Central Banks and other institutions

1 August, 2019

H1 central bank demand at highest level since becoming net buyers in 2010

  • Central bank net purchases rose 47% y-o-y to 224.4t in Q2
  • H1 2019 net purchases totalled 374.1t, the highest level since becoming net buyers in 2010
  • Poland was the largest purchaser in the quarter; reserves grew by 100t (+77%)
Tonnes Q2'18 Q2'19 YoY
Central banks & others 152.8 224.4 47%

Central bank net purchases totalled 224.4t in Q2, 47% higher y-o-y. Total net purchases for the first six months of 2019 rose to 374.1t: 57% higher y-o-y, and the highest level of y-t-d demand since central banks became net purchasers (on an annual basis) in 2010. Buying momentum has continued strongly from last year – 2018 saw the highest level of annual purchases in 50 years – and is a clear indication of the central banking community’s mind-set towards gold.


Highest level of H1 purchases since central banks became net buyers

Highest level of H1 purchases since central banks became net buyers

Sources: Metals Focus, Refinitiv GFMS, World Gold Council; Disclaimer


Ongoing global struggles remain a headache for reserve managers. Sluggishness, exacerbated by trade and geopolitical tensions, continued to cast a dark cloud over the global economy. This was reflected in volatile financial markets during the quarter. Central banks, like other investors, sought safety in gold as they looked to protect themselves in the face of many looming risks.

Central bank demand still geographically diverse. Nine central banks increased their gold reserves by at least a tonne in the first half of 2019. This continues the trend seen over recent quarters, where demand has been spread amongst a larger number of – primarily emerging market – central banks.1

Poland’s giant purchase bumped Russia into a Q2 second place. Poland, which had previously boosted its gold reserves by 25.7t in 2018, followed this up with a whopping 100t purchase in Q2. This is the highest quarterly purchase by a single central bank since India bought 200t from the IMF in November 2009.2 National Bank of Poland President Professor Adam Glapiński indicated that this purchase was strategic, with the intention of further safeguarding the financial security of the country.3

Russian gold reserves rose by 38.7t in Q2, bringing y-t-d net purchases to 94t, compared with 105.3t in H1 2018. Gold reserves stood at 2,207t at the end of June, equal to 19% of total reserves. In May, the central bank discounted its purchase price to encourage domestic producers to export more gold. Governor, Elvira Nabiullina stated: “We introduced discounts for the purchase of gold because we saw that vendors were selling gold mainly to the central bank,” and “The discount aligns domestic conditions with external ones and motivates gold producers to … export too”. This corresponds to a slight dip in purchases by the central bank towards the end of the quarter, but the bank clarified: “It wasn’t a special decision on our part to reduce gold purchases”. 4

Chinese gold reserves grew by 74t in H1. The volume of monthly net purchases increased marginally in Q2, bringing the monthly average for 2019 to 12.3t (11t in Q1). Turkey continued to increase gold reserves, up 60.6t in H1, while Kazakhstan added 24.9t to its gold reserves over the same period. Other central banks to increase their gold reserves by at least a tonne during H1 were: India (17.7t), Ecuador (10.6t), Colombia (6.1t) and the Kyrgyz Republic (2t).

The Central Bank of the Philippines announced at the end of May that it would begin purchasing gold from local small-scale miners. It is anticipated that buying could reach one million ounces a year, significantly higher than the current 20,000-30,000oz.5

Selling remains a fraction of buying. Over the first six months, declared gross sales amounted to 7.2t, just 2% of gross purchases.

But during Q2 more news emerged regarding the swap agreement involving Venezuelan gold reserves. It was reported in April that Venezuela was deemed to have defaulted on a gold swap with Citibank and Deutsche Bank. This meant the banks took control of the US$1.4bn of gold (~28t) that underpinned the swap.

Our treatment of swap agreements is that they are deemed a sale if the agreement ends in default (i.e. the gold is not returned as originally planned). Therefore, the termination of the swap agreement in Q2 is treated as a sale by Venezuela in our figures.  And this is not the first Venezuelan gold swap agreement we have recorded as a sale. In Q4 2017, a swap for approximately 45t ended in default, as did another for around 42t in Q1 2019. The precipitous decline of Venezuelan gold reserves remains closely watched by both market participants and international governments/authorities.

Elsewhere, and as we’ve seen over many recent quarters, declared gross sales by central banks are few and far between. Only four other banks reduced gold reserves by a tonne or more in H1: Uzbekistan (-3.7t), Germany (-3t), Tajikistan (-1.8t) and Mongolia (-1.7t) – with the former relating to a long-established coin-minting programme.


  1. Figures quoted for country-level activity are based on latest available data at the time of publication.

  2. China’s declared purchase of 604.3t in June 2015 has been excluded as this represents purchases over the preceding five-year period.

  3. The impact of this swap agreement on Venezuelan gold reserves has not yet been reflected in the IMF IFS data.

Important disclaimers and disclosures [+]Important disclaimers and disclosures [-]