Keywords

1 Introduction

Electronic participation is defined as the process of involving citizens in politics through ICT, for decision making processes that allow the public administration to be participatory, inclusive, collaborative, and deliberative for intrinsic and instrumental purposes. Electronic participation is a strategic factor to improve citizen participation in digital governance [1], to promote a more efficient society [2], and improve the competitiveness of firms [3]. Despite the above, there is a recurring problem called the “digital divide” which is generally known as “unequal diffusion” between different socioeconomic levels, or countries, gender, age [4, 5], race, region, or geography in terms of access or use [6, 7].

According to the information from “Banca de Oportunidades” [8], 79.1% of adults in Colombia had at least one financial product in a credit establishment, that is, 26.6 million Colombians, 1.8% more with respect to September 2016. Also, as of September 2017, 738,432 adults entered the financial system for the first time and had 106,701 points of access to the financial system, that is, financial offices or banking correspondents.

52% of the population with some financial product were women and the remaining 48% men. The financial product with the highest penetration in the Colombian population as of September 2017 continued to be the savings account, since 73.5% of the adult use this financial product. Within the modalities of savings account, there are deposits and electronic accounts. However, they still have a lower penetration, since, by September 2017, there were approximately 3 million adults with some electronic deposit and 2.8 million actively used them (92.5%).

Electronic savings accounts exist in Colombia since 2007, in accordance with the provisions of Law 1151 of 2007 [9], regulated by Decree 2555 in 2010 [10]. In accordance with the mentioned regulations, electronic savings accounts are aimed at people belonging to level 1 of the Identification System of Potential Beneficiaries of Social Programs -Sisben-, and displaced people registered in the Single Registry of Displaced Population, and meeting the following conditions:

  • Transactions may be made through cards, cell phones, ATMs or any other means and distribution channel that is agreed upon.

  • The entity recognizes an interest rate.

  • The credit establishments and the authorized cooperatives cannot charge the holders for the management of the account or for one of the means enabled for their operation. At least two (2) withdrawals in cash and a balance inquiry made by the customer per month do not cause any payment.

  • Clients must be informed of the benefits, the cost of transactions, or additional queries.

  • No minimum initial deposit nor minimum balance is required for opening accounts.

  • The debits cannot exceed the two (2) legal monthly minimum wages in force.

  • Only one (1) electronic savings account may be held in the financial system.

  • The accounts are not subject to any type of mandatory investment.

According to the above, the purpose of this paper is to apply a multivariate spatial statistics technique: cluster analysis or spatial agglomeration, to classify departments into groups based on behavior in the number of active electronic savings accounts, inactive, and active traditional savings accounts. The importance of addressing this issue lies in the fact that in Colombia, it was stated that based on statistics on cell phone use, the simplest form of banking for the population was the use of electronic savings accounts. Despite this, the figures indicate that it is a financial mechanism that has had low penetration and in most of the national territory it continues to prefer the traditional savings account.

2 Method

2.1 Data

In this work, the Banca de las Oportunidades reports (2018a, 2018b) are used as the primary source. The reports contain information at the departmental level, as of September 2017, on the total of electronic and traditional savings accounts1. The source of the statistics “Banca de las Oportunidades”, is a program of the Government of Colombia, administered by Bancoldex, whose main objective is to promote financial inclusion in Colombia, through access to financial services to families in poverty, small entrepreneurs, small and medium-sized companies and entrepreneurs. All this, in accordance with its mission, to reduce poverty, promote social equality and stimulate economic development in Colombia.

Table 1 shows the data, sources and units used for the following sections. The software used for calculating the models corresponds to Philcarto.

Table 1. Sources and database units used.
Full size table

2.2 Spatial Autocorrelation

For the calculation of spatial autocorrelationFootnote 1 we use the univariate Global Moran index [13,14,15] for each of the variables in Table 1. Given the set of spatial entities and an associated attribute (variable), this index evaluates whether the pattern expressed is grouped, scattered or randomized. The index oscillates between −1 (strong negative spatial autocorrelation) and 1 (strong positive spatial autocorrelation), if it is close to 0 it indicates the absence of spatial autocorrelation. It is interpreted as moderate spatial autocorrelation from 0.3, up to 0.6 and thereafter a high spatial autocorrelation. The formula is in Eq. 1.

$$ I = \frac{{N\sum \left( 2 \right)W_{{y\left( {Y_{J - {\bar{Y}}} } \right)\left( {Y_{J - {\bar{Y}}} } \right)}} }}{{S_{0} \mathop \sum \nolimits_{t - 1}^{N} \left( {Y_{J - {\bar{Y}}} } \right)}} $$
(1)
\( W_{y} \)::

It is the element of the matrix of special weights, corresponding to the \( \left( {i,j} \right) \)

\( Y \)::

Average expected value of the variable \( y \)

\( N \)::

Number of observations or sample size

\( S_{o} \)::

Corresponds to: \( S_0 = \sum i\sum i W_{j} = \sum 2 W_{j} \).

2.3 Space Agglomeration

The purpose of this section is to apply a hierarchical ascending classification model (interdependence method), to identify which behaviors allow to create group clusters in conglomerates according to the number of electronic savings accounts, traditional savings accounts, and balances in millions of pesos. This in order to establish in which departments could be presented focalization programs that encourage the use of electronic accounts and improve the mechanisms of bancarization in Colombia.

Based on the data, an empirical study of the ex post facto type was carried out, using a cluster analysis (Gallardo, 2015, p. 1). The Minimum Distance strategy is used (Langrand and Pinzón, 2009 and Gallardo, 2015). In this way, when performing stage K, n-K clusters are obtained, and the distance between the clusters Ci (with ni elements) and Cj (with nj elements) would be:

$$ d\left( {C_{i} ,C_{j} } \right) = \mathop {\text{Min}}\limits_{\begin{subarray}{l} x_{l} \in C_i \\ x_{m} \in C_j \end{subarray} } \left\{ {d\left( {x_{l} ,x_{m} } \right)} \right\} \quad l = 1, \ldots ,n_{i} ;m = 1, \ldots ,n_{j} $$
(2)

In this case, the clusters Ci and Cj will be joined:

$$ \begin{aligned} d\left( {C_{i} ,C_{j} } \right) & = \mathop {\text{Min}}\limits_{{\mathop {i_{1} ,j_{1} = 1, \ldots ,n - k}\limits_{{i_{1} \ne j_{1} }} }} \left\{ {d\left( {C_{{i_{1} }} ,C_{{j_{1} }} } \right)} \right\} \\ & = \mathop {\text{Min}}\limits_{{\mathop {i_{1} ,j_{1} = 1, \ldots ,n - k}\limits_{{i_{1} \ne j_{1} }} }} \left\{ {\mathop {\text{Min}}\limits_{{\mathop {x_{l} \in C_{{i_{1} }} }\limits_{{x_{m} \in C_{{j_{1} }} }} }} \;\left\{ {d\left( {x_{l} ,x_{m} } \right)} \right\}} \right\}\quad l = 1, \ldots n_{{i_{1} }} \,;\,m = 1, \ldots ,n_{{j_{1} }} \\ \end{aligned} $$
(3)

3 Results

Figures 1, 2 and 3 show the departmental behavior of the use of traditional and electronic savings accounts. As shown in Fig. 1, the differences between traditional and electronic savings accounts are significant, especially in the western part of the country. Also, in the Figs. 2 and 3Footnote 2, it is evident that there is no degree of special autocorrelation between levels 1 and 7. For all cases, it is a random spatial distribution, considering the variables independently.

Fig. 1.
figure 1

Comparison of electronic and traditional savings accounts

Full size image
Fig. 2.
figure 2

Distribution of active electronic savings accounts in Colombia

Full size image
Fig. 3.
figure 3

Distribution of traditional active savings accounts in Colombia

Full size image

When applying the hierarchical ascending classification, the following results were obtained, using the Philcato software, specialized in cartographic analysis. The histogram shows that the significant number of classes is 6 for the 33 spatial units analyzed (departments), explaining 96.94% of the inertia. Figures 4 and 5 show the results of the departments grouping in 6 different significant classes. This grouping explains 96.94% of the data variability, which implies an excellent explanatory level. Figure 6 contains the spatial representation of the departments and is detailed in Table 2. The importance of this analysis lies on the fact that differentiated strategies can be applied, according to the characteristics of each statistically identified group, and not to a homogeneous policy.

Fig. 4.
figure 4

Tree diagram of the clusters.

Full size image
Fig. 5.
figure 5

Diagram of average distance of each class with respect to the general average.

Full size image
Fig. 6.
figure 6

Spatial representation of the application of hierarchical ascending classification to the number of savings accounts in Colombia (September 2017).

Full size image
Table 2. Spatial agglomeration with respect to the number of savings accounts in Colombia (September 2017).
Full size table
figure a

In space units 1 and 5, given that electronic savings accounts have been more popular than in the rest of the country, it is worth focusing efforts with specific programs, which further encourage the use of this type of financial mechanisms. In groups 3 and 4 there is a clear preference for traditional savings mechanisms, which is striking, since most of these departments have good access to the Internet. Finally, it is important to note that in none of the cases is there a simultaneous behavior of being below the national average in electronic and traditional savings accounts. This indicates that at least some of the mechanisms are preferred and there is no evidence of a process of unbanking, at least in terms of saving.

4 Conclusions

In the departments of San Andrés and Providencia, Putumayo, Valle del Cauca, Quindío, Risaralda, Antioquia, Córdoba, Santander, Boyacá and Atlántico, the strategy of active electronic accounts as a financial inclusion mechanism has not been favored, since a preference for traditional accounts persists. For its part, the departments of Nariño, Caquetá, Guaviare, Cundinamarca, Tolima, Bolívar, Magdalena, are in an intermediate stage, since the number of traditional accounts is just a little higher than the average and the number of active electronic accounts, despite being below average, tend to increase.

On the other hand, the departments where the electronic accounts are most used are Cauca, Bogotá, Meta, Casanare, Arauca, Vichada, Huila, Amazonas, Guainía, Vaupés, Caldas, Chocó, Sucre, La Guajira, César, Norte de Santander. These findings are evidence of a breach within the country of the penetration of ICT and the use of alternative banking products such as electronic deposits.

The findings indicate that although it was proposed that electronic savings accounts were one of the mechanisms that would facilitate banking penetration, after 10 years of having been created by national legislation, they do not have a high penetration in the Colombian population, nor they have a homogeneous distribution in the national territory.

To reach the goal of 84% of the banking population in Colombia in 2018, more attractive mechanisms or programs are required for the population, which respond to the needs of users and whose costs are not so high.