1. INTRODUCTION
During the eighties several authors have defined the concept of organizational slack, or slack resources in articles published in business periodicals. This concept plays a theoretical role in both management and financial literature.
Bourgeois (1981, p. 30) defined slack as ‘that cushion of actual or potential resources which allows an organization to adapt successfully to internal pressures for adjustment or to external pressures for change in policy as well as to initiate changes in strategy with respect to the external environment’. Sharfman et al. (1988) add that ‘in order for resources to be considered slack, they must be visible to the manager and employable in the future’. Additionally these authors also brought forward an extra dimension: the level of discretion and flexibility that is provided by the slack resources. They anchored slack along a continuum of managerial discretion, ranging from low-discretionary slack, which can only be used under specific circumstances, and high-discretionary slack that ‘can be used in a wide variety of situations or can give managers a number of options’ (Sharfman et al., 1988, p.602).
There is a significant amount of research concerning the effects of slack resources on different firm outcomes. The slack-performance relationship has been extensively studied (Bourgeois, 1981; George, 2005; Tan & Peng; 2003). Other research includes the effect on innovation (Bromiley, 1991; Damanpour, 1987; Greve, 2003) and risk-taking (Wiseman & Bromiley, 1996).
In which circumstances slack resources are created and what drives the level of slack in an organization during an environmental jolt has been the subject of far less research. This paper will focus on the environmental jolt as a trigger influencing the creation and consumption of slack. We will examine the creation and consumption of slack around the environmental jolt of the internet bubble.
The internet bubble is one of the most recent environmental jolts. Broadly defined, it was the hype surrounding new technologies in the late nineties of the previous century. We will elaborate on the historical circumstances in the section ‘Research Context’. This so called internet bubble is an interesting jolt given its exceptional circumstances. It created unique and exceptional financial possibilities during several years for a limited set of industries focused on IT and new technologies. During the build-up of the internet bubble, the exuberance surrounding these sectors allowed them to issue equity or obtain other forms of external financing with more ease than would traditionally have been the case.
There are two competing theories concerning the level of slack resources during an environmental jolt:
‘ According to organizational theorists, slack is used to stabilize a firm’s operations by absorbing excess resources during periods of growth and by allowing firms to maintain their aspirations and internal commitments during periods of distress (Cyert & March, 1963; Levinthal & March, 1981; Meyer, 1982). This theory predicts that firms will consume slack during an environmental jolt to protect their technical core (Thompson, 1967).
‘ Economic theory on the other hand, states that firms will maintain an optimal level of slack based on a trade-off between the costs and benefits of holding financial slack (Opler et al., 1999). We will expand on these costs and benefits in the ‘theory’ section, but environmental factors should lead to a higher optimal level of financial slack during an environmental jolt. The economic theory predicts that firms will accumulate financial slack during an environmental jolt.
This contradiction leads to our research question: Did Belgian computer-related firms accumulate or consume financial slack during the build-up, crash and aftermath of the internet bubble?
We limit the study to computer-related firms since the internet bubble and resultant jolt were felt more strongly in this industry.
Limited empirical research suggests that whether firms consume or stockpile slack during environmental jolts largely depends on firm characteristics. Old, low-growth firms stockpiled cash during the financial crisis of 2008, while young, high-growth firms consumed financial slack (Paeleman & Vanacker, 2015). This research has the intention to contribute to the literature by providing further empirical evidence concerning the antecedents of slack resources surrounding the environmental jolt of the internet bubble.
We do this by analyzing financial data for Belgian computer-related firms, before, during and after the internet bubble of 2000. This paper will focus on financial slack, because of the availability of financial data, and to improve consistency with previous research (Kim et al., 2008; George, 2005; Paeleman & Vanacker, 2015). Financial slack is high-discretionary slack as defined by Sharfman et al. (1988) and provides more insight into managerial decisions than other slack measures.
2. RESEARCH CONTEXT
In financial markets, a market bubble is ‘a self-perpetuating rise or boom in the price of shares or goods (most likely in a specific new less known or new area). This term bubble can only be used with certainty in retrospect after the value of the specific shares or goods have dropped/crashed’ (Anderson et al., 2010).
A bubble occurs when the market has an, in hindsight, exaggerated perception of potential value or profit making capacity of goods, shares or industry. When buyers, in this context often referred to as speculators, note the fast increase in price of specific goods or shares, they decide to buy in anticipation of further rises in price. Typically, during a bubble, specific goods, shares or companies in a specific industry become extremely overvalued. Finally when the bubble ‘bursts’ the value of those goods or shares will drop very rapidly.
The dot-com or internet bubble was a historic speculative bubble during the second half of the 1990s. In the run-up of the internet bubble, there was a wave of speculation surrounding the technology sector, mainly in America, but also in other industrialized nations. This is best demonstrated using the American NASDAQ index, which rose by almost 200% in the two years leading up to the peak, and which lost 78% of its value during the 2 years following the peak of the internet bubble.
This was all made possible through a combination of increasing stock prices, a market confidence that these computer related companies had profit potential and vast availability of venture capital and new listings in which environment investors were willing to overlook traditional stock exchange metrics for these industries.
The beginning of the bubble is often linked to the introduction of the software company Netscape in 1995 onto NASDAQ. Originally the shares were valued for introduction at $14. Due to high demand they decided to double the introduction price. Finally, due to high demand, the shares went to $ 75 at closing of the introduction day.
During the internet bubble, American news media, including respected business publications such as Forbes and the Wall Street Journal, encouraged the public to invest in risky companies, despite many of the companies’ disregard for basic financial and even legal principles (Lowenstein, 2004). Even the most traditional brokers and investment banks set aside the notion that stock price should reflect potential future profits and urged investors not to miss out on the gold rush (‘The Dot-Com Bubble Bursts’ 2000).
The tech bubble is often referred to as a US phenomenon, but overall the bubble had a worldwide impact on stock exchanges. All European countries including Belgium, e.g. Lernout&Hauspie, and the Netherlands, e.g. Worldcom, had similar patterns.
3. THEORY
In this section we will use existing theories on slack resources to define hypotheses for the research. There are two competing views in the theory on slack resources around environmental jolts:
‘ The organizational view on managing slack resources perceives the organization as an entity that is actively managing the organization before, during and after the environmental jolt (in some way through this specific economic cycle). As such the organization is taking advantage of the different opportunities that this specific cycle is offering. The organizational view perceives the organization as actively managing itself through the cycle by anticipating the next cyclical periods for the organization (long term survival view).
‘ The economical view on the theory on slack resources perceives the economic environment as an external factor and the different economic entities are optimizing economic decisions based on the day to day environment. As such firms are making decisions based on the economic and financial situation of that moment (credit environment, cost of credit, return on investments, profit environment), to determine the optimal level of slack resources.
3.1. The organizational view of slack resources
According to the organizational view, managers use slack resources as a buffer. Cyert & March (1963) view the organization as a coalition of different actors, including shareholders, employees and management. When the firm is able to produce more resources than strictly necessary to maintain these coalitions, the firm can use these to create different kinds of slack. This can take the form of a buffer of cash and cash equivalents, excess employees, higher pay for existing employees or higher dividends for shareholders than strictly necessary.
In times of economic duress, i.e. an environmental jolt, managers will consume slack to absorb the external shock. In this way, organizational slack plays an adaptive and stabilizing role (Cyert & March, 1963) and allows the organization to protect its technical core (Thompson, 1967).
This organization theory leads to several predictions concerning the level of slack resources surrounding the internet bubble. We will use these predictions as our first three hypotheses, and later we will use the competing economic theory to construct three alternate hypotheses.
During the run-up of the internet bubble, ‘the good times’, according to the organizational view we would expect the firm to accumulate financial slack. This leads to our first hypothesis:
H1a: Firms will accumulate financial slack during the build-up of the internet bubble.
When the internet bubble crashed, the economic environment worsened significantly for Belgian computer-related firms. Access to external financing dried up, leading to times of economic duress. The organizational theory predicts that managers will consume slack to absorb the external shock, this leads to our second hypothesis, 2a.
H2a: Firms will consume financial slack during the distress period of the internet bubble.
After the crash and subsequent distress period, the environment for Belgian computer-related firms stabilized and improved again. The organizational theory predicts that firms would replenish their depleted financial slack after the distress period:
H3a: Firms will accumulate financial slack during the recovery stage after the internet bubble.
There is some empirical research that suggests slack has a positive impact on performance during an environmental jolt (Wan & Yiu, 2009). Though examining the performance of firms with more or less slack is not the goal of this paper, this finding is in line with the organizational theory. It reinforces the claim that building up a buffer of slack resources is in fact helpful during an environmental jolt.
3.2. The economic view of slack resources
According to the economic view, the decision to maintain a certain level of slack is a trade-off between the costs and benefits of holding slack (Opler et al., 1999). The costs of holding slack include the low return of liquid assets, lowering the overall return to shareholders. The benefits of holding financial slack include the reduced need of costly external financing (Paeleman& Vanacker, 2015). It also improves credit-worthiness, thus improving the access to external finance when it is required.
During the build-up of the internet bubble, external financing was easy to obtain, as such the benefits of holding financial slack were lower. This leads to a lower optimal level of slack during this period. This leads to our first alternate hypothesis:
H1b: Firms will consume financial slack during the build-up of the internet bubble.
During and shortly after the crash of the internet bubble, external financing was very hard to obtain. Credit had dried up for computer-related firms. Thus the benefits of holding financial slack increased, since only firms with very high credit-worthiness were still able to obtain external financing. On the other hand the costs of holding slack were relatively low due to a less inviting investing environment. We expect firms to increase their financial slack during this period to a higher optimal level following the economic theory:
H2b: Firms will accumulate financial slack during the distress period of the internet bubble.
After the crash and following distress period, we expect a period of stabilization and recovery. Access to external financing improved, once again reducing the benefits of holding financial slack once again. The economic theory predicts firms following the trade-off theory will lower their level of financial slack:
H3b: Firms will consume financial slack during the recovery stage after the internet bubble.
3.3. Summary
Table 1: Summary of hypotheses
‘Run-up’ (1997-1999)
‘Distress’ (2000-2002) ‘Recovery’ (2003-2005)
Organizational theory Accumulation of slack Consumption of slack Accumulation of slack
= Hypothesis 1a = Hypothesis 2a = Hypothesis 3a
Economic theory Consumption of slack Accumulation of slack Consumption of slack
= Hypothesis 1b = Hypothesis 2b = Hypothesis 3b
3.4. The internet bubble as environmental jolt in our research
As a final element in this theoretical analysis we would like to introduce some elements that are more specific to the internet bubble.
The internet bubble and its consequences have led to an environmental jolt that is focused on specific industries and market segments (IT, Telecom, internet, and other related industries). As such any theory on slack should try to take into account the specific characteristics of these industries at that specific moment in time. Some of the specific characteristics of these industries at that specific moment in time could favor one or the other of the theoretical assumptions on the accumulation and consumption of financial slack.
First we would like to indicate that a significant part of these companies were startups or relatively young and new entities. Startups and firms developing new technologies will often be disproportionately focused on research and short term growth. They often fail to create financial returns for an extended period, but are trying to limit any cash drains (Ofek&Richardson, 2003).
As such these organizations are focused on research results, performance of new technologies and most of all fast growth. Long term goals and a long term strategy as preparing for sufficient slack for the next cycle will be less of a priority. As such the organizational view, certainly in the build-up of the bubble, seems less intuitive for these industries.
Moreover the 1997-2000 bubble environment was so unique or exceptional that many entities were often taking more intuitive short term decisions on proposals of financial brokers and investment banks than on logical organizational behavior.
These exceptional circumstances focused on short term profits was not a very natural environment for organizational long term strategic thinking on financial slack, but should favor the economic view. All these specific elements should, in the build-up period, support the use of an economic view on slack (Lovink, 2002).
After the bubble did burst a lot of these companies experienced financial distress. For firms in distress the focus on survival will more than ever favor a short term economic view to avoid further trouble. Also for these weaker firms and industries the economic view on slack will remain the most likely result.
For those industries with less distress the focus could shift to a more long term organizational view away from a short term direct economic environment.
4. METHODOLOGY
4.1. Data
The dataset used for this paper is a set of all Belgian limited liability firms, with a number of selection criteria used. Our sample was obtained using the BelFirst database, a database containing information, including accounts and other financial information on 1.2 million companies in Belgium and Luxembourg.
The data was limited in time from 1995 to 2005, to provide sufficient margin around the events of the build-up of the internet bubble, and subsequent crash.
We selected for firms with computer-related activities using NACE activity codes, this is the ‘Statistical Classification of Economic Activities in the European Community’. We selected only firms with NACE activity code 72, ‘Computer related activities’, because the internet bubble and resultant jolt were felt more strongly in this industry.
We selected only independent firms, i.e. no other firm holds more than 25% of equity. Dependent firms could be strongly influenced by their ownership structure and were therefore excluded.
Only firms with a minimum of 1 employee were selected, in this way we exclude firms that only exist on paper.
To summarize, our selection criteria were as follows:
1) Belgian firms.
2) 1995-2005
3) Firms with NACE activity code 72, ‘Computer related activities’.
4) Independent firms.
5) Firms with at least 1 employee
These criteria led to a dataset of 7848 firms, and 34432 firm-year observations.
4.2. Dependent variable
We use financial slack as the dependent variable. This is consistent with most existing literature. Though slack can exist in other organizational capabilities and assets, it is difficult to obtain a longitudinal sample of private firms for these dimensions. (George, 2005)
Financial slack is often the preferred measure of slack because of the relatively high availability of financial data compared to measures of social or human slack resources. It also offers the highest level of managerial discretion, thereby providing more insight into decision making than other slack measures.
We define financial slack as the sum of Cash and Cash Equivalents. Cash is the most easily deployed resource and provides managers the greatest degree of freedom in allocation to alternate uses. (George, 2005)
Researchers have argued that the influence of slack is relative to target levels of slack rather than absolute levels of resources (Bradley, 2011; Bromiley, 1991; March and Shapira, 1987).
In line with Brav (2009) and Paeleman and Vanacker (2015) we scale Cash and Cash Equivalents by Net Assets, defined as Total Assets minus Cash and Cash Equivalents. This controls for scale effects and mitigates heteroscedasticity (Brav, 2009), i.e. the difference in variability for different sizes of firms.
4.3. Independent variables
To test the hypotheses of the effect of the environmental jolt, we created four distinct periods, that can be identified by 3 dummy variables. A summary can be found in table 1 below.
These periods are: The ‘Baseline’ period from 1995 to 1996 serves as a reference period with a normal level of financial slack. The ‘run-up’ period from 1997 to 1999, is the period where the actual internet bubble was formed. The ‘distress’ period from 2000 to 2002 is the period where the internet bubble crashed, and includes the subsequent period where firms in relevant sectors would have experienced economic duress. The ‘recovery’ period from 2003-2005 is where we expect the negative consequences of the crash to have disappeared and where the situation would have normalized.
They are identified by three dummy variables, ‘Bubble run-up dummy’, ‘Bubble Dummy’ and ‘Post-Bubble Dummy’. These dummy variables were coded as follows. The independent variable, ‘bubble run-up dummy’ is set to 1 for the period 1997 to 1999 and to 0 for all other periods. The independent variable, ‘Bubble Dummy’, is set to 0 for the years before and after the crash of the internet bubble and to 1 during the distress period of the internet bubble, 2000-2002. The second independent variable, ‘Post-Bubble Dummy’ is set to 0 before and during the distress period of the internet bubble and to 1 after, 2003-2005.
This creates four distinct periods that can be identified by these dummy variables.
‘
Table 2: Coding of dummy variables
Bubble run-up dummy Bubble dummy Post-bubble dummy
1995-1996 ‘Baseline’ 0 0 0
1997-1999 ‘Run-up’ 1 0 0
2000-2002 ‘Distress’ 0 1 0
2003-2005 ‘Recovery’ 0 0 1
4.4. Control variables
4.4.1. Potential slack
Previous research in the field has often used Low Discretionary Slack as a control variable (Bromiley, 1991; George, 2005; Greenley&Oktemgil, 1998). It represents the remaining borrowing capacity of a firm or resources not yet put into operations. We chose to use a similar measure under the term ‘Potential slack’. This variable is operationalized as the ratio of Debt and Equity. This is in line with more recent research of Bradley (2011). When this ratio is low, the firm has a higher potential to access additional debt for future investments (George, 2005).
4.4.2. Recoverable slack
Recoverable Slack is used as a control. It is operationalized as Accounts Receivable plus Inventory divided by Net Assets (Steensma&Corley, 2001; Bradley, 2011). We include this measure, because it represents the relative level of resources that is contained in operations, and these resources are considered a substitute for cash (Bigelli&S”nchez-Vidal, 2012). By changing the payment terms for customers, or reducing inventory, the firm can (temporarily) increase their level of cash.
4.4.3. Firm size
Numerous studies have shown that the size of a firm has an effect on the availability of external financing, so we will also use firm size as a control variable. We use the natural logarithm of total assets as an operational definition.
4.4.4. Firm performance
We include Firm performance, operationalized as Profit or Loss after taxes, divided by Total Assets, since the performance of the firm influences the ability of a firm to accumulate slack.
4.5. Variables overview
Table 3: Overview of dependent, independent and control variables
Name Operational Definition
Dependent Financial slack (Cash + Cash Equivalents) / (Total Assets – Cash – Cash Equivalents)
Independent Bubble run-up dummy
Bubble Dummy 1995-1996=0 ; 1997-1999=1 ; 2000-2002=0 ; 2003-2005=0
1995-1996=0 ; 1997-1999=0 ; 2000-2002=1 ; 2003-2005=0
Post-Bubble Dummy 1995-1996=0 ; 1997-1999=0 ; 2000-2002=0 ; 2003-2005=1
Control Size Ln(Total Assets+1)
Performance Profit or Loss after Taxes / Total Assets
Potential slack Debt / Equity
Recoverable slack (Accounts Receivable + Inventories) /
(Total Assets – Accounts Receivable – Inventories)
4.6. Econometric approach
4.6.1. Regression characteristics
We ran a fixed effects regression analysis, since there may be unobserved characteristics of the firm that influence the level of slack resources, using VAT numbers to uniquely identify individual firms. We also used the built-in robustness check in the statistical software, Stata, this should make the results more resistant to a certain degree of data contamination. We clustered standard errors at the firm level to adjust for heteroscedasticity.
We used an F-test to test if the differences between the coefficients of the dummy variables were significant.
4.6.2. Initial regression
Our initial regression used all 344432 observations for 7848 firms and provided coefficients with very low statistical significance. The low significance could be explained by the relatively low average number of observations of 4.3 per individual firm out of a potential maximum of 11 observations per firm. The low average number of observation stems from firms entering and exiting the dataset through natural means, such as new incorporations and bankruptcies. But this could also stem from inconsistencies in reporting, or other sources of missing data. We took measures to increase the average number of observations.
‘
4.6.3. Intermediary regression
We ran a new regression with the same characteristics and methods as the initial regression, while removing all firms with less than 7 observations from the dataset. This left 13746 observations for 1452 firms. This greatly increased the statistical relevance of the obtained coefficients.
4.6.4. Final regression
Finally we also ran the regression while removing all firms without observations for each year from the dataset. This left 5792 observations for 528 firms. These selection criteria risk introducing a survivorship bias, since all firms that exited the dataset throughout the period, are no longer accounted for. This also eliminates firms that were incorporated after 1995, so the average age of the sample is higher. However, with this sample, we are able to gain more insight into how financial slack evolves over a longer period of time. In conclusion, there may be a trade-off, between the undesirable effects of removing a large fraction of the data due, and the statistical advantages to the obtained results.
5. RESULTS
5.1 Initial Regression
The results of our original regression are summarized in table 4. The regression coefficients were all statistically insignificant. The lack of statistically significant coefficients makes it impossible to reject any hypotheses. We cannot reject the null hypothesis that the bubble had no significant effect on the level of financial slack, since our regression coefficients are not statistically significantly different from zero. Similarly, our F-tests for the difference in regression coefficients were statistically insignificant, as shown in table 5.
‘
Table 4: Results of inital regression
Fixed-effects (within) regression Number of observations: 34014
Group variable: VAT number Number of groups: 7848
Observations per group: min = 1
avg = 4.3
max = 11
financialslack Coef. Std. Err. t P>t [95% Conf. Interval]
constant 3.019381 10.2543 0.29 0.768 -17.07962 23.11838
potentialslack .0000604 .0543417 0.00 0.999 -.1064523 .106573
recoverableslack -6.13e-18 8.24e-15 -0.00 0.999 -1.62e-14 1.61e-14
performance .0027136 .182612 0.01 0.988 -.3552159 .3606431
totalassetslog -.2913983 1.742834 -0.17 0.867 -3.707448 3.124651
bubblerunupdummy -.3295519 2.944998 -0.11 0.911 -6.10191 5.442806
bubbledummy -.7446706 2.514339 -0.30 0.767 -5.672913 4.183572
postbubbledummy 2.021579 2.650823 0.76 0.446 -3.174179 7.217336
Table 5: F-test on coefficients of initial regression
F-test for runupbubbledummy and bubbledummy
F (1, 26159) = 2.14
Prob > F = 0.1438
F-test for bubbledummy and postbubbledummy
F (1, 26159) = 0.02
Prob > F = 0.8946
5.2 Intermediary Regression
The results of our second regression are summarized in table 5, and the results of the F-tests for the statistical difference between regression coefficients are shown in table 6.
The coefficient for the variable Bubble run-up dummy was negative (”=-0.39), but statistically insignificant (p=0.10). We cannot reject the hypothesis that the period we defined as the run-up to the internet bubble, 1997-1999, has no significant effect on the level of financial slack. We cannot reject either hypothesis 1a or 1b. We find only weak support for hypothesis 1b, stating that firms will consume slack in the run-up to the internet bubble, following the economic theory.
The coefficient for the distress period of the internet bubble is positive (0.44), significant (p=0.03) and significantly higher than the coefficient for run-up of the internet bubble (p=0.00), as shown in table 5. This allows us to reject hypothesis 2a which resulted from the organizational theory, and predicted that firms would consume financial slack during the environmental jolt and distress period. We find strong empirical evidence in support of hypothesis 2b, which predicted that firms would accumulate financial slack according to the economic theory.
The coefficient for the period after the internet bubble is positive (0.76), and significant (p=0.00). However, it is not significantly higher than that of the distress period (p=0.10), so we cannot reject the hypothesis 3b, that stated that firms consumed slack during the period after the internet bubble. However, this does provide weak support for hypothesis 3a, that firms accumulated financial slack during the recovery stage after the internet bubble.
Table 6: Results of intermediary regression
Fixed-effects (within) regression Number of observations: 13746
Group variable : VAT number Number of groups: 1452
Observations per group: min = 7
avg = 9.5
max = 11
financialslack Coef. Std. Err. t P>t [95% Conf. Interval]
constant 4.626116 .9785448 4.73 0.000 2.708014 6.544217
potentialslack -.0016218 .0065062 -0.25 0.803 -.0143749 .0111313
recoverableslack -.0012808 .0025492 -0.50 0.615 -.0062776 .0037159
performance .2699239 .2402677 1.12 0.261 -.2010385 .7408862
totalassetslog -.5749336 .1589208 -3.62 0.000 -.8864433 -.2634239
bubblerunupdummy -.3939764 .2404237 -1.64 0.101 -.8652446 .0772917
bubbledummy .4361361 .1962255 2.22 0.026 .0515032 .820769
postbubbledummy .7582269 .2065197 3.67 0.000 .353416 1.163038
Table 7: F-test on coefficients of intermediary regression
F-test for runupbubbledummy and bubbledummy
F (1, 12287) = 10.47
Prob > F = 0.0012
F-test for bubble dummy and postbubbledummy
F (1, 12287) = 2.63
Prob > F = 0.1047
5.3 Final Regression
The results of our final regression are summarized in table 7, and the results of the F-tests for the statistical difference between regression coefficients are shown in table 8.
The coefficient for the variable Bubble run-up dummy was negative (”=-0.002), but statistically insignificant (p=0.96). We find strong evidence that the period we defined as the run-up to the internet bubble, 1997-1999, has no significant effect on the level of financial slack. We cannot reject either assumption 1a or 1b, and find no evidence for either of them.
This means that either the period of 1997 to 1999 was not significantly different than that of the baseline years, 1995 and 1996 for computer-related firms in Belgium, or that the differentiating factors during this period had no significant effect on the level of slack resources.
The coefficient for the distress period of the internet bubble is positive (0.09), significant (p=0.04), however it is not significantly higher than the coefficient for run-up of the internet bubble (p=0.09), as shown in table 5. This does not allow us to reject hypothesis 2a which resulted from the organizational theory, and predicted that firms would consume financial slack during the environmental jolt and distress period. We find weak empirical evidence in support of hypothesis 2b, which predicted that firms would accumulate financial slack according to the economic theory.
The coefficient for the period after the internet bubble is positive (0.18), and significant (p=0.00). It is significantly higher than that of the distress period (p=0.02), so we can reject the hypothesis 3b, that used the economic theory to predict that firms would consume slack during the period after the internet bubble. We have found strong empirical evidence for hypothesis 3a, that firms accumulated financial slack during the recovery stage after the internet bubble, this is in line with the organizational theory.
Table 8: Results of final regression
Fixed-effects (within) regression Number of observations: 5792
Group variable : VAT number Number of groups: 528
Observations per group: min = 11
avg = 11
max = 11
financialslack Coef. Std. Err. t P>t [95% Conf. Interval]
constant .307992 .2325744 1.32 0.185 -.1479504 .7639344
potentialslack -.0057912 .0070311 -0.82 0.410 -.0195751 .0079926
recoverableslack -.0044704 .0020029 -2.23 0.026 -.0083969 -.0005439
performance .1729347 .076732 2.25 0.024 .0225082 .3233612
totalassetslog .0312587 .0354354 0.88 0.378 -.0382095 .1007268
bubblerunupdummy -.0024942 .0494313 -0.05 0.960 -.0994 .0944116
bubbledummy .0852101 .042271 2.02 0.044 .0023414 .1680788
postbubbledummy .1823728 .0426323 4.28 0.000 .0987959 .2659498
Table 9: F-test on coefficients of final regression
F-test for runupbubbledummy and bubbledummy
F (1, 5257) = 2.82
Prob > F = 0.0929
F-test for bubble dummy and postbubbledummy
F (1, 5257) = 5.43
Prob > F = 0.0198
6. CONCLUSION
This study aimed to contribute to existing literature by providing further empirical evidence for the relationship between financial slack and environmental jolts. We did this by examining a sample of Belgian firms with computer-related activities before, during and after the internet bubble, with a total period ranging from 1995 to 2005.
Our hypotheses were based on two competing theories, the organizational theory and the economic theory. These theories predicted different behavior concerning the accumulation or consumption of slack before, during and after an environmental jolt. We tested these hypotheses for 3 periods, the run-up of the internet bubble, and the distress period after the crash and the period after the distress period, referred to as the recovery stage.
Our research found no significant accumulation or consumption of financial slack for the run-up of the internet bubble. Belgian computer firms did not accumulate slack in line with the organizational theory, and they did not consume it in line with the economic theory.
Our research has found strong empirical evidence that Belgian computer firms accumulated financial slack during the distress period of the internet bubble. This is in line with the expected behavior according to the economic theory.
For the recovery period after the distress period, we found empirical evidence that Belgian computer firms accumulated financial slack. This is the behavior that was predicted by the organizational theory.
This research has a number of limitations. The use of a solely Belgian sample enables us to examine a well-defined and homogenous market with a consistent regulatory framework, however it also reduces the reproducibility and relevance for other geographical and regulatory contexts.
The reduction of our sample to include only firms with a certain number of observations, has likely introduced some biases into our dataset. Survivor bias being one of these, more research would be welcome examining the relationship between slack and firm survival during environmental jolts. This also increased the average age of firms, which may have skewed our results further in the direction of support for the economic theory of slack, since research by Paeleman&Vanacker (2014) has shown that old, low-growth firms have a tendency to stockpile financial slack during an environmental jolt.
In conclusion, this research has provided further empirical evidence in support of the economic theory, that firms accumulate slack during an environmental jolt. It also provided evidence that shows that firms will continue accumulating financial slack for an extended time after an environmental jolt, showing learning and precautionary behavior in line with the organizational theory.