Macrosectoral analysis of the impact of a VAT increase
This study was commissioned by the Central Economic Council (CEC), and in particular by the ‘Construction’ Special Advisory Commission. It presents the sectoral results of a report that was produced in 2011 by the National Bank of Belgium and the Federal Planning Bureau. As requested by the CEC, we comment here in detail on the impact of a VAT increase without additional measures, (variant 1) and the impact of a VAT increase with transitional neutralization of the effect of that increase on indexation (variant 2).
For each variant, the shock in the model represents a fixed percentage of the GDP in the baseline simulation (i.e. 0.5% of GDP, approximately EUR 1.8 billion in the first year): the shock simulated in year t is maintained during the whole seven-year simulation period. The variants were produced with HERMES, an econometric model used for medium-term macrosectoral simulations. The aim of this exercise is not to deliver results in absolute levels but to measure the specific impact of various measures. The results are thus reported as a deviation relative to the baseline.
At the macroeconomic level, a VAT increase (not neutralized in the index) has a direct impact on prices and triggers an upward spiral of prices and wages that becomes self-sustaining owing to the increase in unit labour costs. For enterprises, an increase in costs means a slight erosion of their competitiveness on the foreign markets and thus of their exports. Households, which bear the main burden of this indirect tax, experience a significant loss in purchasing power due to price increases and job losses, leading to a decrease in private consumption and in residential investments. The decreasing domestic demand also leads to a drop in imports. In total, GDP should be reduced by 0.3% over the medium-term and employment should fall by 25 000 persons.
Under the second variant, which includes a transitional neutralization of the effect of a VAT increase on indexation (neutralisation introduced in year t), the impact of the VAT increase on prices and costs is reduced, but leads in the medium-term to a downturn in economic activity that is identical to that calculated in the first variant. On the one hand, exports stabilize (due to a slower increase in costs), but on the other hand, private consumption decreases more sharply (despite a less significant drop in employment) because of the lack of wage indexation weighs more on the disposable income of households. Similarly, the fall in investments, especially in housing investments, is stronger. Despite an identical decline in activity, job losses are smaller.
As for branches of the economy, the VAT increase has a negative impact on economic activity and employment in all branches. But branches that have production more broadly oriented towards domestic demand experience a greater impact. Over the medium term, the volume of production is expected to decrease by 0.6% in the 'credit & insurance' branch, by 0.52% in the ‘energy’ branch, by 0.51% in the building sector and by 0.42% in the consumer goods industry. The decrease is identical or even stronger under variant 2, except for the consumer goods industry. The combination of an increase in unit labour costs and of a significant decline in domestic demand is particularly harmful for these branches.
For the branches that are more oriented towards foreign trade, the negative impact of a VAT increase on their production volume is less pronounced. Even if their unit labour costs increase, their market (volume of trade worldwide) by assumption should remain unchanged compared to the baseline. External demand is not affected, unlike domestic demand. It is for these reasons that the fall in production is systematically weaker under the second variant since only the unit labour costs, which increase less rapidly, play a role.
As regards job losses, the ranking differs compared to that for output losses. Besides the cost/price and demand effects, another factor comes into play: the capital intensity of the branch. Over the medium term, the largest job losses occur in the (labour intensive) building sector (-1.07%), followed by the other market services, the retail trade, hotels and restaurants and credit and insurance. By contrast, the energy sector, whose production is strongly affected, loses only 0.04% of its jobs compared to the baseline. Under the variant with transitional neutralization of the impact of a VAT increase on wage indexation, job losses are, quite logically, smaller but still significant in the building sector (-0.71%).