On VAT Thresholds

VAT thresholds exist so that micro-businesses do not have to keep formal accounting records when selling goods, works and services.

In practice, however, entrepreneurs use the current threshold system to extract money from companies and to optimize their taxes, especially in the B2B segment.

It works as follows. Suppose a company generates taxable profit. For example, a large company earns KZT 1 billion in profit over the year and must pay corporate income tax (CIT) of 20%, or KZT 200 million.

Instead, the company sets up 10 companies alongside it under a special tax regime (the simplified declaration regime or the retail tax regime), where tax is paid at 3% of turnover. To avoid paying 12% VAT, turnover at each company is kept below KZT 78 million a year. The main company then pays these entities for various services that enjoy preferential taxation (office rental, car rental and so on), moving KZT 780 million out of the company. The entities pay KZT 23.4 million in tax on that sum, while the main company reports gross profit of KZT 200 million and pays KZT 40 million in CIT. The total tax paid to the state comes to KZT 63.4 million, so the effective CIT rate falls from 20% to 6.3%.

If we set a tax-free threshold of KZT 15 million for VAT registration, companies will of course split up less. Even so, some small and medium-sized businesses will keep exploiting the KZT 15 million loophole.

The system is used especially widely in construction. Housing is not subject to VAT, so even honest contractors try to use the exemptions and split their businesses in order not to burden the developer with VAT.

I propose abolishing thresholds altogether. Every company registered as an individual entrepreneur (IE) or a limited liability partnership (LLP) would automatically become a VAT payer. At the same time, any LLP or IE would be automatically exempt from VAT if it uses a special tax regime (Sshe).

In effect, the special tax regime should be available only for the B2C segment, where services, goods and works are sold to the public. This applies above all to services with no input VAT, such as fitness, dentistry, spas and beauty salons.

To deal with the problem of occasional customers from the B2B segment, I propose capping sales under this regime at 10–20% of total revenue.

For the B2B segment, it is not critical to give up the VAT exemption, provided the company is not a cash-out operator. VAT charged on top of the sale price is offset further down the chain by the next legal entity. If it is not, the next company in the chain will end up charging and paying that VAT anyway.

For example:

Without exemptions:

A supplier sells services worth KZT 50 million plus 16% VAT = KZT 8 million, for a total of KZT 58 million.

The buyer, a legal entity, then produces goods worth KZT 100 million and sells them to individuals for KZT 100 million plus 16% VAT = KZT 116 million.

VAT payable is 16 − 8 = KZT 8 million.

Total profit = KZT 50 million (100 − 50), and total cash flow = 116 − 58 − 8 = KZT 50 million.

With a VAT exemption, for example:

The supplier sells services worth KZT 50 million.

The buyer produces goods and sells them for KZT 100 million plus 16% VAT = KZT 116 million.

VAT of KZT 16 million is paid to the budget, since there is no deduction.

Total profit = 100 − 50 = KZT 50 million. Cash flow = 116 − 50 − 16 = KZT 50 million.

In other words, for a supplier in the business segment it makes no difference whether it sells with or without VAT. But if everyone is required to pay VAT, there will be no schemes for avoiding CIT.

As for the argument that small businesses find bookkeeping difficult, we should understand that those who work with legal entities (the B2B segment) keep accounts in any case, because businesses need closing documents and the like anyway. Moreover, the current electronic invoice (ESF) system lets companies issue all invoices fully online, and the amount payable is calculated automatically. In practice, neither legal entities nor individuals have any difficulty with VAT accounting.