Lies, damned lies and statistics

Popularised by Mark Twain, this phrase highlights how statistics can be used to paint a different picture, or obscure the real one, whether by design or error. It is wielded as a defence against those who use numbers or statistics out of context, to hammer home a flawed point.

Professor Jean Tirole, in his lecture “Economics for the common good” given at the London School of Economics, said that one of the problems for economists in communicating with non-economists is that of “cognitive bias”, and that the facts are interpreted “through the prism of one’s own belief”.

The tax buoyancy ratio is merely a measure of how tax revenues will move in relation to changes in output. For example, a tax buoyancy of one would mean that an extra 1% of gross domestic product (GDP) would lead to an increase in tax revenues of 1% while the tax-to-GDP ratio would remain constant.

The tax buoyancy is calculated on an annual basis, and is derived by dividing the expected tax revenue by the GDP. Obviously this involves very complex calculations, and awfully long formulas. Tax revenue would include personal and company income tax, value added tax (VAT), customs duties and excise taxes.

Different economists come up with different tax buoyancies depending on the assumptions used. According to National Treasury, the tax buoyancy for 2017 is 0.88 (2016 – 1.47, 2018 – 1.41). The IMF, in their 2017 paper “How buoyant is the tax system”, calculates the tax buoyancy for South Africa to be 1.072 (long run), 1.610 (short run). The excellent paper by BNP Paribas, “South Africa: Buoyancy blues”, gives much depth to the factors taken into account in calculating the nominal tax buoyancy, which they estimate is likely to average 1.1 over the next three years.

The size of the tax base is only one of the variables in total revenue. Others include levels of tax rates, mix of taxes, and efficiency of tax incentives.

The GDP is a widely used indicator of a country’s economic health, and equates to the total expenditure for all final goods and services produced in a country in a particular period (GDP = consumption + investment + government spending + (exports-imports)). There are various different methods of calculating the GDP, and they all apparently come to the same result.

GDP gives an imperfect result. Only VAT is related to final expenditure. Income taxes, customs duties, and excise duties, are not. Further, GDP does not include income earned offshore (which should be included in a resident’s worldwide income) and which may be liable for SA tax.

Many factors can influence the tax buoyancy calculation. If a significant amount of VAT or PIT refunds are withheld, this money is prevented from going back into the economy, thus negatively impacting the GDP. A structural change in the tax mix will also affect tax buoyancy, resulting in a greater or lesser amount of tax revenue.

Any difference between the expected revenue (tax buoyancy ratio x GDP) and the actual revenue collected, indicates the tax gap. National Treasury, in Chapter 4 of the 2017 budget document, incorrectly refers to the tax gap as the compliance gap. These terms are not interchangeable. A compliance gap implies that any gap in revenue collected is purely as a result of an omission by the taxpayer. The tax gap includes tax avoidance as well as legal ways of reducing taxes paid, for example, a payment into a retirement annuity fund, or arbitraging the differences in domestic tax laws. National Treasury is responsible for amending legislation to keep up with tax avoidance and arbitration techniques. It is incumbent on Sars to identify loopholes in the law through carrying out efficient audits.

It is a moot point that declining economic growth would mean lower GDP growth, and hence less taxes. However, the drop in taxes would not be directly proportional and cannot be solely attributed to declining income. There would be many other reasons, such as an incorrect mix of taxes, tax incentives granted to declining industries, PAYE collected but not paid over to Sars, tax avoidance, tax evasion, and a less equitable distribution of national income. Tax collection inefficiencies cannot be ignored. Incompetent revenue staff will clog administrative decisions. They may not be able to make decisions, delaying an assessment or refund. They may not be competent to verify a document, and may simply disallow the expenditure claimed. An inefficient tax system may give rise to a disincentive to work, for example, a high earner may decide to work fewer hours, or retire earlier.

Not to be forgotten, the unregulated and untaxed shadow economy occupies a sizable chunk of the tax gap. The shadow economy consists of all those untaxed enterprises and business endeavours such as home industries, barter transactions, the taxi industry, street vendors and flea markets. The Association of Chartered Certified Accountants (ACCA), in 2017, published estimates on the global shadow economy in a report titled “Emerging from the shadows”. ACCA estimated the South African shadow economy at 23.33% of GDP in 2017 (2016 – 23.29%). Using a nominal GDP for 2016 of R4.327 trillion, this amounts to an untaxed amount of R1 trillion.

Lastly, there is the much mentioned, but relatively untouched “industry” of illicit financial flows. This, of course, is that dark underground economy of cigarette smuggling, drug money, money laundering, illicit arms trade, and the sex trade. Sadly, some tax officials confuse this with transfer pricing and trade mispricing, which is something completely different.

In conclusion, must one get tied up in regression analysis to explain a drop in tax revenues? We can surely agree that tax buoyancy does have relevance, actual tax revenues have declined, but there is a tax gap. Sars should buck up and go after the real tax criminals.

Brought to you by Moneyweb