Page 20 - Invited Paper Session (IPS) - Volume 2

IPS178 Sana Antoine S. J.
regression allows us to check that, when considering the core CPI measure,
the impact of external variables is reduced.
Estimating each variable’s importance
When explanatory variables are uncorrelated, the contribution of each
variable in explaining the variations of CPI (improving R2) is unambiguous.
However, when these variables become correlated, it is no longer clear how
each one is impacting CPI. We decided to adopt the LMG (Lindeman, Merenda,
and Gold) measure for assessing the relative importance of each variable in
the linear regression model. The way this measure works is pretty simple:
adding a variable (WAIR, for example) to the regression improves the
goodness-of-fit R2 metric by a given amount (R2 jumps from 0.49 to 0.55,
improvement = 0.06). When the explanatory variables are correlated, the
amount of such an improvement (0.06) depends on which variables are already
included in the model before we added the new variable (WAIR). The LMG
measure averages these improvements over all possible orderings of
regressors. In general, the LMG measures do not add up to 1 since variations
in CPI are not perfectly explained by our variables.

3. Results
The results of the sectorial approach are presented in figure 4 and it can
be pointed out that:
 The portion of Lebanese inflation that comes from non-tradable items
such as health, housing and education represents only 26.6%. These
items are purely domestic and are not directly impacted by global
trade. The majority of other items are tradable especially food and
energy items.
 Food and non-alcoholic beverages constitute a major component of
Lebanon’s consumer price index, with the highest weight of 20%. The
food price index in Lebanon has been increasing faster than the
consumer price index indicating rising food prices. In fact, food prices
have increased by an average annual rate of 5% during the period
2010-2013 against 2% during the period 2012-2018. This comes in line
with trends in international food prices which also increased by an
average annual rate of 6% during the same period. Since Lebanon
imports most of its food demand, it is consequently highly exposed to
international food price fluctuations.
 The items that contribute the most to inflation are Energy (36.1%),
Food and non-alcoholic beverages and Housing (15.9% each). Such a
result confirms our main intuition that Food and Energy prices
fluctuations have significant repercussions on headline inflation. Even
though the share of the energy item in the CPI basket is only equal to
11.8%, it explains 36.1% of the volatility of headline CPI. If we add to

7 | I S I W S C 2 0 1 9