By the Numbers: Estimates for The Missing Cryptoqueen

How much did UK investors put into OneCoin?

This article looks at the assistance I provided to the BBC Sounds programme, The Missing Cryptoqueen. Due to legal concerns, I cannot publish the underlying data-sets.

The Sale of OneCoin

OneCoin is a fraudulent marketing enterprise, posing in the guise of a crypto-currency. The company was founded by Dr Ruja Ignatova, who disappeared in 2017.

The story of OneCoin is the focus of Jamie Bartlett’s programme for BBC Sounds: The Missing Cryptoqueen. The programme is produced by Georgia Catt.

Where is Dr Ruja? (Image: BBC)

I should note that my accounting experience mostly relates to an accounting assistant role I held back in 2004–05. This work involves spreadsheet calculations, to verify what the journalists had done. These estimates were used in the BBC’s marketing and articles regarding the programme.

There were three different sets of data provided:

  • monthly sales micro-data from August 2014 to June 2015 (with September 2014 missing) and weekly sales micro-data from 1st January to 5th June 2016 (with 18th — 31st January 2016 missing);
  • a sales dashboard from August 2014 to July 2017, with a month-to-date figure and projection for the incomplete month of August 2017, broken down by region;
  • a Profit & Loss report from September 2014 to March 2017, with projections from April 2017 to December 2018.

All values are given in euros (€), and are treated in cash terms throughout — with no adjustment for inflation.

Whilst the monthly sales reports align with the sales dashboard, there are major differences with those reports and the Profit & Loss documentation. Between October 2014 to March 2017, the sales dashboard say there were €4.2bn of sales, whilst their P&L report suggested revenues of €4.5bn.

These differences vary widely from month to month, and elevated at the start of 2016: from €0.3m in December 2015 to €63.2m in January 2016.

There is no accompanying documentation to explain these differences: an enterprise of this nature may not be adhering to the highest standards of accounting procedure. These figures may also only provide a partial view of the OneCoin operation.

OneCoin across the world

A person may have multiple accounts, and accounts may make multiple investments. Consequently, it is difficult to estimate the number of people who invested in OneCoin — despite clear intrigue in that figure.

The figures here will refer to the spotted collection of sales reports: covering August 2014 to June 2015 (without September 2014), plus 1st January to 5th June 2016 (without 18th — 31st January missing).

The first recorded UK sale was in December 2014. Across the sales reports, there were around 24,000 investments from the UK, totally €30.4m (2.5% of the total).

China has made, by far, the most investments — with 57.6% of the total investment value (€689.2m across the sales reports).

In the sales reports, there were €1.197bn investments made in over 200 listed areas. To clarify, the sales reports contain duplicate countries, e.g. ‘Netherlands’ and ‘Netherlands [Note 1]’ as separate entries.

Estimating uncertainty in the UK figure

It is of plain interest what the value of UK investments were.

The estimate is for the period October 2014 to June 2017 (11 quarters). We can use the real monthly values from October 2014 to June 2016. We can apply the UK share of Europe and CIS regional sales from the sales reports to the regional values in the sales dashboard.

The central estimate is that UK investors put €106m into OneCoin over this time period. As OneCoin is still operating, this is an underestimate for the total investment to the present day.

Plausibly the true value for that time lies between €101m and €109m, based on how much that share varied in the sales reports.

How was this interval constructed?

The idea is that we should treat each month as a random variable. The actual Europe and CIS value is then multipled by a random number.

That random variable is assumed to be Normally distributed with a mean approximately equal to 12.99%, and a standard deviation equal to 1.30%. These simulations then allows us to estimate a plausible interval for the true value.

Each month from July 2015 to June 2017 is simulated 10,000 times, and added to the recorded value from October 2014 to June 2015.

Two such simulations are shown here:

These are two Monte Carlo simulations.

The central 90% interval is then: €101m — €109m.

Could this be under-estimating uncertainty?

Yes. Choices matter, and different analysts could choose:

  • Wider intervals: I have used a 90% confidence interval, but 95% or 99% intervals could be constructed;
  • More deviation: Based on the limited data, I used a standard deviation of 1.3% for the UK monthly share. Higher deviation may be justifiable.
  • Different distribution: I used the Normal distribution, but other distributions with thicker tails could be chosen instead.

Using a different standard deviation of 2.1% means our plausible range (using a 95% interval) is between €98m and €113m. Alternately, the 90% interval is between €99m and €112m. Note, here, our central estimate is slightly smaller, at €105m.

Simulation error is important to recognise when quoting a single set of results.

Legal concerns mean that I cannot publish the underlying data-sets.

This blog looks at the use of statistics in Britain and beyond. It is written by RSS Statistical Ambassador and Chartered Statistician @anthonybmasters.

Get the Medium app

A button that says 'Download on the App Store', and if clicked it will lead you to the iOS App store
A button that says 'Get it on, Google Play', and if clicked it will lead you to the Google Play store