2988 Views
679 Downloads |
To what extent do fiscal spending rules affect budget composition?
Wildmer Daniel Gregori
Article | Year: 2018 | Pages: 325 - 345 | Volume: 42 | Issue: 3 Received: November 29, 2017 | Accepted: June 29, 2018 | Published online: September 3, 2018
|
FULL ARTICLE
FIGURES & DATA
REFERENCES
CROSSMARK POLICY
METRICS
LICENCING
PDF
Municipalities with inhabitants:
|
Year
|
Up to 5,000 (Group A)
|
More than 5,000 (Group B)
|
2004 (t1)
|
None
|
Fiscal gap:
zero growth
|
2005 (t2)
|
None
|
Total
expenditure cap
|
2006 (t3)
|
None
|
Consumption
and Investment caps
|
Source: Ministry of Economics and Finance – Financial laws.
|
Group A
|
|
Obs. per year
|
Mean
|
Std. dev.
|
Min.
|
Max.
|
Revenue side
|
|
|
|
|
|
Taxes
|
916
|
199.24
|
114.70
|
1.31
|
2252.18
|
New loans
|
916
|
128.95
|
246.93
|
0.00
|
3744.68
|
Expenditure
side
|
|
|
|
|
|
Current spending – total
|
916
|
664.42
|
221.14
|
344.79
|
2780.25
|
Wages
|
916
|
205.57
|
73.29
|
58.15
|
902.95
|
Services
|
916
|
254.79
|
126.03
|
7.15
|
1040.88
|
Capital spending – total
|
916
|
434.80
|
607.00
|
3.58
|
15810.63
|
Infrastructure
|
916
|
340.97
|
430.53
|
0.00
|
5905.50
|
|
Group B
|
|
Obs. per year
|
Mean
|
Std. dev.
|
Min.
|
Max.
|
Revenue side
|
|
|
|
|
|
Taxes
|
495
|
204.18
|
118.19
|
2.24
|
1754.92
|
New loans
|
495
|
115.52
|
183.34
|
0.00
|
2529.52
|
Expenditure
side
|
|
|
|
|
|
Current spending – total
|
495
|
609.24
|
232.99
|
317.67
|
3349.18
|
Wages
|
495
|
194.07
|
71.76
|
76.32
|
677.60
|
Services
|
495
|
236.77
|
123.16
|
21.18
|
1899.68
|
Capital spending – total
|
495
|
306.59
|
374.37
|
10.20
|
5496.12.
|
Infrastructure
|
495
|
285.91
|
290.17
|
0.00
|
5468.95
|
Notes: In Group A, there are municipalities with a number of inhabitants in the range 3,000-4,999, while in Group B the range is 5,000-7,000.
Budget item
|
Revenue
|
|
Case A
|
Case B
|
Taxes
|
-0.50
(1.84)
|
0.11
(1.78)
|
0.10
(1.75)
|
-1.86
(1.39)
|
-1.34
(1.41)
|
-1.51
(1.43)
|
New loans
|
-42.03
(37.06)
|
-43.26
(36.68)
|
-43.75
(36.69)
|
62.98**
(30.93)
|
62.28**
(30.86)
|
62.21**
(30.95)
|
|
Expenditure
|
|
Case A
|
Case B
|
Current –
total
|
-3.99
(12.53)
|
-0.76
(9.89)
|
-0.71
(9.93)
|
-30.38***
(10.58)
|
-31.45***
(10.94)
|
-33.40***
(10.82)
|
Wages
|
-1,66
(3.91)
|
-2,14
(3.58)
|
-2.33
(3.60)
|
4.31*
(2.35)
|
2.81
(2.56)
|
2.11
(2.66)
|
Services
|
-10,03
(7.63)
|
-6,38
(7.04)
|
-6.37
(7.05)
|
-22.84***
(6.82)
|
-22.32***
(6.67)
|
-22.66***
(6.75)
|
Capital –
total
|
25.81
(67.11)
|
16.70
(65.07)
|
14.97
(64.17)
|
126.96**
(62.97)
|
114.87*
(64.79)
|
114.55*
(65.29)
|
Infrastructure
|
-21,28
(36.68)
|
-25,92
(35.94)
|
-24.30
(47.55)
|
93.08*
(48.72)
|
84.21*
(50.14)
|
84.01*
(50.06)
|
Time-invariant
controls (X)
|
No
|
Yes
|
Yes
|
No
|
Yes
|
Yes
|
Time-variant
controls (Z)
|
No
|
No
|
Yes
|
No
|
No
|
Yes
|
Obs.
|
2,822
|
2,822
|
2,822
|
2,822
|
2,822
|
2,822
|
Notes: Case A considers t1 and t2 while Case B analyses t2 and t3. The estimations are performed implementing a local linear regression method. Values are in euro per capita. Robust standard errors are reported in parenthesis, clustered at the provincial level. *** (**, *) indicates statistical significance at the 1 (5, 10) percent level.
Budget item
|
LLR
(b=1,500)
|
LLR
(b=1,500)
|
SPA
(order 2)
|
SPA
(order 2)
|
|
Revenue
|
Taxes
|
-1.57
(1.90)
|
-1.07
(1.93)
|
-0.98
(2.64)
|
-1.33
(2.64)
|
New loans
|
70.42**
(35.16)
|
67.70*
(35.17)
|
77.48*
(44.46)
|
78.59*
(44.42)
|
|
Expenditure
|
Current –
total
|
-24.17*
(13.37)
|
-27.33*
(13.79)
|
-24.29
(18.70)
|
-20.77
(17.73)
|
Wages
|
6.51**
(2.80)
|
3.66
(3.23)
|
3.48
(4.45)
|
4.81
(4.93)
|
Services
|
-21.44**
(8.24)
|
-21.44**
(8.05)
|
-18.73*
(10.78)
|
-19.08*
(10.65)
|
Capital –
total
|
154.72**
(68.98)
|
134.62*
(70.88)
|
156.01*
(87.34)
|
149.20*
(88.65)
|
Infrastructure
|
118.53**
(52.15)
|
102.25*
(52.85)
|
-115.51*
(64.02)
|
111.83*
(64.93)
|
Controls (X
and Z))
|
No
|
Yes
|
No
|
Yes
|
Obs.
|
1,976
|
1,976
|
2,822
|
2,822
|
Notes: Case B refers to municipalities between 3,000 and 7,000 inhabitants in the period t2 and t3. The estimations are performed implementing a local linear regression (LLR) method with a bandwidth (b) equal to 1,500. SPA is the spline polynomial approximation method of order 2 and the analysis is performed considering municipalities between 3,000 and 7,000 inhabitants. Values are in euro per capita. Robust standard errors are reported in parentheses, clustered at the Provincial level. *** (**, *) indicates statistical significance at the 1 (5, 10) percent level.
Figure 1Italian municipalities’ consumption, investment and total spending as a percentage of GDP DISPLAY Figure
Table 1Fiscal rules imposed by the Domestic Stability Pact on Italian municipalities DISPLAY Table
Figure 2Level of Consumption over Investment for municipalities under 5,000 inhabitants (Group A) and above (Group B) DISPLAY Figure
Table 2Descriptive statistics. Budget items used in the analysis by group, average of the period 2004-2006 (euro per capita) DISPLAY Table
Table 3Effects of the Domestic Stability Pact on budget items DISPLAY Table
Figure 3Difference-in-discontinuities on consumption and investment spending, Case A (t1 and t2) and Case B (t2 and t3) DISPLAY Figure
Figure 4Difference-in-discontinuities on wages, services, infrastructure and new debt, Case B (t2 and t3) DISPLAY Figure
Table 4Effects of the Domestic Stability Pact on budget items, robustness analysis in Case B DISPLAY Table
Figure 5Density tests DISPLAY Figure

* I would like to thank Massimo Bordignon, Julie Cullen, Matteo Falagiarda, Elena Giarda, Giorgio Gulino, Fabian Kruger, Paolo Manasse, Massimiliano Marzo, Katja Neugebauer, Giovanni Prarolo, Valerie Ramey, James Rauch, Nelio Rosa, Alberto Zanardi and three anonymous referees for useful comments. I would also like to thank participants at the University of Bologna internal seminar, the Europaeum workshop in Prague, the University of California San Diego seminar, the OECD-Oxford University Joint conference on "Economics for a Better World" in Paris, the IWH workshop in Halle, the ZEW conference of Public Finance in Mannheim, the ECB seminar in Frankfurt and the Italian Society of Public Economics conference in Pavia. Part of the work has been done at the University of Bologna, Italy. Responsibility for any errors lies solely with the author. The views expressed are purely those of the author and may not in any circumstances be regarded as stating an official position of the European Commission.
1 For further information about the Stability and Growth Pact see here.
2 Law no. 367/2000. In particular, the specific functions are presented by the 167/1996 Presidential Decree and cover a wide range of subjects, such as general administration, justice, local police, state education (up to primary school and part of secondary school), culture, sport, tourism, local public transportation, urban development, social sector, economic development, productive local services.
3 Ambrosanio and Bordignon ( 2007) discuss the internal application of the European Stability Pact with local governments in some selected European countries (i.e. Germany, Belgium, Spain and Italy), showing that there is not necessarily a link between decentralisation and financial instability.
4 This study focuses on the years 2004-2006 and therefore the decline in investment spending after 2007 goes beyond the findings detailed in section 3.
5 The special-statute regions are Sicily, Sardinia, Aosta Valley, Trentino-Alto-Adige and Friuli-Venezia-Giulia.
6 Further details are shown in the Finance law no. 311, December 30, 2004 and Document of Ministry of Economy and Finance “Circolare della Ragioneria Generale dello Stato” n. 4, February 8, 2005.
7 Further details are shown in the Finance law no. 266, December 23, 2005 and Document of Ministry of Economy and Finance “Circolare della Ragioneria Generale dello Stato” n. 8, February 17, 2006.
8 The fiscal gap had to be equal to zero in 1999-2000, while it could grow for a maximum of 3 and 2.5 per cent in 2001 and 2002, respectively, and again had to be equal to zero in 2003 and 2004.
9 Giurato and Gastaldi ( 2009) summarise the evolution of the DSP in Italy and the fiscal items included by the different DSPs. For an extensive review of different fiscal rules, see Budina et al. ( 2012) and Cordes et al. ( 2015).
10 In any case, all the municipalities are subject to the Law for Local Authorities.
11 The DiDisc approach is also implemented, as a robustness check, by Asatryan et al. ( 2017).
12 The DiDiscThe Diff-in-disc approach needs comparable groups of Municipalities (treated vs not-treated). It may be affirmed that small and large municipalities might have different behaviours in terms of budget policies, therefore a specific (and not excessively wide) distance from the cut-off level needs to be imposed. Nevertheless, a robustness analysis is performed at different threshold to support the results (Table 4).
14 Standard errors are clustered at the provincial level in order to have a sufficient number of municipal units to avoid unreliable standard errors, as it would with clusters at the municipal level (Angrist and Pischke, 2008).
15 Results have been tested for outliers, trimming values greater than 97.5th percentile. Another check has been done considering fixed effects at the regional level. In both cases results still hold (results available upon request). Furthermore, a test for the parallel trend assumption has been implemented considering the three year of the analysis, 2004-2006. Specifically, given that the fiscal rules’ variation in 2005 is not effective, it is possible to test the parallel trend assumption for 2005 using as baseline 2004, and the outcome is that the assumption holds. Repeating the test for the years 2005-2006, the outcome is, as expected, that when the fiscal rules variations is proven to be effective, the assumption does not hold (Bellucci, Pennacchio and Zazzaro, 2018; Angrist and Pischke, 2009; Cerulli, 2015).
|
|
September, 2018 III/2018 |