Tyranny of numbers

The anniversary of the Islamic Revolution 40 years ago this month coincided with the deepest economic crisis Iran has experienced since the war with Iraq in the 1980s.  As top Trump administration officials, who wished the crisis on ordinary Iranians in the hope of enlisting their help in regime change, excitement among the Iranian opposition abroad is palpable.   The occasion has also stimulated discussion of success and failure of the revolution concerning a wide range of issues and metrics.  Much of the discussion involved comparison of living standards in Iran between now and in the 1970s (read my own comparison in Project Syndicate here.)

On the whole, the pieces that I read (or watched) left much to be desired. They generally suffered from bad use of numbers and metrics, resulting in misleading conclusions.  Demands of my day job (teaching economics) did not allow me to provide timely commentaries, but now I think I should get back to the topic.  After all, an important purpose of this blog is to use numbers to shine a light on important issues concerning Iran’s economy, like living standards comparisons.  I plan to write a few short blogs on this topic, and start here with gold as a benchmark for measuring change in living standards (spoiler alert — it fails miserably). 

A dubious gold standard

As before-after comparisons go, how much gold you can buy now and then offers a more edifying standard than the price of cucumbers, but it is more misleading.  Prices of assets are notoriously volatile and therefore should not be used for real income comparisons. You can find several Iranian sites like this one (link in Persian) with such comparisons, but I will discuss the most visible one, the news analyses of the BBC Persian service.  BBC Persian offers many informative programs and generally upholds the high standards of its parent news service, BBC World News. But in this particular cases has erred  badly.  A report which aired a couple weeks ago asked how much gold a teacher would buy now and in 1974.  This year is perhaps not the best choice for a before-after comparison as Iran is in the middle of an economic shock engineered by the US and not the result of the normal operations of the Iranian economy (which is quite poor on its own). Nor is 1974 good as a start year because Iran received nearly four times as much oil revenues per person — a gift from the rest of the world –in 1975 as it does now.  Taking 1972 and 2017 would have offered a more informative comparison, but this is a secondary point.

So, how much gold could a teacher’s salary buy then and now? The BBC news clip (also available here) offers the following comparison (see image below captured at about 3 min. in):

The infographics shows that a teacher’s salary in 1974, reasonably assumed to be 7500 rials per month, could buy 18 grams of gold compared to one-third as much today (assuming a salary of 20 million rials). The graph actually compares how much salary it would take to buy the same amount of gold.

Now let’s apply the same metric to teacher salaries in the Unites States. The Digest of Educational Statistics puts the average teacher salary at $9258 in 1970, when the price of gold was $38.9, and $58,950 in 2017, when gold was about $1300 (down from a high of $1900 in 2011, which would make the comparison even more misleading).  The average teacher in the US could buy about 18.50 ounces (524.5 grams) of gold in 1970 and 3.78 ounces (107.2 grams in 2017), a five fold decrease!  

No one in the US would use such statistics to decry the decline of teacher’s pay in the US.  Using the US consumer price index, teacher salaries actually show a 27 percent real salary increase over the 40 year period.  Gold is clearly a misleading benchmark for comparing living standards over time, the reporter’s assurance that it is reliable because it does not depreciate and is not affected by technological change notwithstanding. 

What would be a good benchmark for comparison? Teachers obviously care about how much housing, food, health, and other useful things they can buy, not gold.  The comparison should therefore take many goods into account, a representative basket rather than any one thing like meat or chicken.  This information happens to be readily available (though perhaps not at the fingertip of a reporter unfamiliar with economics).  Dozens of economists in universities at Penn, UC Davis, and Groningen, to name the top providers, and at the World Bank and the IMF, produce GDP data to make such comparisons possible.  Unfortunately, most journalists are not aware of these data. 

The International Comparison Project ICP) at Penn, which later migrated to the World Bank and to other universities, originated the Purchasing Power Parity (PPP) calculation of GDP per capita that does just this. The PPP series measure the value of the largest bundle of goods and services produced in a country (what we call the GDP) in different countries and at different points in time using uniform  prices in the United States.  This is the type of data (from the Maddison Project Database 2018) that I used in my PS piece. They show that PPP GDP per capita in Iran was $8,492 in 1974 and $15,529 in 2016 (and probably unchanged in 2018/19), an 80 percent increase.  More on different measures of the GDP later.