My post at The Long Room, Financial Times Alphaville
Huge losses from gold holdings among central banks are a reminder that international reserves must be properly diversified, otherwise not serving its capital preservation purpose which is supposed they are there for.
Today´s news from the Swiss National Bank (SNB) said it expected to report a USD 9.9 billion loss in 2013, a year when the precious metal price fell 28%. Financially, losses are equivalent to less than 2% of total international reserves however; a tangible consequence is the SNB will not be able to distribute dividends to the Swiss Confederation and regional cantons.
It could have been worse. A plan contained in a popular referendum that would require it to keep at least 20% of its assets in gold has not been passed yet. In fact, Swiss authorities are suggesting a no-vote on that. Gold holdings are at a lower, but still risky, 8.3% of the central bank reserves.
Indeed, there may be good reasons to have gold as a reserve asset: the lack of credit risk, long term store of value and safe heaven behavior are key features for central banks. More so, after 2008. The question is of course, not ´if´ but ´how much´ to hold.
But in the ´how much´ issue, the consequences of overdoing it may be terrible. Venezuela may turn the Swiss case as a minor anecdote in this respect. Its later government decided the central bank to move away drastically from USD holding into gold, ending up with 68.9% of less liquid, loss making international reserves. Therefore, even if sound macroeconomic management were in place in Venezuela-(which of course, is not)- the sole lack of diversification of its international reserves would have cost Venezuela USD 4.9 billion last year, that is about 20% of its portfolio.
Bottom line: Switzerland and Venezuela may have sidereal differences but its central banks are not free from the adverse effects of asset missallocation of its international reserves. Indeed, a new lesson in humility to investment policy authorities worldwide.
Traditional portfolio optimisation may be evolving to allow for more dynamics in parameter estimation but the basic fact remains: a fairly high degree of asset diversification is both fundamental and sound as a guiding principle for risk management. Portfolio concentration is undesirable and even a naive [1/N] rule is a perfectly robust heuristic when, as it is usually the case, the assumed stability of asset returns and correlations in Markowitz-type models does not make sense.