In multiregional input-output analysis, monetary flows between or within economic sectors are given in a transaction matrix T. Both, the rows and the columns, have the same index (Region 1 Sector 1, Region 1 Sector 2, ... Region 2 Sector 1, Region 2 Sector 2, ... ). The diagonal of the T matrix contains the intra-sectoral monetary flows. Analagously, the diagonal of the "regional blocks" (all sectors from one region) contains the intra-regional monetary flows between sectors.

I figured that in the database with which I am working, the share of intra-regional monetary flows (all flows belong to region 1) compared to all monetary flows of a region (all flows from region 1 to region 2 or vice versa) are rather low. Let's say around 20% on average. To me (non-expert), this seems plausible. However, I would like to crosscheck my findings with "real world data".

Question: Is there an indicator which compares intra-regional monetary flows with total extra-regional monetary flows? ... in other words: How much monetary flows between sectors take place within a country and how much monetary flows between sectors of a country take place with sectors from other countries?

I came across "Trade (% of GDP)" (https://data.worldbank.org/indicator/NE.TRD.GNFS.ZS). However, 1) in input-output analysis I understand GDP to be given by the final demand vector/matrix and 2) the given figures of about 50 to 60% are completly different to my findings of around 20%.


1 Answer 1


The exports and imports in an input-output analysis should correspond to imports and exports as components of GDP so you can use those or some measures derived from them.

For example, trade as a $\%$ of GDP is the sum of exports plus imports over GDP (i.e. $\frac{E+M}{Y}$, where $E$ is export, $M$ import and $Y$ GDP/output).

Hence to calculate any comparable measure from input output table you would have to take the import and export variables from your study (your input-output table should have import row and export column), sum them together and then divide them by output.

You can also find data for $E$, $M$ and $Y$ separately and compare those to the measure you constructed. Based on the question I guess you came to $20\%$ by perhaps doing something like $\frac{E-M}{Y}$?

In addition input-output analysis is not necessary accurate and you should not be surprised to find discrepancies between what your input-output model shows and what holds in real world. Especially, if your input-output model has too many sectors small errors introduced by rounding and linearization can add up and produce results that will not be accurate.

  • $\begingroup$ Thank you for your comment! This already helps a lot. Is this data (intraregional monetary flow compared to extraregional monetary flow) recorded and published somewhere? ... for comparison $\endgroup$
    – Stücke
    Nov 29, 2020 at 10:31
  • $\begingroup$ @Stücke you are welcome. Also, these flows are exports and imports as mentioned in the first paragraph of my answer. Sorry if I did not made that clear. You can find these in any dataset that has national accounts - world bank including. $\endgroup$
    – 1muflon1
    Nov 29, 2020 at 10:33
  • $\begingroup$ @Stücke sorry, I misread the comment for intraregional statistics you can just use GDP for an area as an aggregate/summary for all monetary transaction in the region, if you would want to see sectoral breakdown you can have look at breakdowns of GDP by sectors in that area. Usually statistical offices will provide breakdown of GDP by sectors as well or even GDP put into input output table like here . $\endgroup$
    – 1muflon1
    Nov 29, 2020 at 10:51

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.