Skip to main content

Subsidies, the Shadow of Death and Labor Productivity


Our panel data from over 10,000 Finnish firms during the years 2003–2010 elucidates the effect of different business subsidies on firm productivity performance and on the relationship between firms’ lagged labor productivity and market exit. We find that none of the subsidy types have statistically significant positive short-term or longer-term impacts on the firms’ labor productivity. It appears that employment and investment subsidies, in particular, tend to be allocated to relatively less efficient companies. We further observe that declines in the firm’s lagged labor productivity levels are clearly more weakly related to the subsidized firms’ exit than to the exit of firms that have not received any subsidies. Our empirical findings thus suggest that the allocation of subsidies to relatively inefficient firms increases their liquidity, making their market exit less likely than it would be otherwise. In other words, our data indicate that subsidy allocation weakens the shadow of death phenomenon observed in the previous empirical studies and hinders the process of creative destruction in the economy.

This is a preview of subscription content, access via your institution.

Fig. 1
Fig. 2
Fig. 3


  1. 1.

    See Statistics Finland ( for further information on the firm level datasets used in the analysis.

  2. 2.

    Relatedly, various prior empirical studies conclude that notable and persistent productivity differences exist among firms (see, e.g., Fox and Smeets 2011).

  3. 3.

    It is possible that subsidies also affect the order of magnitude of capital and labor. However, it seems likely that such an effect would appear with a time lag similarly as the effects of a firm’s R&D expenditures that are typically treated as an independent factor explaining variation in firms’ labor productivity.

  4. 4.

    All monetary variables have been deflated by a 2-digit industry level producer price index.

  5. 5.

    Dearden et al. (2000), Ilmakunnas and Maliranta (2005), and Maliranta and Rouvinen (2006), among others, consider a similar type of model in other contexts.

  6. 6.

    The source of business subsidies as well as other data is Statistics Finland. See Statistics Finland ( for detailed information on the business subsidies database. In the empirical analysis, we have excluded from the data R&D subsidies smaller than 30,000 euros per year as, according to the Finnish Funding Agency for Innovation (Tekes), these are used merely for the planning and feasibility studies of R&D projects and do not represent actual R&D subsidies. Also, we use the threshold of 5,000 euros per year for employment subsidies, which reflects the average minimum subsidy for employing one person per year, to remove potentially erroneous recordings from the data (the smallest recorded annual employment subsidy was 19 euros). Similarly, other subsidies are limited to those above 5,000 euros per year. We use a transformation of x = x + 0.001 for the subsidy variables before taking logs to keep firms with no subsidies in the estimation sample.

  7. 7.

    We tested the endogeneity of the three subsidy variables by first estimating a model that explains the potentially endogenous variable with all exogenous variables and instruments. The saved residual from the estimated model was subsequently included as an additional explanatory variable in the model explaining labor productivity as a function of the set of exogenous and potential endogenous variables. The estimated coefficient for the residual was statistically significant in the cases of all three subsidy types. In addition, the endogeneity of all subsidy variables together was strongly supported by Wooldridge’s (1995) score test.

  8. 8.

    The major alternative to the conditional difference-in-differences method would be the pair-wise matching approach, which is rather commonly used for analyzing the causal effects of an industrial policy. The use of a pair-wise matching method—which pair-wise compares identical firms with respect to their characteristics—leads into the use of a greatly limited number of control variables. Each additional control characteristic leads to fewer identical pairs and thus, in practice, because firms are highly heterogeneous, a major loss of data can only be avoided by controlling relatively few factors. Because our database provides a rich set of control variables potentially affecting a firm’s labor productivity, we prefer to use the conditional difference-in-differences method, which enables controlling variation in a multitude of relevant factors.

  9. 9.

    Therefore, the variable RD_BUDGET, which covers government budgets’ for three different R&D subsidy types (i.e., direct subsidies, loans and capital loans), may take six different values annually depending on the types of R&D subsidies that a firm has applied for.

  10. 10.

    Here, we follow the prior work of Koski in defining the empirical model for the firm’s propensity to obtain subsidies (see Koski and Tuuli 2010).

  11. 11.

    The major methodological problem of the empirical studies aimed at evaluating the effectiveness of different public policies is that the selection to the subsidy programs is usually not random.

  12. 12.

    Due to data limitations, our exit variable includes all types of exits, i.e., it includes liquidations, mergers and acquisitions, etc.

  13. 13.

    We also estimated the IV models for different sub-groups of firms such as different geographical locations (i.e., for different provinces and for firms located in cities, in urban areas and in the countryside) and different industries and similarly found either not statistically significant or negative effects between subsidies and labor productivity.

  14. 14.

    As a robustness test, we also estimated the model for each subsidy type separately (i.e., in each estimated equation, we multiplied the lagged labor productivity variable by the reception and non-reception of one subsidy type only). These estimation results led to similar conclusions concerning the difference between the orders of magnitude of coefficients for the lagged labor productivity variable for subsidized and non-subsidized firms.

  15. 15.

    Our data, indeed, show that there is greater variance in labor productivity among firms that are active in R&D than among those that are not.

  16. 16.

    Relatedly, the study of Buts and Jegers (2012) suggests that business subsidies may increase market concentration. Buts and Jegers (2013) find that business subsidies are positively related to the subsequent increase in the market shares of subsidized firms, suggesting another potential way in which subsidies distort competition.


  1. Almus M (2004) The shadow of death—an empirical analysis of the pre-exit performance of new German firms. Small Bus Econ 23:189–201

    Article  Google Scholar 

  2. Baghana R (2010) Public R&D subsidies and productivity: evidence from firm-level data in Quebec. UNU-MERIT working paper series 2010-055

  3. Buts C, Jegers M (2012) A note on state aid and concentration: the case of Belgium. Eur Compet J 8:153–162

    Article  Google Scholar 

  4. Buts C, Jegers M (2013) The effect of ‘State aid’ on market shares: an empirical investigation in an EU Member State. J Ind Compet Trade 13(1):89–100

    Article  Google Scholar 

  5. Carreira C, Teixeira P (2011) The shadow of death: analysing the pre-exit productivity of Portuguese manufacturing firms. Small Bus Econ 36:337–351

    Article  Google Scholar 

  6. David PA, Hall BH, Toole AA (2000) Is public R&D a complement or substitute for private R&D? A review of the econometric evidence. Res Policy 29:497–529

    Article  Google Scholar 

  7. Dearden L, Reed H Van Reenen J (2000) Who gains when workers train? Training and corporate productivity in a panel of British industries. CEPR Discussion Papers 2486

  8. Ebersberger B (2011) Public funding for innovation and the exit of firms. J Evol Econ 21:519–543

    Article  Google Scholar 

  9. Fox JT, Smeets V (2011) Does input quality drive measured differences in firm quality? Int Econ Rev 52:961–989

    Article  Google Scholar 

  10. Gelabert L, Fosfuri A, Tribó JA (2009) Does the effect of public support for R&D depend on the degree of appropriability? J Ind Econ LVII:736–767

    Article  Google Scholar 

  11. Girma S, Görg H, Strobl E, Walsh F (2007) Creating jobs through public subsidies: an empirical analysis. Labour Econ 15:1179–1199

    Article  Google Scholar 

  12. Griliches Z, Regev H (1995) Firm productivity in Israeli industry: 1979–1988. J Econ 65:175–203

    Article  Google Scholar 

  13. Ilmakunnas P, Maliranta M (2005) Technology, labour characteristics and wage-productivity gaps. Oxf Bull Econ Stat 67:623–645

    Article  Google Scholar 

  14. Imbens G, Wooldridge JM (2009) Recent developments in the econometrics of program evaluation. J Econ Lit 47:5–86

    Article  Google Scholar 

  15. Irwin DA, Klenow PJ (1996) High-tech R&D subsidies: estimating the effects of Sematech. J Int Econ 40:323–344

    Article  Google Scholar 

  16. Koski H, Pajarinen M (2013) The role of business subsidies in job creation of start-ups, gazelles and incumbents. Small Bus Econ 41:195–214

    Article  Google Scholar 

  17. Koski H, Tuuli J (2010) Business subsidies in Finland: the dynamics of application and acceptance stages. Etla Discussion Papers 1225

  18. Kösters S (2010) Subsidizing start-ups: policy targeting and policy effectiveness. J Ind Compet Trade 10(3–4):199–225

    Article  Google Scholar 

  19. Maliranta M, Rouvinen P (2006) Informational mobility and productivity: Finnish evidence. Econ Innov New Technol 15:605–616

    Article  Google Scholar 

  20. Managi S (2010) Productivity measures and effects from subsidies and trade: an empirical analysis for Japan’s forestry. Appl Econ 42:3871–3883

    Article  Google Scholar 

  21. Morgan SL, Harding DJ (2006) Matching estimators of causal effects. Prospects and pitfalls in theory and practice. Sociol Methods Res 35:3–60

    Article  Google Scholar 

  22. Oh I, Lee J-D, Heshmati A, Choi G-G (2009) Evaluation of credit guarantee policy using propensity score matching. Small Bus Econ 33:335–351

    Article  Google Scholar 

  23. Ottaviano G, Kangasharju A, Maliranta M (2009) Local innovative activity and regional productivity: implications for the Finnish national innovation policy. In: Evaluation of the Finnish National Innovation System—full report, The Ministry of Education and the Ministry of Employment and the Economy. Taloustieto, Helsinki, pp 203–238

    Google Scholar 

  24. Wallsten SJ (2000) The effects of government-industry R&D programs on private R&D: the case of the Small Business Innovation Research program. Rand J Econ 31:82–100

    Article  Google Scholar 

  25. Wooldridge JM (1995) Score diagnostics for linear models estimated by two stage least squares. In: Maddala GS, Phillips PCB, Srinivasan TN (eds) Advances in econometrics and quantitative economics: essays in honor of Professor C. R. Rao. Blackwell, Oxford, pp 66–87

    Google Scholar 

Download references

Author information



Corresponding author

Correspondence to Heli Koski.

Rights and permissions

Reprints and Permissions

About this article

Verify currency and authenticity via CrossMark

Cite this article

Koski, H., Pajarinen, M. Subsidies, the Shadow of Death and Labor Productivity. J Ind Compet Trade 15, 189–204 (2015).

Download citation


  • Labor productivity
  • Business subsidies
  • Firm exit
  • Enterprise policy
  • Technology policy

JEL Classification

  • D24
  • J23
  • L10
  • L53
  • O25