Table 3.

Heterogeneity by firm size and creditor status.

	By Firm Size				By Firm Size and Creditor Status
	Taxable	Taxable	VAT	VAT	Taxable	Taxable	VAT	VAT
	sales	purchases	liabilities	payments	sales	purchases	liabilities	payments
Treatment (first year) interacted with
$\times$ SME dummy	0.106∗∗∗	0.101∗∗∗	0.116∗∗∗	0.196∗∗∗
	(0.0183)	(0.0197)	(0.0251)	(0.0465)
$\times$ large firm dummy	0.0379∗∗	0.00114	0.0432	−0.109∗∗
	(0.0183)	(0.0204)	(0.0339)	(0.0453)
Treatment (first year) interacted with
$\times$ SME dummy $\times$					0.122∗∗∗	0.0940∗∗∗	0.166∗∗∗	0.343∗∗∗
no VAT credit dummy					(0.0201)	(0.0217)	(0.0249)	(0.0492)
$\times$ SME dummy $\times$					0.0611	0.123∗∗∗	−0.0278	−0.229∗∗
VAT credit dummy					(0.0393)	(0.0420)	(0.0629)	(0.105)
$\times$ large firm dummy $\times$					0.0626∗∗∗	0.0104	0.0844∗∗∗	0.00277
no VAT credit dummy					(0.0214)	(0.0230)	(0.0302)	(0.0504)
$\times$ large firm dummy $\times$					−0.0158	−0.0194	−0.0457	−0.353∗∗∗
VAT credit dummy					(0.0343)	(0.0406)	(0.0839)	(0.0907)
Observations	1,010,380	1,010,380	1,010,380	1,010,380	1,010,380	1,010,380	1,010,380	1,010,380

Note: Results for the balanced sample of firms mandated to adopt e-invoicing before 2019. The inverse hyperbolic sine transformation is applied to all dependent variables, which are originally measured in constant 2014 Peruvian soles. The treatment indicator is equal to one in the quarter of mandatory e-invoicing adoption and the following four quarters. This treatment indicator is interacted with two dummy variables indicating if a firm had sales lower than 2,300 UIT (about $''\$''3$ million) or a positive stock of VAT credits in 2013Q2, respectively. All specifications include quarter fixed effects, firm-specific trends and controls for the effects in the fifth and following quarters after the date of mandatory e-invoicing adoption. Appendix Table A.9, Table A.10 additionally shows estimates with these controls. Firm-clustered standard errors are shown in brackets. ∗ $0.10$, ∗∗ $0.05$, ∗∗∗ $0.01$.