Forecasting Tax Revenues in Latvia: Analysis and Models Velga Ozolina, Astra Auzina-Emsina, Remigijs Pocs Riga Technical University, Latvia
CSB data Data Analysis Ministry of Finance data State Revenue Service (SRS) data Eurostat data
Tax Burden in Latvia in 1995-2012, % of GDP 35 % 30 25 20 15 10 5 0 Data Source: CSB database
Tax Burden in the EU Countries in 2012, % of GDP 60 50 40 30 20 10 0 % Denmark Belgium France Austria Sweden Italy Finland EU-28 Germany Luxembourg Netherlands Hungary Slovenia United Greece Croatia Cyprus Czech Republic Portugal Malta Spain Estonia Poland Ireland Romania Slovakia Latvia Bulgaria Lithuania Data Source: Eurostat database
Tax Revenues in Latvia (ESA95 methodology), m EUR m EUR 2500 m EUR 25000 2000 20000 1500 15000 1000 10000 500 5000 0 0 Value Added Tax Customs Duties Excise Taxes Personal Income Tax Corporate Income Tax Social Contributions Other Taxes Nominal GDP (right axes) Data Source: CSB database
Social Contributions in m EUR 2500 2000 1500 1000 500 0 1995 1996 Latvia, m EUR 1997 1998 1999 2000 2001 2002 2003 2004 2005 ESA95 methodology National methodology Ratio (right axes) Data Source: CSB database, Ministry of Finance data 2006 2007 2008 2009 2010 2011 2012 2013 5 4.5 4 3.5 3 2.5 2 1.5 1 0.5 0
Analysis of Legal Aspects The main laws in the group of direct taxes are: On State Social Insurance, On Personal Income Tax, On Corporate (Enterprise) Income Tax, Micro-enterprise Tax Law.
Analysis of Legal Aspects Employed persons by professional status, thsd thsd 80 70 60 50 40 30 20 10 0 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 Employers (owners) Family workers Employees (workers), right axes Self-employed thsd 1000 Self-employed in the second job 800 600 400 200 0 1000 900 800 700 600 500 400 300 200 100 0 thsd thsd 100 90 80 70 60 50 40 30 20 10 0 I III I III I III I III I III I III I III I III I III I III I III 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 Employed Self-employed, right axes Overall statistics (CSB) Taxpayers (SRS) Data Source: CSB database, State Revenue Service data
Analysis of Legal Aspects The main laws in the group of indirect taxes are: Value Added Tax Law (before 2013 law On Value Added Tax), On Excise Duty.
Seasonality Analysis Revenues of Direct Taxes, m EUR m EUR 250 200 150 100 50 0-50 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 Social Contributions Personal Income Tax Enterprise Income Tax Data Source: Ministry of Finance data
Quarterly Seasonal Indexes for Social Contributions 1.2 1.15 1.1 1.05 1 0.95 Q1 Q2 Q3 Q4 0.9 0.85 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 Data Source: Ministry of Finance data
Quarterly Seasonal Indexes for Personal Income Tax 1.2 1.15 1.1 1.05 1 0.95 0.9 Q1 Q2 Q3 Q4 0.85 0.8 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013
Quarterly Seasonal Indexes for Corporate Income Tax 1.7 1.5 1.3 1.1 0.9 0.7 Q1 Q2 Q3 Q4 0.5 0.3 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013
Seasonality Analysis Revenues of Indirect Taxes, m EUR m EUR 200 180 160 140 120 100 80 60 40 20 0 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 Value Added Tax 2005 2006 2007 2008 Excise Duty Data Source: Ministry of Finance data 2009 2010 2011 2012 2013 2014
Quarterly Seasonal Indexes 1.2 1.15 1.1 1.05 1 0.95 0.9 0.85 0.8 0.75 for Value Added Tax 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 Data Source: Ministry of Finance data Q1 Q2 Q3 Q4
Quarterly Seasonal Indexes for Excise Duty 1.6 1.4 1.2 1 0.8 Q1 Q2 Q3 Q4 0.6 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 Data Source: Ministry of Finance data
Productivity and Economic Activity Analysis Labor productivity and real GDP growth rate (2004-2007) (2008-2010) (2011-2012).
Methodology Modelling Approaches Models and Equations Monthly Quarterly Annual
Monthly Data Seasonality Indexes Corporate Income Tax Revenues
Corporate Income Tax Revenues CIT revenues = coef monthly * CIT revenues lag *(1 + + PCI infl /100)/12 + coef may *PROF lag /100 2.5 2 1.5 where CIT revenues corporate income tax revenues, CIT revenues lag annual corporate income tax revenues with 17-month lag, coef monthly corporate income tax advance payments coefficient, PCI infl annual growth rate of private consumption price index in the previous year, coef may corporate income tax revenues coefficient applied only in May, PROF lag annual profit in the previous year. 1 0.5 0-0.5 % 5 4 3 2 1 0-1 Corporate income tax revenues coefficient, % m EUR 3500 3000 2500 2000 1500 1000 500 0-500 -1000 Annual profit in the previous year, m EUR
Corporate Income Tax Revenues SEE = 0.32 RSQ = 0.7593 RHO = 0.45 Obser = 196 from 1997.009 SEE+1 = 0.28 RBSQ = 0.7530 DW = 1.11 DoFree = 190 to 2013.012 MAPE = 9.70 Variable name Reg-Coef Mexval Elas NorRes Mean Beta 0 LTAX_UIN - - - - - - - - - - - - - - - - - 2.89 - - - 1 intercept 5.45808 5.9 1.89 4.15 1.00 2 @log(pci[12]) -2.37844 18.5-3.97 2.90 4.82-0.904 3 @log(im[12]) 0.48841 2.1 1.03 1.47 6.09 0.467 4 @log(im[6]) 0.80975 7.9 1.72 1.26 6.15 0.761 5 D_5*@log(IM[6]) 0.09099 11.2 0.02 1.04 0.51 0.241 6 @log(w_nom[8]/pci[8]) 0.86536 1.8 0.31 1.00 1.05 0.375
Identities Quarterly Data Econometric Equations
Identities tax_rev = taxr_coef*taxr*tax_base, 0.95 0.9 0.85 Tax rate coefficient of social contributions where tax_rev tax revenues, Tax rate coefficient of personal income tax 0.6 taxr_coef tax rate coefficient, 0.5 0.4 taxr tax rate, 0.3 0.2 0.1 tax_base tax base. 0 0 0.8 0.75 0.7 I III I III I III I III I III I III I III I III I III I III I III I III I III 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 I III I III I III I III I III I III I III I III I III I III I III I III I III I III I III I III I III I III I III I III 19951996199719981999200020012002200320042005200620072008200920102011201220132014 Tax rate coefficient of the value added tax Tax rate coefficient of excise duty, right axes 0.1 0.09 0.08 0.07 0.06 0.05 0.04 0.03 0.02 0.01
Corporate Income Tax CIT revenues = coef q * CIT revenues lag *(1 + PCI infl /100)/12 + + coef II *PROF lag /100, where CIT revenues corporate income tax revenues, CIT revenues lag annual corporate income tax revenues with 2- year lag (quarter 1), with 1- year lag (quarters 3 and 4) or weighted average of the 1-year and 2-year lag (quarter 2), coef q corporate income tax advance payments coefficient, PCI infl annual growth rate of private consumption price index in the previous year, coef II corporate income tax revenues coefficient applied only in the quarter 2, PROF lag annual profit in the previous year. 5 4.5 4 3.5 3 2.5 2 1.5 1 0.5 0 I III I III I III I III I III I III I III I III I III I III I III I III I III I III I III I III I III 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 Advance payments coefficient Tax revenues coefficient, right axes
Social Contributions Revenues SEE = 12.80 RSQ = 0.9911 RHO = 0.43 Obser = 48 from 2002.100 SEE+1 = 11.82 RBSQ = 0.9907 DW = 1.14 DoFree = 45 to 2013.400 MAPE = 2.39 Variable name Reg-Coef Mexval Elas NorRes Mean Beta 0 TAX_SOC - - - - - - - - - - - - - - - - - 394.63 - - - 1 intercept -15.67795 5.9-0.04 112.06 1.00 2 TAXR_SOC*((EMPL*W_NOM*3)/100000) 0.82441 929.9 1.03 1.22 492.62 1.007 3 D_EU 13.39648 10.3 0.01 1.00 0.31 0.046
Personal Income Tax Revenues SEE = 10.34 RSQ = 0.9841 RHO = 0.14 Obser = 48 from 2002.100 SEE+1 = 10.34 RBSQ = 0.9834 DW = 1.72 DoFree = 45 to 2013.400 MAPE = 3.24 Variable name Reg-Coef Mexval Elas NorRes Mean Beta 0 TAX_INC_PERS - - - - - - - - - - - - - - - - - 246.31 - - - 1 intercept -15.42434 9.4-0.06 62.85 1.00 2 TAXR_IIN*((EMPL*(W_NOM-TAX_NMIN))/1000-TAX_SOC*TAX_SOC_E) 3.39666 687.7 1.08 2.44 78.20 1.022 3 D_10-46.54480 56.2-0.02 1.00 0.08-0.157
Corporate Income Tax Revenues SEE = 0.26 RSQ = 0.8398 RHO = 0.21 Obser = 72 from 1996.100 SEE+1 = 0.25 RBSQ = 0.8303 DW = 1.58 DoFree = 67 to 2013.400 MAPE = 5.23 Variable name Reg-Coef Mexval Elas NorRes Mean Beta 0 LTAX_INC_CORP - - - - - - - - - - - - - - - - - 3.92 - - - 1 intercept -8.75285 40.2-2.23 6.24 1.00 2 @log(pi_cons_pr[4]) -2.03484 27.4 0.19 3.47-0.36-0.916 3 @log(im_cp[4]) 0.77342 6.9 1.43 1.45 7.26 0.817 4 @log(im_cp[2]) 0.85733 10.0 1.60 1.13 7.32 0.876 5 D_2*@log(INV_CP[1]) 0.03463 6.4 0.01 1.00 1.58 0.150
Value Added Tax Revenues SEE = 0.10 RSQ = 0.9723 RHO = 0.13 Obser = 76 from 1995.100 SEE+1 = 0.09 RBSQ = 0.9712 DW = 1.74 DoFree = 72 to 2013.400 MAPE = 1.37 Variable name Reg-Coef Mexval Elas NorRes Mean Beta 0 LTAX_VAT - - - - - - - - - - - - - - - - - 5.27 - - - 1 intercept -2.56866 55.4-0.49 36.14 1.00 2 @log(gdp_cp) 1.03000 201.5 1.54 1.28 7.89 1.094 3 @log(time) -0.09040 6.8-0.06 1.13 3.37-0.143 4 D_EU 0.09103 6.5 0.00 1.00 0.20 0.063
Excise Duty Revenues SEE = 0.08 RSQ = 0.9767 RHO = 0.06 Obser = 60 from 1999.100 SEE+1 = 0.08 RBSQ = 0.9759 DW = 1.87 DoFree = 57 to 2013.400 MAPE = 1.25 Variable name Reg-Coef Mexval Elas NorRes Mean Beta 0 LTAX_EXC - - - - - - - - - - - - - - - - - 4.70 - - - 1 intercept -3.29918 170.7-0.70 42.90 1.00 2 @log(gdp_cp) 0.98389 514.3 1.70 1.39 8.11 0.958 3 D_0910 0.14311 17.7 0.00 1.00 0.13 0.098
Annual Data Calculations in ESA95 Identities Transformation Coefficients to forecast national data
Dynamics of Estimated II Pillar Rates in 2003-2013 0.3 0.25 0.2 0.15 0.1 0.05 0-0.05 20032004200520062007200820092010201120122013 Estimated II pillar rate (based on assets) Estimated II pillar rate (based on differences in national and ESA95 methodologies) Estimated II pillar rate (combined) Max II pillar rate
Tax Rate Coefficients 2.5 2.0 1.5 1.0 0.5 0.0-0.5 Tax rate coefficient of personal income tax Tax rate coefficient of corporate income tax Tax rate coefficient of social contributions 0.6 0.5 0.4 0.3 0.2 0.1 0.0 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 Tax rate coefficient of the value added tax Tax rate coefficient of excise duty, right axes 0.10 0.08 0.06 0.04 0.02 0.00
Forecasts Evaluation of precision Numbers
Corporate Income Tax (Monthly Data) 2014.001 2014.007 MAPE = 14.3% Modified dummy MAPE = 3.6%
MAPE Values, % (Quarterly Data) Tax Type 2014 I 2014 I and II Social Contributions 3.7 2.2 Personal Income Tax 6.8 - Corporate Income Tax 0.4 2.5 Value Added Tax 15.4 16.5 Excise Duty 0.7 2.0
Comparison of Forecasts Tax Type Identity-based approach Econometric equations Monthly Quarterly Annual Monthly Quarterly Social Contributions - 2271.3 2139.6-2237.2 Personal Income Tax - 1379.3 1415.1-1385.0 Corporate Income Tax 342.3 359.2 350.0 359.5 354.6 Value Added Tax - 1885.1 1738.7-1638.5 Excise Duty - 757.1 731.2-777.9
Conclusions Using annual data, identity-based approach should be prefered, however quarterly and monthly data can give similar results and thus the choice is in hands of the model-user Identity-based approach allows for a greater flexibility in scenario-building process. Econometric approach involves less assumptions and thus may seem to be more objective Forecasts depend very much on the values of exogenous indicators, therefore modelling approaches should be tested regulary to find the most reliable ones
LAIMA Latvian Interindustry Model (Aggregated/Annual) Goddess of destiny
Thank You for Attention Questions?