The drivers and impacts of subsidies to steel firms

Annex F. Subsidies impact on capacity - multivariate regression results

Results from estimations are presented below.

Table A ‎F.1. Capacity regressions, all sample

	Equation 1	Equation 2: without trend	Equation 3: lagged	Equation 4: lagged and without trend
max⁡(D.grants,0)	10.507(***)	11.985 (***)	8.704 (**)	10.117 (***)
- min(D.grants,0)	.	.	.	.
max⁡(D.BMB,0)	.	.	.	.
- min(D.BMB,0)	.	1.474 (**)	2.342 (***)	2.748 (***)
D.(debt / asset)	.	.	.	.
D.revenue	0.215 (***)	0.213 (***)	.	.
D.EBITDA	-0.180 (*)	-0.181 (*)	0.206 (*)	0.212 (**)
D.GDPsteel	-2.39e-06 (***)	-1.92e-06 (***)	.	.
Profit margin	.	.	-7172 (**)	-7772 (***)
trend	358.49 (***)		348.192 (**)
Sample size	646	646	603	603
R2 stat	11%	13%	5%	8%

Note: Using grants instead of its positive and negative parts separately also result in a significant coefficient at the 5% significance level (hence, somehow lower), but also in terms of magnitude the coefficient is halved to 5, instead of the 10.507 above.

Table A ‎F.2. Capacity regressions, OECD Member countries observations

	Equation 1	Equation 2: without trend	Equation 3: lagged	Equation 4: lagged and without trend
max⁡(D.grants,0)	.	.	.	.
- min(D.grants,0)	.	.	.	.
max⁡(D.BMB,0)	.	.	.	.
- min(D.BMB,0)	.	.	.	.
D.(debt / asset)	.	.	.	.
D.revenue	0.129 (***)	0.130 (***)	.	.
D.EBITDA	-0.270 (***)	-0.270 (***)	0.294 (***)	0.295 (***)
D.GDPsteel	.	.	.	.
Profit margin	.	.	.	.
trend	.		.
Sample size	178	178	165	165
R2 stat (adjusted)	10%	9%	6%	0.0073%

Note: Proper inspection of this table results shows that EBITDA is negatively correlated with current capacity changes, but this is mostly due to revenue becoming significant and taking part of the EBITDA coefficient. Re-estimating by dropping the variable “revenue” in those two equations yields a non-statistically significant EBITDA variable (for explaining contemporary changes in capacity), consistent with the absence of link between current EBITDA and current capacity. Overall, we can say that the model does not “work” in any way shape or form for explaining capacity changes in OECD Member countries: none of the proposed variables explain capacity changes.

Table A ‎F.3. Capacity regressions, partner economies

	Equation 1	Equation 2: without trend	Equation 3: lagged	Equation 4: lagged and without trend
max⁡(D.grants,0)	7.448 (**)	9.440 (**)	7.504 (*)	9.505 (**)
- min(D.grants,0)	.	8.844 (*)	.	.
max⁡(D.BMB,0)	.	.	.	.
- min(D.BMB,0)	.	1.473 (**)	2.249 (***)	2.907 (***)
D.(debt / asset)	.	.	.	.
D.revenue	0.280 (***)	0.280 (***)	0.144 (***)	0.142 (***).
D.EBITDA	.	.	.	.
D.GDPsteel	-3.36e-06 (***)	-2.77e-06 (***)	-1.70e-06 (*)	.
Profit margin	.	.	-6838 (**)	-7833 (**)
trend	542.24 (***)		564.28 (***)
Sample size	468	468	438	438
R2 stat	16%	18%	7%	8%

Note: In Equation 1 D.EBITDA is barely significant at the 10% level (p-value of 10.05%) and becomes non-significant once the equation is re-estimated without the other non-significant variables, and this holds for all 4 equations estimated above.