How large are fiscal multipliers in Turkey?

ABSTRACT

Using the augmented version of the Blanchard-Perotti’s SVAR model, this article seeks to estimate the size of fiscal multipliers in Turkey for the period 2002:q3–2016:q2. Unlike many previous papers that use aggregate data in estimating the size of the fiscal multiplier, we use disaggregated data on taxes and government spending for the same purposes. Our empirical findings indicate that the size of the short-run fiscal multipliers for taxes much differs from that of government spending. Depending on the disaggregated tax and government spending instruments, it ranges from −0.83 to −0.27 for taxes, and from 0.02 to 0.98 for government spending, respectively. Overall, these findings corroborate the idea that a shock to taxes produces a non-Keynesian effect on GDP whereas government spending creates a (weak) Keynesian effect.

KEYWORDS:

• Fiscal policy
• fiscal multipliers
• SVAR
• Turkey

Acknowledgments

This paper is the revised and further improved version of the discussion paper published in EconStor. The discussion paper version is available on-line at: https://www.econstor.eu/bitstream/10419/162763/1/PAPER–Fiscal%20Multipliers–04July2017.pdf

Disclosure statement

No potential conflict of interest was reported by the authors.

Note on contributors

Hüseyin Şen is full professor at Ankara Yıldırım Beyazıt University, where he is at present the chairperson of Public Finance Department. His principal research areas are public finance, public economics, fiscal policy, and development macroeconomics. Published many scholarly papers in international peer-refereed journals; including Panoeconomicus, European Journal of Economics and Economic Policies, Intervention, and The Journal of International Trade & Economic Development.

Ayşe Kaya is an associate professor in the Department of Public Finance at İzmir Kâtip Çelebi University. Her research interest falls into the areas of fiscal policy, tax policy, and public economics. Her research papers have been published in international peer-refereed journals including The Journal of International Trade & Economic Development, European Journal of Economics and Economic Policies, Intervention, and Economic Analysis and Policy.

Notes

1 Riera-Crichton, Vegh, and Vuletin, “Fiscal Multipliers in Recessions.”

2 The effects of fiscal multipliers on output are usually defined in terms of either its level or growth in the literature. Although there is no difference between the two for the first period, in the subsequent periods they become substantially different. We consider the former in this paper.

3 Blanchard and Perotti, “An Empirical Characterization.”

4 Venturini, “Are Estimates.”

5 Çebi, “The Effects,” and Çebi, “The Government Spending Multiplier.”

6 Acemoglu and Ucer, “The Ups and Downs.”

7 Makin, “The Limits of Fiscal Policy.”

8 Arestis, “Fiscal Policy is Still an Effective Instrument.”

9 Hemming, Kell, and Mahfouz, “The Effectiveness of Fiscal Policy.”

10 For more details, see Hemming, Kell, and Mahfouz, “The Effectiveness of Fiscal Policy,” and Arestis, “Fiscal Policy is Still and Effective Instrument.”

11 Giavazzi and Pagano, “Can Severe Fiscal Contractions.”

12 Alesina and Perotti, “Fiscal Adjustments.”

13 Alesina and Ardagna, “Tales of Fiscal Adjustments.”

14 See in particular Alesina and Perotti, “Fiscal Adjustments”; Alesina and Ardagna, “Tales of Fiscal Adjustments”; and Alesina and Ardagna, “Large Changes in Fiscal Policy.”

15 Mountford and Uhlig, “What are the Effects.”

16 Romer and Romer, “The Macroeconomic Effects of Tax Changes.”

17 Auerbach, Gale, and Harris, “Activist Fiscal Policy.”

18 Barro and Redlick, “Macroeconomic Effects.”

19 Auerbach and Gorodnichenko, “Measuring the Output Responses.”

20 Ilzetzki, Mendoza, and Végh, “How Big (Small?) are Fiscal Multipliers?”

21 Cerda, González, and Lagos, “Is Fiscal Policy Effective?”

22 Restrepo and Rincón, “Identifying Fiscal Policy Shocks.”

23 Yadav, Upadhyay, and Sharma, “Impact of Fiscal Policy Shocks.”

24 Tang, Liu, and Cheung, “Changing Impact.”

25 Jha et al., “Effectiveness.”

26 Matheson and Pereira, “Fiscal Multipliers for Brazil.”

27 Ilzetzki, “Fiscal Policy.”

28 Çebi, “The Effects of Fiscal Policy Shocks.”

29 Çebi, “The Government Spending Multiplier.”

30 Ibid.

31 Among various ranges of estimations, for instance, Krogstrup, “Should We Pay Attention” predicts the magnitude of government spending multiplier as well-below of −2, whereas Perotti, “Public Investment” and Mountfold and Uhlig “What are the Effects” estimate it to be above 4 in very extreme cases. In this regard, Christiano, Eichenbaum, and Rebelo, “When is the Government” argue that its size could be in the range of 3–5 under the binding zero lower bound on the interest rate. In all other cases, the size of government spending multiplier is in a range of between 0.5 and 2.

32 For the full-coverage of these arguments, see, inter alia, Hemming, Kell, and Mahfouz, “The Effectiveness of Fiscal Policy”; Ilzetzki, Mendoza, and Végh, “How Big (Small?)”, Whalen and Reichling, “The Fiscal Multiplier”; and Matheson and Pereira, “Fiscal Multipliers.”

33 Arestis, “Fiscal Policy is Still an Effective Instrument,” 144.

34 Whalen and Reichling, “The Fiscal Multiplier.” For other uses of this approach, see Baxter and King, “Fiscal Policy”; Christiano, Eichenbaum, and Rebelo, “When is the Government”; and Coenen, Straub, and, Trabandt, “Gauging the Effects” for DSGE models-based studies; Hemming, Kell, and Mahfouz, “The Effectiveness of Fiscal Policy” for macroeconomic forecasting models-based studies; and Ramey and Shapiro, “Costly Capital Reallocation”; de Castro, “The Macroeconomic Effect”; Mountford and Uhlig, “What are the Effects”; Romer and Romer, “The Macroeconomic Effects”; and Ramey, “Identifying the Government Spending” for time series models-related studies attempting to estimate the magnitude of fiscal multipliers, respectively.

35 In the following two paragraphs, we have heavily benefited from Chinn, “Fiscal Multipliers,” and Whalen and Reichling, “The Fiscal Multipler.”

36 DSGE models are typically built on assumptions, as such that: (i) people have full information regarding the present economy and future economic developments; (ii) people logically base their current decisions on a full life-time plan; (iii) people are assumed to have full access to credit markets so that they can borrow to maintain their consumption in the face of a temporary loss of income; (iv) resources are fully utilized in the economy. For a comprehensive criticism of DSGE models, see also Stiglitz, “Where Modern Macreconomics.”

37 It arises in case of that isolating exogenous movements in government spending and taxes becomes difficult.

38 For further details of these models and their strengths and limitations in estimating the size of fiscal multipliers, see Parker, “On Measuring the Effects,” Chinn, “Fiscal Multipliers,” and Whalen and Reichling, “Fiscal Multiplier.”

39 To our best knowledge, one exceptional study belongs to Mertens and Ravn, “A Reconciliation,” which attempts to reconcile SVAR and narrative techniques.

40 Blanchard and Perotti, “An Empirical Characterization.”

41 What we mean by fiscal policy shocks is unexpected changes induced by the government interventions in fiscal policy – that is, an increase or decrease in taxes or government spending or both.

42 The lag problem, which is very often associated with discretionary fiscal policy, makes up of lags, such as information, recognition, implementation, and reaction lags.

43 Ben-Sliamne, Ben-Tahar, and Essid. “Effects of Discretionary Fiscal Policy.”

44 Blanchard and Perotti, “An Empirical Characterization.”

45 Arestis, “Fiscal Policy is Still an Effective Instrument,” and Jha et al., “Effectiveness of Countercyclical Fiscal Policy.”

46 Yadav, Upadhyay, and Sharma, “Impact of Fiscal Policy Shocks.”

47 Blanchard and Perotti, “An Empirical Characterization.”

48 Ibid.

49 The discretionary fiscal policy shocks are identified by exploiting decision and implementation lags in fiscal policymaking, which means that the coefficients on output in government spending and tax equations only reflect the automatic stabilizers’ effect. Hence, using quarterly data, which reflects a relatively high frequency, allows us to eliminate the effect of the systematic part of discretionary fiscal policy. See Yadav, Upadhyay, and Sharma, “The Impact of Fiscal Policy Shocks,” and Çebi, “The Government Spending Multiplier.”

50 Blanchard and Perotti, “An Empirical Characterization.”

51 Appendices are available in the on-line version or by contacting the corresponding author.

52 Ploberger and Krämer, “The Cusum Test.”

53 The unit root test results can be obtained from the authors upon request.

54 All the quantitative results reported above are available upon request.

55 Johansen, “Statistical Analysis.”

56 Blanchard and Perotti, “An Empirical Characterization.”

57 The model is estimated 1000 times using bootstrapped data. Following Matheson and Pereira, “Fiscal Multiplies in Brazil,” the bootstrapping process can be summarized as follows: (i) estimate the model parameters and obtain the residuals; (ii) resample (with replacement) from the estimated residuals and simulate the model using the parameters from (i); (iii) re-estimate the model, saving the parameters; and (iv) repeat steps (ii) and (iii) number of times.

58 Matheson and Pereira, “Fiscal Multipliers for Brazil.”

59 For further details, see Matheson and Pereira, “Fiscal Multipliers in Brazil.”

60 Even though the fiscal multiplier value of 0.98 corresponds to the weak Keynesian effect, we should not ignore that it is very close to one (1) that refers to the lower value of the Keynesian effect.