1. Introduction
The relationship between profitability and working capital has remained a critical issue since the late 1990s. The literature defines working capital as the value of current assets after excluding current liabilities [
1,
2,
3] and refers to the management of current assets, current liabilities, and inventories for profit maximization and sustainable growth [
4,
5,
6,
7,
8]. Working capital management significantly contributes to firm value by maintaining a balance between risk and profitability [
5,
7,
9,
10,
11,
12,
13]. Depending upon managers’ preferences, this balance may have a range of strategies including high risk–high profit (aggressive strategy) or low risk–low profit (conservative strategy) [
14].
The literature discusses how investment in working capital influences profit [
15,
16,
17,
18]. However, the financing of these investments is equally essential for profit maximization. Therefore, studies not only describe the importance of investment but explain how investments should be financed. Financing decisions play an essential role in firm performance. For instance, leverage explains the financing details of a firm and is frequently used in the literature for evaluation of firm performance [
15,
19,
20,
21,
22]. A firm can have multiple sources financing working capital. These sources may be internal, including retained earnings and debt collection, or external, via short- and long-term borrowing [
15].
One important external financing decision involves selecting short- or long-term borrowing. When investing in working capital, making good financing decisions are crucial because short-term and long-term borrowing have advantages and disadvantages that significantly affect profit and risk. For instance, long-term financing may be a safe strategy as it is free from refinancing uncertainties and interest rate fluctuations. Refinancing uncertainty means the lender, on unsatisfactory firm performance, may refuse to renew the loan on the maturity date. However, short-term financing takes advantage of low interest rates and favorable credit conditions when compared to long-term debt [
23]. Short-term financing reduces possible agency problems between creditors, shareholders, and managers of the firm [
24]. All these factors support that both short- and long-term debts should be used to finance working capital. Against this background, their best combination should be determined for maximal profit.
Several studies have examined the impact of working capital finance (WCF) on firm performance. Banos et al. [
23] analyzed Spanish companies regarding short-term financing with working capital and reported an inverted U-shaped relationship between WCF and profitability. Here, WCF refers to the financing of working capital via short-term borrowing. If a firm finances its working capital by short-term borrowing, a positive WCF–profitability relationship exists, where WCF affects profit positively. However, as short-term borrowing increases, this positive relationship gradually diminishes, and ultimately the firm achieves a break-even point where short-term financing has zero effect on profitability. After this break-even point, a negative WCF–profitability relationship, where WCF affects profit negatively, starts to dominate. This positive and negative combination is cumulatively called an inverted U-shaped relationship. Banos et al. [
23] evaluated the influence of financial flexibility on the WCF–profitability relationship and revealed that the WCF–profitability break-even point changes for highly flexible and inflexible firms. Financial flexibility is the capability of a firm to access its financing at low cost [
23]. Break-even point shows the proportion of short- and long-term debts in WCF. If its value is 0.60, WCF carries 60% short-term and 40% long-term debts. At the break-even point, WCF has zero effect on profitability. Panda and Nanda [
25] analyzed six manufacturing sectors in India and reported changes in the WCF–profitability relationship for each sector. They evaluated the influence of markup (sales to profit margin) on the WCF–profitability relationship and devised a different financing strategy for each manufacturing sector.
From this discussion, small and big companies may behave differently in the WCF–profitability relationship due to different ownership structures, financial flexibility, and tax provisions. Also, other factors such as leverage may influence the WCF–profitability relationship, but have been ignored in previous studies. The current study fills the gap in the literature by evaluating the influence of firm size and leverage on the WCF–profitability relationship. To the best of our knowledge, the moderating effects of these two factors on this relationship have not yet been studied.
Apart from this fundamental contribution, we contribute to the existing literature in numerous ways. First, we observed the changes in the break-even point of the WCF–profitability relationship as the size and leverage level of firms change. Second, we considered a long time period of 18 years with a large number of observations (12,610). Third, this is the first study in China to evaluate the WCF–profitability relationship in the context of firm size and leverage. China is a world-leading economy, with private firms growing remarkably since the introduction of the opening up policy [
26]. Finally, we used panel data and conducted analyses using the generalized method of moments (GMM), a modern dynamic technique used to handle numerous data problems including endogeneity and heterogeneity.
The rest of the paper is organized as follows.
Section 2 explains the relationship of WCF with firm performance along with the potential influence of firm size and leverage on this relationship.
Section 3 discusses the research models and variables used.
Section 4 presents the analysis and results.
Section 5 summarizes and concludes the study.
3. Empirical Model and Variables
We designed Equation (1) to test the relationship between WCF and profitability. This model uses the WCF square variable along with WCF to capture the break-even point. The model is
where
is a dependent variable, return on equity, which is a proxy for profitability measured by the ratio of net profit to equity [
8,
9];
(working capital finance) is an independent variable measured as (short-term borrowing/(current assets − accounts payable)) [
8,
11]. The control variables repeatedly used in the literature include
is the log of total assets,
is the sales growth rate, and
is the ratio of total debts to total assets [
2,
3,
6,
8,
9,
24]. Control variables are used to keep the firm performance free from other possible influences and are lagged by one level to address the problem of endogeneity.
is the time dummy variable, which changes over time but remains unchanged in the selected time period;
represents unique features of firms, such as geographical position, which remain constant over time and enable us to manage these individual effects; and
is the error term of the equation. All variables used in the equation are frequently used in the literature, and we also measure these variables through proxies, which are common in the literature.
Firm size is an important variable in this study. Its various proxies, including value of assets, value of sales, market capitalization, log of assets, log of sales, and log of market capitalization, are available in the literature [
38]. These proxies have different levels of sensitivity to firm performance [
39]. We reviewed the literature and found log of assets to be a more frequently proxy used in the working capital context. Following this, we used the same proxy in our study.
We evaluated changes in the break-even point of the WCF–performance relationship under the moderating effects of size and leverage, so coefficients of WCF and its square were the primary focus. The break-even point in this equation is calculated by . Both the signs and values of these coefficients are important for drawing a concrete conclusion about the WCF–performance relationship and its break-even point.
3.1. Firm Size
We evaluated the influence of firm size on the WCF–profitability break-even point. Firm size is calculated as the log of the total assets of the firm. We calculated the median of the firm size variable to divide the sample into small and large firms. Firms with a size value lower than the median were considered small firms, whereas firms with a higher value than the median were considered large firms. We incorporated these effects into Equation (1) by introducing a size dummy variable. This variable has values of 0 and 1 for large and small firms, respectively.
So, Equation (1) was modified to determine the moderating role of firm size on the WCF–profitability relationship:
This equation has a new break-even point of the WCF–profitability relationship under the moderating effect of firm size, which is calculated as .
3.2. Leverage
We also evaluated the influence of leverage on the WCF–profitability relationship. The literature defines leverage as the ratio of total debts to total assets [
22]. We calculated the median of the leverage variable to divide firms into high- and low-leverage firms. All firms with a leverage value lower than the median were considered low-leverage firms, whereas firms with higher leverage than the median were considered high-leverage firms. To capture the moderating effect of leverage, we propose a leverage dummy variable. It has values of 0 and 1 for low- and high-leverage firms, respectively.
We therefore modified Equation (1) to capture the moderating effect of leverage on the WCF–profitability relationship:
This equation introduces a new break-even point of the WCF–profitability relationship under the moderating role of leverage, which is calculated as
.
We used panel data and the GMM estimator to address heteroskedasticity, unobserved heterogeneity, and endogeneity. Endogeneity refers to the correlation of explanatory variables with error terms [
40]. There are various techniques used to address endogeneity, including lagged dependent variables, lagged independent variables, control variables, GMM, and fixed effects [
41]. Among these, the best is the GMM estimator, which has the highest power to deal with endogeneity [
41]. We estimated all models using GMM, a dynamic panel data estimator introduced by Arellano and Bond [
42]. GMM transforms data to remove the effects of all sources of endogeneity, including unobserved heterogeneity, simultaneity, and dynamic endogeneity [
43]. Transformation occurs when endogenous variables are converted into instrumental variables by taking their lag values. GMM offers multiple lags, and we selected the lags that best address the endogeneity. The Sargan–Hansen test was used to evaluate the effectiveness of these instrumental variables. GMM is also robust to heterogeneity and heteroskedasticity issues with data.
We observed an endogeneity problem in the control variables and therefore took their lags at the first level in all models to remove its potential effects. We used STATA software for analysis.
3.3. Data and Sample
The current study is based in China, and we collected 18 years’ worth of secondary data (2000–2017). The source of data was the China Securities Market and Accounting Research (CSMAR) database, which has extensive detail of financial statement data of Chinese companies on a quarterly and annual basis [
44]. The initial sample size had 18,445 observations of manufacturing firms. We condensed the data in the following ways. First, we eliminated all observations with negative or WCF values more than 1. Second, we removed observations with negative values of assets, liabilities, inventories, accounts receivable, accounts payable, and short-term borrowing. We also excluded the extreme top and bottom values of each variable used. We obtained a final sample with 12,609 observations.
5. Conclusions
In this study, we evaluated the WCF–profitability relationship in Chinese companies under the moderating effects of firm size and leverage during a period of 18 years (2000–2017). The literature explored this relationship under financial flexibility and markup. The panel data technique, i.e., GMM, which handles the potential issues of heterogeneity and endogeneity, was used in the main analysis.
The results of the study explain that an inverted U-shaped relationship exists between WCF and firm performance. The results of this study confirm a strong moderating role of firm size and leverage in the WCF–profitability relationship. More specifically, the results revealed that small firms have an inverted U-shaped relationship and their break-even point is lower than the break-even point of the full sample. We observed the same results for high-leverage firms. However, large firms and low-leverage firms show a U-shaped relationship and their break-even points are also lower than the break-even point of the full sample. This means the break-even points of all subgroups are lower than the break-even point of the full sample. The direction of the WCF–profitability relationship shifts from U-shaped to inverted U-shaped only in some subgroups.
The current study provides practical information for managers and policy makers for achieving an optimum WCF–profitability relationship. Specifically, small firms and high-leverage firms should adopt a conservative WCF strategy to maximize profit. In contrast, large firms and low-leverage firms should follow an aggressive strategy for profit maximization. The results could guide managers during modifications of WCF strategy when firms expand or change their leverage level.
We controlled for firm-specific elements in our results and did not consider macroeconomic factors like gross domestic product (GDP) growth, monetary policy, and inflation. These factors may be incorporated in future research to determine their potential influence on the break-even point of the WCF–profitability relationship. We also used data of manufacturing firms only. Non-manufacturing firms’ WCF–profitability behavior can also be analyzed in future studies. All the research on the WCF–profitability relationship in the literature has been conducted on unbalanced data. So, another important direction for future research is to confirm these results using balanced panel data, which produce comparatively more authenticated results.