Bankruptcy, Investment, and Financial Constraints: Evidence from the Czech Republic

Abstract

Using investment–cash flow sensitivity to analyze financial constraints over the period 2006–2011 in the Czech Republic, we find that healthy companies were financially constrained both before and after the 2008 crisis. There is robust evidence that both the cash flow and the level of debt have a positive and significant impact on the investment rate. Although companies going bankrupt had significantly higher levels of external debt and bank loans, they did not manifest any investment–cash flow sensitivity in the pre-crisis period, which indicates that they were probably not financially constrained at all. After the 2008 crisis, companies that would later declare bankruptcy began to become financially constrained as well.

Keywords:

• bankruptcy
• cash flow
• financial constraints
• investment
• post-transition economy

JEL Classification:

• D22
• D92
• E22
• G32

ACKNOWLEDGMENTS

We are grateful to Jan Hanousek, Randall Filer, Oleksandr Talavera, Miroslav Plašil, Vladimír Benáček, Niclas Berggren, Christian Bjørnskov, Josef Brada, and the two anonymous referees for a thorough review of the paper. We would like to thank Jan Kmenta, Evangelia Vourvachaki, Evžen Kočenda, Jakub Kastl, Jacek Cukrowski, Štěpán Jurajda, Andreas Ortmann, Jakub Kastl, Beyongju Jeong, Jan Švejnar, Lubomír Lízal, Iryna Momotenko, and Avner Shaked, and the seminar participants at the 2013 IES Economic Meeting, the Bratislava Economic Meeting 2012, CERGE-EI, and the Czech National Bank for their useful comments and suggestions. Martin Pospíšil would like to thank Organizational Dynamics at the University of Pennsylvania for their hospitality during his research stay. The views expressed in this paper are those of the authors and not necessarily those of the Czech National Bank or other institutions with which the authors are affiliated.

Notes

1. Loans constituted 18% of the total Czech non-financial corporations’ liabilities over the 2006–2011 period, and this figure had temporarily increased to 20% by the beginning of 2009. Loans constituted approximately 40% of these firms’ external funds.

2. Even though BEEPS 2009 was effectively surveyed in 2008 and asked about the fiscal year 2007, the interviews among the Czech firms were conducted from late 2008 to early 2009. Therefore, answers regarding the current strongest obstacles should already cover the onset of the crisis. The financial constraints remained relatively stable between 2002 and 2009 in the Czech Republic, while they increased dramatically in Russia, for example. In most countries, we observe a temporary increase in the financial constraints in the 2009 survey.

3. The 2008 crisis hit firms in the Czech Republic hard. For example, in 2009, they were complaining that they were being hit by a credit crunch with banks reluctant to lend money (http://goo.gl/r2ulZA). It was observed that the decline in credit after the events of 2008 was due to higher economic uncertainty, more prudent lending because of pressure from parent banks from abroad, and lower demand for credit in general due to low aggregate demand.

4. As lenders cannot perfectly distinguish between good and bad borrowers, relatively good borrowers drop out of the market when the interest rate increases. Then, lenders’ profits may decrease, because this “drop out” can lead to an increase in the default probability on loans made.

5. The Modigliani-Miller theorem in its basic form states that under a certain market price process, in the absence of taxes, bankruptcy costs, agency costs, and asymmetric information, and in an efficient market, the value of a firm is unaffected by how that firm is financed.

6. Bernanke, Gertler, and Gilchrist (1999) show how credit-market frictions can significantly amplify both real and nominal shocks in the economy. Other economists who stand outside the standard macroeconomic models and focus on the importance of credit in the economy include Keynes, Minsky, and Stiglitz.

7. Even though the degree of regulatory uncertainty, connected, for example, to frequent changes in the tax laws, is higher than in the developed Western economies (Havranek, Irsova, and Schwarz 2016b).

8. Havranek, Irsova, and Lesanovska (2016a) find that the long-term pass-through of Czech monetary policy to the client rates was close to complete for most products before the financial crisis, but has weakened considerably afterward.

9. For the discussion about this so-called monotonicity hypothesis, see the original paper by Fazzari, Hubbard, and Petersen (1988), the Kaplan and Zingales critique emanating from Kaplan and Zingales (1997), the reply by Fazzari, Hubbard, and Petersen (2000), and the answer by Kaplan and Zingales (2000). Kaplan and Zingales (1997) theoretically show that even in a one-period model, investment–cash flow sensitivities do not necessarily increase with the degree of financial constraints. They also claim that in a multi-period case, for example, precautionary savings make it even more difficult to justify this relationship. They finally argue that this relationship may be more complicated, with overly risk-averse firms preferring to invest their own cash flow. Using simulated data, Bond and Söderbom (2013) find that the relationship between financial constraints and the sensitivity of the investment with respect to the cash flow, conditional on a measure of the marginal Q, is monotonic.

10. The ratio of the market value of the existing capital to its replacement cost. Usually, we can only observe the average Q (though this observation can be difficult, especially for non-listed companies). The marginal Q, on the other hand, is the ratio of the market value of an additional unit of capital to its replacement cost. It is possible to estimate the marginal Q (Gugler, Mueller, and Yurtoglu 2004), but most empirical work uses the average Q as the proxy for the marginal Q. Berglund (2011) points out that the proposed methods of estimating the marginal Q are likely to produce biased estimates.

11. We add the lagged investment rate and the squared investment rate to control for this variable’s autoregressive nature, as past investments influence today’s investments. The squared term will capture the potential non-linear relationship. We also use the turnover instead of the sales growth in one of the robustness checks.

12. In the Albertina database, the cash flow is calculated as the current year’s profit (loss) + the depreciation of tangible and intangible fixed assets. There are generally two approaches to calculating the cash flow: direct and indirect. For the direct calculation of the cash flow, the balance-sheet information is clearly insufficient. The indirect calculation starts with the information available on the balance sheet, but usually adjusts for the revenues and costs, which either should be included in the net income but are not, or should not be included but are. However, such adjustments cannot be made based on the data available to an outside observer who has access only to the financial statements of firms. As a consequence, for the purposes of financial constraints analysis, the cash flow is usually calculated in the same way as in Albertina.

13. Operated by Bisnode Česká Republika, a.s. (http://www.albertina.cz). We would like to thank the Czech National Bank for access to this dataset.

14. A similar problem would arise if Amadeus, a European-wide, firm-level dataset, was used. Amadeus is compiled by Bureau van Dijk by harmonizing companies’ annual reports obtained from various European vendors. As in our case, the information on bankruptcy is not accurate for the above-mentioned reasons.

15. As noted, due to the length of the bankruptcy process, it makes more sense to assign the bankruptcy to the fiscal year just preceding the year when the firm formally goes bankrupt. Therefore, companies in our dataset can go bankrupt in the years 2007–2012.

16. We cannot use the possibly most relevant deflators based on the EU KLEMS database to obtain country-sector-specific output and intermediate input deflators, as the EU KLEMS data are available only until 2007. We believe that using EBRD deflators will be sufficient. However, these deflators vary only on the country level, which is a drawback.

17. Observations of companies that will go bankrupt at some point in the future constitute 1.93% of our final dataset after our data management (including observations older than 2006, which are used as instruments). Companies that will go bankrupt the following year constitute 0.14% of the sample. Due to our data management, we drop 44,030 observations from before 2011 or from that year. Observations of companies that go bankrupt at some point constitute 5.81% of the dropped observations, and companies that go bankrupt the following year constitute 0.4% of it. Thus, we lose a slightly above-proportional number of observations related to companies that go bankrupt, which is probably due to a higher error rate in their financial statements’ data. The unreliability of the data that have been dropped makes it impossible to seriously estimate the scope and the direction of the potential selection bias.

18. The numbers differ because “going bankrupt at some point” are the yearly observations for the companies that we know will go bankrupt (i.e., firms that go bankrupt between 2007 and June 2013).

19. The hypothesis of equality of means between these two types of companies is rejected at the 10% confidence level only in the pre-crisis period.

20. A negative cash flow means that the cash inflow from the sales is lower than the cash outflow of the cash payments. The common reasons for a negative cash flow are usually thought to be low sales, high operating expenses, wrong investments, or unattractive financing conditions.

21. The data are available in the ARAD data series system of the Czech National Bank (http://www.cnb.cz/docs/ARADY/HTML/index.htm).

22. As Oliner and Rudebusch (1996) note, the research on the credit channel stresses that the central bank’s actions affect the output in part by causing shifts in the supply of loans.

Additional information

Funding

We acknowledge support from the Czech National Bank (project #C6/12). Martin Pospíšil acknowledges support from the Grant Agency of Charles University (grant #616812).

Notes on contributors

Jiří Schwarz

Jiří Schwarz is an advisor to the board at the Czech National Bank in Prague and an assistant professor at the Institute of Economic Studies, Faculty of Social Sciences, Charles University, Prague, Czech Republic.