CAF - Corruption, Custom Reform and Firm Growth

CORRUPTION, CUSTOM REFORM AND FIRM GROWTH: QUASI-EXPERIMENTAL

EVIDENCE FROM COLOMBIA

15/11/2017 N° 2017/25

Laajaj, R. Eslava, M. Kinda, T.

CORRUPTION, CUSTOM

REFORM AND FIRM

GROWTH: QUASIEXPERIMENTAL EVIDENCE FROM COLOMBIA2 CORRUPTION, CUSTOM REFORM AND FIRM GROWTH: QUASI -EXPERIMENTAL EVIDENCE FROM COLOMBIA

CORRUPTION, CUSTOM REFORM AND FIRM GROWTH:

QUASI-EXPERIMENTAL EVIDENCE FROM COLOMBIA

Laajaj, R. Eslava, M. Kinda, T.

CAF – Working paper N° 2017/25 15/11/2017

ABSTRACT

Customs are often prone to corruption because it concentrates a lot of discretionary power in the hand of custom agents who take decisions with high economic stakes for the fi rms, providing an opportunity for custom agents to extract a rent from the firms. Communication technologies offer the possibility to limit this discretionary power by reducing direct interactions between firms and custom agents. Combining firm level panel data on about 6,000 manufacturing firms with custom level data, we assess the effects on firm level outcomes of a computerization of import transactions that occurred sequentially in the 26 Colombian customs between 2000 and 2005. We apply a triple difference strategy that makes use of the variation between customs, time and the firms' exposure to the reform, based on whether it was an importing firm before it started. We find large effects of the computerization of the custom on the growth of importing fi rms' inputs,

We apply a triple difference strategy that makes use of the variation between customs, time and the firms' exposure to the reform, based on whether it was an importing firm before it started. We find large effects of the computerization of the custom on the growth of importing fi rms' inputs, investments and value added. We also provide evidence of a large increase of imports declared, taxes collected, and a reduction in corruption cases following the custom reform.

Small sections of text, that are less than two paragraphs, may be quoted without explicit permission as lo ng as this document is stated. Findings, interpretations and conclusions expressed in this publication are the sole re sponsibility of its author(s), and it cannot be, in any way, attributed to CAF, its Executive Directors or the countries they represent. CAF does not guarantee the a ccuracy of the data included in this publication and is not, in any way, responsible for any consequences resulting from its use.

© 2017 Corporación Andina de Fomento3 CORRUPTION, CUSTOM REFORM AND FIRM GROWTH: QUASI -EXPERIMENTAL EVIDENCE FROM COLOMBIA

CORRUPCIÓN, REFORMA ADUANERA Y CRECIMIENTO

DE LAS FIRMAS: EVIDENCIA CUASIEXPERIMENTAL

PARA COLOMBIA

Laajaj, R. Eslava, M. Kinda, T.

CAF - Documento de trabajo N° 2017/25 15/11/2017

RESUMEN

Las aduanas son a menudo propensas a la corrupción porque concentran gran parte del poder discrecional en manos de los agentes aduaneros quienes toman decisiones que implican grandes intereses económicos para las empresas. Esto brinda a los agentes de aduana la

Las aduanas son a menudo propensas a la corrupción porque concentran gran parte del poder discrecional en manos de los agentes aduaneros quienes toman decisiones que implican grandes intereses económicos para las empresas. Esto brinda a los agentes de aduana la oportunidad de extraer rentas de las empresas. Las tecnologías de comunicación permiten limitar este poder discrecional al reducir las interaccio nes directas entre las empresas y los agentes aduaneros. Combinamos datos de panel a nivel de firma para 6.000 empresas manufactureras con los datos de aduana para evaluar los efectos de una informatización de las transacciones de importación que se aplicó secuencialmente en las 26 aduanas colombianas entre 2000 y 2005 sobre resultados a nivel de firma. Aplicamos una estrategia de triple diferencia que aprovecha la variación entre aduanas, tiempo y exposición de las empresas a la reforma, en función de si s e trataba de una empresa importadora o no antes de la política. Encontramos grandes efectos de la informatización en el crecimiento de la cantidad de insumos, de las inversiones y del valor agregado de las empresas importadoras. También proporcionamos evid encia de un gran aumento de las importaciones declaradas y los impuestos recaudados, y de una reducción en los casos de corrupción tras la reforma aduanera.

Small sections of text, that are less than two paragraphs, may be quoted without explicit permission as lo ng as this document is stated. Findings, interpretations and conclusions expressed in this publication are the sole responsibility of its author(s), and it cannot be, in any way, attributed to CAF, its Executive Directors or the countries they represent. CAF

document is stated. Findings, interpretations and conclusions expressed in this publication are the sole responsibility of its author(s), and it cannot be, in any way, attributed to CAF, its Executive Directors or the countries they represent. CAF does not guarantee the a ccuracy of the data included in this publication and is not, in any way, responsible for any consequences resulting from its use.

© 2017 Corporación Andina de FomentoCorruption, Custom Reform and Firm Gro wth: Quasi-Exp erimen tal Evidence from Colom bia Rac hid Laa ja j ∗, Marcela Esla v a †& Tidiane Kinda ‡Ÿ No v em b er 15, 2017 Abstract Customs are often prone to corruption b ecause it concen trates a lot of discretionary p o w er in the hand of custom agen ts who tak e decisions with high economic stak es for the rms, pro viding an opp ortunit y for custom agen ts to extract a ren t from the rms. Comm unication tec hnologies oer the p ossibilit y to limit this discretionary p o w er b y reducing direct in teractions b et w een rms and custom agen ts. Com bining rm lev el panel data on ab out 6,000 man ufacturing rms with custom lev el data, w e assess the eects on rm lev el outcomes of a computerization of imp ort transactions that o ccurred sequen tially in the 26 Colom bian customs b et w een 2000 and 2005. W e apply a triple dierence strategy that mak es use of the v ariation b et w een customs, time and the rms' exp osure to

the reform, based on whether it w as an imp orting rm b efore it started. W e nd large eects of the computerization of the custom on the gro wth of imp orting rms' inputs, in v estmen ts and v alue added. W e also pro vide evidence of a large increase of imp orts declared, taxes collected, and a reduction in corruption cases follo wing the custom reform. 1 In tro duction Corruption is b eliev ed to tak e a toll on most of the dev eloping w orld and ha v e consequences b ey ond a simple transfer to corrupt individuals. It can discourage in v estmen t (Samphan tharak and Malesky, 2008) and capital accum ulation (F erraz et al., 2012; Reinikk a and Sv ensson, 2004), lead to the misallo cation of capital (Kh w a ja and Mian, 2005), or talen ts (Eb ek e et al., 2015). It aects the economic viabilit y of in v estmen ts in the priv ate sector (Sv ensson, 2003) and of public exp enditures Olk en (2006), and harms ∗Univ ersit y of Los Andes, corresp onding author r.laa ja j@uniandes.edu.co †Univ ersit y of Los Andes ‡In ternational Monetary F und Ÿ W e thank particpan ts at the ...Univ ersidad T orcuato di T ella, the 4th DIAL Dev elopmen t Conference in P aris Dauphine, the 16th Journées Louis-André Gérard-V aret at the at the Aix-Marseille Sc ho ol of Economics, for their commen ts. W e thank

Daniel Espinoza Castro for his relen tless w ork as a researc h assistan t. W e are grateful to D ANE and DIAN for making their data a v ailable to us, and to the Banco de Desarollo de America Latina for the funding. 1the go v ernmen t's abilit y to correct externalities (Olk en and Barron, 2009) 1 . Customs in particular are a h ub for corruption. In unreformed customs, rev en ue collections rely mostly on ph ysical insp ections. T ax rates v ary b y t yp e of go o d, whic h ma y b e misclassied, in ten tionally or not. Hence custom agen ts ha v e a cen tral role in inuencing the amoun t to b e c harged, the sp eed of con trols and other decisions with costly consequences for the rm. As explained b y Sequeira (2015), customs ocers ha v e signican t discretion to stop cargo and a broad bureaucratic to olkit from whic h to dra w on to justify dela ys and unan ticipated h urdles. This discretionary p o w er creates an opp ortunit y for ren t extraction, whic h can act as an additional tax for the rms. And y et w e kno w relativ ely little ab out the eects of corruption at customs on the econom y . A large literature aims at understanding ho w trade aects the econom y , traditionally fo cusing on the eect of tari and non-tari barriers. Sac hs et al. (1995),Dollar et al. (2003) and W acziarg and W elc h (2008) nd evidence that trade increases gro wth and reduces p o v ert y , whereas Ro drik et al. (2004)

claims that once con trolling for the qualit y of institutions, trade has an insignican t or negativ e eect on income lev els. More recen tly , rm lev el studies emphasized the b enets of trade lib eralization for lab or pro ductivit y and gro wth in the Canada-US free trade agreemen t (T reer, 2004), Armenia (Halp ern et al., 2011) or Indonesia Amiti and K onings (2007), explaining the ndings with learning, v ariet y , and qualit y eects of the c heap er imp orted inputs. Using the same Colom bian man ufacturing plan ts census data as the one used in this pap er, Kugler and V erho ogen (2009) pro vide evidence that inputs on the imp ort mark et are higher qualit y than the ones on the domestic mark et, and they are used b y rms that are larger and more pro ductiv e. Giv en the abundan t evidence of the negativ e eects of imp ort barriers on rms economic activit y , if corruption acts as an additional tax for rms, then it should lead to similar eects. Ho w ev er, the eect of corruption on trade costs remains relativ ely unexplored and can go in b oth directions dep ending on whether it is collusiv e or co erciv e (Sequeira, 2015). Y et a n um b er of studies p oin t to w ards costs of corruption that ma y b e greater than the ones of an ocial tax. Analyzing data from 45 host coun tries,

W ei (2000) conclude that an increase in the corruption lev el from that of Singap ore to that of Mexico discourages in w ard FDI b y as m uc h as a t w en t y p ercen tage p oin ts raise in tax. Using Ugandan rms data, Sv ensson (2003) nds that a one p ercen tage p oin t increase in brib es reduces ann ual rm gro wth b y three p ercen tage p oin ts, whic h is three times the negativ e impact of equiv alen t formal taxes on rm p erformance. If the same pattern holds at the customs then the customs ocers use of their discretionary p o w er to extract a ren t w ould ha v e disprop ortionate eects on the econom y . Some studies ha v e tak en a closer lo ok at the role of customs in trade facilitation. Using the W orld Business En vironmen t Surv eys across 80 coun tries, Batra et al. (2003) iden tify customs and foreign trade regulations as the second most serious tax and regulatory constrain t. Guasc h and K ogan (2001) nd that in v en tory holdings in man ufacturing rms in dev eloping coun tries are 3 to 6 times higher than the ones in the United States, in part b ecause of the uncertain t y in the timing and cost of deliv ery , and they predict 1 F or an o v erview of the empirical literature on corruption in dev eloping coun tries see Olk en and P ande (2012) 2that in dev eloping coun tries, halving in v en tories could reduce pro duction costs b y 20 p ercen t. Clark et al.

(2004) nd that p ort eciency is an imp ortan t determinan t of shipping costs. It sho ws that p orts with p o or infrastructure and cum b ersome regulations, are equiv alen t to b eing 60% farther a w a y from mark ets for the a v erage coun try . (Y ang, 2006) sho ws that hiring priv ate rms for pre-shipmen t insp ections of imp orts increases tari rev en ues. Dollar et al. (2003) compare rms across 8 dev eloping coun tries to nd that exp ortation is negativ ely aected b y custom clearance time. Customs are a strategic place of whic h the eciency has high stak e for the econom y . Despite signican t trade lib eralization o v er the last decades, rev en ues collected b y customs remain a k ey comp onen t of public resources in man y dev eloping economies and is exp ected to remain so for the foreseeable future. A dequately collecting this rev en ue is therefore an essen tial task for the coun try , but equally imp ortan t is to ensure that collection do es not act as an additional tax b y imp eding trade o ws. Information tec hnologies can reduce the discretionary p o w er of customs ocers b y limiting their interaction with rms. A w are of it, the Arusha Declaration of the W orld Customs Organization Concerning Go o d Go v ernance and In tegrit y in Customs (1993) puts automation of transactions in customs as one of the k ey pillars to reduce corruption. Computerization of custom transactions has b een adopted b y the v ast ma jorit y of OECD coun tries, and it is commonly recommended b y in ternational organizations (suc h

as the IMF, UNCT AD or the In ter-American Cen ter for T ax A dministrators) whic h ha v e in v ested in the pro vision of training and soft w are for the computerization of custom transactions in middle and lo w income coun tries. A n um b er of cases studies do cumen t the b enets and c hallenges of custom computerization, suc h as the ones of Moro cco (De W ulf and Sok ol, 2004) and Angola (Engman, 2005), p erceiv ed as some of the most successful examples. Ho w ev er, these reforms are not exogenous to other p olicies and institutions and rarely come on their o wn, adding to the iden tication c hallenge. Finally trac king the eect on rms economic activit y requires ric h data, hence the need to bring rigorous evidence to this debate. This P ap er examines the eects of a custom reform on the economic activit y of man ufacturing rms in Colom bia. The DIAN (Spanish acron ym for Direction of T axes and National Customs) computerized the imp ortation pro cess. After the reform, imp orters w ould declare their imp orts online rather than at arriv al at the custom, whic h limits the in teraction with the customs ocer and impro v es trac king of eac h transaction, from transp ortation in the foreign coun try to pa ymen t of the taxes. The explicit in ten tion of the reform w as the reduction of opp ortunities for customs ocers to exploit their decision p o w er to extract a ren t from imp orters. It aimed at reducing bureaucracy , allo wing the custom to ecien tly apply the necessary con trols while reducing imp edimen ts to the rms.

The limited capacit y of the team that put in place the computerization led to a sequen tial computerization of the reform, from the b eginning of 2000 (b efore an y imp ortation w as computerized) to the end of 2005, when more than 99% of transactions w ere computerized (see gure 1). Figure 2 presen ts the share of imp orts that w ere digital in eac h one of the 15 largest customs in Colom bia. W e see that the y ear of computerization v aries b y custom and that computerization is c haracterized b y a sudden jump from 3none to almost all transactions in the custom b eing computerized. Of course the computerization is asso ciated with some degree of reorganization and c hanges in resp onsibilities of customs ocials (including a reduction of in teraction and narro wing of the role of custom agen ts), and the reform that is assessed in this pap er includes this en tire set of c hanges made p ossible b y the computerization. This sequen tial implemen tation oers a unique opp ortunit y to isolate the eects of the reform on the econom y . It b ears little risk to b e confounded with an y other c hange in p olicy or taxation, whic h w as implemen ted at the coun try lev el and w ould th us b e captured b y y ear xed eects in the analysis. T o ev aluate the eect of the reform, w e com bine y early panel census data on (appro ximately 40,000) man ufacturing plan ts in Colom bia, with administrativ e data from the DIAN, co v ering all imp orts going

through the 26 customs of Colom bia, from 1997 to 2008. W e assign eac h man ufacturing plan t to its closest custom (dened as the custom that w as most used for imp ortations b y rms from the same m unicipalit y prior to treatmen t). W e consider that an imp orting rm is treated once the assigned custom underw en t the computerization. T o analyze the role of corruption as a c hannel w e also matc hed it with administrativ e data o v er the same p erio d, from the Pr o cur aduría Gener al de la Nación (National General Sup erin tendency), whic h gathers all in v estigated cases, from whic h extracted the cases that w ere related to corruption at the DIAN (a total of 32 cases during the study p erio d). The k ey c hallenge to a prop er iden tication of the eects of the reform is that the timing of computerization of eac h custom is related to mo dernization capacit y , whic h is lik ely to b e correlated with the pro ductivit y of the plan ts that are closer to the custom. T o address this issue, w e use a triple dierence based on customs, time and rms exp osure to the treatmen t. The plan t's exp osure to the treatmen t dep ends on whether they w ere imp orting in 1999 (b efore an y computerization). Using this approac h, if w e nd that, in the y ears follo wing the reforms, imp orting rms tend to gro w faster than non-imp orting rms, then it w ould b e attributed to the reform. This relies on the assumption that the date of the

custom reform is orthogonal to the dierence in what the gro wth b et w een imp orting and non-imp orting asso ciated plan ts w ould ha v e b een without the reforms. This w ould b e violated if the DIAN prioritized not only areas with higher gro wth p oten tial, but with a greater dierence in gro wth p oten tial b et w een exp orters and non-exp orters. Ho w ev er the parallel trend analysis do esn't reject that the trends in the dieren t outcomes observ ed w ere similar b efore the reform. Hence for this to b e a concern, one w ould need to assume that the DIAN had access to some information ab out gro wth p oten tial that go es b ey ond what w e w ere able to capture in our analysis of the ric h census data on all man ufacturing rms and all imp ortations. The empirical strategy shares similarities with Amiti and K onings (2007), whic h claims to b e the rst study to isolate the eects of reducing input taris on imp orting plan ts from other plan ts. W e nd that the custom reforms led to a gradual and substan tial increase in the man ufacturing rms v alue added, input use, sales and capital. Lab or pro ductivit y and the prop ortion of exp orting rms also increased signican tly . The custom lev el data sho w that the reform led to a large increase in the n um b er of imp ortations that go through the custom, and in the total ann ual v alue of imp orts declared. The total of taxes increased as a result of the increase in imp orts, but the a v erage tax rate did not c hange.

4Smoking gun evidence of the corruption c hannel is alw a ys c hallenging, ho w ev er m ultiple evidence p oin t to w ards it. First, in the comp osition of taxes, w e nd a reduction in sanctions and other an in the dierence b et w een the amoun t that is due and the amoun t that is actually paid b y the imp orter. These three elemen ts are less predictable and more sub ject to the discretionary p o w er of the customs ocers. They ma y just b e the tip of the iceb erg, the carrot and stic k for customs ocials to extract brib es from imp orters. W e nd an increase in transp ortation costs of merc handises reac hing treated customs (b oth in absolute lev els and as a share of the v alue of imp orts). It rev eals that imp orting rms are willing to tra v el further in order to go through the computerized customs. This is in line with Sequeira and Djank o v (2014), who sho w that rms in Southern Africa are willing to tra v el signican tly larger distances in order to go through a less corrupt p ort. Finally , w e also sho w that the custom reforms are follo w ed b y a signican t drop in the n um b er of judiciary cases related to corruption at DIAN in the corresp onding m unicipalit y . T o the b est of our kno wledge, there is curren tly no solid empirical evidence on ho w automatization of transactions at customs aects the imp orting rms. This pap er pro vides no v el evidence ab out the

ineciency costs of corruption at customs, but also pro vides an example of a successful, practical solution to tac kle this complex issue. It ts in to a gro wing literature on the p oten tial for information tec hnologies to impro v e eciency and reduce corruption ((Giné et al., 2012; Muralidharan et al., 2014; Banerjee et al., n.d.; Lewis-F aup el et al., 2016)). A theoretically in teresting feature of this in terv en tion is that, b y con trast with the studies men tioned, the b enets seem to result from cutting a harmful comm unication link rather than an increase in the information o w. In the case of Colom bia, the cost of the computerization is dw arfed b y the observ ed b enets on the econom y . This can serv e as an example for a large n um b er of lo w income coun tries that ha v e not y et undergone a similar reform. Of course replicabilit y in dieren t con text is not a giv en, De W ulf and Sok ol (2004) argue that p olitical will and institutional capacit y are strong determinan ts of its implemen tation and success. F uture in v estigation should explore the conditions for successful implemen tation of similar reforms. 2 Con text and the Computarization of Customs in Colom bia 2.1 The Colom bian Con text Colom bia is a Middle income coun try with a Human Dev elopmen t Index in 2014 that ranks 97th out of the 188 coun tries for whic h it is a v ailable. The coun try has suered from an armed conict since the

1960s, whic h escalated in the 1980's and 90's, but then decreased from 2005 on w ard. It remains one of the coun tries with the highest lev el of inequalities; the Human Dev elopmen t Rep ort of 2013 giv es Colom bia the 11th highest Gini out of the 138 coun tries for whic h the data is a v ailable. F rom 1990 to 2015, the GDP p er capita of the coun try increased b y ab out 2.3% p er y ear, and the trade op enness as a share of 5GDP has sligh tly increased from ab out 33% to 39%. In 2015, the man ufacturing sector represen ted 11% of the v alue of Colom bian GDP . 2 Colom bia has its share of corruption and go v ernance issues, fueled among other things b y money from narco-trac king. Ho w ev er, during the p erio d of the study , its ranking according to the W orld Go v ernmen t Indicators increased from 34th p ercen tile of coun tries for whic h the data are a v ailable in 2000 to the 41st p ercen tile in 2006 (and reac hing the 45th p ercen tile in 2015) with the greatest impro v emen ts seen in the Rule of La w and Con trol of Corruption indexes. The mo dernization of customs w as one of the reforms tak en b y the Colom bian Go v ernmen t during this p erio d in order to tac kle Corruption. T o illustrate this, Figure 3 sho ws that the a v erage custom clearance time for imp ortation w en t do wn from ab out 17 da ys in 1995 to a bit less than 10 da ys in 2014. The general impro v emen t in Colom bian custom facilitation,

trade and institution are compatible with the main ndings of this pap er, and raise the imp ortance of dra wing lessons from the p olicies that ma y ha v e generated this progress. 2.2 Description on the Custom Reform and Qualitativ e Evidence The decision to fo cus on the computerization rst comes from discussion with the managemen t and op erations of the DIAN. When ask ed ab out the c hanges that w ere most inuen tial to tac kle corruption and facilitate the transits for the rms, the computerization w as brough t forw ard as a game c hanger in customs abilit y to facilitate transactions and to monitor them. The reform allo w ed more c hec ks, while reducing the in teraction with the customs ocer, hence reducing opp ortunities for the customs ocers to exploit their decision p o w er to extract a ren t from the imp orter. The new program w as named Siglo XXI. The DIAN started its dev elopmen t with the pro ject Sistema Global de Información y Mejoramien to de la Gestión A duanera Siglo XXI in 1997, and its rst implemen tation w as in the y ear 2000. It allo w ed users to declare their imp orts online rather than at arriv al at the customs. F or the imp orting rm, an extended duration at the b order can ha v e large economic costs due to the immobilization of merc handises. Moreo v er, it can also trigger sanction fees c harged to the imp orter b y the custom, further increasing the negotiating p o w er of customs ocers. An explicit goal of the computerization of the imp ortation pro cess is to tak e a w a y some of the decisions from the customs ocers who directly in teract with the imp orter,

in order to increase eciency and reduce corruption opp ortunities. The in ternal do cumen tation mak es it clear that the reform go es b ey ond a digitalization of the declaration. It emphasizes the need to bring the tec hnological strategy in line with dev elopmen ts in regulation and pro cedures. F or this reason, this study m ust b e in terpreted as an ev aluation of the en tire reform that w as made p ossible b y the computerization. The main in ternal do cumen t describing the pro ject, pro duced in 2001, is named A dv ances in the Con trol of Corruption Through the Customs System Siglo XXI. It includes a summary table with a remark ably detailed thinking of the prior a ws that w ere lea ving ro om for corruption, and ho w the reform 2 "NA TIONALES TRIMESTRALES -PIBComp osición del PIB Colom biano p or demanda y Comp osición del PIB Colom biano Oferta". dane.go v.co 6w ould address it. App endix 1 pro vides a translation of this table, rst b ecause it pro vides evidence that corruption w as a cen tral goal of the reform, and that qualitativ e evidence p oin ts to w ards a n um b er of c hannels through whic h the reform can ha v e addressed it. Second, it pro vides an example for practitioners in terested in the details of the implemen tation of suc h a reform. Belo w w e highligh t some k ey features ab out ho w, in 2001, the DIAN b eliev ed that the reform w as tac kling corruption: • Before the reform, The only criteria to insp ect the c harge w as the judgmen t of the customs insp ector, but after the reform, the p ossibilit y of corruption acts has b een diminished since the transp orter in teracts with the customs ocer only when the computer system determines it • Prior to the reform, there w as a risk of a cargo not b eing presen ted to the Custom oce, or that the quan tities declared are b elo w what w as imp orted. After the reform, customs is automatically informed of deliv eries. • The risk that the customs insp ector declare conformit y b et w een the cargo manifest and the ph ysical cargo when there is a dierence that should cause its apprehension w as reduced b y the full con trol and iden tication of the actors who p erform eac h one of the c hec ks generating traceabilit y and b etter accoun tabilit y from the customs ocers, who can b e concerned ab out a risk that the customs authorit y v eries the corresp ondence b et w een the form and the real cargo. • In the previous sc heme, there w as no transmission of the message with pa ymen t notication b y the Bank, whic h generated m ultiple problems that w ere dicult to detect b efore the release (falsications, adulteration, etc.) whereas the in the new sc heme, the DIAN receiv es the online conrmation of pa ymen ts and once the pa ymen t is made, the release request is executed without an y in terv ention b y customs authorities , ab out 86% of the merc handise obtain automatic release compared

to 30 to 40% b efore the reform. • Siglo XXI allo w ed the instan taneous registration, and comparison of declarations from the transp orter, the imp orter, the w arehouse and the bank. the discretion of the ocials for the selection of do cumen ts for insp ection purp oses is completely eliminated. The ph ysical or do cumen tary insp ection pro cesses are triggered b y inconsistencies and risk proles rather than arbitrary decisions of customs ocers. In the new sc heme, the custom executes ph ysical insp ections to not more than 9% of the do cumen ts. This is an impro v emen t since b efore, the rate w as greater than 50%. Automatic risk proles w ere already in place in 2001, but man y additional elemen ts w ere exp ected to b e progressiv ely incorp orated. • Eciency gains from the rapid pro cessing of information and a v oiding duplication of pro cesses (suc h as t yping) for users • T o ols are pro vided so that users can do managemen t con trol of their pro cesses 7The rep ort concludes that the pro ject's main impacts w ere: • Decrease of the discretion in the exercise of the customs function, through the automation of a large part of the con trols and the exercise of ph ysical reviews based on risk proles, and with the supp ort of the computer application, pro viding greater transparency in the pro cedures • Con v enience [...] with the exception of the ph ysical insp ection and pa ymen t, all the pro cesses

related to the declaran t's obligations can b e dev elop ed from its managemen t cen ter through the In ternet • reduce the p ossibilit y of errors that imply indirect costs to the user suc h as onerous sanctions or dela ys in the pro cedures • Impro v emen t in the con trol of the fulllmen t of the obligations b y the customs authorit y and sp ecically those of declaration and pa ymen t of t