge each image.
Replications files are available here.
1. Projected relationship between Covid-19 and output
) |
Source: Authors’ calculation.
Note: Black line:the most recent week. Red line:one week earlier. Blue line:two weeks earlier. The vertical axis shows the number of cumulative deaths by the end of the next 12 months. The horizontal axis shows the average output loss over the next twelve months. The horizontal dashed line indicates the total number of Covid-19 deaths during 2020. The darkest and the second darkest grey areas indicate 20- and 40-percent confidence sets, respectively. The second lightest and the lightest grey areas indicated 60- and 80-percent confidence sets, respectively.
2. Forecast Errors
i. One-week horizon
“Conditional” forecast from last week |
Actual |
“Conditional” forecast error |
|
---|---|---|---|
New Cases | 8,335 | 8,911 | -576 |
New Deaths | 208 | 241 | -33 |
Source: Authors’ calculation.
Note: Our model provides projections of Covid-19 conditional on paths of output but does not provide the projection of output. Thus, we compare the actual Covid-19 outcomes with the projection of them if the model had known the realized path of output.
3. Real-time Evaluation of the Model’s Forecasting Performance
i. One-week horizon
Source: Authors’ calculation.
Note: The red lines—“Forecast”—are what the model would have predicted given the data available up to that point.