Presentation on theme: "An editing strategy for annual VAT-turnover Montreal June 18-21, 2007 * Jeffrey Hoogland Grietje van Haren."— Presentation transcript:
An editing strategy for annual VAT-turnover Montreal June 18-21, 2007 * Jeffrey Hoogland Grietje van Haren
–Register-based business statistics –Micro-editing of VAT-data for fiscal units –Conversion of fiscal units to enterprises –Macro-editing of VAT-turnover for enterprises –Objectives Outline
- Pressure from Dutch government and business community -Use available register information instead of questionnaires - Decrease response burden for small and medium sized companies Register-based business statistics
1.micro-editing of VAT data for fiscal units 2.conversion of VAT-turnover for fiscal units to VAT-turnover for enterprises 3.macro-editing of VAT-turnover for enterprises An editing strategy for annual VAT-turnover
–Fiscal units submit a VAT-declaration per month, quarter, or year –Those with large turnover tax or those that try to mislead the IRS have to report per month –For a fiscal unit i compute VAT-turnover for each time period t in year r and year r-1 –Detect outliers within a fiscal unit that (usually) report a substantial VAT-turnover Micro-editing of VAT-data for fiscal units
VAT-turnover of a fiscal unit substantial and possibly too large for a time period in year r if: > d million euro and > c × Median value of VAT-turnover across time periods of fiscal unit in year r and year r-1 Micro-editing of VAT-data for fiscal units Example: d=1; c=9; year out of monthly declarations 535 out of quarterly declarations
VAT-turnover of a fiscal unit substantial and possibly too small for a time period in year r if: Median value of VAT-turnover > d million euro and VAT-turnover < Median value / c Example: d=1; c=9; year out of monthly declarations 272 out of quarterly declarations
Conversion of fiscal units to enterprises fiscal unitfiscal personCBS personenterprise fiscal unit fiscal person CBS person enterprise
Selectivity in available VAT-turnover
Figure 1. Box plots for raw and edited SBS-turnover (thousands of euros) for enterprises in building industry with 2-4 employees. There are two groups: SBS-records with VAT-turnover and SBS-records without VAT-turnover.
Macro-editing of VAT-turnover – VAT-data incomplete – Weight VAT-turnover – Direct weighting if available VAT-data not selective – Reference aggregates –in design phase: VAT-, SBS-, and CT-turnover –in production phase: VAT-turnover of last year – Detect influential inconsistencies
- For a specific publication cell - Stratify by enterprise size - Detect outliers per stratum - Using box plot criterium for i = log |y i -y med | - - Direct weighting of outliers per stratum - Direct weighting of non-outliers per stratum Direct weighting of VAT-turnover
Table 3. Net turnover for 17 publication cells for builders with 1-99 employees.
Detecting influential inconsistencies Example of score function for A B, e.g. PS VAT Score functions are used to detect influential inconsistencies
Objectives – Improve cooperation with tax authorities – Reduce differences in definitions and filling-in behaviour – Improve weighting and outlier detection for VAT-turnover – Use tax assessments and surveys to develop a treatment strategy for VAT-declarations – Reduce use of business surveys