2020.12.25

The economic effect of the restriction by Japanese government under COVID-19

Research and Development by

Hiroyasu Inoue, Graduate School of Simulation Studies, University of Hyogo

Corresponding Research Area

Simulation and designing countermeasures against possible COVID-19 resurgence: predicting spreading of infection, estimating and verifying the effectiveness of countermeasures, and predicting deman


Supply chain data of Japan

  • Covers almost all firms
  • Includes trade partners.
  • Also includes standard attributes (sales, profit, location, sector, financing banks etc.)
  • Has snapshots of 2006, 2007, 2011, 2012, 2014, and 2016
  • Has 1,668,567 firms and 5,943,072 supplier-client relationships in 2016
  • (Provided Tokyo Shoko Research Ltd.)

Supply-chain model

No price. No market. Try keeping production volume before shock.

Overview of model

Parameter calibration: The Great East Japan Earthquake (2011)


Actual restrictions

Restriction strength is incorporated from literature. (By industry.   See next page.)
(Guan et al. 2020 and Bonadio 2020)

Vertical axis: GDP, Horizontal axis: day

Blue line: benchmark (U.S.)

Pink area: GDP loss estimated by IAIA

Green line: Mitigated benchmark restriction that fits pink area in terms of total GDP loss

 Restriction of green line is 35% weaker than blue.

As a result, restriction of Japan was 35% economically efficient than benchmark (U.S.)

Benchmark restriction

In Japan, there is no data about how firms lowered the production due to restrictions.
Therefore, we incorporate it from the survey in U.S. (Guan et al. 2020 and Bonadio 2020)
For example, work-at-home rate in 0.134 and exposure level is 0.1 in agriculture. The production reduction is (1-0.134)*0.1 = 0.0866.
All firms in a industry has the same reduction rate.

Lifting restriction in one prefecture

All prefectures impose restrictions but one prefecture is lifted.
How much would its GRP recover?

GRP recovery rate = increase in GRP of pref a when a lifts restrictions / decrease in GRP of pref a when all prefs are restricted
(GRP: Gross Regional Product)

1. Prefs with large GRP have large recovery rate.
2. Prefs with small GRP have diverse recovery rate

Prefecture with high substitutability recover well

If suppliers of firms in pref a are substitutable by firms in pref a, GRP recovery is high.

Blue: Top 10 pref w.r.t. GRP
Red: Middle 37
Black: Bottom 10

Helmholtz-Hodge Decomposition

Decompose network flows into potential and loop flows

Relation between recovery and decomposed flows

A prefecture with strong loop flows recover well
→ Isolated economy is strong

A pref with low potential recover well
→ Low potential means downstream firms. Demand decrease propagates fast but supply is slower than that because of inventories

Blue: Top 10 pref w.r.t. GRP
Red: Middle 37
Black: Bottom 10

Lifting lockdown in two prefectures simultaneously

A pref (a)'s lockdown is lifted with another pref (b)
How much would a's recovery depend on b?

Relative GRP recovery rate = increase in GRP of a with b's lift / increase in GRP of pref a when only a is lifted

Large recoveries are associated with large GRP prefs
If adjacent prefs are important, the red and black colors should have more recovery rate. But blue (large GRP) prefs are dominant

Blue: Top 10 pref w.r.t. GRP, Red: Middle 37, Black: Bottom 10

(a) Potential flows from a to b is positively associated with recovery of pref a

(b) Inversed flows of (a) is also positively associated with recovery of pref a

(c) Loop flows between a and b is positively associated with recovery of pref a

(d) If suppliers to firms in pref a are highly substitutable by firms in pref b, recovery of pref a is high