We can now proceed with some stylized facts

related first to our components and then to the index overall.

We'll consider components one by one,

and then move to the overall index.

So, if we look at the first components,

the most striking fact is that the share of direct connections

between any country has been declining over the 2006-2015 period.

This share where it was close to 20 percent in 2006,

and it is now equal to 18.5 percent more or less.

Another striking fact is that correspondingly,

the share of connections involving

two or three transhipments has increased quite significantly over the same period.

It was close to 13 percent in 2006,

and it is now above 18 percent.

So, we're going to keep that in mind for a while,

and now move to the second component,

and there we see a bit of different picture.

So the second component,

remember is the number of common direct connections.

So, the connections which are common to any countries in any pair in the sample.

And there, whether we look at the average or dominion figures,

we see an improvement.

The number of common direct connection has moved from

five as far as the median is concerned into sudden six to seven in 2015,

and for the average,

we move from eight to 11.

But an important fact that must be noticed is that

the standard deviation of this component has increased meaning,

that the dispersion of the distribution of the values has increased over the same period,

and that would be an important fact to actually interpret some further results later on.

The third component, remember which is an indicator,

which is a combination of the two centrality indicators related to each of the countries,

this component has also seen some improvement.

So, the number of direct connections has on

average increased over the 2006-2015 period.

So, we moved from 24 as an average in 2006 to 32 in 2015.

But there once again,

standard deviation has also increased.

So, there is an idea that, well,

we get a higher value on average and perhaps also at the median,

but at the same time,

we get a higher dispersion.

So, if we move to the last two components,

which relates more to the competition perhaps that

characterizes specific bilateral trade connection,

there we also have data.

Well, the number of companies active on the route has increased on average,

but the maximum number has decreased.

We find the standard deviation,

which has remained more or less the same over the whole period.

So, we do have on average more companies which are active on any single routes,

but the maximum number of companies has decreased.

As far as the size of the vessels active on those specific maritime routes,

well, the size has increased,

and this is already something that was put forward

by Yang when talking about the LSBCI components.

Okay. But there also,

we have seen the tendency of an increase of maximum value.

I remember that Yang mentioned the fact that

the maximum size of the vessel now deployed on

maritime routes has significantly increased over the last five to 10 years.

If we move now keeping the evolution

of the variance indices or the various components of

the index in mind to overall figures,

we see that both the mean and the median of

our LSBCI values has been increasing over the whole period.

So, we were on average of values around 0.22 in 2006,

and we are in 2015 around values close to a 0.25.

As to the median,

we move from 0.20 to 0.23.

But there, so we have seen

an overall improvement in the connectivity of countries around the world.

Remember that our index can be compared across time not only across countries,

so we can deduce from the evolution of average values

that connectivity has improved over the whole period.

But at the same time, we also see that

the standard deviation of connectivity has increased.

So, we may be faced with a phenomenon

where a smaller number of country is getting more connected.

So, the half periphery structure has been sharpened over the last 10 years.

And that's exactly what our stylized facts suggest.

And that explains why also the share of

direct connection has been decreasing over that period of time,

while it's standard deviation has been increasing.

So, once again, we may have fewer countries which

are better connected than they were 10 years before,

and some others which are getting more and more into the periphery.

We can represent all the story I've been talking about with some graphs here,

and the story with correspond to the fact that we have a higher mass of

our distribution which has moved a bit to the left,

while the right and tail has been extended compare in 2015,

compared to the same tail in 2006.

So, we have a more dispersed distribution of value of our LSBCI

with a higher concentration around the average and the median values.

Obviously, we have been talking about components and the index overall,

but we can look at country specific experiences.

And for instance, we can pick up,

I should have said country per specific experiences.

And we, for instance,

can pick up the top 10 country pairs.

So, we have almost 6,000 country pairs in the country,

and we will just focus on the top 10, okay?

So, may not be a fully representative of what has been going

on for the rest of the country pairs in December, but at least,

we can retrieve some stylized facts,

and I'll leave you the rest of the distribution

as the observation and analysis of the rest of the distribution as an exercise.

So, if we look at the year 2006, well,

that was essentially dominated by Western European countries such as Great Britain,

Belgium, and the Netherlands,

also Germany was there.

2010, what happened is that Hong Kong and China have been moving up the ranking,

and in 2015, well,

the predominance of European countries is back again.

But if we look at the relative performance of country pairs,

the one which has improved most of that period is

the Hong Kong and mainland China country pair.

We also taking from there,

a look at the very bottom of our ranking,

and we'll find essentially small islands and

vulnerable states populating the country pairs in the bottom 10.

Notice that we also have countries like Albania,

which are country highly relying on some other harbour or suburb of some regions.

In the case of Albania,

Italy would be the harbour or suburb of the region.

So, most of it's connections go through Italy,

and that could explain why it ends up very low in the ranking.

We can use our bilateral data to construct

some actually unilateral indicator of connectivity in the wake of the LSBCI,

and we will find that well,

the countries populating the top of the ranking,

here we are only looking at the top 25 countries,

are also the countries that populate the top 10 country pairs.

So, we find that China,

Great Britain, Hong Kong,

the US are appearing here where they are not part of any top 10 country pair,

but they are obviously as a country taken alone,

big connected country, any other country exports to the US.

So that's why they are so high in the country level ranking.

These rankings and the results for the LSBCI are pretty much

consistent with the results of the LSBCI presented previously.

So, the bilateral connectivity index is fully coherent

with the results that would be generated by the LSBCI in the same type of environment.