“actual quantitative proxy indicators, seasonality analysis, econometric projections, and a financial programming model. The report used the counterfactual methodology to estimate the economic loss, where the “crisis scenario” is the estimation of the actual GDP, and the “continuing scenario” is the projection of GDP during the same period using past historical data to project what would have been likely had the crisis not occurred.”

Even thought the SCPR is a fascinating report in many ways, it is not very specific about the methodology (something that can be easily accommodated in future reports by a more detailed methodology section). At the same time, there are few other options for macroeconomic data available for Syria, as the World Development Indicators (WDI) series run only until 2007, and the IMF’s World Economic Outlook (WEO) database only until 2010. Two exceptions are the Conference Board’s Total Economy Database, and the World Bank’s Global Economic Prospects (GEP) annual publication series, which both publish real GDP growth rates for Syria.

With actual GDP of Syria, we can document current output during the crisis compared to how it did before the crisis, say in 2010. But as Syria’s economy was growing before it started, such a measure would understate the loss, and a more accurate estimate of the real loss would take into account what Syria’s GDP would have been in a counterfactual, non-crisis, scenario. For the purpose of this blog posts I will use pre-crisis forecasts by SCPR and IMF as estimates for such scenarios.

As for longer-horizon forecasts, the IMF (at least in published work) appears to have given up trying to forecast Syria’s economy sometime in 2012 due to “the uncertain political situation”. But as the previous years forecasts are over a six-year horizon, the 2010 forecast, done before the crisis started, runs through 2014 and can thus be used as an alternative “counterfactual” scenario to that constructed by the SCPR. This does not necessarily mean that the difference between the observed and the forecasted series represent the causal effect of the Syrian crisis (a composite of the Arab Spring, the civil war, and whatever the Western policy in the country is), but in absence of better identification strategies it offers meaningful alternatives of the Syrian economy as perceived before 2011.

I thus have two types of GDP series available; 1) estimates of actual GDP and 2) forecasts of GDP from before the conflict (IMF) or as-if the conflict hadn’t occurred from SCPR. Among the two, SCPR forecasts real GDP growth at 6.6 percent per year 2010-2014, with the IMF’s World Economic Outlook (WEO) in 2010 forecasting 5.6 percent growth per year for the same period. Both sources forecast Syrian GDP growth above the preceding decade, which was 4.8 percent. As a more conservative candidate I extrapolate past GDP as a third alternative for a non-conflict scenario. (I have no prejudice as to whether the IMF is better or worse than the SCPR one, but what is clear is that both expected Syrian GDP to perform better than it had in the recent past.This is in itself quite interesting, but outside the scope of this post)

For both the series on actual GDP estimates as well as the forecasts, these are undoubtedly fraught with various measurement problems that could fit in (and perhaps deserves) a blog post on its own. Starting off with the IMF’s 2009 Article IV consultation for Syria already complaining about the quality of data and reporting standards, add to that also the question of how reliable statistics for the country is to be collected in the midst of a civil war, when the government controls just half its territory.  Moreover, the size of the informal or black market could increase in wartime, and if Syrian government statistics are forced to omit from recording output occurring in rebel-held areas, this could overestimate GDP losses – that is, if the relevant constituency for data collection remains the Syrian state borders as they were in 2010 (and even if we were more interested in just the area controlled by the Syrian government, it would still be unclear whether we’re measuring lower GDP losses as a result of losing territory as opposed to lower output for a given territory because of the crisis.) Neither the Conference Board or the SCPR estimates divulge whether they are estimates for the entire 2010-border state of Syria or exclusively for areas controlled by the government.

The degree to which Syria’s GDP losses could be overestimated would be greater depending on how much output is produced in rebel-held areas. It may be fair to assume that the Syrian government remains in control of the wealthier parts of the country (for example, the government controls a larger share of the population than the share of territory). Notwithstanding, the government’s loss of some oil fields to rebels could be a sign that there is some output unaccounted for, even though any oil output by rebels using Syrian oil fields is likely to receive lower prices (perhaps as sales would probably have to involve smuggling), suffer from lack of skilled workers and – especially if oil fields were severely contested – these could be places with significant destruction of the capital stock and under risk of infighting between rebel factions. But even if those parts of Syria not controlled by the government constitutes more of the economic periphery, or if those resources become significantly less productive, these potential limits in the quality of statistics collection need to be kept in mind.

These considerations aside, the below figure shows the relevant series: the historical IMF series of Syria’s GDP from 1990-2010; the estimates for Syria’s more recent actual GDP from SCPR (solid blue), the Conference Board (solid red line), and the World Bank (solid green line); and the forecasts for Syria’s GDP from SCPR (dashed blue line), the IMF (dashed red line), as well as the trend from the last ten years of the historical IMF series extrapolated to 2014 (dashed gray line). All series are in constant Syrian Pounds, or local currency units (LCU).