Determinants of Banking Profitability in Portugal and Spain: Evidence with Panel Data

Abstract

This article aims to study the determinants of banking performance in the countries of the Iberian Peninsula, Portugal and Spain. To achieve the proposed objective, the methodology of panel data was used, specifically the estimation method Generalized Method of Moments (GMM-system). An unbalanced panel of 267 banks was used, of which 122 belong to the Portuguese banking sector and 145 to the Spanish banking sector. Two variables were used as performance measures, the average return on total assets (ROAA) and the average return on equity (ROAE). The results show that bank profitability is generally influenced by internal variables, and not so much by sector-specific or macroeconomic variables. Therefore, the results suggest that management decisions are the ones that most influence performance. We conclude that bordering countries, despite having different economies, have very similar influences on bank profitability.

1. Introduction

Financial institutions and in particular banks capture savings from economic agents that have higher levels of liquidity to lend to those that lack liquidity [1]. When these transactions are efficient, the economy and the financial sector of the countries tend to become more solid and stable [2].

According to the financial literature ([2, 3] among others), banking performance is affected both by internal determinants and by factors external to the bank. Thus, it is consensual that the internal factors result from policies applied by their managers. Meanwhile, the external determinants that, as they are exogenous to the institution, are not within the reach of the bank’s direction and management. However, they can be predicted. And if the external factors are anticipated by the banks, they will be able, on time, to face the less favorable situations.

External factors can also, according to the literature ([2, 3], among others), be divided into two categories, industry-specific factors, and macroeconomic factors. These variables are determined by characteristics inherent to the types of institutions, as well as the economic and legal environment of the country in question [4].

Thus, through the estimation technique used by Arellano and Bond [5], Arellano and Bover [6], and Blundell and Bond [7], the Generalized Method of Moments (GMM), models that will allow obtaining more efficient results will be estimated. to possible endogeneity problems.

Thus, this study aims to determine the profitability of banks operating in Portugal and Spain in the period between 2011 and 2016. The sample consists of 267 banks in the Iberian Peninsula, of which 122 are Portuguese and 145 are Spanish. Overall, the results show that internal factors are the ones that most affect bank profitability in the three samples. The variables capital, operational efficiency, and the annual growth of deposits are the factors that best explain the profitability of the banking sector, both when considering individual countries and in the joint sample. The rest of the work is organized as follows: The second section presents the most relevant studies on the subject and hypotheses accordingly. Next, the research design is presented, which includes the sample data, the variables, and the estimation method. In Section 4 the main results are discussed and finally, in Section 5 the conclusions, limitations, and lines of future research are presented.

2. Literature review

Several authors [4, 8, 9, 10, 11, 12, 13] showed that the determinants of bank profitability can be internal and external, and it is also possible that external factors can be subdivided into specific industry factors and macroeconomic factors.

Internal factors are specific to banks and are normally controlled by management [8, 14]. While exogenous factors derive from the country’s economic and legal environment, they do not depend on the manager [3, 4, 14]. According to these authors, internal determinants can include asset structure, asset quality, capital, operational efficiency, revenue diversification, annual deposit growth, and size.

Likewise, among the external factors we can highlight ownership, whether the banks are listed or not, the inflation, and economic growth. All these variables were used in the aforementioned studies.

2.1 Bank specific determinants

2.1.1 Asset structure

Banks tend to diversify their loan portfolio and increase liquidity to reduce risks, particularly in times of crisis [14]. Much of the literature [9, 10] agrees that this action, to other safer assets, should cause profitability to increase more quickly. These operations tend to increase operational maintenance costs, however, García-Herrero et al. [15] argue that profit should also increase.

It should be noted that the increase in the level of credit can cause a high liquidity risk, if the manager does not effectively reduce its liabilities or if it does not know how to properly finance the increase in assets [10, 11]. If these operations are carried out well, the increase in loans will increase the bank’s revenues, therefore it will also increase profitability [16]. Therefore, Saona [12] and Trujillo-Ponce [10] found a positive relationship between the relative percentage of loans in a bank’s assets and its profitability. Also, Tan et al. [16] confirmed that liquidity risk exerts a positive influence when considering ROAA as a measure of bank profitability in Chinese banks, however, there is a negative relationship when considering ROAE. On the other hand, Trabelsi and Trad [17] empirically showed that asset structure negatively influences ROAA and positively influences ROAE. Guru et al. [8] and Rumler and Waschiczek [9] emphasize that the asset structure negatively influences bank profitability.

Thus, it can be seen that there is no consensus regarding the sign and significance of this variable with profitability. On the one hand, more loan amounts mean higher turnover and, in principle, more results. However, more loans also translate into more processing costs, higher chances of credit losses, and the cost of maintaining required capital reserves.

Thus, according to the literature, we expose hypothesis one, with no pre-defined sign.

Hypothesis 1—There is a significant relationship between the composition of banks’ assets and their profitability (with no defined sign).

2.1.2 Asset quality

According to Trabelsi and Trad [17], this variable indicates the economic and financial situation of banks, as it warns us of financial vulnerability, assessing their resilience to financial shocks.

In fact, in unfavorable times, there may be an increase in bad debt assets, causing banks to distribute a portion of their gross margin for provisions, to cover any loan losses [10]. These operations are associated with a credit risk that affects bank profitability [3]. Thus, with an increase in impairment losses on loans and accounts receivable, the quality of the assets of banking institutions may be negatively affected [10]. Mester [13] also showed that the increase in loan quality is associated with an increase in bank operating costs, which may have an opposite effect to that expected.

Empirical analyzes by Alshatti [18], Athanasoglou et al. [2], Trabelsi and Trad [17], and Trujillo-Ponce [10] found a negative association between the quality of bank assets and their profit. Similarly, Dietrich and Wanzenried [3] showed a negative influence of asset quality on bank profitability during the time of crisis (2007–2009). Garcia and Guerreiro [19], on the other hand, were faced with a negative relationship when this variable was associated with ROAA, but when they use ROAE, this relationship is positive.

In contrast, Saona [12] showed a positive relationship between asset quality and profitability of Latin American banks. The author argues that this sign is observed because Latin American banks charge their customers with paying higher prices for services provided to combat the costs associated with credit risk. He also claims that these transactions are possible because the interest of investors is not protected in those countries. Following this literature, we propose the following hypothesis:

Hypothesis2—There is a significant relationship between the quality of assets and their profitability (with no defined sign).

2.1.3 Capital

Capital refers to the amount of own funds available to support banking activity, exerting a safety net in case of hostile developments [14]. Banks with a high net worth on assets are seen as safer and less risky banks compared to institutions with lower capital, that is, well-capitalized banks are able to cope with times of crisis [3, 8]. In fact, according to Iannotta, Nocera, and Sironi [20], a better capitalization of banks may reflect a higher quality of management. This association can help banks finance their assets with lower interest rates, as the risk of bankruptcy is reduced [3, 15], thus making increase your profitability.

However, Djalilov and Piesse [14] suggest that the increase in financing costs due to the high level of capital could negatively affect bank profitability. Thus, the authors found a positive relationship between capital and profitability in the countries that made the initial transaction, however, the countries of the former USSR did not show any relationship between capital and the profitability of their banks. Dietrich and Wanzenried [3] also showed that in the pre-crisis period, capital did not influence the profitability of the Swiss banking sector, but between 2007 and 2009 they had a significantly negative ROAA. On the other hand, Knezevic and Dobromirov [11] show that the profits of Serbian national banks are not influenced by capital, on the other hand, foreign banks are negatively influenced.

Trujillo-Ponce [10] showed that Spanish banks are positively influenced by capital when profitability is calculated using ROAA, however, when it is related to ROAE, they present a negative relationship. In contrast, the Portuguese banks analyzed by Garcia and Guerreiro [19] showed a negative association with ROAA, but insignificant with ROAE.

Other studies, such as Alshatti [18], Athanasoglou et al. [2], Rumler and Waschiczek [9], Saona [12] and Trabelsi and Trad [17] show a positive relationship between return and equity on assets. While the studies by Guru et al. [8] and Shehzad, De Haan and Scholtens [21] have a negative sign.

According to the exposed literature, the following hypothesis is proposed:

Hypothesis 3—There is a significant relationship between the capital ratio of banks and their profitability (with no defined sign).

2.1.4 Operational efficiency

Beccalli et al. [22] argue that efficiency represents the minimization of inputs (that is, consuming fewer inputs for the same level of results) or the maximization of outputs (producing more outputs for the same amount of inputs). To this, authors such as Beccalli et al. [22] and García-Herrero et al. [15], call it X-efficiency (best practice indicator). According to some studies [2, 4], operational efficiency is one of the indicators that most influence bank profitability. Thus, for profitability to be high, the degree of efficiency of the financial institution’s management must also be high [2, 3], that is, the reduction of operating costs (administrative expenses, employees’ salaries, property expenses, among others) and, simultaneously, the increase in income, lead to a high level of bank profitability [11].

Traditionally, the operational efficiency of the banking sector is calculated using the cost-to-income (CIR) ratio, that is, expenses to income, with a high value reflecting more inefficiency. Therefore, it is expected that expenses will be lower than revenues so that efficiency will positively influence banks’ profitability.

Thus, some showed a negative association between bank efficiency and profit [3, 8, 10, 11, 19, 21].

For example, Ding et al. [4] concluded that the Chinese banking sector is more efficient than US banking institutions in times of crisis, however, after the crisis, the US overlaps China. Tan et al. [16] found that efficiency in Chinese banks negatively influences ROAA and positively ROAE.

According to the literature cited, the fourth hypothesis is presented:

Hypothesis 4—There is a positive relationship between operational efficiency and your bank profitability.

2.1.5 Revenue diversification

Banking activities can be divided into traditional activities and non-traditional activities, both of which are important for bank profitability. According to Trujillo-Ponce [10], non-traditional activities arise for diversification, trying, in this way, to generate new sources of income complementing traditional activities. In this sense, Stiroh and Rumble [23] follow in stating that financial institutions have to make more profitable sources that are generated by non-traditional activities so that they increase profitability levels and manage to survive the competition.

However, DeYoung and Rice [24] argue that one cannot put all the emphasis on non-traditional activities, due to the consequent increase in profitability, since, if they are not associated with traditional activities, they become an unsound strategy, thus putting, concerned the possible profit.

Even so, studies have concluded that revenue diversification has a positive impact on profitability above the spread [25]. While Saona [12] presented a negative sign for this relationship. Tan et al. [16] showed a positive relationship between non-traditional activities and ROAA, but a negative one with ROAE. However, Elsas, Hackethal and Holzhäuser [26], Stiroh and Rumble [23], and Trujillo-Ponce [10] did not find significant differences to be able to state that diversity affects profitability.

As per the provisions, it appears that there is a relationship between the diversity of revenues and the profitability of the banking sector. Accordingly, the following hypothesis arises:

Hypothesis 5—There is a significant relationship between revenue diversity and bank profitability (with no defined sign).

2.1.6 Deposit growth

In general, deposits represent stable and cheaper resources than other types of financing, and, to this extent, they contribute to increasing bank profitability [15]. But the global financial crisis led banks to adopt aggressive policies, mortgaging their margins at the expense of paying higher rates, which contributed to the decrease in profitability [10].

Dietrich and Wanzenried [3] state that an increase in deposits also implies attending to numerous factors, such as operational efficiency, as banks must be able to convert deposit liabilities into revenue-generating assets, taking into account good credit quality. However, high deposit growth rates also attract additional competitors, affecting the profitability of banking institutions.

Thus, Trujillo-Ponce [10] did not find any relationship between the growth rate of bank deposits and Spanish bank profitability. However, Garcia and Guerreiro [19] found that the growth of deposits intervenes positively in ROAA, but that it has no statistical significance in ROAE. In contrast, Dietrich and Wanzenried [3] are faced with a negative influence on ROAA and a positive influence on ROAE.

In harmony with the exposed literature, the following hypothesis is put forward:

Hypothesis 6—There is a significant relationship between the growth rate of deposits and bank profitability (no defined sign).

2.1.7 Bank size

The size of banks is one of the characteristics that have traditionally been used to determine their levels of profitability because, in principle, the bigger the bank, the greater the use of synergies and economies of scale, leading to a reduction in expenses and, consequently, an increase in results and profitability [14, 20]. Saona [12] claims that a large bank will incur in large operations, therefore, it will be associated with a higher risk, which, consequently, will cause the institution to charge higher margins, positively influencing profit.

However, a bank that is too large may incur diseconomies of scale as it will have an increase in variable costs, such as operating, bureaucratic and marketing expenses, negatively affecting bank profitability [2, 3]. According to García-Herrero et al. [15], the increase in size can make bank management difficult due to the occurrence of aggressive competitive strategies.

Therefore, empirical investigations [12, 16, 17] have found a positive and significant relationship between profitability and size.

Dietrich and Wanzenried [3] showed that in Switzerland the largest banks are the least profitable, following Berger and Mester [27] who had concluded the same.

In another sense, Ding et al. [4] showed that the large US banks after the crisis were the ones that were able to restructure the fastest and obtain higher levels of profitability. Once the authors had obtained a negative relationship during the crisis. Also, Elsas et al. [26] and Knezevic & Dobromirov [11] found a negative and significant relationship between size and profitability. Other empirical research does not find any significant relationship between profitability and bank size [2, 9, 10, 18, 28].

Following the exposed literature, we proposed hypothesis 7:

Hypothesis 7—There is a significant relationship between the bank’s size and its profitability (with no defined sign).

2.2 Industry-specific determinants

2.2.1 Ownership

Banks can be private or public institutions, the private ones belong, essentially, to private entities (more than 50% of these institutions) and the public ones, mainly, to the State. Berger and Mester [27] argue that the more external investors there are, the greater the control, the greater the efficiency, consequently the greater the profitability.

However, there is much empirical evidence that this variable does not influence the institution’s profit [2, 18, 28].

DeYoung and Rice [24] and Knezevic and Dobromirov [11] found a negative relationship between ROAE and ROAA, respectively. Dietrich and Wanzenried [3] also concluded that ROAA is negatively influenced by the type of property, however, this is not statistically significant when the performance index is the ROAE.

In contrast, Rumler and Waschiczek [9] show that banks with public capital positively influence ROAE. Under these points of view the following hypothesis is placed:

Hypothesis 8—There is a significant relationship between the nature of bank ownership and its profitability (with no defined sign).

2.2.2 Stock exchange quotation

According to Beccalli et al. [22], information on the earnings of institutions can be incorporated into stock prices, however, changes in stock prices do not properly reflect the extent of changes in earnings. Dietrich and Wanzenried [3] argue that the fact that listed banks negatively affect institutions’ profits makes them subject to greater requirements, such as additional reporting and greater market scrutiny. This fact may affect the profitability of banks with the additional costs that they entail. However, financial institutions listed on the stock exchange that has a positive influence on performance will suffer greater pressure from the financial market (shareholders, financial analysts, etc.…).

Iannotta et al. [20] showed that the stock market positively affects bank profitability.

On the other hand, Dietrich and Wanzenried [3] faced a negative influence when using ROAA. García-Herrero et al. [15] showed that banks present on the stock exchange are not more profitable than those that are not.

By the way, the hypothesis to be tested will be:

Hypothesis 9—Banks listed on the stock exchange have greater profitability than unlisted ones.

2.3 Macroeconomic determinants

2.3.1 Inflation

Inflation can influence profitability depending on how it interferes with operating income and costs. Thus, if management manages to forecast the inflation rate, it can regulate interest rates appropriately, to increase revenues faster than costs [11]. Otherwise, banking costs will be higher than revenues and will negatively affect the profitability of banking institutions.

Athanasoglou et al. [2], García-Herrero et al. [15], Guru et al. [8], Rumler and Waschiczek [9], Saona [12], and Tan et al. [16] confirmed that the relationship between inflation and bank profitability is positive. However, Djalilov and Piesse [14] and Shehzad et al. [21] did not find a significant relationship between profitability and this variable. Trabelsi and Trad [17] and Trujillo-Ponce [10], in turn, showed that ROAA is positively influenced by the inflation rate while ROAE is negatively affected.

Following the literature, we propose the following hypothesis:

Hypothesis 10—There is a direct relationship between inflation and bank profitability (with no defined sign).

2.3.2 Economic growth

Economic growth varies over the years as the economy goes through several economic cycles. On the one hand, if the country’s economic conditions are unfavorable, this could mean an increase in banks’ provisions due to the loss of credit and the poor quality of assets mortgaged to profitability. On the other hand, if the country’s economic conditions are favorable, the demand for credit from households and companies will increase and, consequently, so will profitability [10, 14].

However, Saona [12] concluded that in periods of strong economic growth, banks may tend to adjust their margins, leading to lower results and profitability. That said, Athanasoglou et al. [2], Dietrich and Wanzenried [3], Lee and Kim [28], Rumler and Waschiczek [9], Trabelsi and Trad [17], and Trujillo-Ponce [10] show a positive relationship between economic growth and economic profitability. While other investigations [11, 14, 15] have not found any relationship between economic growth and bank profitability. However, Saona [12] and Shehzad et al. [21] showed a negative association between Gross Domestic Product (GDP) and bank profitability. Finally, Garcia and Guerreiro [19] concluded that GDP negatively affects the banking sector if it is analyzed with the ROAE profit indicator, however, if it is related to ROAA, the sign becomes insignificant.

In line with the provisions, we propose the following hypothesis to be tested:

Hypothesis 11—The country’s GDP influences bank profitability.

3. Research design

3.1 Sample data

The sample, for the period between 2011 and 2016, is composed of 267 Iberian banks, of which 122 are Portuguese and 145 are Spanish. All databases without complete data for at least four consecutive years were excluded, a necessary condition for second-order correlation estimation [5]. As the second-order correlation is a GMM assumption, and this will be the estimation method used, this correlation must be tested [29]. Data to calculate bank- and industry-specific variables are sourced from Orbis Bank Focus, Bureau van Dijk database. While the macroeconomic variables come from The World Bank¹.

3.2 Sample variables

3.2.1 Dependent

ROAA and ROAE have traditionally been used as measures of banking performance and to that extent are also the variables that we will use as dependent variables.

The average return on total assets (ROAA) is the ratio between Ebit and total assets [2, 30]. Garcia and Guerreiro [19] state that ROAA portrays management efficiency and is, therefore, an imminently economic indicator. Return on average equity (ROAE) is the ratio of net income to equity [2]. Rumler and Waschiczek [9] suggest that ROAE is a more popular performance measure among financial analysts.

This indicator translates into shareholder returns and, as such, there may be pressure from shareholders to distribute results, threatening the capitalization of banks. we know that the asset may be valued at acquisition cost, which can lead to the undervaluation or overvaluation of the elements that comprise it, thus influencing the ROAA ratio.

3.2.2 Independent variables

Table 1 presents the explanatory variables to be used in the regression models, highlighting their internal origin, intrinsic to management, or external, without direct influence from the manager.

Variables	Designation	Proxy	Theoretical foundation
Asset Structure	AS	Total Loans /Total Assets	Alshatti [18], García-Herrero et al. [15], Sharma et al. [1] and Trujillo-Ponce [10]
Asset Quality	AQ	Provisions for Loan Losses/Total Loans	Athanasoglou et al. [2], Dietrich e Wanzenried [8], Saona [12]
Equity	EQ	Equity/Total Assets	Alshatti [18], Garcia and Guerreiro [19], Knezevic and Dobromirov [11] and Trabelsi and Trad [17]
Cost to Income ratio	CIR	Total Expense/Total Revenue	Dietrich and Wanzenried [3], Knezevic and Dobromirov [11] and Saona [12]
Revenue Diversification	RD	Non-interest income/gross income	Tan et al. [16]
Annual Deposit Growth	ADG	(Total Depositst—Total Depositst-1)/Total Depositst-1)	Dietrich and Wanzenried [3], Garcia and Guerreiro [19] and Trujillo-Ponce [10]
Size	Size	Ln (total Assets)	Alshatti [18], Djalilov and Piesse [14], Knezevic and Dobromirov [11], Tan et al. [16]; Trujillo-Ponce [10]; Neves, Proença and Dias [31]
Owner	Owner	Dummy variable—takes value 1 if it has Public ownership and 0 if Private
Quoted	Quoted	Dummy variable—1 if quoted 0 otherwise
Inflation	INF	Annual Inflation Rate at Constant Prices	Knezevic and Dobromirov [11]; Trujillo-Ponce [10]
GDP	GDP	Real Gross Domestic Product Growth	Athanasoglou et al. [2], Dietrich e Wanzenried [3], Garcia e Guerreiro [19] Rumler e Waschiczek [9]

Table 1.

Specific characteristics of banks, industry-specific characteristics, and macroeconomic factors.

3.3 Methodology

Considering ROAE and ROAA as the dependent variables and the independent variables as defined previously, we obtain the following models:

$$\begin{array}{rcl}{\text{ROAE}}_{\mathrm{it}}& =& {\mathrm{\beta }}_0+{\mathrm{\beta }}_1{\mathrm{AS}}_{\mathrm{it}}+{\mathrm{\beta }}_2{\mathrm{AQ}}_{\mathrm{it}}+{\mathrm{\beta }}_3{\mathrm{EQ}}_{\mathrm{it}}+{\mathrm{\beta }}_4{\mathrm{CIR}}_{\mathrm{it}}+{\mathrm{\beta }}_5{\mathrm{RD}}_{\mathrm{it}}+{\mathrm{\beta }}_6{\mathrm{ADG}}_{\mathrm{it}}\\ & & +{\mathrm{\beta }}_7{\text{SIZE}}_{\mathrm{it}}+{\mathrm{\beta }}_8\text{Owner}+{\mathrm{\beta }}_9\text{Quoted}+{\mathrm{\beta }}_{10}{\mathrm{INF}}_{\mathrm{it}}+{\mathrm{\beta }}_{11}{\mathrm{GDP}}_{\mathrm{it}}++{\mathrm{\mu }}_{\mathrm{it}}\hfill \end{array}$$E1

$$\begin{array}{rcl}{\text{ROAA}}_{\mathrm{it}}& =& {\mathrm{\beta }}_0+{\mathrm{\beta }}_1{\mathrm{AS}}_{\mathrm{it}}+{\mathrm{\beta }}_2{\mathrm{AQ}}_{\mathrm{it}}+{\mathrm{\beta }}_3{\mathrm{EQ}}_{\mathrm{it}}+{\mathrm{\beta }}_4{\mathrm{CIR}}_{\mathrm{it}}+{\mathrm{\beta }}_5{\mathrm{RD}}_{\mathrm{it}}+{\mathrm{\beta }}_6{\mathrm{ADG}}_{\mathrm{it}}\\ & & +{\mathrm{\beta }}_7{\text{SIZE}}_{\mathrm{it}}+{\mathrm{\beta }}_8\text{Owner}+{\mathrm{\beta }}_9\text{Quoted}+{\mathrm{\beta }}_{10}{\mathrm{INF}}_{\mathrm{it}}+{\mathrm{\beta }}_{11}{\mathrm{GDP}}_{\mathrm{it}}++{\mathrm{\mu }}_{\mathrm{it}}\hfill \end{array}$$E2

To estimate these models, the GMM dynamic model was used, initially proposed by Arellano and Bond [5] and improved by Arrellano and Bover [6] and Blundell and Bond [7]. By using the GMM method, we solve two fundamental problems such as endogeneity and unobserved heterogeneity [14, 15, 29].

4. Results

4.1 Descriptive statistics

This chapter describes descriptive statistics (mean, minimum, maximum, and standard deviation) for the variables used in the sample. From what can be seen in Table 2, the study shows a positive mean of the dependent variables for all the years under observation. It is also observed, in the three samples, that the means of the independent variables are mostly positive, except GDP in the samples from the Iberian Peninsula and Portugal. Regarding the standard deviation, it can be seen that the annual growth of deposits is the variable that presents the highest value for the samples from the Iberian Peninsula and Spain. While Portugal presents the asset quality variable with the greatest discrepancy to the average.

			Iberian Peninsula		Portugal					Spain
Variables	Mean	Minimum	Maximum	Standard Deviation	Mean	Minimum	Maximum	Standard Deviation	Mean	Minimum	Maximum	Standard Deviation
ROAE	2.815	−596.314	516.159	37.512	2.196	−131.397	516.159	36.318	3.173	−596.314	348.454	34.985
ROAA	0.415	−17.502	19.236	2.394	0.136	−17.502	17.239	2.519	0.576	−14.037	19.236	2.299
AS	51.070	0.000	99.967	23.947	47.727	0.143	98.125	22.333	53.066	0	99.967	25.958
AQ	14.442	0.000	1884.362	137.350	33.507	0.002	1884.362	220.676	2.650	0	140.937	19.738
EQ	13.987	−7.831	99.987	18.277	13.827	−5.607	93.253	13.128	14.078	−7.831	99.987	20.56033
CIR	67.954	0.200	872.730	49.619	10.582	1.22	620.690	43.900	66.363	0.200	872.730	51.072
RD	44.500	0.000	550.000	34.369	44.428	0	179.604	24.204	44.570	0	550.000	42.111
ADG	393.336	−98.914	184025.500	8323.199	6.649	−77.243	172.352	22.625	577.918	−98.914	184025.5	8970.098
SIZE	6.119	4.113	9.127	1.146	5.807	4.360	8.139	0.991	6.299	4.113	9.127	1.159
OWNER	0.023	0.000	1	0.151	0.030	0	1	0.171	0.018	0	1	0.142
Quoted	0.038	0.000	1	0.191	0.030	0	1	0.171	0.044	0	1	0.199
INF	1.289	−0.500	3.653	1.478	1.426	−0.278	3.653	1.524	1.182	−0.500	3.196	1.342
GDP	−0.319	−4.028	3.432	2.304	−0.877	−4.023	1.822	2.054	0.117	−2.928	3.432	2.26

Table 2.

Descriptives statistics.

4.2 Discussion results

Table 3 presents the results for the banks of Portugal and Spain as the Iberian Peninsula, and Table 4 an individual analysis of the determinants that affect the profitability of these two border countries. Thus, economic growth exhibits a statistically significant and positive sign for the joint analysis (Table 3). Such evidence may be due to the increase in credit on the part of families and companies after the financial crisis. As well as, a decrease in bad debt assets, during some favorable economic growth of the countries. Since in favorable growth situations, borrowers can meet their debts. By the provisions, hypothesis 11 is corroborated, following the results found by Lee and Kim [28], Rumler and Waschiczek [9], and Trabelsi and Trad [17]. Regarding inflation, this influences positively and significantly both the Iberian banking sector and the Portuguese banking sector (Table 4). However, banking in Portugal is only influenced by the operating performance index, which is the ROAA. In this way, hypothesis 10 is supported by the studies by Rumler and Waschiczek [9], Tan et al. [16], and Trujillo-Ponce [10]. This reveals that the managers of the banks under analysis can predict the rate of inflation, properly regulating interest rates so that revenues increase faster than expenses. On the other hand, an increase in inflation can also translate into an increase in the purchasing power of the population in general, so this increase can mean more deposits, more credit compliance, safer, therefore more profitability. The asset structure presents a negative and statistically significant result concerning the ROAE of the sample of Iberian and Portuguese banks. In fact, the asset structure of Spanish banks (Table 4) also shows the same sign for both dependent variables. By the provisions, it is clear that it is possible to corroborate hypothesis 1. This evidence is supported by the results of Rumler and Waschiczek [9] and Tan et al. [16]. This means that the banking sector is not able to efficiently manage and increase the loan portfolio in the period under review. A period that is characterized by strong competition and banking competitiveness, as banks were under great pressure to attract customers. However, an increase in the loan portfolio implies increases in operating costs, so if interest rates are not well adjusted, they will become incapable of supporting operating expenses, harming banks’ profitability. As well, the increase in the loan portfolio can also lead to a high risk of credit defaults. It appears that the sign of the variable capital is positive in both Tables. That said, hypothesis 3 is corroborated. This evidence shows that Portuguese and Spanish banks are well-capitalized.

	ROAE					ROAA
	Coefficient	St. Deviation	Z	P-value		Coefficient	St. Deviation	Z	P-value
Constant	−16.763	(77.231)	−0.22	0.828		−7.288	(7.646)	−0.95	0.341
AS	−0.222	(0.103)	−2.15	0.032	**	−0.011	(0.008)	−1.36	0.174
AQ	−0.176	(0.149)	−1.18	0.239		−0.001	(0.019)	−0.06	0.995
EQ	1.173	(0.570)	2.06	0.040	**	0.191	(0.059)	3.24	0.001	***
CIR	−0.352	(0.042)	−8.38	0.000	***	−0.021	(0.004)	−5.11	0.000	***
RD	−0.019	(0.025)	−0.78	0.434		0.001	(0.003)	0.19	0.847
ADG	0.018	(0.007)	2.48	0.013	**	0.003	(0.001)	4.10	0.000	***
SIZE	6.515	(10.884)	0.60	0.549		1.255	(1.129)	1.11	0.266
OWNER	99.912	(95.047)	1.05	0.293		1.815	(4.439)	0.41	0.683
Quoted	−35.332	(30.531)	−1.16	0.247		−6.798	(5.114)	−1.33	0.184
INF	2.452	(1.002)	2.45	0.014	**	0.255	(0.091)	2.79	0.005	***
GDP	1.218	(0.290)	4.20	0.000	***	0.064	(0.029)	2.16	0.031	**
Sargan			16.01318 (11)	0.1406				10.57061 (11)	0.4799
Wald			215.79 (12)	0.0000				69.20 (12)	0.0000
AR (1)			0.43101	0.6665				−1.9816	0.0475
AR (2)			−1.5599	0.1188				−0.13714	0.8905

Table 3.

Estimation results of models 1 and 2 for the Iberian Peninsula.

In the table above, all the variables are those included in the literature review. In order to understand the reading of the results, we still need to know that: (i) the values in parentheses represent the asymptotic standard errors compatible with heteroscedasticity problems; (ii) **,*** represent the statistically significant coefficients at the level of 10%, 5% and 1%, respectively; (iii) The Sargan test has a p value greater than 5%, which means that the instruments are valid and the values between relatives represent degrees of freedom; (iv) The Wald test has a p-value less than 5% indicating that the set of coefficients is asymptotically distributed as χ2 under the null hypothesis without significance, the degrees of freedom are represented in parentheses; (v) The Arellano-Bond test is asymptotically distributed as N (0,1) under the null hypothesis of no serial correlation. The AR(2) test indicates that there is no second-order serial correlation.

	Portugal										Spain
	ROAE					ROAA					ROAE					ROAA
	Coefficient	St. Deviation	Z	P-value		Coefficient	St. Deviation	Z	P-value		Coefficient	St. Deviation	Z	P-value		Coeficiente	Erro padrão	Z	P-value
Constant	173.441	(166.815)	1.04	0.298		1.824	(7.023)	0.26	0.795		87.620	(24.922)	3.52	0.000	***	10.126	(4.790)	2.11	0.035	**
AS	−0.512	(0.214)	−2.39	0.017	**	−0.009	(0.008)	−1.11	0.266		−0.454	(0.084)	−5.42	0.000	***	−0.034	(0.009)	−3.83	0.000	***
AQ	−0.366	(0.479)	−0.76	0.455		−0.081	(0.036)	−2.23	0.026	**	−4.471	(0.387)	−11.55	0.000	***	−0.448	(0.049)	−9.19	0.000	***
EQ	1.934	(0.405)	4.78	0.000	***	0.133	(0.033)	3.98	0.000	***	0.775	(0.251)	3.09	0.002	***	0.092	(0.043)	2.14	0.032	**
CIR	−0.543	(0.103)	−5.24	0.000	***	−0.032	(0.003)	−10.07	0.000	***	−0.370	(0.027)	−13.62	0.000	***	−0.031	(0.003)	−11.89	0.000	***
RD	−0.051	(0.026)	−1.91	0.056	*	−0.004	(0.001)	−2.51	0.012	**	−0.031	(0.025)	−1.23	0.217		−0.004	(0.003)	−1.16	0.247
ADG	0.248	(0.064)	3.90	0.000	***	0.018	(0.003)	5.85	0.000	***	0.044	(0.025)	1.73	0.084	*	0.003	(0.001)	2.25	0.025	**
SIZE	−22.012	(24.628)	−0.89	0.371		−0.050	(0.934)	−0.05	0.957		−5.009	(3.175)	−1.58	0.115		−0.796	0.586	−1.36	0.174
OWNER	−27.070	(73.281)	−0.37	0.712		0.772	(6.379)	0.12	0.904		−700.091	(920.457)	−0.76	0.447		−119.784	(211.781)	−0.57	0.572
Quoted	−80.435	(67.069)	−1.20	0.230		12.742	(13.500)	0.94	0.345		3.412	(3.327)	1.03	0.305		−0.279	(0.296)	−0.94	0.347
INF	2.128	(1.514)	1.41	0.160		0.214	(0.052)	4.12	0.000	***	−1.592	(1.143)	−1.13	0.260		0.005	(0.099)	0.05	0.960
GDP	0.571	(0.575)	0.99	0.321		−0.013	(0.042)	−0.30	0.762		0.106	(0.477)	0.22	0.824		0.044	(0.044)	0.99	0.323
Sargan			7.240062 (11)	0.7793				7.007189 (11)	0.7985				7.754619 (9)	0.5591				12.96967 (9)	0.1640
Wald			1720.97 (12)	0.0000				12162.61 (12)	0.0000				5941.12 (13)	0.0000				784.38 (13)	0.0000
AR (1)			0.5725	0.5670				−2.1586	0.0309				−1.1767	0.2393				−1.6326	0.1025
AR (2)			−0.80537	0.4206				−0.4231	0.6722				−0.78984	0.4296				0.00352	0.9972

Table 4.

Estimation results for Portugal and Spain.

In the table above, all the variables are those included in the literature review. To understand the reading of the results, we still need to know that: (i) the values in parentheses represent the asymptotic standard errors compatible with heteroscedasticity problems; (ii) *,**,*** represent the statistically significant coefficients at the level of 10%, 5% and 1%, respectively; (iii) The Sargan test has a p-value greater than 5%, which means that the instruments are valid and the values between relatives represent degrees of freedom; (iv) The Wald test has a p-value less than 5% indicating that the set of coefficients is asymptotically distributed as χ2 under the null hypothesis without significance, the degrees of freedom are represented in parentheses; (v) The Arellano-Bond test is asymptotically distributed as N (0,1) under the null hypothesis of no serial correlation. The AR(2) test indicates that there is no second-order serial correlation.

Regarding Table 3, Banks that present better levels of capital denote a lower risk of bankruptcy, they are considered.

with lower financing costs, therefore they can obtain higher gross margins, which leads to higher levels of profitability. These results are in line with the investigations of Garcia and Guerreiro [19], Trujillo-Ponce [10], and Saona [12]. Consistent with hypothesis 2, the results in Table 4 show that asset quality negatively and significantly affects the ROAA of Portuguese banks and both performance indicators relative to Spain. This is due to the increase in impairment losses on loans, which negatively affects the performance of banking institutions. Naturally, in dire economic cycles, which is the case in the sample period, households tend to default on credit. Therefore, the period from 2011 to 2016 is marked by losses of millions of euros in loans to customers, which was noted in the financial statements of institutions. And once banks have a significant increase in bad debt assets, they tend to distribute their gross margin to cover expected losses [10]. This result is in line with the results obtained by Alshatti [18], Athanasoglou et al. [2], Trabelsi and Trad [17] and Trujillo-Ponce [10].

Regarding the operational efficiency variable, it manifests itself with a negative and significant sign in both Tables. The negative and significant coefficient of the cost-to-income ratio shows that poor expenditure management is one of the main contributors to poor profitability performance. In other words, to obtain a higher performance it is necessary to have a decrease in expenses and/or an increase in income [11]. The result obtained is under the empirical analysis by Dietrich and Wanzenried [3] and Shehzad et al. [21], that is, the higher the CIR, the lower the efficiency and therefore the profitability.

The annual growth rate of deposits is showing a statistically significant and positive sign. Indeed, hypothesis 6 is corroborated, following the result obtained in the investigation by Dietrich and Wanzenried [3] and Garcia and Guerreiro [19]. The increased demand for deposits increases bank profitability both in the Iberian Peninsula and in Portugal and Spain. Banks may be benefiting from the increased purchasing power of depositors following the effects of the financial crisis. To that extent, there will be no need to incur aggressive policies (which can negatively affect performance through lower margins) to attract a greater number of depositors. Finally, as happened with the authors Saona [12] and Tan et al. [16], and corroborating hypothesis 5, the results of Portuguese banking show a negative and significant sign for the variable revenue diversification, for both profitability indicators. This result suggests that non-traditional activities, by themselves, do not boost Portuguese banking profitability.

Also, DeYoung and Rice [24], had already warned that if banks did not associate these activities with traditional activities, they could incur losses. Despite this, this result may be due to charges, for example, with securities. As a means of recapitalization, to comply with the rules established by the Basel III agreement, banks were subject to CoCos (Contingent Convertible Bonds) bonds, however, the financial charges with the high-interest rates of these bonds may have harmed the banks’ performance. Portuguese.

It should be noted that the significant variables in the three samples always show the same sign, which somehow gives credibility to the results found. In the sample of Spanish banks, it is possible to verify that the explanatory variables are exactly the same using the ROAA or the ROAE as dependent variables, which similarly suggests that the explanatory variables were well selected. In general, when operating profitability is used as a performance measure, the variables that remain significant in all samples are bank capitalization, operating efficiency measured by the CIR and the growth of deposits.

Likewise, when using ROAE as a performance measure, the signs and significance of these variables remain unchanged and the composition of assets is added as a determinant of return on equity. These results suggest that the role of managers is fundamental in defining and monitoring capital and deposit growth ratios to improve performance. Furthermore, it is necessary to improve operational efficiency as well as maximize the asset structure, which in a highly competitive environment has not been able to increase bank profitability levels.

5. Conclusion

The economic and financial stability of banking institutions is important for the economic stability of the country where the sector is located. Therefore, satisfactory levels of profitability can translate into the stable financial health of countries.

This study aimed to analyze the factors that influence the profitability of Portuguese and Spanish banks, during a period between 2011 and 2016.

The empirical study was carried out considering three sub-samples to observe the Iberian Peninsula as a whole, and each country individually in order to understand the differences in the determinants of profitability in these two border countries. However, our results show that the profitability of Portuguese and Spanish banks is mostly influenced by the same internal variables, which shows that, probably, the fact that they are neighboring countries can lead to similar behavior of managers. These management decisions are those that exert the greatest influence on bank profitability.

In particular, it can be seen that the performance of banks, both in the Iberian Peninsula as a whole and in Portugal or Spain, is positively influenced by demand from depositors. Likewise, it is possible to verify that the more capital the financial institutions of both countries hold, the better their capacity to face adverse situations.

Finally, it is concluded that if Iberian banks do not efficiently manage their expenses and costs, they incur a poorly applied operational policy, negatively influencing the profitability of these institutions.

Therefore, despite different economic systems, Portugal and Spain have similar internal banking policies. This may be due to the strong presence of Spain in the Portuguese financial sector, establishing management methods that are very similar to each other.

Since in this work, only Iberian banks were used and to that extent, the sample is small, which may constitute a limitation in the extrapolation of these results, we propose to analyze in future work a broader set of countries with different legal and institutional environments, using an example to an additional efficiency model such as the DEA.

Acknowledgments

This work is supported by national funds, through the FCT—Portuguese Foundation for Science and Technology under the project UIDB/04011/2020.