Diebold Slides - Applied Econometrics

Click here to load reader

  • date post

    08-Feb-2022
  • Category

    Documents

  • view

    0
  • download

    0

Embed Size (px)

Transcript of Diebold Slides - Applied Econometrics

Entering the Pandemic Recession
December 16, 2020
2 / 26
Real growth factor:
y i t = ci + βixt + εit
Stock and Watson (1989 NBER Macro Annual) Aruoba, Diebold and Scotti (2009 JBES)
4 / 26
t vt
t ZP ′t at+1 = Tat|t
Pt+1 = TPt|tT ′ + RQR ′
vt = yt − Zat − Γwt
Ft = ZPtZ ′ + H
at|t ≡ E (αt |Yt), Pt|t = var (αt |Yt), at ≡ E (αt |Yt−1), Pt = var (αt |Yt−1), Yt ≡ {y1, ..., yt}, t = 1, ..., T
Likelihood Evaluation · Model Estimation · Optimal Extraction
6 / 26
High-dimensional state
8 / 26
9 / 26
What do we believe ex post?
What was believed in real time and how does it compare?
Relation to COVID-19?
Relation to other indexes?
Expanding-sample estimation, vintage data (Pseudo Real Time)
Expanding-sample estimation, vintage information (Real Time)
11 / 26
Vintage Information
Economic theory and empirical economic experience evolve. Over time this may prompt, for example, the removal or re-weighting of some component nowcast indicators and/or addition of others, as well as deeper changes in the nowcasting model.
Econometric/statistical theory and experience evolve, prompting changes to the estimation procedure; the frequency and timing of re-estimation and its interaction with benchmark revisions; the estimation sample period; allowance for parameter variation and breaks; the treatment of outliers; the strength of regularization employed; the predictive loss function employed; etc.
Exact times and reliability of nowcast/forecast calculation and release may differ due to technological problems; outright mistakes in nowcast/forecast construction; evolving or changing software algorithms and associated bugs; parallel problems at the agencies responsible for the underlying data and decisions regarding how to deal with them in forecast/nowcast construction; etc.
ADS enables vintage-information examination of Great Recession exit and Pandemic Recession entry
12 / 26
13 / 26
14 / 26
15 / 26
16 / 26
18 / 26
19 / 26
23 / 26
2018 2019 2020 2021
Chicago Fed National Activity Index
What is the National Activity Index? The index is a weighted average of 85 indi- cators of growth in national economic activity drawn from four broad categories of data: 1) production and income; 2) employment, unemployment, and hours; 3) personal con- sumption and housing; and 4) sales, orders, and inventories.
Why are there three index values? Each month, we provide a monthly index (the CFNAI), its three-month moving average, and a diffusion index. Month-to-month movements can be volatile, so the monthly index’s three-month moving average, the CFNAI-MA3, provides a more consistent picture of national economic growth. The CFNAI Diffusion Index instead captures the degree to which a change in the monthly index is spread out among its 85 indicators, averaged over a three-month period.
What do the numbers mean? A zero value for the monthly index has been associated with the national economy ex- panding at its historical trend (average) rate of growth; negative values with below-average growth (in standard deviation units); and positive values with above-average growth. Periods of economic expansion have histor- ically been associated with values of the CFNAI-MA3 above –0.70 and the CFNAI Diffusion Index above –0.35. Conversely, periods of economic contraction have his- torically been associated with values of the CFNAI-MA3 below –0.70 and the CFNAI Diffusion Index below –0.35. An increasing likelihood of a period of sustained increasing inflation has historically been associated with values of the CFNAI-MA3 above +0.70 more than two years into an economic expansion. Similarly, a substantial likelihood of a period of sustained increasing inflation has historically been associated with values of the CFNAI-MA3 above +1.00 more than two years into an economic expansion.
The next CFNAI will be released: December 21, 2020 8:30 am Eastern Time 7:30 am Central Time
News Release Embargoed for release: 8:30 am Eastern Time 7:30 am Central Time The next CFNAI will be released: November 23, 2020
Contact: Michael Adleman Media Relations Federal Reserve Bank of Chicago
Led by improvements in production-related indicators, the Chicago Fed National Activity Index (CFNAI) increased to +0.83 in October from +0.32 in September. Three of the four broad categories of indicators used to construct the index made positive contributions in October, and three of the four categories increased from September. The index’s three-month moving average, CFNAI-MA3, fell to +0.75 in October from +1.37 in September.
The CFNAI Diffusion Index, which is also a three-month moving average, edged down to +0.51 in October from +0.55 in September. Sixty-one of the 85 individual indicators made positive contributions to the CFNAI in October, while 24 made negative contributions. Fifty-four indicators improved from September to October, 30 indicators deteriorated, and one was unchanged. Of the indicators that improved, eight made negative contributions.
Current and Previous values reflect index values as of the November 23, 2020, release and October 26, 2020, release, respectively. N/A indicates not applicable.
CFNAI, CFNAI-MA3, and CFNAI Diffusion for the Latest Six Months and Year-Ago Month
Index suggests economic growth increased in October
Chicago Fed National Activity Index, by Categories
Production and income Personal consumption and housing Employment, unemployment, and hours Sales, orders, and inventories
Oct ’20 Sep ’20 Aug ’20 Jul ’20 Jun ’20 May ’20 Oct ’19 CFNAI Current +0.83 +0.32 +1.10 +2.68 +5.87 +4.29 −0.41 Previous N/A +0.27 +1.11 +2.61 +5.93 +4.22 −0.41
CFNAI-MA3 Current +0.75 +1.37 +3.22 +4.28 −2.52 −5.93 −0.20 Previous N/A +1.33 +3.22 +4.25 −2.53 −5.96 −0.19
CFNAI Diffusion Current +0.51 +0.55 +0.72 +0.75 +0.16 −0.42 −0.17 Previous N/A +0.51 +0.71 +0.74 +0.16 −0.42 −0.17
Dynamic factor model (of course) but otherwise very different: Different people, different indicators (approx 100), different estimation (PC), different data frequency (monthly), different nowcasting frequency (monthly)
24 / 26
25 / 26
Conclusions
Credible real activity growth nowcasting (but noisier in the Pandemic Recession)
Credible real activity growth chronology ex post
Above-average real activity growth resumed in May
Tighten your seat belts for the next two months!
26 / 26