ECON3006/4437/8037: Financial Economics Assignment 2, Week 7-9 金融经济学作业代写 General Information: • 100 % = 27 points. • Maximal 3 students can submit one solution of Assignment 2. All...View details
Eco 519 Take home project#2: Evaluating SPF Probability forecasts for GDP Declines
概率预测作业代写 The time series data on the probability forecasts for real GNP/GDP declines in the next quarter (called Anxious Index) from 1969:Q1 based on the
The time series data on the probability forecasts for real GNP/GDP declines in the next quarter (called Anxious Index) from 1969:Q1 based on the Survey of Professional Forecasters (SPF) can be downloaded from
The data ends at 2021:Q1. The probabilities are the averages over all respondents in the survey in each quarter.
Based on the actual real GNP growth in the target quarter (real time preliminary 1st month
announcements), a negative GDP growth quarter is denoted as 99999 (in Col D under RECESS). Also, due to Government shutdown, real time GDP growth has a missing value for 1995Q4. Since this quarter didn’t have negative growth, so use “0” for this quarter and all other quarters not identified as 99999.
Define a dummy variable Y = (0, 1) in your analysis, 1 when a quarter is 99999 in the Excel file and 0 otherwise.
Based on these two series, do the following: 概率预测作业代写
1) Compute the QPS, and the Skill Score (base forecast is the proportion of times GDP declined over your sample period.)
2) Run an OLS regression of the dummy (1, 0) variable on the probability forecasts to test if these forecasts are unbiased.
3) Draw the Receiver Operating Characteristic (ROC) curve with values of the threshold ranging from 0 to 0.5 at an increment of 0.05. Search over 0.5 to 1.0 is not necessary because these threshold values are irrelevant for this data series.
4) One popular criterion to choose the optimal threshold to convert probability forecasts into binary forecasts is Kuiper’s or Peirce Skill (PS) score. It is defined as hit rate – false alarm rate.
What desired threshold does PS suggest for these forecasts? What are the associated hit rate and false alarm rates for this optimal threshold? Are these values reasonable to you given your personal risk appetite for type I and type II errors in forecasting recessions?
Please submit your (clean and typed) report in pdf format as an email attachment before December 15 (Wednesday). Projects 1 and 2 will count 30% towards the final. Because I have to submit the final grades before deadline, I cannot accept late submissions.