Coronavirus disease 2019 (COVID-19) became the main concern in numerous fields in the globe.1 Benford’s law (BL) is mathematical theorem used as a detection technique to reveal accounting fraud,2 election fraud,3 in manipulated health data4 and even social media bots.5 Several studies have already applied BL to detect misreported data of COVID-19.6 Recently, the governmental statistics agency in Russia (ROSSTAT) announced that the reported death cases in the country are much higher than the official figures. Previous studies confirmed that Russia has indeed manipulated figures based on BL.7 Succinctly, this law dictates that there is a certain frequency of first digit distribution that occur in the natural system where digit 1 occurs more often than the subsequent digits (i.e. 2–9). Many good-to-fit tests can be applied to check if the observed distribution deviate from the theoretical distribution frequency.
This study focuses on the former Soviet states to determine if there is also misreported data. We apply BL using three good-to-fit tests (chi-square > 17.54, Kuiper > 1.75 for α = 0.05 and MAD > 0.015—marginally acceptable conformity) on the daily reported cases, daily reported death, cumulative cases and cumulative death in 14 former Soviet state (we excluded Turkmenistan because they do not release any information).
In results, we found two interesting outcomes. First, all the countries showed misreported data in the daily death counts. Second, former Soviet states which transitioned into full democracy (Lithuania, Estonia and Latvia) showed no data manipulation in the reported cases; whereas the most undemocratic regimes (i.e. Belarus, Tajikistan and Russia) showed the highest deviation from BL. Regarding the first outcome, there has been numerous media reports indicating that patients with chronical diseases who dies from COVID-19 are not registered under COVID-19 deaths. This practice was mainly noted in Belarus and Russia. We assume that it’s also a practice conducted in the other former Soviet states because of the centralization of healthcare in the past. As for the second outcome, we anticipate that when a country establishes free press and have an accountable government, there will be more transparency in the release of accurate information related to COVID-198. Figure 1 illustrates the difference in the distribution between democratic and undemocratic state.
Fig. 1.
First digit distribution of the number of confirmed cases in Russia (left) and in Latvia.
References
- 1. Georgiou GP, Georgiou C, Kilani A. How the language we speak determines the transmission of COVID-19. Ir J Med Sci 2021. Published online. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2. Nigrini MJ. Benford's Law: Applications for Forensic Accounting, Auditing, and Fraud Detection. New Jersey: Wiley Corporate F&A, 2012. [Google Scholar]
- 3. Deckert J, Myagkov M, Ordeshook PC. Benford's law and the detection of election fraud. Polit Anal 2011;19:245–68. [Google Scholar]
- 4. Crocetti1 E, Randi G. Using the Benford’s law as a first step to assess the quality of the cancer registry data. Front Public Health 2016;4. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5. Golbeck J. Benford’s law can detect malicious social bots. First Monday 2018;24. doi: 10.5210/fm.v24i8.10163. [DOI] [Google Scholar]
- 6. Koch C, Okamura K. Benford's law and COVID-19 reporting. Econ Lett 2020;196:109573. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7. Kilani A, Georgiou G. Countries with potential data misreport based on Benford’s law. J Public Health 2021. Published online. [DOI] [PubMed] [Google Scholar]
- 8. Kilani A. Authoritarian regimes have a propensity to manipulate their COVID-19 data based on a statistical analysis of Benford’s law. Commonw Comp Polit 2021;59. [Google Scholar]