The Moorish Wanderer

Ramadanomics

Posted in Dismal Economics, Flash News, Moroccan Politics & Economics, Morocco, Read & Heard by Zouhair ABH on August 19, 2012

Last day of the fasting month. Can’t find anything spiritual about it, other than a sharp increase of overt religious behaviour. Incidentally:

عيد فطر سعيد للجميع

I was browsing through the New Statesman twitter feed (great newspaper by the way) when I stumbled upon a great article on Ramadan and its impact on productivity, GDP and other macroeconomic aggregates.

You see, one of the big ideas about Islam as a major monotheist religion, is its supposedly ready-made “الإسلام دين الفطرة، صالح لكل زمان ومكان” its fitness is timeless, omniscient and ubiquitous ; in economics, in particular, Islamic banking, with the big money it raises annually, seems to point to a revival of some Islamic Capitalism. I remain however much more sceptical, because the basic model (for academics anyway) in Arrow & Debreu provides a general setting for risk sharing and market completeness, which means all the trouble Hallal financial products go through to avoid paying interest end up useless: in technical terms, the stochastic discounting factor is still there. Enf of story. As for the supposedly eminent benefits of Ramadan, it seems our Moroccan society, through its behaviour during daylight and after sunset, points to an unhealthy cross-over between mass consumption keynesianism and an over-indulging economy. Hardly the stuff of austere Muslims.

Back to The New Statesman story, I wonder how Ramadan affects the national economy, not just market volatility. The trouble is, monthly – or even quarterly- GDP data is usually hard, if not impossible to find. We have to find a device to generate it; but for the time being, let us just focus on the immediately available daily data, the stock market, and then work our way up the economy to assess the ‘Ramadan effect’. Let us before just list some (obvious ?) assumptions about it:

– Sharp increase in household and public sector consumption and low productivity: variations in household consumption are understandable, the public sector on the other hand, has a different set of dynamics to it, since public servants are paid the same for fewer worked hours; finally, lower productivity can be observed everywhere.

– Contrary to the article on the Egyptian Stock Market, I posit market volatility decreases, simply because volume and price turnouts are low.

– Balance deficit is higher than usual: we would tend to import a lot more than usual during or before Ramadan.

MASI Stock Index tends to perform little growth during the Ramadan Month: annual growth is .03%, while it grows .05% in normal days.

Let us start with the stock market from the early 1990s (available data from the Casablanca Stock Exchange trace it as far back as 1992) with 5,140 days, including about 599 Ramadan days. We first observe daily trades tend to generate a modest profit in non-Radman days, up to .11%. On the other hand, Ramadan days observe either quasi-stable index value, or more often than not, small index decreases, as high as .03% in absolute value.

Overall, the differences in daily growth between Ramadan and Non Ramadan trades is statistically significant, and because of the sample size, it would be safe to validate the earlier assumption. On the other hand, we should state that respective standard deviations are not statistically different from one and another, which further buttress the claim not only the stock exchange does not do as well as normal day as during Ramadan, but more importantly, volatility does not increase – contrary to the findings of the New Statesman article on the Egyptian stock exchange. as far as the MASI goes, Ramadan is just a set of 30-odd slow days, with less money to make, and thus higher risk to run into losses.

What about the real economy? Though daily or monthly data is unavailable, we can use quarterly data to proceed with similar analysis: if the assumption holds, we should expect quarters including Ramadan to be significantly different from the remaining 3 others: differences need to be accounted for in GDP, Consumption and Government Expenditure (Inflation could be included as well).

(Note: Statistical results are either t-tested or sample-based variance comparison tests, and these can be found further down the post)

We consider hard data from 1989:I to 2011:IV to check these assumptions: because of the 11-days rule-of-thumb to match Christian and Islamic calendars, we assume Ramadan belongs to the quarter were most of its days are located; alternatively, we could also look at two quarters instead of one, but that would dilute an already blurry picture (it might also explain why the second graph does not provide equal distances between quarter)

Two observations can made:

1/ Inflation increases during or shortly after Ramadan:over the observed 92 quarters, Ramadan-linked inflation tends to be significantly higher than either aggregate or non-Ramadan-linked quarters. Quarterly inflation during Ramadan tends to increase 2.5 times that of normal days, and about twice as much as the aggregate CPI; furthermore, because Ramadan and Non-Ramadan inflation exhibit essentially the same levels of volatility, the Ramadan effect on CPI is undeniable: it contributes on average .77 percentage point to regular inflation. This is robust evidence for price increases during Ramadan. The second point deals with the source of this CPI-related inflation, in this case household and public sector consumption.

Keynesian month indeed: CPI inflation and Consumer-led growth are stronger during Ramadan (dark Blue for GDP, light blue 1989 base year CPI)

2/ Consumption-based GDP components do a lot better than overall output: Government expenditure tends to be a lot more volatile than (and grows at higher pace) GDP. Contrary to other aggregates, there is a pitfall to be aware of; government expenditure tends to increase at the end of the year (departmental budgets need to be fully used to get the next annual budget) but even with that bias, we observed government expenditure increases during the Ramadan month, although at a steadier pace, when compared to aggregate and non-Ramadan quarters.

As for Household Consumption, results are similar to those of GDP, with an additional effect: Consumption typically increases a couple of months before the Ramadan quarter, and is a lot more significant when both compared to aggregate and non-Ramadan household consumption. Though there is need to provide an adequate framework model to link them, we can at least assert Ramadan inflation is tightly linked to a higher than usual levels of consumption for both public and private agents.

I quipped early Ramadan it was a Keynesian month, turns out I was right on the Consumption and Price Level:

So there goes the statistical evidence Ramadan in Morocco is no spiritual month: productivity is low in the economy as a whole, consumption is high (aren’t we supposed to fast out of sympathy for the hungry?) and if anything, we are losing anything between 3.1% and 3.2% in annual productivity per effective worker. In the aggregate economy, Ramadan means productivity takes a (roughly) 28% to 40% dive. Well, the figure seems a little far-fetched, and computed on quarterly basis, but it would be fair if we posit the overly optimistic assumption of productivity as a linear combination of worked hours. Either way, the Moroccan economy is brought to a productivity near stand-still every quarter.

The last assumption has to do with the trade balance: over the past decade (2005-2010) in theory, monthly imports are supposed to be more or less uniformly distributed, i.e. Morocco receives about 8.3% of its annual import each month. We observe however this is not the case; in fact, we tend to import less than monthly average on Ramadan, but more than make up for it in the month before, up to 20% of annual imports are concentrated on two months (Ramadan and the month before).

Variance ratio test MASI
------------------------------------------------------------------------------
Variable |     Obs        Mean    Std. Err.   Std. Dev.   [95% Conf. Interval]
---------+--------------------------------------------------------------------
g_rama~n |     437    .0003361    .0003469     .007251   -.0003456    .0010179
 g_non_r |    4702    .0004557    .0000995      .00682    .0002608    .0006507
---------+--------------------------------------------------------------------
combined |    5139    .0004456    .0000957     .006857     .000258    .0006331
------------------------------------------------------------------------------
    ratio = sd(g_ramadan) / sd(g_non_r)                           f =   1.1304
Ho: ratio = 1                                   degrees of freedom = 436, 4701
---------------------------------------------------------------------------
    Ha: ratio < 1               Ha: ratio != 1                 Ha: ratio > 1
  Pr(F < f) = 0.9625         2*Pr(F > f) = 0.0751           Pr(F > f) = 0.0375

Variance ratio test IPC/CPI
------------------------------------------------------------------------------
Variable |     Obs        Mean    Std. Err.   Std. Dev.   [95% Conf. Interval]
---------+--------------------------------------------------------------------
icv_nonr |      69    .0052085    .0012316    .0102307    .0027508    .0076661
icv_Ra~n |      23    .0129398    .0021269    .0102004    .0085289    .0173508
---------+--------------------------------------------------------------------
combined |      92    .0071413    .0011166    .0107097    .0049234    .0093592
------------------------------------------------------------------------------
    ratio = sd(icv_nonr) / sd(icv_Ramadan)                        f =   1.0060
Ho: ratio = 1                                    degrees of freedom =   68, 22
---------------------------------------------------------------------------
    Ha: ratio < 1               Ha: ratio != 1                 Ha: ratio > 1
  Pr(F < f) = 0.4831         2*Pr(F < f) = 0.9662           Pr(F > f) = 0.5169

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s

%d bloggers like this: