In 1994, Denis S. Karnosky, Ph.D. and Brian D. Singer, CFA published a monograph entitled “Global Asset Management and Performance Attribution” (KS). They presented the idea that – due to the arbitrage known as interest rate parity – some contribution to the total return of a multi-currency portfolio is known, and ‘baked into’ a foreign-currency investment at the time it is made. Because this contribution is knowable and hedge-able, it should accrue to the currency market allocations of the manager – whether or not these are ultimately hedged.
While hugely influential among asset managers globally, these ideas remain largely unexploited in current performance practice. Though KS fully detail an attribution approach based on an expansion of the Brinson-Fachler method, and though this method is implemented in several commercially available performance systems – it is rarely adopted in the field.
We posit this circumstance to have arisen from several causes: misalignments between the paper’s formulation and practical investing reality, as well as inaccurate readings of, and consequently flawed implementations of the attribution method it sets out. We explore these causes in detail, and how they contribute to attribution results that fail to explain portfolio performance, obscuring the otherwise substantial value of KS’ central premise.
Finally, we develop a re-statement of KS that addresses those issues, producing an accurate decomposition of multi-currency effects that precisely explains the portfolio’s performance, while preserving the original paper’s essential insight. We go further to generalize this method and demonstrate its applicability to any investment attribution methodology.
by Mark R. David, CFA, Director of Performance, Risk and Analytics
Read Full Article