Since January 2021, even without temporary measures in response to the pandemic, uncertainty still seems to have decreased a lot.
Baseline resultsusing EPU (all items) and COVID-19 deaths per 1,000 people, as well as all dummies for presidential administrations and recessions.
figure 1: economic policy uncertainty Index (blue) and average of the presidential dummy, logarithmic (1 + Covid death rate per 1,000 people), presidential and recession dummies (tan). NBER-defined recession peak-to-trough dates appear gray. Source: policyuncertainty.com, Department of Disease ControlNBER, and author’s calculations.
Robustness test: using alternative covid dummies.
reader Economical Criticize my sloppy empirical analysis. I repeat this analysis by (1) testing for unit roots, (2) estimating all presidential dummies, (3) using COVID-19 deaths per capita as a proxy for the COVID-19 effect, and (4) including recession dummies.
figure 2: COVID-19 death rate per 1,000 population (CDPM, blue, left scale) and ln(1+CDPM) (tan, right scale). NBER-defined recession peak-to-trough dates appear gray. source: Department of Disease Controlpopulation from FRED, and author’s calculations.
I got the results estimated using OLS on the baseline EPU.
Epp = 119.9 + 411.7ln(1+CDPM) + 36.0economic recession – 15.7GW Bush – 34.4clinton – 30.6GW Bush + 11.2Obama + 21.4trump card + 2.7Biden
Adjust-R2 = 0.48, SER = 28.7, DW = 0.73, NObs = 464. Bold indicates significance at 10% msl using HAC robust standard errors.
The F-test for equal coefficients of Trump and Biden is rejected with a p-value of 0.32.
The Trump administration is characterized by an EPU that is 21.4 percentage points higher than the Reagan administration. Add a new dummy variable (Coronavirus 20m02-22m02) During the Covid period as suggested by reader Econned (2020M02-2022M02), the estimated coefficient for the Trump dummy would decrease to 20.4 (statistically significant at 5% msl), while the Biden coefficient (6.9) would not be statistically significant .
Epp = 118.5 + 57.4Coronavirus 20m02-22m02 + 33.2economic recession – 13.7GW Bush – 32.9clinton – 28.6GW Bush +12.8Obama + 20.4trump card + 6.9Biden
Adjust-R2 = 0.49, SER = 28.4, DW = 0.72, NObs = 464. Bold indicates significance at 10% msl using HAC robust standard errors.
An F-test with equal coefficients for Trump and Biden would fail to reject the traditional level (p-value of 0.5).
Personally, I would prefer to use a continuous series (deaths per thousand), or a dummy that imposes minimal judgment (covid dummy starting from 2020M02), rather than an arbitrarily chosen series (2020M02-2022M02) .
Uncertainty Indicator: News or World Uncertainty Index
image 3: US EPU (all items) (blue, left scale), EPU (news only) (tan, left scale), Ahir-Bloom-Furceri WUI (green, right scale). NBER-defined recession peak-to-trough dates are shaded gray. Source: NBER, policyuncertainty.com.
There is a question: whether to use the full EPU (including policy changes) or just the EPU for news items. I use the latter, and use the Econned virtual model in response to the Covid pandemic.
Eppmessage = 107.5 + 106.9Coronavirus 20m02-22m02 + 52.5economic recession – 5.6GW Bush – 23.1clinton – 24.4GW Bush + 6.3Obama + 15.5trump card – 16.2Biden
Adjust-R2 = 0.35, SER = 45.2, DW = 0.57, NObs = 464. Bold indicates significance at 10% msl using HAC robust standard errors.
The F test that the coefficients of Trump and Biden are equal is strongly rejected at 0.005.
To cross-check these results, I also used Ahir-Bloom-Fuceri World Uncertainty Index (according to the Economist Intelligence Unit report) during the same sample period. The index is provided on a quarterly basis, so I aggregated the monthly data into quarters and ran a regression to obtain:
Ui = 0.103 +0.024Coronavirus 20m02-22m02 +0.078economic recession – 0.053GW Bush – 0.033clinton +0.065GW Bush + 0.127Obama + 0.224trump card +0.091Biden
Adjust-R2 = 0.30, SER = 0.135, DW = 1.12, NObs = 154. Bold indicates significance at 10% msl using HAC robust standard errors.
The F test with equal coefficients for Trump and Biden is rejected at the 0.33 level. The equality test of the Trump-Obama coefficient is rejected at the 0.10 level.
in conclusion: There is less uncertainty under Biden than under Trump.
