After unilaterally severing the link between gold and the dollar in August 1971, as "friendly" countries like France and the U.K. insisted on exchanging their Treasuries for gold, the U.S. retained the most monetary gold, with about 8,100 tons. This is more than the second largest holder, Germany, with about 3,400 tons.
As of the end of last year, China had the sixth largest, if one were to include the IMF in the rankings. China's gold holdings were a little more than 1,000 tons (street value ~$62.5 billion). Recall China has more than $3.2 trillion in reserves. If it were to double its gold holdings, would that really amount to a meaningful diversification of its reserves away from dollars?
Global currency reserves are valued at about $10.2 trillion. Central banks' monetary gold holdings are roughly 31,000 tons. The value of that gold is about $2 trillion. Simple back-of-the-envelop calculations suggest that it would require a 4-fold increase in the price of gold in order to bring the value of monetary gold to the value of the currency reserves.
As a thought experiment, we can ask what price of gold is needed to back not only U.S. money supply (given that the last gold link was unilaterally broken) but also, say, the eurozone's and Japan's money supply. It appears it would take at least a 10-fold increase in gold prices. Less than that would risk triggering a depression and deflation.
The U.S. and many European countries' gold holdings account for around 3/4 of their total reserves (currencies plus gold). The BRICs' holdings are considerably lower. China's gold holdings account for less than 2% of its reserves. Russia and India's gold holdings account for about 10% of their reserves. Brazil holds practically no monetary gold.
The fact that the possibility of a return to a gold standard is so remote does not mean that countries will abandon gold purchases. Central banks were net sellers of gold between 1990 and 2010. This year will likely be the third year of net purchases.
However, the amounts of money seem so modest compared with the currency reserves that diversification away from paper money (dollars?) does not seem a very satisfying explanation. Status/prestige of rising powers seem to be a more compelling explanation than moving to any sort of efficient frontier of portfolio diversification.