Page 10

AD ALTA

JOURNAL OF INTERDISCIPLINARY RESEARCH

In the regression analysis, there were used the Pooling model
(PM), Fixed effects model (FEM), and Random effects model
(REM), as well as the first difference model and the difference
between model. Based on our testing, we found out that
statistically insignificant are difference between model and first
difference model, the other models were statistically significant
while the significance was determined by Hausmann test. The
statistical test determined as the most appropriate model for
testing the Pooling model (PM). The results of the original
Pooling model in which were included all variables pointed out
that the shadow economy, unemployment, and public debt were
statistically insignificant. Therefore, we removed these variables
from the model correction. Another variable that has also been
removed from the model was the value-added-to-GDP ratio, as
its presence in the model is irrelevant in terms of the presence of
intermediate consumption and total GDP consumption. Finally,
we also removed the GDP per capita variable as it became
statistically insignificant after several adjustments to the model.
The modified model is shown in Tab. 3 and it is statistically
significant. Further testing of the adjusted Pooling model
revealed the following model features: (1) according to Lagrange
Multiplicate test, an individual effect in the model is significant,
while the time effect is insignificant; (2) according to the Chow
pool ability test, it is necessary to take into account the panel
data structure; (3) according to Woldridge test, it is rejected the
presence of autocorrelation; (4) testing the model for absolute
correlation confirmed that correlation is insignificant for the
whole model; (5) Maddala-Wu unit root test confirms the
existence of time series stationarity; (6) according to White test,
there is not confirmed heteroscedasticity in the model; (7)
normal distribution was tested according to Jarque-Bera test,
Shapiro-Wilk test and Kolmogorov-Smirnov test.

6 Results and discussion

Tax gaps in the EU Member States from 2004 to 2017 are shown
in Tab.2. Based on our calculation, the lowest tax gap was
reported in 2004 at the level of 614,000 mils. EUR. On the other
hand, the highest tax gap was quantified in 2017 at the level of
946,000 mils. EUR. If we look at tax gaps in the individual
countries, we can conclude that in Germany and France were tax
gaps for the whole observed period higher than 100,000 mils.
EUR every year (in Germany higher than 200,000 mils. EUR,
from 2006 to 2015 even higher than 300,000 mils. EUR). On the
other hand, the smallest tax gaps were quantified in Malta
(540 mils. EUR on average for period) and Cyprus (980 mils.
EUR on average). To sum up, tax gaps in the EU countries grew
continually from 2004 to 2017, except from 2009 when tax gaps
decreased. In the next part of the contribution, we will analyze
the period of the financial crisis and its consequences on tax
gaps. The highest VAT gaps were measured in Germany,
France, the United Kingdom, and in Italy. The smallest VAT
gaps we quantified in Malta, Cyprus, and Latvia.

In the last observed year 2017, the Czech Republic moved to the
first cluster with higher tax gaps countries. Greece, on the other
hand, moved to the cluster with smaller tax gaps. In France,
Italy, and the United Kingdom, the total tax liabilities were risen
by 28% on average, in Germany even by 79% and in the
Netherlands by 61%. German tax revenues were increased only
by 34%. Generally, we can say that even though tax revenues in
all EU countries rose, but tax liabilities rose at a greater extent.
Therefore, there was an increase in tax gaps each year.

Table 2 Tax gap in EU in period 2004-2017 (in mil. EUR)

Country/Year

2004

2005

2006

2007

2008

2009

2010

9,955

10,809

11,024

11,860

12,831

12,923

13,283

18,003

19,399

20,663

22,155

22,318

21,491

23,302

853

991

1,257

1,319

1,733

1,328

1,310

391

465

537

586

637

592

588

4,071

4,687

5,104

5,901

6,992

6,584

7,259

176,709

181,286

200,171

242,316

252,871

240,435

254,188

12,812

14,094

14,951

16,087

16,004

15,924

16,061

340

385

489

593

580

649

651

32,146

37,682

41,623

42,625

36,069

27,145

37,033

7,973

8,594

9,215

9,582

9,688

9,375

9,546

114,868

119,315

121,616

123,911

124,420

117,383

121,628

5,046

5,366

6,338

7,140

6,994

6,604

6,801

1,758

1,949

2,223

2,455

2,623

2,423

2,422

4,457

4,807

4,559

5,167

5,424

5,200

5,727

8,028

9,106

10,285

10,396

8,957

7,616

6,833

66,523

70,252

78,031

81,291

78,748

68,412

76,738

418

527

660

775

902

709

821

1,257

1,568

1,669

2,015

2,110

2,123

2,379

319

407

519

697

698

478

487

157

182

202

215

255

237

255

32,843

35,671

40,990

44,973

48,039

47,714

49,505

6,500

8,804

10,900

13,452

15,492

11,612

14,686

5,494

6,267

6,487

6,473

6,354

5,448

5,987

1,388

2,269

2,715

3,114

3,681

2,816

3,502

19,143

20,355

22,240

23,619

24,166

21,137

25,244

1,012

1,111

1,250

1,353

1,495

1,359

1,398

1,101

1,323

1,467

1,678

1,943

1,887

1,942

80,607

83,606

90,531

90,475

77,030

61,513

77,723

Total

614,174

651,278

707,716

772,224

769,051

701,117

767,299

Country/Year

2011

2012

2013

2014

2015

2016

2017

13,445

14,098

14,166

14,390

14,873

15,406

16,045

24,761

25,468

25,462

25,711

25,080

25,811

26,905

1,354

1,458

1,675

1,628

1,729

1,823

1,919

559

551

481

472

469

501

562

8,080

8,291

8,117

8,772

9,225

10,382

278,428

276,591

278,765

296,217

309,689

321,820

335,230

16,087

16,569

16,230

16,763

16,911

17,162

17,647

715

778

805

901

997

1,073

1,164

- 10 -