Why FX Currency Indexes Use Trade Weights Instead of Volume

# Is this type of currency index a thing already?


# Is this type of currency index a thing already?












I was considering making some sort of free index data set that would basically attempt to estimate changes in the value of a currency against all other currencies. So I came to thinking if I took daily volume, and daily changes, summed dollar volume for all currencies except the one for the given index, then divided a given currency's estimated volume in dollars lets say for example is CHF, this gives us percent of daily volume, then we would multiply by the percent of daily volume by the change of the currency for the index, for example lets say we are calculating the USD index, we then multiply the price change for USDCHF by the percent of daily volume. Then continue for every currency against the dollar in order to calculate how much the index should move. Is this a thing or should I make it one?

## Answer by q.t.f. (score 3)

https://quant.stackexchange.com/a/27872

It is not a thing. The basic problem is that volume of currency traded is unknown. FX traded almost exclusively over the counter without centralised clearing or reporting. There are no daily volume statistics.

The thing that is most commonly done for FX indexes is trade weighting. Changes in a currency's value are measured against the value of a basket of other currencies, where each currency in the basket is weighted according to its share of the import/export trade. Probably the weighting would be adjusted annually, or in line with how often the government accounting office publishes import/export statistics.

## Answer by alex talei (score 1)

https://quant.stackexchange.com/a/27904

re: fx volume

majority of fx volume is "unknown", but so is much else in markets. there are ways of estimating volume to feed your models - just don't get overly reliant on it IMO.

check out CME T&S and a few spot venues. google Lee/Ready and Tick Rule to get ideas on how to create your own estimation models based on available/accessible data for FX.

## Answer by Phil H (score 1)

https://quant.stackexchange.com/a/28175

An index has to have a purpose. For example, the FTSE100 tracks the performance of the 100 largest capitalised companies on the LSE; it's purpose is to give the expected return of a fund of those largest companies. It is, therefore, replicable; if I wanted to get the same performance as the FTSE100, I just buy the right proportions of shares as the index contains, and keep rebalancing vs the changing portfolio and weightings.

So for a currency index, the question has to be: What is it for? If you want to know how well USD is performing, then performing against what? If I was a US investor in a collection of European companies, where I want my USD initial capital to track the movement of the currencies of those companies, then I would want a basket containing that proportion of currencies. If I was a Swiss investor with a need to maintain a fund of CNY for my imports, then I would want that in the basket.

Ultimately, I think an index which just tracks 'volume' (as measured by whom?) isn't particularly useful. Currency movements are less about how much has moved, and more about the economies and inflationary environments of the countries that issue them.

Consider the recent Brexit move. If I were a US investor into the UK, then obviously GBPUSD is going to be a large component in the risks I'm looking at. But for exporting companies, a drop in GBPUSD is helpful, as it cuts labour, rent and other local costs at a stroke. But that UK capitalisation is also reduced. So then perhaps I care more about the long term interest rates available in the UK, and the market's view on its future. So I might want a basket of Gilts, German bonds and Treasuries, each with their own currency dependencies, to reflect my interest or my risk.