Research

Articles on AI and currency exposure

A reading list of editorial pieces on how machine-learning methods are applied to currency exposure inside international portfolios. Each article sets out one mechanism, the data behind it, and the model's limits. Nothing here is investment advice; nothing is sold.

Translation drag · Method

Decomposing a foreign holding into local return and FX return

A model can split the performance of an overseas position into what the asset did in its own currency and what the exchange rate contributed. This piece walks through the decomposition and why a single blended figure hides the part that matters for hedging decisions.

Hedging cost · Forward pricing

Reading the carry that is already in the forward

The interest-rate difference between two currencies is embedded in the forward price, so a hedge is never costless. Here we look at how that embedded carry is measured and when it makes an unhedged position the more honest choice.

Volatility regimes · Clustering

Marking volatility regimes that move a hedge ratio

Clustering foreign-exchange volatility surfaces periods where a static hedge ratio leaves a portfolio under- or over-protected. The article sets out how the clusters are built, what they miss, and where the method tends to fail.

Base currency · Reporting

The base currency as a moving reporting lens

Reported returns change when the currency results are measured against shifts underneath them. We examine how models account for that movement across periods, rather than freezing the base currency in place.

Questions we hear

Do you give investment advice or manage money?

No. This is an editorial publication, not an advisory firm. We do not take on clients, hold funds, place trades, or tell anyone what to buy or sell — and nothing published here should be read as a recommendation to act on any position.

Is the research free to read?

Yes — every article is open to read, with no account, paywall or sign-up. If something is unclear, the inquiry form and our email are the channel for follow-up questions.

What does AI actually do in this analysis?

The methods cluster periods of foreign-exchange volatility, estimate the carry built into forward rates, and identify which exchange-rate moves most influence a portfolio's reported result. Each article lays out the technique and where it breaks down, so the reader can assess the method on its own terms instead of copying it.

Where is the business based, and who answers inquiries?

The editorial desk sits at No. 93, Tiding Blvd., Neihu Dist., Taipei City, Taiwan. Questions sent to info@meridian-frame.digital are read and answered by the writing team, not routed to a sales queue. The business is Meridian Frame Research Co., Ltd. (Uniform Business Number (統一編號) 10433218), representative Lin Yu-Ting).