Tyranny of numbers

The question of accuracy of reported Covid-19 deaths is not news for most countries, but in the case of Iran it is because it has geopolitical significance. Iran’s foes are interested in it because the higher the number of deaths from Covid in Iran, the more likely Iranian leaders are to stop resisting US economic pressure. For others, who care more about the plight of ordinary Iranians, higher true fatalities signify a greater humanitarian cost of US sanctions.

But how misleading are Iran’s fatality rates from Covid? The fact that they have been skyrocketing since September (see Figures below) is a sign that if Iranian authorities have been lying, as some have claimed, they are not doing enough. Why else would the picture painted by the official statistics keep getting more grim? Turkey and Saudi Arabia, which I use for comparison in this figure, are doing better than Iran in controlling fatalities from Covid, something which, as I have noted previously, is in part due to Iran’s much weaker economic conditions.

Figure. Covid-19 deaths per day per million in Iran, Saudi Arabia and Turkey.

Source: ourworldindata.org.

I have no basis to question the Covid the data from Turkey and Saudi Arabia, though other have questioned them. Whatever is the truth about other countries, doubts about underestimation of deaths from Covid in Iran are justified (though charges of systematic manipulation of data are not). We learn this from the estimates of excess deaths during spring and summer of 2020, which are available from another official source, the quarterly reports published by Iran’s agency for births and deaths registration, sabt ahval.

These data suggest that for two consecutive quarters in 2020 while Covid was raging in Iran the number of registered deaths exceeded the average for previous years by a considerable margin. A comparison of these “excess deaths” with the official Covid deaths reported by the health ministry indicates serious undercounting. (The fact that these data are readily available on the government’s website is another indication that, if Iranian official are lying, they are doing a very bad job of it).

Registered deaths for spring and summer 2020 (20 March to 20 September 2020) in excess of the average for prior years (“excess deaths”) are more than twice the officially reported Covid deaths. We do not know if all the excess deaths are due to the coronavirus, but comparison with seven prior years and the geographic distribution of the gap between the two for last winter (much higher in Qom and Gilan where the virus spread first), is fairly convincing that they are. Table 1 lists the excess registered deaths and the Covid deaths reported by the health ministry.

Notes: The first row of the table shows the number of Covid related deaths that Iran’s health ministry has reported to the WHO during the period in question (downloaded from Ourworldindata.com). Estimated excess deaths are the number of registered deaths in a quarter minus the average for previous-year quarters since 2013).

The data for winter 2019 quarter is only for the month of Esfand (20 February to 20 March) during which the Covid deaths occurred. This allows us to compare the progression of registered deaths per month (from 4,565 in winter 2019 to 6,162 in spring 2020 and to 12,274 in summer 2020) with that of registered deaths (from 1,284 to 7,988 to 14,846). The faster growth of the registered deaths suggests that the undercounting has been rising.