Introduction

In recent years, it has been observed that one of the main objectives of managers is to increase the value of the company for shareholders. Naturally, a company requires assets, whether tangible or intangible, to carry out its business, create value, and achieve its business objectives. The financing of these assets can be done through three options: a) obtaining debt (bank or corporate); b) shareholder contribution; and c) own funds. Deciding on the mix or combination of debt and capital can be a difficult exercise because high indebtedness could increase the probability of bankruptcy, simultaneously reducing the payment of taxes and, therefore, providing a tax advantage. On the other hand, little debt would imply a reduction in yield due to increased tax pressure and issues leading to over-investment problems, followed by a reduction in the returns of shareholders.

The importance of studying the capital structure of companies lies in the generation of models that can summarize and predict the most relevant factors that influence the decision to obtain an “optimal” capital structure that allows the efficient use of resources. It was Modigliani and Miller who in 1958 developed the first theory about capital structure, where they indicated that in a perfect market the value of the firm has no relation to its capital structure, and that taxes, cost of capital, information asymmetry, and bankruptcy costs would not affect either, but that its value is given by the value of the assets it possessed (^{Modigliani & Miller, 1958}).

However, the problem with the theories of Modigliani & Miller is that they assume unrealistic scenarios. In a recent work by (^{Serrasqueiro, Matias, & Salsa, 2016}), they indicate that the tax advantages generated by the use of debt added substantial value to companies, mainly by saving on tax payments. Furthermore, ^{Miller (1977)} included in his work the effect of taxes in the analysis and determined that taxes only affect at a macro level and not individual companies. ^{DeAngelo & Masulis (1980)} indicated that not only could the payment of taxes be reduced through indebtedness, but that there were substitutes (depreciation, amortization, tax credits for investment, among others) that did not increase the liabilities of corporations. On the other hand, Myers (1974) argued that corporations would prefer debt over equity for two reasons: 1) the cost is lower, and 2) investors would consider the company to be overvalued and could restrict the issuance of new capital.

From the above emerged a great number of alternative theories and models, such as including market imperfections and allowing corporate directors to make financial decisions on an objective and proven basis. Several of these have largely failed, since economic and financial conditions may be very different depending on the region studied, leading to a neutralization of the validity of such models (^{Booth, Aivazian, Demirguc-Kunt, & Maksimovic, 2001}). In addition, the theories differ in their emphasis and no consensus has been reached among academics (^{Sheikh & Wang, 2011}) regarding which best synthesizes the financial behavior of the companies. An example is that research with alternative approaches has studied the relationship between the compensation of the directors and the capital structure of the companies they run. It was found that the market value of capital and debt decreases when the director is changed, as well as the expected cash flow is higher in companies that do not change their directors. On the contrary, maintaining the director would have a positive correlation with company size, leverage, salary compensation, and cash flow (^{Berkovitch, Israel, & Spiegel, 2000}).

There are two main theories that seek to address the issue of capital structure from different perspectives and different predictions. The first is the Trade-off Theory of Capital Structure, which emphasizes taxes, financial difficulties, and conflict of interest; and the second is the Pecking Order Theory, which emphasizes information asymmetry among directors, shareholders and creditors, and financial independence as determinants of capital structure.

The objective of this research is to discover the determinants of the capital structure of 304 companies in the manufacturing sector of Guayaquil during the period of 2012-2016 and to identify which theory best summarizes the financing decisions of the companies. The importance of this lies in verifying the determinants of indebtedness in companies in a developing country and comparing the results with the predictions of the theories on Capital Structure and on the Pecking Order. In other words, to check the predictions of these two theories in order to generate evidence to determine their explanatory power in companies in developing Latin American countries. In this sense, it is also expected that the results of this work will contribute to the general knowledge about the financial conduct of companies in developing countries and, through the results obtained, to the theories of capital structure.

Theoretical Framework

Data and Methodology

To achieve the established objectives, an exploratory, empirical, and quantitative study was proposed on small and medium manufacturing enterprises in Guayaquil, during the period of 2012-2016. The population under study has been obtained from the companies associated with the Chamber of Small Industry of Guayas (CAPIG for its acronym in Spanish)¹. In order to carry out the study, the 304 companies for which sufficient information was available from the five years under study were taken.

CAPIG groups its partners in 25 different sectors. Within the final sample, there are only 5 sectors that group more than 50% of the companies: Chemical (18.42%), Food and Beverages (10.20%), Metalworking (8.88%), Services (7.57%), and Plastics and Rubber (6.91%). In 2016, the sales of the sampled companies reached $1.38 billion dollars (1.43% of GDP) as a whole, while the nominal Ecuadorian GDP of the same year, according to data from the World Bank, was $97.8 billion dollars.

This study has considered only companies incorporated according to the regulations of the Superintendence of Companies and Securities, eliminating individual companies as there is no public information on their financial situation. Generally, similar research has focused on analyzing listed companies. However, the Ecuadorian stock market is not representative in comparison to other economies, so the study has focused on SMEs to be more useful, from a comparative point of view, with that of other surrounding countries.

Since the time interval runs from 2012 to 2016, the first step was to eliminate companies with incomplete information and fewer years of financial information than the study required. The financial statements and balance sheets of the available companies were then collected, from which all the data necessary for the study were extracted. All data were obtained through the web page of the Superintendence of Companies, Securities and Insurance of Ecuador² and worked with securities at year-end. Microsoft Excel was used to sort and adapt the data in the form of panel data, for later export to the Stata v. 14³ statistical tool, with which the statistical models for the study were generated together with the corresponding statistical tests.

Results and discussion

This section shows the results of the fixed-effects model, as well as the descriptive statistics of each variable.

Figure 1 shows the historical total debt ratio from 2012 to 2016 of the Ecuadorian companies.

Source: Superintendence of Companies, Securities and Insurance.

Figure 1 Total debt ratio of all Ecuadorian companies. 

It is observed that in all the time between, the ratio did not have a great variation. In 2012, for example, 60.6% of corporate assets were financed by debt, while in 2015 it reached a minimum level of 57.8% and ended up rebounding to 58.8%. Ecuador, being a country with a heavy dependence on oil revenues, experienced a severe blow to its economy by the fall in crude oil prices between 2014 and 2015. In fact, according to reports from the Central Bank of Ecuador, the volume of credit granted by financial institutions had a drastic reduction between 2015 and 2016.

The values obtained from the descriptive statistics are shown in the following table;

The total debt ratio (DR) indicates that 57.4% of the total assets of the companies under consideration are financed through debt, regardless of whether it is long or short term.

Profitability, measured by ROA, shows an average of 8.7% along with a standard deviation of 13.4%, indicating a large variation of returns on assets within the sample. The presence of tax shields not generated by debt represents only 3.8% of balance sheet expenditures, with a standard deviation of 4.2%. Tangibility, on the other hand, indicates that on average companies have 34.8% of net fixed assets in their balance sheets, with a standard deviation of 25.8%. Regarding liquidity, the average indicates that current liabilities are covered 4.7 times by current liabilities, with a standard deviation of 22.9. Finally, the GO (growth opportunities) variable indicates that, on average, for every 1% increase in the total assets of the companies, their sales increase 1.3 times.

Table 3 shows the low correlation of the variables with each other and with the debt ratio.

It can be observed that the correlation between the TSNGD and the TANG is the highest positive of all. This is explained by the fact that the more net fixed assets a company owns, the greater the expense for depreciation of those assets and, consequently, the TSNGD measure increases simultaneously with TANG. On the contrary, the variables DR and PROF show a negative correlation; this could be explained by the fact that a higher level of debt increases interest payments, reducing pre-tax profits.

After the correlation results, Table 4 shows the Variance Inflation Factor for each variable. Remarkably, they are at low levels-1.09 on average-indicating a low correlation between the variables. Thanks to this test it can be stated that there is no significant problem of multicollinearity.

Based on the above data, the Ordinary Least Squares (OLS) model has been tested and the results are presented in the following Table

Before selecting the model that best explains the relationships between the variables, the Hausman test was carried out.

Table 6 indicates a p-value = 0.0034 that prevents rejection of the null hypothesis, which states that there is no correlation between individual effects and explanatory variables. Therefore, it was decided that the appropriate model was the fixed effects model.

According to the results of the Fixed Effects Model, Table 7, at a 5% level of significance the variables of profitability, size, tax shields not generated by debt, tangibility of assets, and liquidity are highly significant. Only the GO (growth opportunities) variable appears to be highly non-significant under the same level of significance.

The value of R2 is 0.1968 and indicates that only 19.6% of the variation of total indebtedness (DR) is explained by the effect of independent variables.

Table 8 shows the results obtained from the Random Effects Model, which measures the effect of each explanatory variable with respect to dependent variable, debt ratio, with a statistical significance of each of them in each model. An error level of 5% (0.05) has been considered, together with a confidence level of 95%.

Empirical results suggest that profitability has a negative and highly significant relationship to the level of corporate debt. This result is consistent with what is expected only under the POT, but not under the TOT. The POT assumes that companies prefer to use their own resources rather than borrow, so the more profitable the company, the less debt it will use as a financial resource. The result obtained is similar to that of other researches (^{Serrasqueiro, Matias, & Salsa, 2016}; ^{Huang & Song, 2006}; ^{Chen & Chen, 2011}; ^{Saeedi & Mahmoodi, 2009}; ^{Acaravci, 2016}; ^{Tanaka, 2008}; ^{Khrawish & Khraiwesh, 2010}; ^{Booth, Aivazian, Demirguc-Kunt, & Maksimovic, 2001}; Eriotis, Frangouli, & Ventoura-Neokosmides, 2011). Hypothesis H2 is accepted.

According to the TOT, size has a positive relationship with indebtedness. The results of the model show a positive and highly significant relationship between both variables, thus verifying the hypothesis of the TOT with respect to this relationship. The explanation, according to the theory, is that the larger the company, the more diversified it is and the lower the risk of a) bankruptcy and 2) income volatility. These two qualities motivate directors to borrow at higher levels and relax the demands of the creditors. This result only supports the prediction of the TOT, not the POT. The result of this study is similar to that of other authors (^{Acedo, Alútiz, & Ruiz, 2012}; ^{Kouki & Said, 2012}; ^{Eriotis, Vasiliou, & Ventoura-Neokosmidi, 2007}; ^{Khrawish & Khraiwesh, 2010}; ^{Serrasqueiro & Caetano, 2015}). Hypothesis H3 is accepted.

The TOT suggests a negative relationship between the level of debt and the TSNGD. The results of the model show a positive and significant relationship between the variables. The explanation for this could be that companies use their fixed assets as collateral for debt, which would increase the expense of depreciation of assets at the time of acquisition, and the level of debt. This result does not support the prediction of the TOT. Other studies showed similar results to this one (^{Serrasqueiro, Matias, & Salsa, 2016}; ^{Serrasqueiro & Caetano, 2015}; ^{DeAngelo & Masulis, 1980}). Hypothesis H5 is accepted.

Both the TOT and the POT predict a positive relationship between asset tangibility and level of debt. Such a prediction would be based on the fact that companies offer their tangible assets as debt collateral, so the more tangible assets the company has, the higher its level of debt. The results of the model suggest the opposite, that is, both variables have a negative relationship, indicating that the level of debt should not necessarily increase in light of the presence of a large quantity of tangible assets. This result is similar to other researches (^{Serrasqueiro, Matias, & Salsa, 2016}; ^{Kouki & Said, 2012}; ^{Saeedi & Mahmoodi, 2009}; ^{Acaravci, 2016}; ^{Booth, Aivazian, Demirguc-Kunt, & Maksimovic, 2001}). One factor that could cause this negative relationship would be that tangible assets do not have a high market value, which would discourage creditors from accepting them as collateral for debt. Hypothesis H7 is accepted.

The model suggests a significant negative relationship between liquidity and the debt ratio. This result is consistent with the predictions in the POT, but not with the TOT. This result is similar to that of other researches (^{Serrasqueiro, Matias, & Salsa, 2016}; ^{Saeedi & Mahmoodi, 2009}; ^{Eriotis, Vasiliou, & Ventoura-Neokosmidi, 2007}). The logic behind this relationship is that companies prefer to use retained earnings as a way of financing themselves. This result is very consistent with what was found in a study of SMEs in Guayaquil: almost half of the directors of these companies indicated that the main problem at the time of financing is the cost of financing, i.e. high interest rate on credit (^{Moran, Jácome, & Rodríguez, 2017}). Hypothesis H10 is accepted.

Lastly, the GO (growth opportunities) variable has no statistical significance in the model. This result confirms that the expected growth of companies, considered as an intangible asset, cannot be used as collateral for debt, which implies that the debt ratio is not affected.

Conclusions

This empirical study has explored the determinants of the capital structure of 304 Guayaquil companies during the period of 2012-2016 using a panel data structure and analyzing it using a Fixed Effects Model.

The results indicate that there is a negative relationship between profitability, tangibility, and liquidity with respect to corporate indebtedness. Conversely, size and tax shields not generated by debt have a significantly positive relationship with indebtedness. Only the variable “growth opportunities” turned out not to have any statistical importance on the level of debt.

The evidence of the explanatory power of the Trade-off Theory of Capital Structure regarding the type of financing preferred by Guayaquil companies is weak. This theory could only predict the relationship of a determinant with the level of debt with the size of the company. In other words, the results of the model suggest a significant and positive relationship between the size and level of debt. On the other hand, the Picking Order Theory was able to predict almost accurately the relationship of each factor and its relationship to the level of debt and the choice of financing. Consequently, it better predicts the financial behavior of the sample. The latter correctly predicted the relationships of profitability, tangibility, liquidity, and growth opportunities (although the latter was not statistically significant), but not the relationship of company size to indebtedness. The results of the model agree with three predictions of this theory: a negative and significant correlation between profitability, liquidity, and the level of debt; and positive, but not significant, between growth opportunities and the level of debt. The results are consistent with four of the five predictions of this theory.

The reason why there is a negative relationship between profitability, tangibility, and liquidity, and a positive relationship between size and tax shields not generated by debt against the level of debt is as follows: companies prefer to use their own resources to finance themselves, so the more profitable and liquid the company, the less debt it requires. With respect to tangibility, companies with lower levels of fixed assets have a higher level of indebtedness. These companies are presumed to 1) provide current assets as collateral and/or 2) use short-term debt as a financing option.

Concerning the size and tax shields not generated by debt, it is presumed that larger companies obtain greater credits thanks to the low risk of bankruptcy and non-payment of the same, by having a more diversified business than a smaller company. While for tax shields not generated by debt, it is presumed that companies do use debt to take advantage of tax deductions due to interest expenditure.

This leads to the conclusion that a company affiliated to CAPIG definitively prefers to finance itself rather than contract debt with banks or issue debt instruments. These results are consistent with those obtained from ^{Tenjo, López & Zamudio (2006)} for Colombian companies; ^{Paredes, Ángeles, & Flores (2016)} for companies in five Latin countries; ^{Booth, Aivazian, Demirguc-Kunt, & Maksimovic (2001)} for companies in ten developing countries; and ^{Acaravci (2016)} for Turkish companies.

According to a study conducted by the Universidad de Especialidades Espiritu Santo, presented in 2017, 90% of existing Ecuadorian companies are family-owned. This means that it is very likely that both shareholders and directors are related to each other, leaving little room for investors outside the family.

The cost of debt could be an important aspect to consider in order to better understand why companies prefer to finance themselves rather than get into debt. In Ecuador, the interest rate for business loans is in the range of 7-10%⁵ per annum on average; Colombia, 8-12%⁶; while in countries such as Chile and Perú it is 5-7%⁷ and 3-5%⁸, respectively.

Future lines of research should integrate the historical interest rate into the models to estimate the effect they have on the choice of financing by the companies. In addition, it would be enriching to know the determinants of the capital structure by sector and industries of the economy, also adding a variable that segregates companies among family or non-family, so as to have more precise estimates of corporate financing preferences with a wide level of detail.