. 2020 Jul 2;15(3):707–740. doi: 10.1007/s11558-020-09385-y

Table 5.

Heterogeneous effects according to sectors and loan sizes

			(1)	(2)	(3)	(4)	(5)	(6)		(7)
	no.^a	mean^a	Loan dummy						Obs		Obs.
VARIABLES			all	1st q.tile	2nd q.tile	3rd q.tile	4th q.tile	no. loans		ln loans
Total	8,772	48.21	0.1712***	0.0179	0.0602*	0.0575	0.1274**	0.7976***	6,442	-0.0256	2,782
			(0.0521)	(0.0421)	(0.0331)	(0.0558)	(0.0625)	(0.2688)		(0.2533)
Agriculture	28	35.62	0.0019	0.0010	0.0011	0.0012	0.0023	32.5068*	108	omitted	17
			(0.0020)	(0.0009)	(0.0011)	(0.0011)	(0.0020)	(16.7606)
Energy	1,812	39.71	-0.0246	-0.0064	-0.0286**	-0.0204	-0.0283	0.6201***	4,667	-0.4318	895
			(0.0346)	(0.0138)	(0.0138)	(0.0222)	(0.0295)	(0.2064)		(0.5369)
Industry	1,609	49.48	0.0091	-0.0275	-0.0429*	-0.0299	0.0431	-0.1088	4,917	0.2310	908
			(0.0431)	(0.0279)	(0.0242)	(0.0251)	(0.0318)	(0.2687)		(0.5254)
Infrastructure	4558	49.61	0.1332**	-0.0017	0.0276	0.0814**	0.1528***	1.2456***	5,968	0.7775***	1,859
			(0.0522)	(0.0288)	(0.0303)	(0.0397)	(0.0539)	(0.3308)		(0.2884)
Non-Market Services	472	70.87	0.0729	0.0027	0.0165	0.0529*	0.0018	0.4562	2,598	-2.2379	341
			(0.0465)	(0.0052)	(0.0125)	(0.0285)	(0.0221)	(0.4374)		(0.6550)
Services	293	36.65	0.0046	-0.0104*	-0.0007	0.0102	0.0157	0.1158	2,494	-0.0313	217
			(0.0329)	(0.0060)	(0.0186)	(0.0095)	(0.0162)	(0.5886)		(0.7934)
Observations			6642	6,642	6,642	6,642	6,642	-		-
Number of regions			266	266	266	266	266	-		-

^a No. corresponds to the number of projects per sector and mean to the mean project size in million Euros after 1990. *** p < 0.01, ** p < 0.05, * p < 0.1. The table presents results for the estimation of the model in Eq. 1. Standard errors are in parentheses and clustered on the level of NUTS 2 regions. The sample start year is 1990 for all regressions. Control variables, region fixed effects, and country-year fixed effects are included in all regressions. The method is OLS for all regressions except column 6, where it is PPML