An "iceberg-type" visualisation of value chains

Background

This visualisation utilises the accounting framework that I describe in my conference paper "Structure and Length of Value Chains" (presented in Osnabruck, Germany and Beijing, China in 2017). The calculations required to build the chart below are not explicitly described in that paper, but may be easily derived. This visualisation does not appear in the said paper, and I only included it in the presentation at the Beijing workshop.

There is a simple idea behind the chart: take the output of one industry in one country and decompose it simultaneously into inputs that come from primary and intermediate suppliers and outputs that go to final and intermediate users. Then show the shares of inputs and outputs that cross one, two, three and more borders. I call this chart "iceberg-type" because it shows how output "melts" as it crosses more borders in a value chain leading to its eventual suppliers or users.

Data

The data input for this visualisation is from the 2015 edition of the OECD Inter-Country Input-Output (ICIO) tables. The tables cover 61 countries plus the rest of the world as a another country and 34 industries. The years covered are from 1995 to 2011, but I only calculate results for 2000, 2005, 2008 and 2011.

Understanding the visualisation

Choose a country, an industry and year. You will see a stacked bar chart. The horizontal axis has zero in the middle which corresponds to our country-industry of interest. The left part of the axis denotes the average number of borders the inputs have to cross before being used to produce our chosen country-industry output. So, don't be surpised that there are no negative values (only for the sake of visualisation!). The right part of the horizontal axis marks the average number of borders the outputs of our country-industry have to cross before being used in production or consumption further down the value chain. Each bar corresponds to the share of output – measured at the vertical axis – that crosses the respective number of borders on the way to the initial origins of inputs or eventual destination of outputs. Bars are split into "direct" (darker) and "indirect" (lighter) parts. "Direct" signifies that, within the borders of an economy, this share of output does not undergo any transformation: it consists of either primary inputs contributed by labour and capital, or final products for consumption or investment. Note that, to the left from zero, only the first bar has the "direct" component, because value added always orginates on the territory of the producing country and cannot be physically traded across borders (if you are confused, please see the national accounting conventions). Of course, this is not the case for final products that can be produced in one country and consumed in another one. "Indirect" signifies that this share of output is made of intermediate inputs or serves as intermediate input in further production. All bars on either left or right part of the chart add up to 100%.

You can see an animated version of the chart here.

An example (read this if you are still confused about the interpretation of the chart)

Let us experiment with the data on the German automotive industry. Choose "Germany" from country list, "Motor vehicles, trailers and semi-trailers" from the industry list and the year 2011.

On the horizontal axis, the interval between 0 and 1 in both directions corresponds to the German economy where the inputs and outputs of the automotive industry are involved in production without leaving its territory. The first bar on the left ( the tallest one) shows you that 67% of the output of automotive industry originates in Germany and has never crossed any borders. This includes 27% direct component that can be recognised as the value added share of this industry. The 40% indirect component is in the form of intermediate inputs provided by all German industries – this can also be understood as indirect domestic value added. The second bar on the left, between 1 and 2, depicts 22% of inputs that have only crossed one border – a notional border betweent the exporting economy and Germany. These inputs have been entirely produced in the exporting economy (call it "first-tier supplier") and constitute indirect foreign value added. The third bar is equivalent to only 8% of inputs that have crossed borders twice – between second-tier supplier and first-tier supplier and between first-tier supplier and Germany. These inputs should also be treated as indirect foreign value added except a special case where Germany itself is the second-tier supplier (I do not address it here). And so on.

Let us look at the right part of the chart. The first bar shows that 37% of the value of automotive products ends up in Germany without leaving its territory: 24% as final products (cars and other types of vehicles) and 12% as intermediate products (parts and components, auxiliary services). The largest part, however, is ultimately destined for customers in other countries after crossing one border: the second bar is equivalent to 51%. Of these, 36% are delivered as ready-to-use vehicles (or similar final products) and 15% as part and components. Look at the third bar: 9% of German automotive industry outputs travel further down the value chain, crossing two borders. This means that the direct importing countries use 6% of German outputs to produce final automotive products for customers in other countries, and the rest 3% are processed and delivered as intermediate inputs. You can see that the outputs that "survive" until the third border crossing, melt down to only 2% of the initial output.

Technical notes

Take note that choosing China or Mexico from the country list and agriculture, mining or any manufacturing industry will not produce any visualisation because these industries are additionally disaggregated in the OECD ICIO tables into those operating in standard mode and those operating in processing exports mode. So you have to choose the country with a specific type of production, for example, "China (domestic sales only)" or "China (processing goods exports)". But the country categories "China" and "Mexico" still work if you want to visualise services value chains.
The interval between 0 and 1 on the horizontal axis corresponds to the domestic economy and the interval from 1 to 2 to partner country. However, after the second border crossing, outputs may return to the domestic economy for further processing or consumption, and inputs can in the same way originate from the domestic economy. Such circular trade along global value chains is not distinguished here, though this is possible with some modification in the underlying accounting framework.
In rare cases, you will see that the direct component of the first bar to the right from zero (dark green rectangle between 0 and 1) is negative. The reason is usually that the changes in inventories that are part of final demand may be negative and may outweigh the output of final products of a particular industry. This means that the output was not suffucient to meet final demand, and the gap was filled from inventories.

Acknowledgment

This visualisation was made in Observable wuth the D3 JavaScript library.