• Nem Talált Eredményt

New short-term macroeconomic indicators and their general usefulness for economic nowcasting

N/A
N/A
Protected

Academic year: 2022

Ossza meg "New short-term macroeconomic indicators and their general usefulness for economic nowcasting"

Copied!
29
0
0

Teljes szövegt

(1)

New short-term macroeconomic indicators and their general

usefulness for economic nowcasting

Klaus Weyerstrass

Macroeconomics and Business Cycles Group Evidence-based Policy Making, 27 May 2021

A collaborative event of WIFO, IHS, and CEU

(2)

Timing of economic data releases

• Quarterly national accounts data are released with several weeks lag

• For example, data for the 1 st quarter is released mid-May (flash estimated) end of May (first full estimates)

• Other data on activity (e.g. industrial production, retail sales) as well business and consumer confidence are published monthly

• Still, also monthly data have at least a publication lag of 1

month

(3)

Source: Eurostat: https://ec.europa.eu/eurostat/web/main/news/release-calendar

Example: Eurostat release calendar

(4)

Source: Eurostat: https://ec.europa.eu/eurostat/web/main/news/release-calendar

Example: Eurostat release calendar /cont’d

(5)

Corona pandemic and economic activity

• In Austria 1 st lock-down starts on 16 March

• 14 April: first easing of containment measures: some shops re-open

• 29 May: end of 1 st lockdown

• Already in March sharp drop in economic activity,

unemployment and short-time work soars within just a

few weeks

(6)

Labour market in Austria during the Corona crisis

Source: Austrian Federal Ministry of Labour, https://www.bma.gv.at/Services/News/Aktuelle-Arbeitsmarktzahlen.html;

own illustration

No. of short-time workers

(7)

Labour market in Austria / cont’d

Source: Eurostat; own illustration

No. of unemployed persons Unemployment rate

(8)

Real economic activity in Austria

Source: Eurostat; own illustration

Seasonally adjusted real GDP Change over the same quarter of the previous year

-15.0 -12.5 -10.0 -7.5 -5.0 -2.5 0.0 2.5 5.0

96 98 00 02 04 06 08 10 12 14 16 18 20 -12 -8 -4 0 4 8 12 16

96 98 00 02 04 06 08 10 12 14 16 18 20

Seasonally adjusted real GDP

Change over the previous quarter

(9)

Economic sentiment in Austria

Sources: European Commission, OECD; own illustration

Economic Sentiment Indicator (European Commission) and OECD Composite Leading Indicator for Austria

60 70 80 90 100 110 120

94 95 96 97 98 99 100

Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec Jan Feb Mar Apr May

2020 2021

ESI AUT OECD CLI AUT

(10)

New short-term economic indicators

(11)

New short-term economic indicators / cont’d

Source: Google COVID-19 Community Mobility

Reports: https://www.google.com/covid19/mobility/;

own illustration

(12)

New short-term economic indicators / cont’d

Source: Google COVID-19 Community Mobility

Reports: https://www.google.com/covid19/mobility/;

own illustration

(13)

Short-term indicators

used by the Austrian

National Bank (OeNB)

for its weekly indicator

(14)

Similarly for Germany

Source: Deutsche Bundesbank:

https://www.bundesbank.de/en/statistics/economic-activity-and- prices/weekly-activity-index/methodology-833982#tar-2

(15)

Short-term indicators

• Similar high-frequency indicators have been developed, e.g., by IHS and by WIFO

• Many of the underlying data are not publicly available (e.g. credit card payments data)

• Others are costly, at least for countries outside the EU (e.g. electricity consumption data published by the IEA)

• Some indicators are country-specific and thus only

partially internationally comparable: motorway toll, rail

and road freight transport

(16)

Empirical analysis

• Evaluation of the forecasting / nowcasting performance of two (three) short-term indicators

• Google Mobility Trends

• Electricity consumption

• Weekly indicator developed by the Austrian Central Bank

• Comparison with traditional, monthly indicators: OECD Composite Leading Indicator, Economic Sentiment

Indicator published by the European Commission.

• Countries: G7 + Austria

(17)

Empirical strategy

• Panel estimations with different sets of explanatory variables:

• 6 Google Mobility Indicators

• Electricity consumption (except for Canada and Japan)

• OECD Composite Leading Indicator

• Single equation estimations for Austria

• Google Mobility data due to shortness of time series only separately

• Electricity consumption

• OECD Composite Leading Indicator

• Economic Sentiment Indicator (ESI)

• OeNB Weekly GDP Indicator

(18)

Forecast evaluation results

Evaluation statistics for US real GDP growth

Forecast RMSE MAE MAPE SMAPE Theil U2

All variables 2.127 1.771 348.443 82.765 0.221 OECD CLI amplitude ajd. 2.679 2.377 166.531 121.893 0.392 OECD CLI trend restored 4.366 3.630 196.099 153.602 0.791 Electricity consumption 5.999 5.200 461.168 185.776 1.102 Google all 1.982 1.565 337.473 85.573 0.065 Google Work places 2.681 2.301 468.084 93.441 0.135 Google Retail & Recreation 2.668 2.108 278.715 102.043 0.494 Google Residential 2.102 1.614 200.664 76.290 0.319 Google Pharmacies, Grocery 3.646 2.746 350.858 89.640 0.593 Google Transit stations 2.831 2.254 300.647 87.482 0.453 Google Parks 3.390 2.514 352.029 103.713 0.673 Simple mean 2.292 1.902 224.357 91.038 0.424

Evaluation period: 2020q1 – 2021q1

(19)

Forecast evaluation results / cont’d

Evaluation statistics for Canada's real GDP growth

Forecast RMSE MAE MAPE SMAPE Theil U2

All variables 1.014 0.819 128.234 84.915 0.147 OECD CLI amplitude ajd. 6.014 4.896 281.534 157.884 0.869 OECD CLI trend restored 7.552 6.421 465.732 195.309 1.066

Google all 1.669 1.467 328.450 94.891 0.064

Google Work places 2.879 2.278 603.388 89.062 0.133 Google Retail & Recreation 4.329 3.755 640.866 105.326 0.510 Google Residential 4.088 3.486 579.584 101.385 0.383 Google Pharmacies,

Grocery 4.620 3.748 712.579 105.674 0.411

Google Transit stations 4.666 3.983 730.243 113.000 0.496 Google Parks 4.553 3.573 573.910 100.610 0.679 Simple mean 3.296 2.576 298.270 71.950 0.462

Evaluation period: 2020q1 – 2021q1

(20)

Forecast evaluation results / cont’d

Evaluation statistics for Japan's real GDP growth

Forecast RMSE MAE MAPE SMAPE Theil U2

All variables 0.983 0.775 44.141 29.480 0.203 OECD CLI amplitude ajd. 4.421 3.570 93.794 156.482 1.156 OECD CLI trend restored 4.588 3.520 84.740 149.583 1.115

Google all 2.094 1.932 92.618 77.198 0.654

Google Work places 2.347 2.086 100.706 84.482 0.920 Google Retail & Recreation 2.608 2.348 96.458 61.891 0.905 Google Residential 2.096 1.870 81.340 52.284 0.783 Google Pharmacies, Grocery 3.486 2.884 118.515 67.924 1.286 Google Transit stations 2.333 2.094 89.361 57.867 0.946 Google Parks 3.380 2.962 121.561 74.674 0.922 Simple mean 2.418 2.027 69.970 55.182 0.702

Evaluation period: 2020q1 – 2021q1

(21)

Forecast evaluation results / cont’d

Evaluation statistics for UK real GDP growth

Forecast RMSE MAE MAPE SMAPE Theil U2

All variables 1.252 1.061 20.721 18.689 0.050 OECD CLI amplitude ajd. 7.912 6.923 74.573 134.562 0.605 OECD CLI trend restored 11.830 10.616 124.640 185.961 0.943 Electricity consumption 11.231 8.950 97.296 155.441 1.052

Google all 1.820 1.669 35.327 50.534 0.072

Google Work places 3.208 2.470 50.502 57.495 0.178 Google Retail & Recreation 4.708 3.972 61.183 46.362 0.423 Google Residential 4.122 3.522 50.412 41.149 0.356 Google Pharmacies,

Grocery 3.704 2.637 36.283 29.525 0.352

Google Transit stations 5.004 3.939 48.274 40.301 0.463 Google Parks 6.466 4.914 89.114 54.257 0.642

Simple mean 4.588 3.198 25.777 30.250 0.447

Evaluation period: 2020q1 – 2021q1

(22)

Forecast evaluation results / cont’d

Evaluation statistics for France's real GDP growth

Forecast RMSE MAE MAPE SMAPE Theil U2

All variables 1.755 1.583 51.253 57.580 0.227 OECD CLI amplitude ajd. 5.965 4.972 93.243 133.227 0.832 OECD CLI trend restored 7.739 5.952 82.663 149.378 1.072 Electricity consumption 8.356 6.280 85.625 155.910 1.180

Google all 2.802 2.485 51.971 65.087 0.355

Google Work places 3.519 2.586 87.174 55.298 0.537 Google Retail & Recreation 5.830 4.309 149.952 70.783 0.895 Google Residential 4.517 3.769 125.089 80.907 0.671 Google Pharmacies, Grocery 4.236 3.791 122.215 86.982 0.585 Google Transit stations 4.759 3.668 129.742 69.920 0.732 Google Parks 6.714 5.000 141.797 78.647 0.983 Simple mean 4.490 3.249 84.597 67.001 0.658

Evaluation period: 2020q1 – 2021q1

(23)

Forecast evaluation results / cont’d

Evaluation statistics for Italy’s real GDP growth

Forecast RMSE MAE MAPE SMAPE Theil U2

All variables 1.637 1.308 25.217 23.568 0.194 OECD CLI amplitude ajd. 5.687 4.914 78.862 125.784 0.882 OECD CLI trend restored 7.790 6.791 117.610 167.844 1.188 Electricity consumption 7.984 6.839 136.218 133.401 1.231

Google all 3.411 3.010 63.080 54.283 0.272

Google Work places 3.313 2.944 87.010 48.599 0.463 Google Retail & Recreation 4.968 4.073 126.386 60.356 0.720 Google Residential 5.155 4.728 135.119 89.588 0.625 Google Pharmacies,

Grocery 3.543 3.420 89.258 63.733 0.417

Google Transit stations 4.768 3.918 124.988 59.699 0.689 Google Parks 5.679 4.321 111.687 58.741 0.873

Simple mean 4.009 3.399 76.052 57.211 0.613

Evaluation period: 2020q1 – 2021q1

(24)

Forecast evaluation results / cont’d

Evaluation statistics for Germany's real GDP growth

Forecast RMSE MAE MAPE SMAPE Theil U2

All variables 1.267 1.177 34.776 36.256 0.201 OECD CLI amplitude ajd. 3.926 3.667 93.429 141.382 0.677 OECD CLI trend restored 5.673 5.280 128.860 178.074 1.073 Electricity consumption 5.464 4.658 103.976 164.361 1.158

Google all 1.726 1.351 27.321 25.188 0.342

Google Work places 2.527 2.191 56.389 49.492 0.490 Google Retail & Recreation 3.903 3.411 84.110 74.773 0.744 Google Residential 3.446 3.124 77.323 69.676 0.659 Google Pharmacies, Grocery 3.115 2.783 69.436 59.758 0.614 Google Transit stations 3.425 2.923 69.783 59.622 0.666 Google Parks 3.246 2.578 57.541 50.442 0.718 Simple mean 2.503 1.748 30.939 35.341 0.562

Evaluation period: 2020q1 – 2021q1

(25)

Forecast evaluation results / cont’d

Evaluation statistics for Austria's real GDP growth

Forecast RMSE MAE MAPE SMAPE Theil U2

All variables 1.490 1.340 30.875 27.664 0.217 OECD CLI amplitude ajd. 5.540 4.893 89.647 157.590 1.030 OECD CLI trend restored 5.922 5.007 87.145 156.030 1.141 Electricity consumption 6.710 5.784 102.599 191.670 1.278

Google all 2.363 1.811 35.774 30.564 0.440

Google Work places 3.134 2.515 58.281 44.434 0.515 Google Retail & Recreation 4.336 3.655 83.483 66.004 0.723 Google Residential 4.031 3.662 85.079 77.685 0.631 Google Pharmacies,

Grocery 3.103 2.512 54.504 41.384 0.541

Google Transit stations 3.561 2.896 65.184 49.422 0.605 Google Parks 4.034 3.141 60.938 50.514 0.746

Simple mean 3.269 2.328 39.200 39.093 0.648

Evaluation period: 2020q1 – 2021q1

(26)

Forecast evaluation results / cont’d

-14 -12 -10 -8 -6 -4 -2 0 2

I II III IV I

2020 2021

Year % Change GDPR_AUT OeNB

Google Parks Google Phamacy Google Residential Google Retail

Google Transit stations Google Work places Electricity

(27)

Forecast evaluation results / cont’d

Evaluation statistics for Austria's real GDP growth

Forecast RMSE MAE MAPE SMAPE Theil U2

Electricity, ESI, OECD AA 1.530 1.192 90.993 84.902 0.884

ESI 2.182 1.289 79.093 70.550 1.237

OECD CLI amp. Adjusted 2.542 1.688 123.464 97.161 0.950 PECD CLI trend restored 2.074 1.455 112.553 88.381 1.089 Electricity consumption 2.466 1.829 121.900 109.572 1.013

Simple mean 1.872 1.174 74.345 68.849 0.952

Evaluation period: 2009q1 – 2021q1

(28)

Summary and conclusions

• Corona pandemic caused unprecedentedly fast reduction in economic activity

• General challenge in economic forecasting of sizeable data publication lag was aggravated

• New high-frequency indicators were created

• Many of the underlying data are not publicly available

• Mobility data will probably no longer have information content once mobility is back to normal

• Some high-frequency data are worthwhile to monitor also

in the future, e.g. electricity consumption, freight transport

(29)

Thank you for your attention

Hivatkozások

KAPCSOLÓDÓ DOKUMENTUMOK

Monthly data on Russian interest rates and macroeconomic indicators for the period July 1995 – October 1997 were used to estimate relationship of GKO interest rate, interest rate

Keywords – Accumulating plants, Ellenberg indicator values, heavy metals, metal uptake, phytoextractors.. Even though the European Environment Agency indicators reveal that

The aim of this paper is to evaluate to what degree, if any, the published data on national competitiveness by the World Economic Forum (WEF), probably the most well-known

Report adopted in July by tlre OECD Tourism Committee / Organisation for Economic Co-operation and Development — Paris : OECD, 1995.. —-

Time Percent of trend.. lags compare the data points at turning points only, while the cross-correlation lag com- pares all the data points.) For example, the long-term credit

– From an economic perspective, the manner in which natural re- sources are used and managed affects a) the short-term costs and the long-term economic sustainability; b) the supply

Az OECD rövid távú statisztikai mutatókból számított kompozit indexe (Composit Leading Indicators – CLI) 2008 júliusában, a májusi adatok alapján már erősödő

In this study a new leading indicator called Lead- ING HUBE (LeadING Index for the Hungarian Busi- ness Economy) is introduced and being calculated on a monthly basis