INTRODUCTION
The problem statement
The 2008 financial crisis marked the worst economic downturn since the Great Depression, significantly impacting global economies, particularly in Southeast Asia Vietnam, however, managed to remain relatively resilient, yet experienced low GDP growth rates from 2009 to 2015, indicating lingering effects of the crisis During this period, inflation surged to an alarming 19% in 2011 before declining, while bad debts became a pressing issue, leading to an increase in the number of enterprises reporting losses and many businesses facing dissolution Consequently, effective cash management gained prominence, as firms recognized that maintaining cash holdings is crucial for survival in a challenging economic landscape where accessing capital markets proves difficult Understanding the factors that influence a firm's cash holdings is essential for successful cash management.
Previous research has extensively examined the relationship between cash flow and cash holdings, yielding varied conclusions Almeida et al (2004) and Khurana et al (2006) identified a positive sensitivity of cash flow, while Riddick and Whited (2009), along with Bao, Chan, and Zhang (2012), reported a negative sensitivity.
The research on cash flow sensitivity of cash holdings is crucial for firms to develop more effective cash management models, leading to more efficient cash utilization Despite its importance, the impact of cash flow sensitivity on cash holdings in Vietnam remains underexplored, with ongoing debates globally regarding this issue Consequently, the author has selected the topic "The Effect of Cash Flow Sensitivity on Cash Holdings: Evidence from Vietnam" to investigate the relationship between cash flow sensitivity and cash holdings in Vietnamese enterprises.
The research objectives
The main purpose is examing the effect of cash flow sensitivity on cash holdings, the paper includes two basic objectives:
- Examining the effect of cash flow sensitivity on cash holdings in Vietnamese enterprises
- Verifying the nonlinear relationship between cash flow sensitivity and cash holdings
To accomplish two research objectives, the paper will solve the following research problems:
- Is there any evidence in the world about the effect of cash flow sensitivity on cash holdings?
- Which method is appropriate to test the effect of cash flow sensitivity on cash holdings in Vietnam?
- What factors affects the cash flow sensitivity on cash holdings?
- What is the effect of cash flow sensitivity on cash holdings in case the firm has positive cash flow and negative cash flow?
Research contribution
Research on cash flow sensitivities is crucial for firms, as it indicates the level of risk associated with changes in cash flow High cash flow sensitivity can complicate a company's financial policies, impacting its ability to maintain a stable capital structure and implement effective dividend policies due to liquidity concerns Understanding the implications of cash flow sensitivity enables firms to enhance their cash management strategies and optimize dividend payments and capital mobilization efforts Furthermore, the findings of this study will contribute to the broader discourse on the relationship between cash flow sensitivity and cash holdings.
The thesis structure
The research is organized in the following chapters:
Chapter 1: Introduction This chapter presents the reasons for choosing topics, research objectives, research methods, research contribution and thesis structure
Chapter 2: Overview of previous studies on the impact of cash flow sensitivity on cash holdings This chapter presents the results of the previous study on the effect of cash flow sensitivity on cash holdings, findings, arguments and limited issues in these studies
Chapter 3: Research Methods This chapter will detail the research model, variables in the model, data as well as expectations about the research results
Chapter 4: Research Results This chapter presents and discusses the results of the study on the effect of cash flow sensitivity on cash holdings of firms in Vietnam, the results of examination with financial constraints and agency problem
Chapter 5: Conclusions and limitations of the study This chapter presents the contributions of the research, the next research direction, and the limitations.
LITERATURE REVIEW
Introduction
Companies maintain cash reserves primarily to settle debts and finance investments When access to capital markets is unrestricted, cash holdings tend to be lower However, seeking external funding often incurs significant capital expenditures, prompting firms to prioritize the use of their own cash to mitigate these costs A crucial source of cash flow for companies is generated from their revenues, and fluctuations in this cash flow directly impact their cash holdings This relationship has been a focal point for researchers examining cash management strategies.
Studies on the effects of cash flow sensitivity on cash holdings
2.2.1 The impact of change in cash flow on cash holdings
Almeida et al developed a liquidity demand model that aligns with Keynes' insights, suggesting that firms facing capital constraints maintain cash reserves derived from cash inflows They argue that financially unconstrained firms exhibit no cash incentives; thus, their cash holdings remain stable despite fluctuations in cash flow In contrast, firms with financial constraints are significantly impacted by changes in cash flow, leading to adjustments in their cash reserves.
Almeida et al conducted a study using a database of manufacturing firms from 1971 to 2000, applying OLS regression to estimate their model They hypothesized that changes in cash holdings would correlate with cash flow shocks and that financially constrained firms' cash holdings would be influenced by the potential for future investment opportunities To capture the unobserved value of long-term growth options, they incorporated the Q variable into their analysis.
The OLS regression analysis revealed that firms with limited access to capital tend to increase their cash holdings when cash flow is positive, and decrease them when cash flow is negative Additionally, the variable Q exhibited a positive and significant relationship for financially constrained firms, indicating that these firms are likely to boost their cash reserves in anticipation of future investment opportunities.
Riddick and Whited (2009) examined cash flow sensitivity by validating various theories and models, including that of Almeida et al They contended that Almeida's model indicates positive cash flow does not lead to higher yields on fixed assets Consequently, when a company experiences positive cash flow, it lacks the incentive to convert its liquid assets into fixed assets, opting instead to bolster its cash reserves.
The results of Riddick and Whited differed from the results of Almeida et al It was caused by correction of the measurement error in the Tobin's Q variable Greene
Measurement errors in Tobin's Q can influence the cash flow variable, as noted by (1997, p 440) Such errors can impact all coefficients in a regression analysis, especially if the regression variable is correlated with other variables.
Riddick and Whited employed fourth-order GMM estimation to address measurement errors in Tobin's Q, revealing that companies with positive cash flow tend to see a decline in cash reserves This decrease occurs due to a positive yield shock that boosts cash flow and capital margin profit, prompting companies to utilize their cash reserves for acquiring more productive tangible assets and investing, ultimately leading to reduced cash reserves.
Bao, Chan, and Zhang (2012) employed the GMM4 estimation alongside an enhanced empirical model to validate the findings of Riddick and Whited They incorporated additional control variables, including firm size, capital expenditure, non-cash flow, and short-term debt, into Riddick and Whited's experimental framework Their results reaffirmed the negative correlation between cash flow sensitivity and cash holdings, consistent with the earlier study.
Researchers are investigating the impact of cash flow sensitivity on cash holdings, focusing on whether the relationship between cash flow and its sensitivity is linear or nonlinear A linear relationship implies that changes in cash flow, whether positive or negative, do not affect the magnitude of cash reserve changes Faulkender and Wang (2006) argue that insufficient cash can lead to a diversification of a company's cash reserves This indicates that the effect of cash flow sensitivity on cash flow may differ significantly between firms experiencing positive cash flow and those facing negative cash flow.
Research by Almeida et al (2004) revealed that financially constrained firms exhibit positive cash flow sensitivity; however, there is no significant difference in the impact of cash flow sensitivity on cash flow between firms with positive cash flow and those with negative cash flow.
Riddick and Whited (2009) highlighted that medium and large firms demonstrate a negative nonlinear relationship between cash flow sensitivity and cash holdings While their study did not specifically examine this nonlinear relationship, they noted that an increase in cash flow typically leads firms to retain cash for investments, as higher cash flow indicates more productive tangible assets As cash flows rise, companies are more likely to allocate reserves towards lucrative projects Conversely, negative cash flow suggests low productivity of tangible fixed assets or projects with negative net present value (NPV), prompting firms to halt these initiatives to conserve cash Thus, Riddick and Whited concluded that the relationship between cash flow sensitivity and cash holdings remains negative, regardless of whether cash flow is positive or negative.
According to Bao, Chan, and Zhang (2012), companies often struggle to immediately terminate projects with negative NPVs during periods of negative cash flow due to three primary reasons Firstly, binding contracts associated with certain projects, particularly tender projects, prevent immediate cessation Secondly, as noted by Kothari et al (2009), managers may have incentives to conceal unfavorable information; terminating a poor project could expose the company's issues, prompting managers to retain such projects to delay bad news until better information can emerge Lastly, the agency problem highlighted by Jensen and Meckling (1976) and Jensen (1986) suggests that managers might continue investing in negative NPV projects to maximize their own profits, which can lead to a depletion of cash reserves rather than an increase, as they prioritize personal gain over the company's financial health.
A company experiencing negative cash flow may not immediately terminate unprofitable projects, as its cash reserves may remain stagnant Research conducted by Bao, Chan, and Zhang on manufacturing firms from 1972 to 2006 explored the impact of cash flow sensitivity on cash flows, revealing a nonlinear relationship in cash holdings between firms with positive and negative cash flow Their findings indicated that cash flow sensitivity is negative for firms with positive cash flow, consistent with Riddick and Whited (2009), while it becomes positive for firms with negative cash flow This asymmetrical influence of cash flow sensitivity on cash holdings supports the hypothesis proposed by Bao, Chan, and Zhang.
2.2.2 Relationship between financial constraints and cash holdings of a company
In corporate finance, two key research areas are the influence of financial constraints on corporate policy and the management of finance by firms These aspects are interconnected; when a company enjoys unrestricted access to external funding, it does not require substantial cash reserves for future investments Conversely, if a firm faces high costs that limit its access to external funding, maintaining adequate cash reserves becomes essential to meet its financial needs.
Kaplan and Zingales (1997) define financially constrained firms as those that differentiate between internal and external capital expenditures However, it appears that all companies face some level of financial constraints, as even minimal transaction costs associated with external funding can categorize them as such Additionally, research by Fazzari, Hubbard, and Petersen (1988) and others suggests that a higher sensitivity of cash flow to investment indicates greater financial constraints.
Almeida et al (2004) revealed that financial constraints significantly influence corporate behavior by affecting a company's payment needs They argued that maintaining high cash reserves is costly, as it necessitates forgoing profitable investment opportunities Consequently, financially constrained firms adopt an optimal cash policy to balance the returns from current and future investments, contrasting with unconstrained firms that do not incur cash holding costs and typically do not utilize cash reserves.
Summary of previous findings
Numerous studies have explored the impact of cash flow sensitivity on cash holdings, with a consensus emerging on its significance Almeida et al (2004) utilized the OLS method to analyze data from US manufacturing firms between 1971 and 2000, revealing a positive correlation between cash flow changes and cash reserves Their findings indicated that firms adjust their cash holdings in response to cash flow fluctuations, particularly highlighting that only financially constrained firms, due to limited access to external capital, save cash from cash flow This conclusion aligns with the subsequent research by Khurana et al (2006), which supports the notion that cash flow sensitivity primarily affects firms facing financial constraints.
Almeida et al suggest that firms with positive cash flow reinvest profits into cash reserves, but Riddick and Whited (2009) challenge this view They argue that Almeida's model shows a misleading positive cash flow sensitivity, as increased cash flow does not necessarily correlate with enhanced productivity of physical assets Instead, Riddick and Whited posit that firms experiencing cash flow growth are likely to invest those reserves, reflecting higher asset productivity They criticize the Ordinary Least Squares (OLS) method used by Almeida, citing measurement errors in Tobin's Q that could skew results By employing the General Method of Moments (GMM) estimation, as proposed by Erickson and Whited (2000), Riddick and Whited found a negative cash flow sensitivity, indicating that increased cash flow leads firms to reduce cash holdings in favor of investments Their findings contradict Almeida's conclusions, emphasizing the need for a reevaluation of cash flow sensitivity in financial models.
Almeida et al argued that firms without financial constraints experience minimal impact on their cash due to fluctuations in cash flow In contrast, Riddick and Whited demonstrated that these unconstrained firms exhibit a higher sensitivity to cash flow changes compared to their financially constrained counterparts.
Sharing the same empirical method with Riddick and Whited (2009), suggesting that the fourth-order GMM (GMM4) was more suitable than OLS, Bao, Chan and Zhang
In 2012, a reassessment of cash flow sensitivity's impact on cash holdings reaffirmed Riddick and Whited's findings However, Bao, Chan, and Zhang advanced the discussion by exploring a nonlinear relationship between cash flow sensitivity and cash holdings They posited that cash flow sensitivity could be positive in certain situations, particularly when firms experience negative cash flow While Riddick and Whited suggested that negative cash flow indicated declining productivity or ineffective investments, leading firms to halt projects and increase cash reserves, Bao, Chan, and Zhang argued that firms might not be able to immediately cease unproductive projects Their empirical findings supported the notion that negative cash flow sensitivity may not always be sustainable.
Research indicates that agency costs significantly influence the relationship between cash flow sensitivity and cash holdings Dittmar, Mahrt-Smith, and Servaes (2003) suggest that firms facing larger agency problems tend to maintain higher cash reserves Conversely, Bao, Chan, and Zhang (2012) highlight that when a firm's shares are predominantly held by institutional investors, indicating stronger external control and reduced agency issues, the impact of cash flow sensitivity on cash holdings diminishes.
RESEARCH METHODOLOGY
The analytical framework
The difficulties in accessing capital
The sensitivity of Cash flow
The econometric models
This research paper addresses three key issues: the influence of cash flow sensitivity on cash holdings, the impact of financial constraints, and the implications of agency costs To explore these problems, the study employs three distinct models for analysis.
3.2.1 Model the effect of cash flow sensitivity on cash holdings
This study utilizes the model proposed by Bao, Chan, and Zhang (2012) to investigate how cash flow sensitivity influences cash holdings, emphasizing the differing impacts of positive and negative cash flows The model is represented by the equation ΔCashHoldings_it = α_0 + α_1 CashFlow_it + α_2 Neg_it + α_3 CashFlow_it * Neg_it + α_4 Q_it + α_5 Size_it + α_6 Expenditure_it + α_7 Acquisition_it + α_8 ΔNCWC_it + α_9 ShortDebt_it-1 + ε_it, which incorporates various factors affecting cash holdings, including firm size, expenditures, acquisitions, changes in net working capital, and prior short-term debt.
CashHoldings: is calculated as the cash in company divided by total asset ΔCashHoldings: the difference of cash between year t and year t−1 over total assets
CashFlow: is the profit after interest, dividends and taxes payments plus depreciation over total assets
Neg: is the indicator variable Neg equals zero if in that year the firm has positive cash flow and one otherwise
Q: represents the market capitalization of the company Q is calculated by total of the market value of the capital and the book value of assets minus the book value of the capital and divided by the book value of assets
Size: represents the scope of the company, calculated as the natural logarithm of total assets
Expenditure: the ratio of capital expenditures to total assets
Acquisition: the indicator variable If the company has no acquisition in that year, Acquisition equals one and zero otherwise
NCWC: is calculated as net non-cash working capital which equals working capital minus cash divided by total assets ΔNCWC: is the difference of NCWC between year t and year t−1
ShortDebt: is the debt in short-term weighted by total assets i and t refers to firm and year respectively, ε is random error term
3.2.2 The effect of cash flow sensitivity on cash holdings is under financial constraints
This study investigates the impact of cash flow sensitivity on cash holdings and analyzes the differences between financially constrained and unconstrained firms To achieve this, the research employs three distinct measures to categorize the sample into these two groups.
The study utilizes the KZ index, derived from Kaplan and Zingales (1997), as it aligns with data specific to Vietnam This index is calculated using specific metrics that reflect the financial conditions within the country.
KZindex = -1.002 x CashFlow + 0.283 x Q + 3.139 x Leverage - 39.368 x Dividends - 1.315 x CashHoldings
CashFlow: is the profit after interest, dividends and taxes payments plus depreciation over total assets
Q: represents the market capitalization of the company Q is calculated by total of the market value of the capital and the book value of assets minus the book value of the capital and divided by the book value of assets
Leverage: represents the capital structure of the firm which is calculated as the ratio of the total debt and total assets
Dividends: represents the dividend policy of the company which is calculated as the cash dividends divided by total assets
CashHoldings: is calculated as the cash in company divided by total asset
The study assesses companies using the KZ index developed by Almeida et al (2004), categorizing firms within the highest 33% of the KZ index over a year as being in the financial constraint group.
Secondly, the study uses an additional index to assess the financial constraints among firms This index is based on the research by Whited and Wu (2006) and is called the
According to Bao, Chan, and Zhang (2012), the WW index is a more suitable measure than the KZ index for assessing a firm's financial constraints, as the characteristics captured by the WW index are more closely aligned with these constraints.
WW does not include Tobin's Q The data for calculating the WW are also consistent with the data in Vietnam The WW index is calculated as follows:
WWindex = -0.091 x CashFlow it - 0.062 x DIVPOS it + 0.021 x TLTD it – 0.044 x Size it
CashFlow: is the profit after interest, dividends and taxes payments plus depreciation over total assets
DIVPOS: the indicator variable If firm i pays dividend by cash in year t, DIVPOS is consider as one and zero otherwise
TLTD: is the debt in long-term weighted by total assets
Size: represents the scope of the company, calculated as the natural logarithm of total assets
ISG: the industry's revenue growth rate
SG: the growth rate of the company
According to Bao, Chan and Zhang (2012), in a year, firms in the top 25% of the highest WW index were considered as firms in the financial constraint group
The study evaluates whether a firm falls into the financially constrained category based on its dividend payments; specifically, firms that do not distribute dividends in a given year are classified as financially constrained, while those that do pay dividends are not.
The study investigates the impact of financial constraints on a firm's cash holdings by modifying the initial equation into a new format The revised equation (2) incorporates various factors, including cash flow, negative cash flow, and constraints, along with their interactions, to provide a comprehensive analysis Key variables such as firm size, expenditures, acquisitions, changes in net working capital, and short-term debt are also included to enhance the understanding of cash holding dynamics under financial constraints.
The variables in equation (2) are defined as in equation (1)
In equation (2), there is a constraint dummy variable (value of one if the firm is considered financially constrained) and Constraint's interactive variables with CashFlow variable, Neg dummy variable
3.2.3 The effect of cash flow sensitivity on cash holdings is in the agency problem
This study examines the agency problem's impact by employing a comprehensive model that incorporates various factors affecting cash holdings The model includes key variables such as cash flow, negative cash flow, institutional ownership, and firm size, alongside interaction terms that capture the relationships between these elements By analyzing these components, the research aims to provide insights into how different financial dynamics influence corporate cash management strategies.
+ γ 10 Expenditure it + γ 11 Acquisition it + γ 12 ΔNCWC it + γ 13 ShortDebt it-1 + ε it (3)
Where: Inst equals one if the number of shares held by institutional shareholders is in the top 10% of the company
The variables in the model are calculated as follows:
In their study on cash flow sensitivity, Almeida et al (2004) defined cash holdings as the sum of a firm's cash and marketable securities divided by its total assets Similarly, Bao, Chan, and Zhang (2012) defined cash as the firm's total cash holdings relative to total assets Utilizing financial statements from Vietnamese firms, the current study calculates cash holdings by taking cash and cash equivalents from the balance sheet, the most significant liquid asset, and dividing it by the firm's total assets.
CashHoldings: is calculated as the cash in company divided by total asset
Cash flow, as defined by Almeida et al (2004) and Bao, Chan, and Zhang (2012), is calculated as earnings before extraordinary items and depreciation divided by total assets In the context of Vietnamese firms, extraordinary items are absent from their income statements, leading the study to adopt the variable definition proposed by Bates et al.
(2009), in which CashFlow variable is the profit after interest, dividends and taxes payments plus depreciation By this definition, the CashFlow variable is calculated as follows:
Where: depreciation is taken from the depreciation of fixed assets on the indirect cash flow statement
Tobin's Q, introduced by James Tobin in 1969, is a crucial financial metric that represents the ratio of a company's market value to its asset replacement value Although the original margin value q is not directly observable, Tobin's Q has become a widely accepted substitute in empirical research, as demonstrated in studies by Bates et al (2009) and Bao, Chan, and Zhang (2012) This value is determined by evaluating the firm's market capitalization against its book value, providing insights into the company's valuation and investment potential.
Book value of debts is taken from liabilities item on balance sheet
Market capilization equals P/B multiplied by book value of the capital ie equals outstanding shares multiplied by price at that time
Total book value of assets is taken from total assets item on balance sheet
Capital expenditure is closely linked to a firm's cash flow, as Riddick and Whited (2009) noted that increased cash flow leads to enhanced productivity of tangible fixed assets Instead of hoarding cash, firms reinvest it to acquire more fixed assets when they experience positive cash flow This spending on fixed assets for investment projects is termed capital expenditure, indicating potential investment opportunities To quantify capital expenditures, the study calculates it as the difference between a firm's fixed assets in the current year and the previous year, adjusted for fixed asset depreciation in both years.
The data
This study analyzes cash holdings by examining data from 274 non-financial firms listed on the Hanoi Stock Exchange (HNX) and the Hochiminh Stock Exchange (HOSE) over the period from 2009 to 2015 Excluding financial sector firms such as those in real estate, securities, banking, and insurance, the research adheres to special accounting standards highlighted by Shadi Farshadfar and Reza Monem (2013) The resulting panel data comprises 1,918 observations, sourced from the financial statements of the firms, which were obtained from Vietstock and CafeF.
Fixed Effects and Random Effects
This study employs the Ordinary Least Squares (OLS) regression method for panel data analysis, utilizing both fixed and random effects models Two quantitative methods are executed concurrently on the collected data Following the regression analysis, the Hausman test is applied to evaluate the outcomes of the two methods, determining which yields more reliable results.
Hausman's test evaluates the performance of Fixed effects and Random effects regression models by analyzing the differences in their regression coefficients The null hypothesis posits that these differences are not systematic; if H0 is accepted, it indicates that Random effects are more suitable Conversely, rejecting the null hypothesis suggests that Fixed effects provide a better fit for the data.
If the Hausman test indicates that the Fixed Effects Model (FEM) is superior to the Random Effects Model (REM), the study will adopt the findings from the Fixed Effects model To address heteroskedasticity, the study will apply Robust and Cluster adjustments to the Fixed Effects model.
When heteroskedasticity is present, Ordinary Least Squares (OLS) estimation remains an unbiased estimator; however, the efficiency of the estimates is compromised due to increased variance Consequently, the variance estimates become biased, rendering significance levels and confidence interval tests based on t and F statistics unreliable This highlights that while the regression coefficients remain accurate, the standard errors are affected by the presence of heteroskedasticity.
Robust adjustment method, the regression coefficient will be retained, and the variance will be adjusted to achieve the smallest value
Cluster, to fix the autocorrelation, the study uses cluster-adjusted By fixing the id value (representing for companies), this measure will fix the autocorrelation phenomenon in the variance
When estimating with FE adding Robust and Cluster, the results will be better, and the forecasting will be better.
GMM4 estimations
To address the issue of error in variable estimation, employing instrumental variables is essential, especially when dealing with endogenous variables that correlate with model errors Instrumental variables should be correlated with the endogenous variables but not with the errors of the model to ensure accurate estimation This paper will explore an alternative method proposed by Erickson and Whited (2000), specifically high-order GMM estimation, to enhance the reliability of the model's results.
Erickson and Whited introduced a method for estimating models that effectively measures variables and accounts for measurement errors without relying on tool variables This approach utilizes available information from high-order moments of the model's variables to enhance accuracy.
In probability theory, if F represents a random variable and k is a natural number, the k-th order moment of F is denoted as E(F^k), while the k-th order central moment is expressed as E((F - E(F))^k) Each moment provides insights into the distribution characteristics of the random variable, with the first-order moment corresponding to the expected value and the second-order central moment representing the variance.
High-order GMM estimation leverages the information from high-order moments of explanatory variables to compute regression coefficients, minimizing bias from measurement errors in the variables The effectiveness of this GMM4 estimation approach is validated through the Monte Carlo simulation model developed by Erickson.
Whited (2000), as well as in some recent studies The high-order GMM estimation is considered to be able to overcome the problem of measurement errors of Tobin's Q
This study employs FEM, REM, and fourth-order GMM as estimation methods for the research model By analyzing regression results from data collected in Vietnam, the paper aims to identify the most appropriate method to serve as the primary estimation for the research model.
Expectations on research results
3.6.1 Effects of cash flow sensitivity on cash holdings
The test equation for assessing the impact of cash flow sensitivity on cash holdings is represented as ΔCashHoldings it = α 0 + α 1 CashFlow it + α 2 Neg it + α 3 Cashflow it * Neg it + α 4 Q it + α 5 Size it + α 6 Expenditure it + α 7 Acquisition it + α 8 ΔNCWC it + α 9 ShortDebt it-1 + ε it Previous studies suggest that the α 1 coefficient of the CashFlow variable is anticipated to be positive in OLS regression, aligning with Almeida et al (2004), who posited that firms accumulate cash during periods of positive cash flow for future needs Conversely, GMM4 regression results indicate a negative α 1 coefficient, suggesting a negative cash flow sensitivity However, given that the data from Vietnam pertains to a post-crisis period with a short sampling timeframe, this study posits that the negative impact of cash flow may not apply to Vietnamese firms, leading to an expectation that the α 1 coefficient could be either negative or positive Additionally, the α 3 coefficient is expected to be positive, as firms experiencing negative cash flow may rely on cash reserves to finance ongoing projects A significant positive α 3 coefficient, alongside a negative α 1 coefficient, would imply a disparity in the effects of cash flow sensitivity on cash holdings.
The coefficients for the control variables Expenditure and Acquisition are anticipated to be negative, indicating that capital spending and mergers diminish the company's cash reserves In contrast, the ShortDebt variable may yield either a positive or negative coefficient, reflecting the cash flow outflows during the year, which could either decrease cash holdings or enhance the manager's cash flow.
3.6.2 Effects of cash flow sensitivity on cash holdings under financial constraints ΔCashHoldings it = β 0 + β 1 CashFlow it + β 2 Neg it + β 3 Cashflow it * Neg it + β 4 Constraintit + β 5 CashFlow it * Constraint it + β 6 Constraint it * Neg it + β 7 CashFlow it * Constraint it * Neg it + β 8 Q it + β 9 Size it + β 10 Expenditure it + β 11 Acquisition it + β 12 ΔNCWC it + β 13 ShortDebt it-1 + ε it (2)
The β 1 coefficient may be either negative or positive, while the β 3 coefficient is anticipated to be positive, reflecting the influence of cash flow sensitivity on cash management, which remains controlled by financial constraints The β 5 coefficient is also expected to be positive, as financial constraints often arise from limited capital access, leading firms to forgo valuable investment opportunities in favor of increasing cash reserves during periods of positive cash flow Consequently, the β 7 coefficient is predicted to be negative, indicating that firms facing financial constraints struggle to sustain existing projects using cash reserves compared to their financially unconstrained counterparts.
3.6.3 Effects of cash flow sensitivity on cash holdings under agency problem ΔCashHoldings it = γ 0 + γ 1 CashFlow it + γ 2 Neg it + γ 3 Cashflow it * Neg it + γ 4 Inst it + γ 5 CashFlow it * Inst it + γ 6 Inst it * Neg it + γ 7 CashFlow it * Inst it * Neg it + γ 8 Q it + γ 9 Size it
+ γ 10 Expenditure it + γ 11 Acquisition it + γ 12 ΔNCWC it + γ 13 ShortDebt it-1 + ε it (3)
The Inst variable indicates the volume of institutional shares in firms, with low representation in large firms The coefficients in the equation suggest that γ1 may be either negative or positive, while γ3 is anticipated to be positive, reflecting the influence of cash flow sensitivity on cash holdings amidst agency costs Conversely, γ5 is expected to be negative, as firms with greater external control experience lower agency costs, leading to reduced cash accumulation in favor of financing projects with positive cash flow Additionally, the γ7 coefficient is projected to be negative, indicating that firms with negative cash flows may increase cash holdings to prevent managers from over-investing in inefficient projects for personal gain.
RESEARCH RESULTS
Data description
(This table shows the number of firms in each industry as well as the percentage of firms in each industry in the sample)
N Mean Median Std dev Q1 Q3 ΔCashHoldings 1,918 0.030 0.011 0.109 -0.014 0.065 Cashholdings 1,918 0.117 0.076 0.124 0.028 0.165
Table 4 3: Compare the mean of variables
The table provides descriptive statistics for the model's variables and compares the mean values between two groups of firms categorized by the WW index CashHoldings is defined as the company's cash divided by total assets, while ΔCashHoldings represents the change in cash from year t to year t−1, also expressed as a ratio of total assets CashFlow is calculated as the profit remaining after accounting for interest, dividends, and taxes, plus depreciation, relative to total assets Additionally, Q is derived from the total market value of capital and the book value of assets, minus the book value of capital, and then divided accordingly.
Constraint Unconstraint Difference t statistics ΔCashHoldings 0.024 0.032 -0.008 1.352
ShortDebt 0.325 0.389 -0.064 1.504 the book value of assets, Size is the natural logarithm of the total asset, Expenditure is the ratio of capital expenditures to total assets, Acquisition is an indicator variable that equals one if the firm makes an acquisition in that year and zero otherwise, ΔNCWC is the difference in non-cash working capital between year t and year t−1, ShortDebt is the debt in short-term weighted by total assets)
Tables 4.2 and 4.3 provide descriptive statistics for the model variables, with Table 4.2 detailing the statistics and Table 4.3 comparing the means between two groups of companies based on the WW index The ΔCashHoldings variable has a mean of 0.030 and a median of 0.011, indicating minimal changes in cash holdings among the sampled companies, which, on average, hold 11.7% of their total assets in cash In contrast, short-term debt constitutes an average of 38.5% of total assets, reflecting a relatively high debt ratio for Vietnamese firms Additionally, non-cash working capital shows slight variation, with a mean of 0.162 and a median of 0.018 Notably, the rate of mergers and acquisitions is low, with only 5.1% of the companies in the sample engaged in such activities.
Table 4.3 indicates that there are no significant differences in cash holdings, acquisitions, working capital, and short-term debt among firms Financially unconstrained firms engage in more mergers and acquisitions and larger capital expenditures, as they have easier access to funding compared to their constrained counterparts Additionally, the Q coefficients for these unconstrained firms suggest they possess greater growth opportunities.
Table 4 4: Pearson and Spearman correlation coefficients
The table illustrates the correlation coefficients among the model's variables, featuring the Pearson coefficient beneath the diagonal and the Spearman coefficient above it Notably, the correlation between ΔCashHoldings and CashFlow is 0.171, indicating a positive relationship, while the correlation with Expenditure shows a negative value of -0.100 Other correlations, such as ΔCashHoldings with Size (0.040) and ShortDebt (-0.033), reveal varying degrees of association, emphasizing the intricate interplay between these financial variables.
Table 4.4 displays the Pearson and Spearman correlation coefficients among the model variables The correlation between Q and CashFlow is 0.178, the highest among the variables, yet it remains low, implying that Q error has minimal impact on CashFlow's coefficient in the OLS regression Additionally, the correlation between ΔCashHoldings and CashFlow is both positive and significant, suggesting that higher cash flow leads to increased cash holdings for companies Furthermore, the positive correlation between CashFlow and Expenditure indicates that when companies experience increased cash flow, they are likely to invest in new projects, a trend that will be further explored in the model estimation results.
Regression results and discussions
4.2.1 Effects of cash flow sensitivity on cash holdings
This study aims to examine the relationship between cash and cash flow sensitivity using the Almeida et al (2004) model, which features fewer control variables than the primary model The Almeida et al (2000) model is expressed as ΔCashHoldings it = α 0 + α 1 CashFlow it + α 2 Q it + α 3 Size it + ε it The findings will be compared to the main model to assess their effectiveness.
Riddick and Whited (2009) demonstrated that measurement errors in explanatory variables can impact the estimates of other variables in the model To address this issue, Toni and Whited (2000) introduced a high-order Generalized Method of Moments (GMM) approach This study will utilize Fixed Effects Model (FEM), Random Effects Model (REM), and GMM4 to estimate the proposed model effectively.
Table 4 5: The results of Almeida (2004)
Dep.Var: ΔCashHoldings GMM4 FEM REM
Hausman specification test between FEM and REM χ2 (df =3) = 18.23
The study analyzes non-financial companies from 2009 to 2015, focusing on the dependent variable ΔCashHoldings, which represents the change in cash as a percentage of total assets from year t to year t−1 CashFlow is defined as profit after accounting for interest, dividends, taxes, and depreciation, also expressed as a ratio of total assets The variable Q is calculated by taking the sum of the market value of capital and the book value of assets, subtracting the book value of capital, and dividing the result by the book value of assets Additionally, Size is measured as the natural logarithm of total assets Significance levels are indicated with symbols: ***, **, and * denote statistical significance at 1%, 5%, and 10%, respectively.
The REM method reveals a positive and significant CashFlow coefficient, aligning with Almeida et al (2004), indicating that an increase in a firm's cash flow leads to higher cash holdings Additionally, the Q coefficient is also positive and significant, supporting Almeida's findings In Almeida's model, the Q variable reflects long-term growth opportunities, suggesting that when growth prospects arise, companies increase their cash reserves to facilitate new investment projects.
The Hausman test indicated significant results, leading to the rejection of the null hypothesis and suggesting that the Fixed Effects Model (FEM) outperformed the Random Effects Model (REM) However, the FEM analysis revealed that the CashFlow variable was positive yet not statistically significant.
The GMM4 method is applied to estimate the Almeida et al model, revealing that Riddick and Whited observed a negative sensitivity of cash flow when utilizing GMM4 to address measurement errors in the Q variable In contrast, the GMM4 estimation of the Almeida model using data from Vietnam demonstrated a positive but insignificant sensitivity of cash flow, aligning with the findings from the FEM method.
To investigate whether the observed characteristics are specific to Vietnamese companies or if the Almeida model is inadequate, the study estimates the equation (1) to analyze the impact of cash flow sensitivity on cash holdings This analysis temporarily omits the asymmetric cash flow sensitivity by excluding the dummy variable Neg and the interaction variable CashFlow * Neg The study employs GMM4, FEM, and REM methods simultaneously Notably, equation (1) does not account for a nonlinear relationship between cash flow sensitivity and cash holdings, as represented by the equation: ΔCashHoldings it = α 0 + α 1 CashFlow it + α 2 Q it + α 3 Size it + α 4 Expenditure it + α 5 Acquisition it + α 6 ΔNCWC it + α 7 ShortDebt it-1 + ε it.
Table 4 6: Estimation results of the model
Hausman specification test between FEM and REM χ2 (df =7) = 22.36
The study examines non-financial companies from 2009 to 2015, focusing on key financial metrics The dependent variable, ΔCashHoldings, represents the change in cash as a percentage of total assets from year t to year t−1 CashFlow is calculated as profit after interest, dividends, and taxes, plus depreciation, also expressed as a percentage of total assets The variable Q is derived from the market value of capital and book value of assets, adjusted for the book value of capital, then divided by the book value of assets Size is measured using the natural logarithm of total assets, while Expenditure refers to the ratio of capital expenditures to total assets The Acquisition variable indicates whether the company made acquisitions during the year, marked as 1 if applicable Additionally, ΔNCWC is the non-cash flow difference relative to total assets, and ShortDebt represents short-term debt weighted by total assets.
The regression analysis using GMM4, FEM, and REM indicates a positive cash flow sensitivity, aligning with Almeida et al (2004) Notably, capital expenditure shows a negative relationship, suggesting that increased spending on fixed assets reduces company cash The negative coefficient for ShortDebt reveals that higher debt ratios in Vietnam significantly diminish firm gross profit due to increased interest payments, ultimately lowering after-tax profits and cash reserves Conversely, firms with lower short-term debt ratios retain more profit for cash holdings To enhance working capital, companies can either increase short-term debt or sell long-term assets for short-term investments However, while borrowing may initially boost cash inflows, it can lead to higher short-term liabilities and reduced liquidity over time Therefore, selling long-term assets is often a more effective strategy, as it not only increases working capital but also promotes investment in non-cash short-term assets like inventory and equipment, resulting in faster growth of non-cash working capital compared to cash holdings.
However, the coefficient of CashFlow in the GMM4 estimator is not significant, which may be a sign that the GMM4 is no longer a suitable estimation method for data in Vietnam
The analysis using the FEM and REM methods reveals that the Q coefficient is positive and statistically significant at a 1% level, while the CashFlow variable also shows a positive and statistically significant relationship at the 5% and 10% levels This indicates that the FEM and REM methods are more suitable for analyzing data from Vietnamese enterprises compared to the GMM4 method The Hausman test results were statistically significant, leading to the rejection of the null hypothesis and confirming that the FEM model provides superior results The positive coefficient of the CashFlow variable in the FEM regression aligns with the study's initial expectations, suggesting that increased cash flow correlates with higher cash holdings in Vietnam However, the small coefficient indicates that cash flow sensitivity's impact on cash holdings is minimal This finding contrasts with previous studies, such as those by Riddick and Whited (2009) and Bao, Chan, and Zhang (2012), which suggested that higher cash flow from well-performing businesses would result in lower cash reserves due to investments in tangible assets and profitable projects To further investigate the effects of cash-flow sensitivity on cash holdings in the context of both positive and negative cash flows, the study will incorporate a dummy variable for negative cash flow (Neg) and an interaction term CashFlow*Neg in the main model.
Table 4 7: Results of the asymmetry sensitivity of cash flow
Dep.Var: ΔCashHoldings GMM4 FEM REM
Hausman specification test between FEM and REM χ2 (df =9) = 28.96
The study analyzes non-financial companies from 2009 to 2015, focusing on the dependent variable ΔCashHoldings, which represents the cash change from year t to year t-1 relative to total assets CashFlow is defined as income before extraordinary items, while Neg indicates whether the company has negative cash flow Additionally, Q refers to the total market value of capital plus the book value of assets, adjusted for the natural logarithm of total assets Expenditure is assessed as capital expenditure relative to total assets, and Acquisition is an indicator for companies engaged in acquisition activities within a given year ΔNCWC denotes the difference in non-cash working capital as a proportion of total assets, and ShortDebt represents short-term debt relative to total assets Statistical significance is indicated by ***, **, and * for p-values of 1%, 5%, and 10%, respectively.
The GMM4 method revealed that the coefficients for the CashFlow variable, dummy variables, and interaction variables were not statistically significant, suggesting that there is no discernible difference in the impact of cash flow between positive and negative scenarios.
The Hausman test was employed to evaluate the choice between Fixed Effects Model (FEM) and Random Effects Model (REM) The significant results led to the rejection of the null hypothesis, indicating that the FEM model outperformed the REM model.
The FEM regression results indicate that the CashFlow variable is positively and statistically significant at 5%, contrasting with the findings of Bao, Chan, Bao, Chan, and Zhang (2012) and Riddick and Whited (2009), who reported a negative correlation between CashFlow and CashHoldings The negative coefficient of the Neg variables is not statistically significant, suggesting no difference in cash holdings changes between firms with negative and positive cash flows However, the negative and statistically significant coefficient of the CashFlow*Neg variable at 10% implies that companies with negative cash flow experience increased cash holdings as their cash flow weakens This aligns with Bao, Chan, and Zhang's (2012) assertion of an asymmetrical relationship between cash flow sensitivity and cash holdings, highlighting that the impact of cash flow sensitivity on cash holdings differs in Vietnam for positive versus negative cash flows.
The impact of cash flow sensitivity on cash holdings in Vietnam differs from previous studies, as evidenced by research data from 274 non-financial companies listed on HOSE and HNX between 2009 and 2015 This period follows the global financial crisis that began in 2008, highlighting unique financial dynamics in the Vietnamese market.