Organizational Resilience and Performance: Analysis of the Relevance of Suppliers' Trade Credit and Bank Diversification in the Spanish Construction Industry

The objective of this study is to determine the effect of relevant variables related to strategic sources of financial resources — in our case, suppliers’ trade credit and use of financial institutions — over performance among Spanish construction firms. By analyzing a large dataset including information for 3590 Spanish construction businesses during 2004-2011, the results of the longitudinal analysis reveal that trade credit granted by suppliers constitute a relevant source of liquidity and financial resources that positively impacts economic performance. During the period of economic downturn that affected Spain after 2008, those construction firms that benefited from longer average payment periods from their suppliers reported superior performance levels. Additionally, we find that bank diversification is conducive to performance but only during the crisis period: performance is significantly higher in businesses that work with a greater number of financial institutions.

In the specific case of Spain, the crisis in the construction sector has been primarily associated with a combination of factors that include fast growing housing supply in the shortterm, high expectations for supply and demand, and high leverage allowances on production and home purchase by financial institutions (Kapelko et al. 2014, Fernández-López andCoto-Millán 2015).
From an organizational perspective, the identified negative consequences of the decline of the construction industry were mostly linked to credit rationing problems that followed the 2008 housing bubble burst that constrained many organizations' operations (Horta et al. 2013).
Notwithstanding the relevance of credit for the construction industry, financial institutions are not the only suppliers of finance for the industry (Cuñat 2007). As a reaction to organizational changes derived from the economic downturn that started in 2008, scholars have recently focused their efforts on analyzing the effect that the financial strategy of construction firms has on performance. In this regard, special attention is paid to the effects on organizational performance of the management of liquidity levels (Bigelli and Sánchez-Vidal 2012); the management of suppliers and the role of the trade credit granted by them (Cuñat 2007, Garcia-Appendini andMontoriol-Garriga 2013); and the dependence on financial institutions as a source of capital (Chava andRoberts 2008, Kahle andStulz 2013).
The analysis of the relationship between strategic variables related to the financing (in our case, access to credit from suppliers and financial institutions) and performance of Spanish construction businesses is the focus of this study.
Prior studies analyzing the situation of the Spanish construction sector during the crisis period have adopted a productive approach in which the allocation of resources and operational efficiency are key aspects of the analysis (e.g., Kapelko et al. 2014, Fernández-López andCoto-Millán 2015). The analysis of how construction businesses use suppliers' trade credit as a source of financial recourses and how this strategy impacts their performance level has been largely sidelined in prior research.
In line with recent contributions in the field of economics (see e.g., Burkart and Ellinsen 2004, Cuñat 2007, Garcia-Appendini and Montoriol-Garriga 2013, this study focuses on the role of suppliers' trade credit, as opposed to financial suppliers, as an alternative source of finance for credit-constrained businesses in the construction sector. In our approach, in periods of economic downturn and credit rationing, businesses modify their financial strategy-making in the short-term, and the longer payment periods granted by commercial suppliers-who represent a potential substitute form of credit-may play a decisive role on performance.
The empirical application uses a sample of 3590 Spanish businesses operating in the construction sector between 2004 and 2011. The Spanish setting is attractive because it offers the opportunity to analyze how businesses in the construction sector respond to changes in economic conditions as a result to the burst of the housing bubble, and how businesses capitalize on the adoption of new financial strategies based on the exploitation of their relationship with suppliers.
The reminder of the paper is organized as follows. Section 2 presents the theoretical underpinning. Section 3 presents a brief description of the characteristics of the Spanish construction industry and the effects of the burst of the housing bubble. Section 4 describes the data and the methodological approach, while Section 5 offers the empirical results. Finally, Section 6 presents the discussion, concluding remarks and implications of the study.

Sources of Financial Resources: Background Literature and Hypotheses Development
There is a vast amount of literature dealing with the role of financial debt on business performance (see, e.g., Diamond 1993, Bolton and Freixas 2000, Denis and Mihov 2003, Sufi 2009, Rauh and Sufi 2010, Lin et al. 2013. This literature emphasizes the relevance of bank lending as a primary source of financial resources that allows most businesses to operate and generate profits. But, this section focuses on the review of the implications of suppliers' trade credit and seeks to explain to what extent systematic differences in economic conditions shed light on the potential benefits of trade credit. Besides the amount of input sold on credit, a supplier's decision to grant trade credit includes the due date and the cost of the trade credit. In this study we focus on the effect of the maturity component of trade credit as it may reflect the supplier's incentives to sell on credit. It has been argued that commercial suppliers may be concerned Electronic copy available at: https://ssrn.com/abstract=3022425 with losing crucial customers and therefore be willing to support these customers when they have temporary financial difficulties (Wilner 2000).
As a result of the increasing imposition by banks of capital constraints as the creditworthiness of the borrower deteriorates (Nini et al. 2009), firms are making more intensive use of suppliers' trade credit s an alternative, less traditional, source of financing.
Trade credit arises when a supplier allows a customer to delay payment for intermediate goods already delivered (Cuñat 2007, p. 491). The intuition underlying the potential benefit of using suppliers as a source of finance mainly comes from the increased level of liquidity derived from the concession of borrowing facilities to cover the costs of raw materials in the short-term (Rauh and Sufi 2010).
This theoretical argument is based on two elements. First, suppliers are better able to enforce debt repayment than banks because their customers are aware that suppliers can stop the supply of intermediate goods if the trade credit is not repaid (Giannetti et al. 2011). Second, suppliers may act as liquidity providers, supporting their customers whenever they experience temporary liquidity shocks, as it happens in periods of economic recession (Garcia-Appendini and Montoriol-Garriga 2013). Therefore, a strong connection between the supplier and the customer is a necessary condition for increasing the supplier's willingness to grant trade credit: a strong commercial relationship makes it costly for the customer to find alternative suppliers and costly for the supplier to lose its current customers.
At this point, it is important to ask why commercial suppliers are more flexible lenders than banks are. Cuñat (2007) proposes that suppliers have a greater capacity than banks to ensure compliance with debt payments because their relationship with the customer involves bargaining power resulting from either high switching costs or high specificity in the good provided by the supplier. Additionally, suppliers have the capacity to freeze the operations of a business that claims trade credit by cancelling the supply of resources to their clients.
Another relevant issue that is worth questioning is whether suppliers extend trade credit to customers irrespective of the market conditions (growth or recession), and whether businesses use trade credit as a substitute for bank credit in periods of economic decline to improve their financial results.
Liquidity safeguards are an indispensable requirement for business performance in times of crisis. In this regard, Kahle and Stulz (2013) indicate that firms whose credit is mainly granted by banks significantly increase their liquidity and do not reduce their financial costs during the first year of crisis. This result implicitly suggests an increased use of trade credit instruments. Therefore, suppliers may play a central role in supporting business survival and performance when the economy is going through bad times. As we indicated above, resource flow (raw materials) from suppliers is a key aspect of any business' value chain, and payment for these resources is associated with planned cash flow reductions in the short term. Following this logic, suppliers may also act as suppliers of liquidity, in an effort to ensure their long-term relationship with their customers by establishing financial relationships that increase their customers' likelihood of survival in periods of liquidity shocks (Cuñat 2007). Empirical evidence by Garcia-Appendini and Montoriol-Garriga (2013) suggests that businesses employing a larger number of suppliers significantly increase the use of trade credit during the period of economic crisis.
Prior research has documented that suppliers tend to be more financially compassionate towards businesses facing financial problems (e.g., Franks andSussman 2005, Huyghebaert et al. 2007). Businesses with a strategy based on the use of suppliers as a source of financing seek to guarantee a short-term financial position. As stated by Garcia-Appendini and Montoriol-Garriga (2013), the most effective commercial relationships are those involving firms that have a history of high liquidity (liquid assets) and a supplier that is willing to sacrifice its own immediate liquidity in exchange for a guaranteed long-term cash flow. Trade credit tends to be used as the last resort, when other kinds of credit have run out. High growth firms, which need more funding, together with firms experiencing liquidity problems, have a higher proportion of credit from their suppliers (Cuñat 2007). The validity of these arguments amplifies in periods of economic difficulties. Therefore, we hypothesize H1: In periods of economic downturn, the use of suppliers' trade credit-in terms of longer maturity-has a positive impact on business performance Now we turn to the other side of the credit coin analyzed in this study: financial credit.
Financial credit is a means for accessing capital and the crisis is a detonator of breaches of trade deals for different reasons. The financial response to breaches of trade deals is expected to be stronger when credit suppliers take different measures to moderate the supply of capital, and access by debtors to banking-based funding is limited or relatively expensive (Nini et al. 2009).
For businesses using short and long-term debt, financial institutions who supply debt will make no concessions when the debtor cannot completely pay back; i.e. debt will entail risk.
A mixture of short-and long-term bank debt is configured in such a way that the bank can continue to make low-risk loans or extend credit maturity rather than making downward adjustments to credit ratings, for which reason creditors prefer liquidation (Diamond 1993).
Financial firms, in their role as capital suppliers, drastically reduce average short-and longterm credit in times of crisis (Kahle and Stulz 2013).
Banks tend to create a reputation of tough creditors to reduce adverse selection problems, that is, reduce the number of low-quality businesses applying for a bank credit.
Therefore, financial institutions are very inflexible in debt renegotiations with financially distressed businesses (Franks and Sussman 2005). Additionally, financial institutions follow strict liquidation protocols when debtors face financial distress. In the context of this study, the probability of default increases in periods of economic recession, and this is particularly so in Electronic copy available at: https://ssrn.com/abstract=3022425 the case of the construction sector which suffered the most after the crash of the Spain's economy in 2008 (Castro et al. 2009, Kapelko et al. 2014.
In this scenario, businesses have incentives to diversify their portfolio of suppliers of financial capital, that is, a likely solution to overcome liquidity problems comes from increasing the number of banks with which the business has credit operations. Although banking is a competitive sector characterized by shared information, credit diversification offers construction businesses the possibility to access credit lines that are essential for the functioning of these organizations. Therefore, increasing the number of creditors constitutes a valid strategy to obtain additional financial resources that contribute to sustain the business' competitive position. Following this theory and evidence we hypothesize:

H2a: A greater creditor diversification-in terms of the number of financial institutions
with which a business operates-is positively associated with business performance.

H2b: The positive relationship between the number of creditors with which a business
operates and performance is stronger in periods of economic recession.

The Context: The Housing Bubble and the Crisis in the Spanish Construction Sector
The beginning of the 21st century witnessed a rapid and increasing growth in the Spanish construction sector that reached 16% of Spain's gross domestic product in 2006, a figure that doubles that reported in the rest of EU countries (Kapelko et al. 2014). Until 2007, the expansion of the construction industry was a driving force behind Spain's economic growth employing 13.16% of the labor force (Spanish Statistics Office: www.ine.es).
During this period a housing bubble was created as a result of a combination of factors including the growing housing supply in the short-term, the high market expectations of both construction businesses (supply side) and their potential customers (demand side), as well as the matching between competitive interest rates with increased leverage allowances on Electronic copy available at: https://ssrn.com/abstract=3022425 construction activity and home purchase by financial institutions (Horta et al. 2013). The amalgamation of these economic factors led to the formation of similar real estate bubbles in, among others, many EU countries, China, and the United States (Shiller 2008, Gimeno and Martinez-Carrascal 2010, Choy, 2011.
Notwithstanding the stricter lending conditions by financial institutions, housing oversupply and the emerging global financial crisis that started in 2008 when Lehman Brothers filed for bankruptcy protection led to the collapse of the Spanish economy. Following the burst of the housing bubble, which generated a notorious global subprime mortgage crisis (Shiller 2008), the construction sector was most affected due to overpriced property and extreme credit reductions. Fueled by the excessive business-level leverage, the financial constraints imposed by financial institutions put Spanish construction businesses in a more disadvantageous position than firms operating in other sectors, which caused the crunch of the whole industry (Gimeno and Martinez-Carrascal 2010). The breaking of the economic crisis negatively impacted the construction sector by drastically shrinking the sector's output and employment levels.  (2000), the importance of the construction industry is not only related to its size but also to its role in economic development. The deductions coming from the analysis of the evolution of the Spanish construction sector before and after the crash of the economy lead to conclude that the scarcity of financial resources is one of the key factors pushing up the number of bankruptcies in the sector. Yet, existing work on the Spanish construction industry mostly focuses on the underlying operational factors that explain productivity loses in the sector or the relationship between bank lending and activity in the sector (see, e.g., Gimeno and Martinez-Carrascal 2010, Horta et al. 2013, Kapelko et al. 2014, Fernández-López and Coto-Millán 2015. These arguments further justify the proposed analysis of the role of alternative sources of finance-in our case, suppliers' trade credit-on performance; looking for a more comprehensive analysis of how construction businesses shape their financial strategy-making in times of economic downturn, and how their credit choices or possibilities impact their performance level.

Data
The database used in this study was obtained from the Sistema de Análisis de Balances Ibéricos (SABI), provided by the Bureau van Dijk©. This database contains detailed organizational and financial information for Spanish businesses. Additionally, the database includes data on the financial institutions (banks, savings banks and other financial institutions) with which Spanish businesses operate. Due to the interest in studying the behavior of businesses in the construction sector in periods of economic growth and during the crisis period, information was collected for the period between 2004 and 2011.
In line with our core objective, data on suppliers and the banks with which the business has relations is critical for the study. Therefore, the initial database obtained from SABI comprises information for a total number of 20312 businesses in the Spanish construction sector with data available on the accounts payable (suppliers) and the names of the banks associated to the sampled construction businesses (NACE codes: 41 and 42). However, in the interest of following a rigorous methodology, we conducted a meticulous sampling procedure aiming to ensure the robustness of the results. First, we excluded 221 businesses that exited the construction market in the first study year (2004) and 341 businesses that were created in the last studied period (2011). Second, construction businesses were included in the sample if the start-up year was available and if their total assets were clearly identified in the database in each analyzed period (from 2004 to 2011). In this second step, 8284 construction businesses were dropped from the sample mostly because their assets were not available in one or more of the analyzed periods. Third, we included in the sample only those businesses whose operating profit and total debt values were reported in all the study periods. Based on this criterion, 7876 businesses were dropped.
Note that the representativeness of the study sample is ensured insofar as it includes businesses from the 17 Autonomous Communities that form Spain. Table 1  ---Insert Table 1 about here ---

Variable Definition
Dependent variable. We measure economic performance using the return on assets (ROA), defined as the ratio of operating profit divided by total assets. Within the economic and strategic management literatures, this variable has been widely used to proxy the economic performance of organizations operating in different industry sectors (see e.g., Sufi 2009, Rauh and Sufi 2010, Epure and Lafuente 2015. In line with the tone of Spain's economy, descriptive statistics in Table 2  capacity to access credit from its commercial suppliers. This variable-computed as the economic amount of accounts payable (suppliers) divided by the daily cost of materials (cost of materials / 365)-allows to proxy the extent to which construction businesses use suppliers' trade credit to finance their operations. In our sample, businesses pay to their commercial suppliers in 160.24 days (5 months and 10 days approximately) (Table 2). Additionally, descriptive statistics in Table 2  Use of financial institutions. Instead of focusing on the economic value of debt, in this study debt heterogeneity is analyzed via a novel variable that captures the business' capacity to access financial resources from multiple financial agents. Our dataset allows at identifying the total number of financial institutions used by each construction business in the sample. This variable not only captures the businesses' capacity to work with more banks and, therefore, increase the likeliness of accessing to more sources of credit; but also measures the willingness of financial institutions for working with businesses operating in the construction sector (Giannetti et al. 2011).
Control variables. We control for business size, business age, type of financial creditor, geographic location, leverage and time in the different model specifications. Business size is defined as total assets and is expressed in millions of constant euro at 2011 prices, while business age is measured in years of market experience. We acknowledge that businesses can access credit from different financial institutions (Sufi, 2009). Thus, to account for the potential effect that access to different types of creditors may have in performance, we introduced the proportion of banks relative to the number of financial institutions used by the sampled construction businesses. We included a set of dummy variables that account for the location of the sample businesses (in all model specifications Madrid is the omitted Autonomous Community). We introduced the leverage ratio, defined as the relationship between total debt and total assets. Finally, we included two time-related variables. First, we used a 'crisis' dummy taking the value of one if for the period 2008-2011, and zero otherwise. Second, we introduced a set of time dummies to rule out potential the effects of other economic and environmental conditions that may affect the economic performance of businesses in the construction industry.

Method
In line with the arguments that underpin this study, we argue that businesses seek to access financial resources from different sources on the basis of expected performance improvements.
We employ panel data techniques to estimate the proposed model which emphasizes a positive relationship between the access to financial resources from diverse sources (suppliers and financial institutions) and businesses' economic performance. Pooling repeated observations on the same organizations violate the assumption of independence of observations, resulting in autocorrelation in the residuals. First-order autocorrelation occurs when the disturbances in one time period are correlated with those in the previous time period, resulting in incorrect variance estimates, rendering ordinary least squares (OLS) estimates inefficient and biased (Wooldridge 2002). Therefore, we estimate random-effects (GLS) panel data models with robust standard errors to correct for autocorrelation of error terms due to constant university-specific effects (Greene 2003). Additionally, the proposed estimation approach allows at evaluating the effect of relevant time-invariant factors on business performance (in our case, the number of financial institutions used by construction businesses).
To evaluate the role of different sources of finance empirically we propose a random-effects model with the following form: (1) In equation (1) ROA is the economic performance variable computed for each business (i) and each time period (t), j are parameter estimates estimated for the independent variables (j), T refers to the set of time dummy variables, while is the normally distributed error term that varies cross-universities and cross-time (t).
In terms of the study hypotheses, we expect that the increased use of suppliers' trade credit as a source of finance during the crisis period yields superior performance (H1: 12 0 ).
Additionally, we expect that businesses with a greater capacity to operate with more financial institutions improve their performance (H2a: 3 0 ), and that this effect is stronger during the crisis period (H2b: 13 0 ). Table 3 reports the estimates of the random-effects regression models linking the sources of financial resources and economic performance. Model 1 is the baseline specification which includes the different sources of finance analyzed in this work (commercial suppliers and financial institutions) and the control variables. Model 2 includes the main effects and the interaction terms between the 'crisis' dummy and the analyzed sources of finance.

Empirical Results
To address the threat of collinearity, we computed the average variance inflation factor (VIF) for all variables. The average VIF value for model 1 1.88 and ranges between 1.01 and 5.36. For model 2 the average VIF is 7.77, and the only VIF values that exceed 10-a generally accepted rule of thumb for assessing collinearity-were observed for the variables linked to the interaction terms. By construction these terms are correlated and-even if computationally correct-this explains the VIF results (Greene 2003). The results for this diagnostic test do not raise collinearity concerns.
The findings for the control variables indicate that performance is positively associated with businesses size, while the relationship turns negative in the case of the business age variable. Additionally, the results indicate that businesses performance is negatively affected by the exposure to high levels of debt, in relation to business assets.
The results in Table 3 indicate that economic performance of the sampled construction businesses drastically declined during the crisis period. More concretely, performance suffers an average fall of 16.28 percentage points during the period 2008-2011, compared to the average performance level in the growth period (2004)(2005)(2006)(2007).
---Insert Table 3 about here ---Concerning the key results of the study, from model 2 in Table 3 we note that the coefficient for the interaction term between the 'crisis' dummy and the average payment period is positive and statistically significant ( 12 0.0069 and value 5% p ). This suggests that longer payment periods granted by suppliers are associated with superior performance in the crisis period (2008)(2009)(2010)(2011). This is in line with our arguments that suppliers play a central role on business performance in crisis periods by extending payment periods that allow the business to obtain financial resources in the short term, and ensuring their long-term relationship with their customers in periods of liquidity shocks (Cuñat 2007). Therefore, we confirm our hypothesis H1 that proposes that the use of suppliers' trade credit as an alternative source of finance in periods of economic recession leads to superior performance levels.
To aid in the interpretation of the results, we plot the interaction terms between the 'crisis' dummy and the sources of finance variables based on estimates from model 2 (equation (1)). The results are presented in Figures 3 and 4    Consequently, we do not find support to our hypothesis H2a that states that businesses with a greater capacity to operate with more financial institutions show higher performance; while we confirm our hypothesis H2b that proposes that the positive effect of working with more financial institutions on business performance is stronger in crisis periods.

Concluding Remarks, Implications and Future Lines of Research
In this study, we proposed that changes in the configuration of the business' suppliers of finance constitute a valid strategy to improve business performance. Furthermore, we argue that, in the crisis period that follows the bursting of a housing bubble, the incentives to use suppliers' trade credit and to diversify traditional sources of finance (i.e., banks) have important performance implications for businesses in the construction sector.
The crisis represents an unexpected negative shock to the supply of external finance for construction businesses (Kapelko et al. 2014 2008. Also, it is found that credit diversification-i.e., number of banks associated to the construction businesses-is conducive to performance but only during the crisis period.
The results of this study have relevant implications for scholars and practitioners. From an academic perspective, the results contribute to extend the growing literature on the relevance of suppliers' trade credit as a means for enhancing business performance (e.g., Burkart and Ellinsen 2004, Cuñat 2007, Garcia-Appendini and Montoriol-Garrig 2013. In the context of this study, we argue that the specific technological characteristics of the production process in the construction sector-e.g., high interactions between suppliers and corporate clients, The finding that increased credit diversification is conducive to superior performance in periods of economic downturn helps to better understand the way through which strategic changes of businesses' debt structure yield superior economic results. By analyzing a novel source of debt heterogeneity based on the number of banks with which construction businesses work, this study also increases the large literature on debt heterogeneity (e.g., Diamond 1993, Huyghebaert et al. 2007, Sufi 2009, Rauh and Sufi 2010, Kahle and Stulz 2013. The construction industry involves a series of cooperative relationships between clients, construction businesses, specialist sub-contractors and suppliers (Keung and Shen 2017). We suggest that construction managers need to turn their attention to the relationships with their suppliers when considering the introduction of strategic changes that will modify the business' debt structure. Bank credit is not the only source of financial resources, and our results underline the relevance of both supplier's trade credit and bank diversification. Because the collaborative incentives between suppliers and construction businesses are stronger in a crisis period, a network analysis seems necessary (Pryke 2005). By conducting a profound analysis of the business' commercial networks, managers of construction businesses will be in a better position to understand the potential value of trade credit as an alternative source of financing as well as to better develop business strategy in periods of economic growth or stagnation.