Trends in reserve management: 2020 survey results

The survey questionnaire was sent to 130 central banks in February 2020. By the end of April, replies had been received from 74 reserve managers, responsible for a total of $6.3 trillion, or 46% of the world’s total, with gold valued at market prices as of August 2019. The average holding of respondents was $84.7 billion. Breakdowns of the respondents by geography, economic development and reserve holdings can be found in the tables below.¹¹ In this survey we follow the World Bank classification of countries for region and income group, see https://datahelpdesk.worldbank.org/knowledgebase/articles/906519-world-bank-country-andlending-groups. Accessed 29 April 2020.

Which in your view are the most significant risks facing reserve managers in 2020? (Please rank the following 1–5 with 1 being the most significant risk.)

Concerns over a Chinese economic downturn and continuing low yields globally dominate reserve managers’ thinking for 2020, but their comments reflect the overwhelming impact of the Covid-19 pandemic³³ In this chapter the terms “Covid-19” and “coronavirus” are used interchangeably. on markets, policies and economies that was unfolding as they submitted replies.

A reserve manager from Europe, whose chief concern is a Chinese downturn and contagion, expressed this view succinctly and also noted wider possible ramifications. “The coronavirus spreading outside China puts global growth at risk and leads to a lower-yield environment. Geopolitical risks in the Middle East are not off the table. Uncertainty remains about the future trading relationship between the US and China, but also between the EU and the UK.” A central banker from sub-Saharan Africa offered a straightforward assessment: “Coronavirus development is the most significant risk in the market,” a view echoed by a reserve manager from the Middle East: “Our main risk is the global recession caused by the pandemic which changes the whole financial market.” A respondent from a high-income country in Europe noted internal as well as external risks: “Recent coronavirus developments pose the most significant risk to markets. In addition to the economic consequences, related health risks and possible operational breakdowns should also be considered.”

The all-encompassing nature of the pandemic meant several reserve managers declined to prioritise the risks listed in the question, but offered general views, again dominated by Covid-19. A reserve manager from Europe summed up this view: “It’s difficult to know which of these are that important at this current time while Covid-19 is around....until this is controlled nothing else matters....and just in the short term (at least) liquidity is the only game in town.” A central banker from an upper middle income country in the Caribbean was clear: “The #1 issue now will be the anticipated adverse impact of Covid-19 on the financial markets.”

Which in your view are the most significant risks facing reserve managers in 2020? (Please rank the following 1–5 with 1 being the most significant risk.)

The 28 reserve managers who selected a Chinese downturn and contagion as the most significant risk are responsible for $2.4 trillion in reserves and were dominated by central bankers from high-income countries. This view was the most prevalent among this economic group: more than 60% chose this option. This was not the case for middle income countries however where majorities saw the continuing low-yield environment as more important. Geography and the spread of Covid-19 may well be a factor here as over half of this group of 16 countries were from Africa and the Americas, which at the time of submission, had yet to be significantly impacted. The 22 who saw the continuing low-yield environment as the chief risk are responsible for $942 billion and were mostly middle-income countries; indeed this was the choice for almost half the reserve managers from this group.

Reserve managers from 24 countries placed geopolitics as first or second in significance. Of these, seven placed it first, a group of responsible for around $20 billion each, with exception of one large holder. In their comments, most made it plain that this was an extension of the Covid-19 impact. As a reserve manager from the Americas put it: “The actual environment suggests the most relevant risks are geopolitics and Coronavirus impact on the global economy.” A large reserve holder from the same region noted: “After the Covid crisis the geopolitics likely will be a hot spot again.”

The euro celebrated its 20th anniversary in 2019 and, according to the most recent data, accounts for 20% of allocated official sector reserves. How would you characterise its performance, versus its potential, as a reserve currency over the past two decades?

A lack of policy coordination, which is associated with crises, and negative rates have weighed on the euro’s performance as a reserve currency in the minds of reserve managers. In the main, they regard the euro as having underperformed in its first 20 years of existence. Fifty-seven reserves managers, responsible for $4.4 trillion in reserves were of this view, a group whose average holding of $77 billion was just below the survey average and which featured several large holders from Europe, Asia and the Americas. It included almost all the reserve managers from middle income countries and all those from low income countries.

In their comments, reserve managers typically stressed a lack of policy coordination and the negative yields of recent years as being the defining factors. This view was neatly expressed by a reserve manager from the Americas: “Participation is still low, especially if take European countries out of the sample. Negative yields and doubt about future of integration weigh on the currency.”

Two reserve managers, one from the Americas, the other a high income country from Europe, explicitly referred to geo-politics in their comments. The first noted representation on inter-governmental forums: “In light of the fact that they also have three members of the G7 countries, it would be expected that the currency would have been doing much better when compared to the US.” The second, from a eurozone country, was blunt: “One of the world’s superpowers is not pulling its weight in economic terms.” In an extended comment, a reserve manager from Europe felt the recent crises and policy actions of the euro’s central bank were unbecoming of a reserve currency and noted, also, the impact of negative yields for central banks in particular:

A greater propensity for crisis was also a theme in the comments offered by a reserve manager from the Middle East: “The euro area is more crisis prone than the US and [a] lack of uniformity in policy puts it at a disadvantage against the USD.”

A reserve manager from the Americas saw the euro as hamstrung by its fiscal set-up: “Specifically in the last decade, the lack of fiscal integration and thus of policy coordination, has limited the potential of the EUR as a reserve currency. Similarly, the low levels of interest rates and poor economic prospects, make the currency not very attractive.” For a reserve manager from Europe, but outside the eurozone, economic differences were at the heart of the problem: “Economic imbalances of the euro area countries contributed to the protracted effects of economic crisis, which resulted in a low/negative level of government bond yields on the European market and undermined the role of the euro as a reserve currency.”

In an extended comment, a reserve manager from an upper middle income country in Europe offered a nuanced assessment noting the performance of the exchange rate and inflation as positives:

Almost all respondents from Sub-Saharan Africa stressed the impact of negative rates in their comments. One reserve manager commented generally that negative rates “has been a major challenge in reserve management,” and two noted that central banks had reduced their holdings to day-to-day needs, eschewing a long-term investment commitment. “It is mostly used by many central banks to meet the euro liabilities,” said a reserve manager from East Africa, adding “The currency would have definitely performed better if interest rates were above zero as this would support capital preservation as well as enhance portfolio returns.” This view was also held by a central banker from a high income country “The potential of the euro to perform better as a reserves currency can be said to be higher than at present. The reason for this is some banks are keeping the bare minimum of the currency for operational needs to not incur increased charges from holding high amounts of the currency.”

Only eight reserve managers saw achievement beyond expectations, a group which included three members of the eurozone. This group was drawn from high or upper-middle income countries and featured single representatives of the Americas, Asia and Middle East regions, but not one from Africa. As a whole, these eight reserve managers were responsible for just over $220 billion, although five of the holdings were below $10 billion. A reserve manager from Asia felt the single currency had performed well given the bloc’s economic performance and the negative yields: “Given EU’s share in the world GDP and negative yield environment in euro area, we think 20% share in reserve currency is outperformance.”

What in your view would make the biggest difference to the attractiveness of the euro as a reserve currency? (Please rank 1–5 with 1 having the most impact. )

Overwhelmingly reserve managers see a return to higher, positive yields as having the biggest impact on attractiveness of the euro as a reserve currency. Thirty-nine reserve managers, 63% of respondents, placed this first in their ranking and a further 18% placed it second. The 39 central bankers were responsible for $1.56 trillion in reserves and included over three-quarters of respondents from lower middle income and low income countries. This choice was, by some distance, the most popular with “closer political union” second as the first choice of only nine reserve managers. “Deeper, more liquid markets” placed third but was the most popular second choice with almost half of the respondents selecting this. A benchmark issuance at the European level was the most popular third choice.

There was however variation in views across levels of economic development and geography. Among high income countries, there was noticeably more support for deeper, more liquid markets, with one in five ranking this first; just under half saw positive yields as having the most impact. Reserve managers in upper middle income countries were more inclined towards political union: one in five placed this first. This was a popular choice among large reserve holders: almost 40% of respondents responsible for more than $50 billion placed this first, as a large holder from Europe noted: “A closer political and/ or banking union might provide more stability and improve the attractiveness of the EUR.” Among lower middle income countries, reserve managers overwhelmingly see higher yields as the most important factor. Outside o positive rates, which was the most popular choice, for reserve managers from Asia a benchmark issuance figured prominently whereas is in Europe more weight was given to deeper more liquid markets. For African central bankers banking union was the second choice.

What in your view would make the biggest difference to the attractiveness of the euro as a reserve currency? (Please rank 1–5 with 1 having the most impact.)

A comment from a reserve manager succinctly summed up the overall sentiment: “As long as the European assets have low levels of returns, it is difficult to allocate a substantially higher exposure to the currency.” For a reserve manager from a central bank in Africa, positive yields would precipitate a return to the market: “ Have a positive nominal return: it’s a basic premise (among others) for us to consider a reserve currency to be included in our portfolio. We do agree that if the euro could have positive interest rates, at least we would have some explorations in the money market.” From within the eurozone, a reserve manager felt political commitment could have the biggest impact: “Closer political union would reinforce security and confidence in the euro as a reserve currency. More liquid markets and higher (non-negative) yields always help in the attractiveness of the currency for reserve management purposes.”

An interesting perspective was offered by reserve managers from central banks in Europe but outside the eurozone, as investors and would-be participants. For one, economic underperformance was the root of the problem:

For a reserve manager from a neighbouring country, it was more a function of the state of the markets: “The negative yield environment combined with the fragmented market structure and significantly decreased liquidity in several asset classes make the euro less attractive as a reserve currency compared to its peers.” In contrast, a large holder from Europe was satisfied with the condition of the markets and, interestingly, felt that a benchmark issuance could confuse: “Negative yields seem to be one of the major factors undermining the role of the euro as a reserve currency as, despite fragmentation, the eurozone financial markets are mature and liquid. Paradoxically European benchmark issuances could blur their transparency and make it more difficult to assess credit risk.”

Has the low-yield environment, notably in major reserve currencies, changed reserve management policy and practices at your central bank?

The low-yield environment has had significant and far reaching effects on reserve management policy and practices. More than 80% of respondents, responsible for $3.6 trillion in reserves, said low yields have led to changes and significant majorities of this group increased the assets classes and geographies they invest in and the currencies they are exposed to. Significant minorities also said they had reduced their minimum acceptable credit rating and changed their active/passive approach. This overall view with respect to change was not affected by level of economic development or geography significantly, although among high income countries, at 72%, it was somewhat lower than the survey average. Similarly for large reserve holders, 79% of those responsible for more than $25 billion said the low yield environment had led to changes, whereas for those responsible for less than $25 billion, it was 82%.

Has it led to changes in the following?

In several comments, reserve managers noted the pervasive and systematic effects of negative yields, as a respondent from Europe explained: “The negative interest rate environment turned the possibility of loss on reserves into certainty making capital preservation an unachievable goal. It led to a major reallocation of reserves into new currencies and asset classes. Yield earned on reserves got a more pronounced attention in strategic asset allocation models.” For a reserve manager from the Americas, diversification broadly has been a major response to low yields: “Throughout the last decade, diversification is, internally, playing a key role when dealing with the current low-yield environment, by means of adding exposure to new asset classes, such as equities and MBS [mortgage-backed securities], and increasing allocation in other economies, such as China.” In an extended comment, a reserve manager from an upper middle income country set out the changes they had made in the past two years:

Such an all-encompassing approach was not uncommon: 18 reserve managers, responsible for more than $1.6 trillion in reserves said they had increased asset classes, geographies and currencies.

In terms of the changes made, 40 reserve managers or 78% of those who replied said they had increased the number of markets/geographies they invested in. This group was responsible for $2.2 trillion in reserves, an average of $61 billion. Upper middle income countries featured prominently, indeed two-thirds of this economic category could be found in this group. An African reserve manager from an upper middle income country commented: “Although asset classes and currencies have remained largely unchanged in recent years, there have been a notable increase in geographies.” The second most selected choice (although the most popular by percentage), was an increase in the number of asset classes invested in. Thirty-six reserve managers, 86% of those who responded, who are responsible for $2.2 trillion said they had added asset classes as a result of lower yields. A reserve manager from the Americas said their “Main change has been in SAA looking for higher yielding, yet high quality assets.” A second reserve manager, also from an upper middle income country, commented generally: “Central banks have had to find ways to increase reserve by going into more riskier asset classes such as corporate bonds or allowing for exposures below investment grade in order to gain higher yields to boost reserves.” The 36 reserve managers who had increased the number of currencies they invest in were on average larger reserve holders, and this group did not include any central bankers from low income countries. Changes to minimum credit rating were split almost evenly: 19 reserve managers responsible for $880 billion said they had increased this, while a smaller number, 16, albeit responsible for just over $1 trillion said they had reduced this.

Broadly, the extent of diversification by a central bank showed a positive relation with the level of economic development of the country: among central banks from high income countries the percentages for increasing asset classes, geographies and currencies were, respectively: 95%, 85% and 76%. For upper middle income country central banks they were 85%, 85% and 75%; for lower middle income countries: 83%, 75% and 43%. For low income countries, only 33% had increased asset classes, 67% had increased markets/ geographies and none had increased currency exposure. Minimum acceptable credit rating presented a different vista however: among both high income and low income countries, two-thirds of respondents said they had increased this rating. Among middle income countries, just over half of respondents said they had reduced this.

Has it led to changes in the following?

While there was considerable change in portfolios, this is less the case with strategy and use of external managers and collateral management. Just under half of respondents said they had made a change in strategy: this group had an average holding of $33 billion, below the survey average, and featured a number of high income and upper middle income respondents. The 15 reserve managers who made a change in external management were larger reserve holders, on average, at $68 billion and change in collateral management was mostly found in central banks from high income countries.

Have you added (removed) any asset class or currency to (from) your portfolio in the past 12–18 months?

Change in asset classes has been a common feature of reserve portfolios over the past 12–18 months, with addition being most prevalent. There was less of a trend with respect to change in currencies, although as with asset classes, reserve managers are adding to their portfolios. This pattern was more pronounced among large holders, with 90% of reserve managers responsible for more than $50 billion adding asset classes or currencies. Overall, 46 reserve managers, 62% of survey respondents, responsible for $3.4 trillion said they had changed either asset class or currency. Almost all reserve managers from lower middle income countries noted a change as did two-thirds of respondents from high income countries. The shares of upper middle income and low income were lower at 50% and 60% respectively.

The 32 reserve managers who said they had added an asset class were responsible for almost $2.4 trillion and had an average portfolio size just below that of the survey. There was a clear link with the previous question: 30 of these 32 reserve managers said they had changed reserve management practices because of the low yield environment. Interestingly only two of these 32 also said they had removed an asset class as well. A reserve manager from the Middle East said they had instituted a “broader strategic asset allocation”, and a respondent from the Americas said they had “added exposure to investment grade corporate bonds.” A reserve manager from Europe said they had added equities, and a reserve manager from an upper middle income country explained why they had ventured into mortgage-backed securities: “Exposure to MBS has recently been added to our portfolio in such a way to build capacity for accessing liquidity at a broader range of markets.” Of the 32 who said they added an asset class, ten also added a currency. These were in the main large holders, with an average portfolio size of $115 billion. A reserve manager from the Middle East explained their change: “We have added some currency exposure to currencies which we did not invest in the past and added two MBS external mandates” and a large holder from the Americas said they had added Korean government bonds.”

There was less evidence of change with respect to currency composition. Twenty-three reserve managers reported a change in currency of which the largest group, 16, said they had added a currency (one said that they had also removed a currency). These reserve managers were responsible for $1.3 trillion in reserves and included several large holders. Indeed, the 13 reserve managers responsible for less than $50 billion were split almost evenly between adding and removing a currency. Several reserve managers mentioned removing the euro and two the Canadian and Australian dollars. A reserve manager from the Caribbean said they had added the renminbi and one reserve manager, from Asia, said they had removed the dollar and added China’s currency.

In light of the continuing low-yield environment are you considering investing in less liquid investments in order to generate extra yield?

Reserve managers are not in the main considering less-liquid investments, in part influenced by the impact of the Covid-19 pandemic, which was apparent in several of the comments received. Within this response there was little variation across economic classifications of countries. Only low income countries differed significantly: 80% said they were not considering less-liquid investments. By geography, European and Central Asian reserve managers bucked the trend: just over half, 55%, said they were in fact considering less-liquid investments.

The 46 reserve managers not considering investing in less-liquid assets were responsible for just under $4.5 trillion and included several very large holders. A reserve manager from the Middle East was clear this view was “Because of the current situation ‘Covid-19’,” and an African reserve manager said they had heightened liquidity requirements “at the moment”. A reserve manager from Central Asia was “Putting more weight on liquidity compared to return generation,” which prevented them from investing in less liquid instruments. A reserve manager from the Americas noted they were not focused on returns at present: “Liquidity of investments and capital preservation are the main concerns in our reserves management operation.”

Comments from those considering less-liquid investments noted their market outlook and in several cases the assets under consideration. A reserve manager from an upper middle income country saw an opportunity: “The markets are volatile and the future yield on investments are unpredictable so investing in less liquid investment is beneficial in this instance.” A reserve manager noted their investment outlook was long term: “The investment horizon of [the central bank’s] reserve assets is long, which provides room somewhat to invest in less liquid investments.” Two reserve managers, from Central Asia and Africa respectively, noted the asset classes they were considering: “We think about adding corporate bonds, MBSs, emerging markets bonds in local currencies,” and “The Chinese bond market is attractive as it still offers relatively high yield.” A reserve manager from Europe noted they were considering diversification but did not view this as sacrificing liquidity: “We are considering eg, investment in equities, MBS. However we perceived them not less liquid but definitely less traditional reserve asset class.”

Are you able to purchase private placements?

Private placements are the preserve of a minority of reserve managers. Just over a quarter of respondents, responsible for just over $1 trillion, said they were able to invest in private placements. This group was 40% high income countries, with 30% lower middle income and 25% upper middle income. In terms of limits, eight respondents said they did not have any restrictions, five gave percentage figures from 10–33%, two said percentages existed but did not reveal them and one said that it depended on the average rating. Two reserve managers offered interesting observations, one, from the Americas, noting their preference for company: “Not restricted, although we greatly value not being sole investor in issue. Maximum allocation decided on a case by case basis,” and the second, from Europe, noting difficulties in tracking the process: “Yes, however in practice it is hard to monitor such limitation as details on issuance are not available – just the general assumptions/rules behind the financing program are published.”

Do you apply any form of tranching in the management of your central bank’s reserves?

Tranching is a common means of managing reserves in 2020 especially in central banks from middle and low income countries, and for those with reserves of $25 billion or less. The 49 reserve managers who said they tranched their reserves were responsible for $3.7 trillion, an average of $76 billion. The group was dominated by middle income countries, and there was a notable divide when looked at through the lens of economic development. While reserve managers from all lower middle income countries, 90% of upper middle income and all but one low income countries, said they tranched, those from high income countries were split 50:50. The amount of reserves was also a factor with just over 80% of those with less than $25 billion tranching, but for those with more than that amount the percentage fell to 52%. The eight reserve managers responsible for less than $25 billion who said they did not tranche were, with one exception, from high income countries; over half were from Europe.

In their comments, several reserve managers referred to the number of tranches (see Box 2.2 on page 24) and the rationale for this. Liquidity was the key for most, as an Asian reserve manager explained: “In line with our bank’s strategic asset allocation approach, reserves are bifurcated into three tranches based on liquidity.” A “classical” approach was employed in the Middle East: “Reserves are tranched into classical tranches reflecting the difference in liquidity needs and investment horizon.” Somewhat against the grain, an increase in reserves had been the trigger for a large reserve holder in Europe: “With higher reserves accumulated recently we have decided to tranche them. The reason is that it helps to anchor the risk profile because the tranches we have are designed for very different investment horizon.” A reserve manager in the Caribbean uses “three tranches, each with their own objective and risk tolerance. Two reserve managers from high income countries offered variations on the typical liquidity-investment split. An Asia-Pacific portfolio was split by hedging: “We have two portfolios, hedged and unhedged. It is not exactly tranching per se. Tranching often refers to a liquidity tranche and an investment tranche.” In Europe, it was a combination of activity (and accounting convention): “For most of our reserve currencies, we segregate our holdings into two different portfolios: trading portfolio and held-to-maturity portfolio.”

Has this changed in the past year?

Tranching arrangements for reserve management exhibit a high degree of stability. Just over a quarter of respondents said they had made changes in the past year. Comments from respondents could be divided into three groups: small changes in line with SAA, an increased focus on liquidity, and, conversely, an emphasis on investment. An Asian reserve manager noted adding to liquidity as they had “increased the proportion of the operations tranche,” and for a reserve manager in the Americas the reason was a change in the local market: “because of more liquidity needs of the domestic financial system.” For an African reserve manager: “Heightened liquidity requirements resulted in reduction of portfolio allocation to investment tranche.” In contrast, three reserve managers spoke to a greater emphasis on investment. A large holder from the Americas explained: “Throughout the last years, we have been increasing the long term tranche by decreasing the short term one,” and a European reserve manager said simply they were making “More long term investments.” In an extended comment, a reserve manager from an upper middle income country explained how an increase in reserves was driving changes:

The 18 reserve managers who reported a change were responsible for $721 billion in reserves and dominated by middle income countries: indeed 48% of upper middle income country respondents said their tranching had changed in the past year.

Does your central bank use a quantitative model for strategic asset allocation (SAA) in its reserve management?

Quantitative modelling is a mainstay of SAA in reserve management. Eighty per cent of respondents said they made use of these techniques, a group responsible for $4 trillion in reserves. There was not a great deal of variation across levels of economic development with central banks from lower middle income countries exhibiting the highest percentage at 85% and low income countries with 75%. This was not the case with respect to regional groupings however where only half of the samples from Middle East and North Africa, and South Asia, said they used quantitative modelling. All respondents from central banks in Sub-Saharan Africa said they used quantitative modelling, one reserve manager described their two-step approach: “The SAA optimisation uses a multi-factor model developed in Matlab to generate the optimal portfolios and the corresponding risk/return metric, while another model is used to simulate the efficient frontier and validated by yet another model for SAA simulation and stress testing.” Another combined a World Bank model with historical data and projections: “The [central bank] makes use of the Workbench Tool for optimisation and looks at historical payments and expected foreign obligations when tranching under the SAA.” A reserve manager from a large holder in Europe described the suite of models they use: “We use different quantitative methods in our decision process, but not one single model that gives us a final allocation. Methods used include: mean/ variance optimisations; optimisation with downside risk measures; simulation methods such as bootstrapping or Monte Carlo analysis; combination of the methods stated above, eg, minimisation of the shortfall probability using bootstrapping simulations.” Conditional value at risk (CVaR) was the risk measure used by one respondent from an upper middle income country: “We find the portfolio that maximises the expected return subject to a level of risk, which we determine as the CVaR (95%) of the portfolio. Along this process, we use the market data to construct the probability distribution function of any given portfolio consisting of assets in our eligible universe.”

Two respondents were at pains to stress that the quantitative analysis was part of a larger process: “We use quantitative models, running complex optimisation and simulation analyses, but they are to support not determine the decision on strategic asset allocation. Quantitative analysis is just one element of the process,” said a European reserve manager. For a reserve manager from the Americas, the analysis generated inputs to a qualitative process: “Although our SAA process does not rely merely on quantitative models, they play an important role on the whole allocation process. They are mainly used to provide important insights, such as allocation trends, which will be used as inputs to the more qualitative models.”

If yes, is the model:

Central banks tend to use their own resources to build quantitative models for strategic asset allocation, although this is most common in central banks from high income countries or large reserve holders. Thirty-four central banks, responsible for $3.7 trillion in reserves, said their models were built in-house. Eight of these did however select another additional option: five selected “built by another third party”, two “by a commercial supplier” and one selected all three options. The 34 central banks were dominated by high income countries which made up two-thirds of that group. Indeed, among high income countries, 73% said they built in house. This percentage was significantly lower among upper middle income countries (35%) and lower middle income countries (45%). No reserve manager from a low income country said they built their model inhouse. The contrast was more pronounced when looked at through the prism of reserve holdings: 80% of those responsible for more than $50 billion said they built in house, whereas the figure was 40% for those responsible for less than that amount.

If yes, does the model contain dynamic elements?

Overwhelmingly, reserve managers look to incorporate dynamic elements into their quantitative SAA models. Just over 80% of respondents said they did so, a group in which high income countries figured prominently. Two reserve managers from Europe explained their approaches: “We have a risk budget that is relative to the risk of the policy portfolio. As the risk budget is linked to the policy portfolio, it becomes dynamic (or adaptive) in the sense that the acceptable risk level shifts through time with the risk of the policy portfolio,” and for a large holder: “Depends on what you mean with ‘dynamic’. Inputs to the models are ‘dynamic’, so eg when interest rates change, this will have an effect on return expectations. Also, different time periods can be used to examine correlations and covariance matrices.”

How often does your central bank review its strategic asset allocation and strategic asset allocation process?

Reserve managers typically review their SAA on an annual basis though a significant minority do this on a more frequent basis, especially those with larger reserve holdings. The 43 reserve managers who said they carried out their reviews annually were responsible for $2.49 trillion in reserves, with an average of $58 billion, well below that of the survey overall. This was by far the most popular selection across economic groups, although lower middle income with 71% and low income countries with 80% were noticeably higher than the survey overall. A somewhat sharper contrast was provided however by reserve holdings: broadly speaking those with more reserves review more often. For those responsible for less than $25 billion, 67% review their SAA annually and 22% review more frequently than that. For those with more than $25 billion, 50% review the SAA annually and 39% review it at shorter intervals.

In their comments three reserve managers hinted at an accelerated process of SAA review. For a reserve manager from the Asia–Pacific, a review was under way to that end: “We are currently reviewing the process to increase the frequency.” A reserve manager from an upper middle income country in Europe had been asked to look into this: “We review our strategic asset allocation at least once annually. However, having in mind the dynamics of market developments in recent years, we were challenged to review our SSA more frequently. The chosen model currently meets our requirements and we do not plan to alter the SAA model.” A central banker from the Americas noted they had some flexibility in this area: “The formal review is annually; but we can conduct reviews more frequently if we consider that we need to review the model or propose a change to an allocation given a specific situation.”

Have these changed in the past two years?

Change is common feature of strategic asset allocation, but much less so with respect to the processes that produce this, which exhibit a high degree of continuity. The 34 reserve managers who said they had changed their SAA in the past two years were responsible for $2.89 trillion and were larger holders on average than the survey. The group was dominated by high and upper middle income countries, who made up just under three-quarters of the group. Reserve managers from three out of the five low income countries who took part in the survey said they had changed their SAA in the past two years. There was a trend towards greater change as reserve levels increased: 74% of those with reserves worth more than $50 billion said they had changed their SAA in the past two years. For several, the change was part of a regular process, as reserve manager from a high income country in Europe explained: “Each year, the asset allocation process incorporates the most relevant techniques used in investment management.”

In an extended comment a reserve manager from an upper middle income country explained the changes that had taken place:

Only 17 reserve managers however said they had changed the processes around their SAA in the past two years, a group dominated by high income countries. In terms of reserve size, the group featured several large holders, but was on average responsible for $79 billion, just under the survey average.

Which factors are most important to your central bank in counterparty selection? (Please rank 1–8 with 1 being most important.)

A counterparty’s credit rating is unambiguously the most important factor for reserve managers when making their selection. Just over 80% of respondents placed this first, a group responsible for $3.5 trillion in reserves. Broadly speaking this increased in importance with reserve holdings: among those responsible for more than $50 billion, 87% placed it first. There was some variation across income groups. Among high income countries, both pricing and existing relationship and performance are more important: 15% and 12%, respectively, of respondents placed this first in their ranking. For lower middle income and low income countries credit rating is more important. An African central banker summed up this viewpoint: “The credit rating is stipulated in the policy document to ensure capital preservation while the rest of the factors are more flexible.”

Which factors are most important to your central bank in counterparty selection? (Please rank 1–8 with 1 being most important.)

Three factors stood out in terms of second place: pricing, existing relationships and performance, and breadth of markets covered. Pricing was the most popular choice of the three with just over one-third of respondents placing it second. This group of 21 reserve managers was responsible for just over a trillion in reserves, giving them an average below that of the survey sample. A reserve manager from the eurozone neatly encapsulated this view: “If credit rating of the counterparty meets our credit rating threshold, then only pricing is important. Other factors are much less important.” Existing relationships and performance was placed second by just over a quarter of respondents, a group which included a number of large reserve holders. This group had an average holding of just over $100 billion. A large holder from Europe explained their process: “Process of counterparty approval is based on objective criteria (most important is the creditworthiness assessment), pricing and operational issues (soft factors) are less important, as the aim is to make the whole process transparent and unbiased.” The 12 reserve managers who placed breadth of markets second were responsible for $723 billion as a whole. Over half of this group was from Asia, and one reserve manager commented: “Credit rating and breadth are our most important factors.”

Degree of specialisation in a particular market and advisory and research garnered some support as third most important factors, but training was mainly selected as a 6th or 7th most important factor. ESG credentials do not figure prominently in reserve managers’ calculations: three-quarters of respondents put this last. A reserve manager from the Americas offered a criterion not listed in the table: “On top of those criteria we include a back office performance evaluation (which would be ranked as the third most important criteria), which evaluates the counterparty in terms of its timeliness and accuracy.”

Do you apply any new technologies, such as advanced learning algorithms or artificial intelligence, to portfolio construction or risk management?

The utilisation of new technologies in the areas of portfolio construction or risk management remains very low among reserve managers. Among those survey participants who stated that they employ new technologies, all but one were from high income and upper middle income economies, with one reserve manager from a lower middle income country. The average reserve holding of this group was $88 billion, which included one very large holder. A European reserve manager noted the techniques they used: “We use AI techniques in the construction of our in-house quantitative duration model for active management. Additionally, we use data mining to assess market impact of monetary policy changes, among other uses in market intelligence.” A reserve manager from Africa described applications for front and back offices: “[the central bank] has started applying machine learning techniques in interest rate modelling and risk analysis in the portfolio management area (front office), while a trial robotic process automation (RPA) is in the process of being implemented in the settlements area (back office). The new methods employed by those reserve managers included machine learning, perception models, data mining and trial robotic process automation while the platforms used included Bloomberg Portfolio Analytics and BlackRock Aladdin.

Even though an overwhelming majority (92%) of reserve managers indicated that they do not apply new technologies when it comes to portfolio construction or risk management, some comments suggest that this might change in the near future. A central banker from Latin America noted that their fintech projects are in early phases while a respondent from Africa said that: “This is still under consideration and we expect to apply these methodologies in the near future.” Further, some central bankers, while not using the new technologies themselves, were benefitting indirectly through external managers and counterparts. A reserve manager from a central bank based in Europe explained: “We have not used these new technologies directly. However, our counterparties execute our FX orders by using algorithms.” Only one comment was dismissive of the idea: “We do not have any plans to apply new AI technologies in our FX reserves management framework.”

How many custodians do you use?

Reserve managers use a range of custodians with, broadly speaking, the higher the income group of the country’s central bank, the greater the number of custodians employed. The most popular single category was five or more, chosen by 30% of respondents, but over half said they use three or fewer custodians. The group of central bankers who use five or more custodians was dominated by high income countries: more than a half (52%) were from that group, while 33% came from upper middle income economies. Indeed among high income countries overall, 37% said they used five or more, a share which fell to 33% among upper middle income country respondents and 23% among lower middle income countries. Not one respondent from a low income country said they use five or more custodians.

A similar pattern can also be seen among reserve holdings with over half of those responsible for more than $25 billion in reserves using five or more custodians, while for those with less than $25 billion it was one in six. A majority of reserve managers who indicated that they were using more than five custodians came from Europe and Central Asia which was followed by those who were located in East Asia and Pacific with a 19% share. Just over half of respondents said they used three or fewer custodians.

These 37 reserve managers tended to be smaller holders, with an average of $22 billion, though they did include three holders with more than $100 billion. The central banks with one custodian, the second most popular single choice, were in the main upper middle income countries. This group of 16 featured one large holder; removing this central bank from the group brought the average holding down to $11 billion. The 11 reserve managers who use four custodians featured five holders of more than $25 billion and were predominantly from Europe and Central Asia.

Of these, how many are private sector/commercial organisations?

Where a central bank has one custodian, this is typically a commercial entity. Of the 16 reserve managers who said they used one custodian 12 said this was a commercial entity (two said ‘other’ and two did not reply). These were typically small holders, with $13 billion on average, from middle income countries. Among the 11 who said they employ two custodians, just under half said they used one commercial custodian. Of the three who said they used two, one was a very large holder with more than $100 billion. Of the 10 who said they use three, only one said all three were commercial organisations: four said that one custodian only was a commercial entity. Among those with four custodians, the picture was more mixed. Seven reserve managers said they used 1 or 2 commercial custodians. Three said they use three commercial custodians.

The most common arrangement for those with five or more custodians is to have two commercial partners. Nine of the 21 respondents, which included several large holders, with an average of $72 billion, said this was their set-up. Eight respondents from the group of 21 said they engaged three or four commercial entities. This group included very large holders and was responsible for more than $1.5 trillion. One manager, from a central bank in the Americas summed up their view of the trade-off between commercial and official sector custodians: “Although commercial custodians may charge more than central banks, their broader range of services may compensate the price difference.”

Has this changed in the past 2–3 years?

Reserve managers’ custody arrangements exhibit a high degree of stability. A substantial majority of survey respondents said the number of custodians they engaged had not changed in the past 2–3 years. Fourteen respondents, one-fifth of those who answered, did however say they had changed their custody arrangements. This group was responsible for just over $1 trillion and included four very large holders each responsible for more than $100 billion. It was dominated by high income countries who made up just over half of the 14 and, indeed, 27% of high income respondents said their custody arrangements had changed. This was not the highest share among income groups however: 31% of respondents from lower middle income countries had changed the number of custodians employed. Among regions, East Asia was notable as half the respondents from this group said they had changed. In terms of number of custodians, almost half (six) said they employed five or more. A common thread running through several of the comments received was that an increase in currencies was a key motivation for a change in custody. “We increased our custodians because we decided to diversify our basket of reserve currencies,” said a European reserve manager while another said this was “by adding CNY bonds into the portfolio.” In an extended comment, a reserve manager from a high income country who had changed custody arrangements recently offered what they saw as drivers of change in this area:

Which factors are most important to your central bank in custodian selection? (Please rank 1–7 with 1 being most important.)

Safety of assets is paramount for reserve managers when choosing custodians. The vast majority, 84%, of respondents ranked this as their number one option. The second most popular was pricing where 40% of central bankers ranked it as their number two choice. A reserve manager from a central bank in Europe summed up the prevailing view:

The safety of assets and pricing were followed by accounting and reporting where 29% of respondents ranked it as third. There was little variation among sectors or by reserves size: as a priority, the safety of assets transcends. Looking at second-placed choices however yields differentiation. Among those with less than $25 billion, pricing matters a great deal: 46% placed this first or second, whereas the figure for those with more than $25 billion was 37%. For those larger holders, accounting and reporting (21% vs 10%) and technology (16% vs 8%) are more important than for those with less than $25 billion. For the group with less than $25 billion, a custodian’s track record with other central banks (21% vs 11%) was more important than for larger holders. Direct market access matters more to larger holders: 22% placed this first or second compared to 16% of smaller holders (of this 16%, 13 percentage points was assigned to second place). A securities lending proposition was mostly ranked 7^th by smaller holders and 6^th and 7^th by those with more reserves. Of the 14 reserve managers who said they had changed their custody arrangements in the past 2–3 years, safety and pricing were the most important to the ten who answered the question. Securities lending did however figure more prominently, with two of the ten placing this second.

Which factors are most important to your central bank in custodian selection? (Please rank 1–7 with 1 being most important.)

Does your central bank engage in securities lending?

Securities lending is an important element of reserve management especially among central banks from higher income countries and with larger reserves. More than half of the respondents, a group responsible for $3.3 trillion in reserves, said that their central bank engages in securities lending. Reserve managers in this group were on average responsible for $80 billion, compared to an average of $54 billion for those who do not. Securities lending appears positively associated with reserve size: just under half of those with less than $25 billion said they used this service whereas just over three-quarters of those with more than $25 billion said they did.

The group of 41 was dominated by high income countries with 59% from this category and 29% from upper middle income economies. Lower middle income and low income countries contributed 10% and 2% respectively. Europe and Central Asia has the greatest representation of the regions. Use of securities lending and economic classification appears positively correlated with 77% of high income countries using this, a percentage which declines with the level of income to 52%, then 33% and finally 20% for low income countries.

The 30 reserve managers who said they do not engage in securities lending were mostly from upper and lower middle income countries. They were typically smaller holders: 70% of respondents held less than $10 billion in reserves, with most of these in fact holding less than $5 billion. Two comments received explicitly mentioned returns, with one reserve manager noting that use of this service was one way to bring the cost of commercial and official sector custody closer together: “Securities lending is used in such a way to narrow the cost gap between what is charged by commercial custodians and central banks.”

If yes, how is this managed?

For securities lending services most reserve managers look to a 3^rd party provider. Just over 70% of respondents, responsible for $2.7 trillion, said they outsource this function. The remaining 14 who managed this in house were on average much larger holders with an average of $114 billion. The likelihood of third party use is negatively associated with economic development: 62% of high income countries said they looked outside the central bank, a percentage which rose to 82% among middle income countries. The one low income country who responded used a third party provider. Just over half of those who chose a third party provider option were from high income countries while the upper middle income countries made up just under one-third. A reserve manager from the Americas commented: “The central bank is involved in securities lending programmes carried out by our custodians, which have had positive results. This helps us to compare the results between them.” A European reserve manager, one of six who selected both inhouse and third party, explained their three-pronged approach: “Our security lending activity is primarily based on three pillars: third party lending, custodian’s automatic lending and in-house lending, which is limited due to shortage of internal resources.”

Do you allow collateral in the following:

Government bonds and cash are the most commonly accepted collateral for reserve managers, see table above. Half of respondents said they accepted investment grade bonds and 10% (four respondents) said they accepted equities. The 25 central banks that accept investment grade bonds included several very large holders and were as a group responsible for just over $2 trillion. This group was almost all high income and upper middle income countries. In their comments several noted that eligible collateral was the same or nearly the same as the eligible investment instruments. As one reserve manager from the Americas, who accepted government bonds only, explained: “The asset classes that are accepted as collateral on securities lending must comply with the same investment guidelines that apply to the reserves management. In other words, we cannot receive any collateral on which we can’t invest in.” Those for whom equities are acceptable included one very large holder, but also two with reserves of less than $10 billion. They all also accept investment grade bonds. One reserve manager commented that equity collateral was only for equity lending.

The seven who did not accept government bonds were smaller holders with $6 billion on average. The 13 who did not accept cash included several large holders included two very large holders; without these the average holding was $11 billion.

Do you make use of derivatives in your reserve management?

Derivatives are a common tool in modern reserve management, especially among those central banks in high income countries. More than two-thirds of respondents, responsible for $3.8 trillion, said they made use of these financial instruments. This group included most of those responsible for more than $25 billion and had an average holding of $78 billion. A European reserve manager from this group explained their thinking:

In contrast, the reserve managers who do not use derivatives tend to have smaller portfolios: the 23 reserve managers had an average holding of $50 billion, and removing two very large holders from this group reduced the average to $21 billion. This group of 48 was dominated by high income and upper middle income economies; indeed four out five reserve managers from high income countries said they use derivatives. Almost half of the 48 reserve managers were based in Europe and Central Asia. The regions of Latin America and the Caribbean and Sub-Saharan Africa were second and third in prominence with 17% and 15% shares respectively. Larger reserve holders tend to be more inclined to derivative use: 75% of those with more than $25 billion said they use these financial instruments, but the distinction was not that stark as 63% of those with less than $25 billion said they do as well.

If yes, which derivatives do you use?

Interest rate futures are the most prevalent form of derivative among reserve managers. Four out of five respondents said they made use of these. Use of other interest rate instruments was notably lower, with swaps at 43% and options only used by 6%. The 39 reserve managers who said they used interest rate futures were responsible for $3.1 trillion and had an average holding of $87 billion. This group was dominated by high income countries; indeed almost all of the 25 high income countries who answered this question use this financial instrument. Only around half of upper middle income use derivatives but they were large holders: almost twice that of the average for that income category as a whole. The 38 reserve managers who said they FX swaps were responsible for $3.6 trillion in reserves and had an average holding of $94 billion, well above the survey average. This group was also dominated by high income countries. The combination of FX futures and interest rate swaps was a popular one: 32 reserve managers, two-thirds of the sample, use both.

The 20 reserve managers who use interest rate swaps were on average large holders with a mean holding of $112 billion. Eighteen of this group, just over one third of the sample, also use interest rate futures. Cross currency swap users are on average large holders: the mean holding of the 14 reserve managers who said they use this instrument $136 billion. Options are the preserve of a minority: only seven reserve managers said they were writing options. Also mentioned by respondents were credit default swaps, gold options and FX forwards.

Broadly, speaking derivative use was more likely with greater reserves: 95% of those with more than $25 billion said they use FX swaps compared to 69% among those with less. For interest rate futures the proportions were 90% and 77%. Over half of the larger holders use interest rate swaps compared to 35% among the smaller holders; 43% use cross-currency swaps compared to 19%.

If you use derivatives please say for what purpose(s):

Hedging is the most common use for derivatives among reserve managers, with almost 90% including this in their responses. These 37 reserve managers were responsible for $2.75 trillion in reserves and just over half were from high income countries, as indeed was the sample of 42. This was typically mentioned with other uses however: only three respondents checked this answer alone. Reserve managers can be characterised as using suite of derivatives: only six respondents gave a single answer; six gave five or more, 12 gave four, nine gave three and a further nine gave two. Broadly speaking this was related to reserves size: 11 of the 19 respondents responsible for $25 billion or more use derivatives for four or more reasons. For those with less than $25 billion the figure was 11 from 23 respondents. Duration management and trading are popular: 76% and 64% of respondents, respectively, chose these. Overlay and yield enhancement are minority interests, although at nearly 40% it would not be fair to describe them as insignificant. These were almost entirely the preserve of central banks from high income countries, with these making up 80% of the 26 who said they used derivatives for either overlay or yield enhancement. The reserve managers who use derivatives for overlay tended to be much larger holders, with an average holding of $148 billion and were almost 90% high income countries. Those looking to enhance yield were ‘only’ 75% high income countries and had an average holding of $85 billion.

Has your use of derivatives changed in the past year?

The standout group for change in derivative use was high income countries where two-thirds said this had increased. Overall 14 reserve managers, responsible for just over $1 trillion said their use of derivatives had increased. These were on average considerably larger holders than those who had reduced their use. Two comments – one for increase and one for decrease mentioned that this move was due to change in the size of the portfolio, and in an interesting comment a reserve manager from Europe said their reduction in activity was a reflection of calmer markets in 2019:

Does your central bank incorporate an element of socially responsible investing (SRI) into reserve management?

There is considerable growing and active interest in incorporating SRI into reserve management. Almost half of respondents (45%) said that they considering this and 27% said they already did. Twenty-eight per cent of respondents said they have no interest in incorporating SRI.

The group of 32 who said that they are considering incorporating SRI was dominated by upper middle income (44%) and high income (34%) countries. As a group they are responsible for $1.97 trillion in reserves, with five responsible for more than $100 billion. Europe and Central Asia and Latin America and the Caribbean regions were the lion’s share with 38% and 25% respectively. The 27% or those who said that they were already incorporating the SRI into reserve management was largely dominated by central banks from high income countries with 74% while the rest was shared between lower middle income and upper middle income economies. These 19 reserve managers were responsible for $2.4 trillion and included three with more than $200 billion in their portfolio.

There was considerable variation across levels of economic development with high income countries much more likely to have incorporated SRI (47%) compared to middle income countries (9% and 23%). No central bank from a low income country had done so. There was however considerable interest among the upper middle income (61%) and the lower middle income groups (46%) in moving to this. In both these groups around 30% reported no interest. Reserves size was also a notable factor: 9% of those with less than $25 billion said they had incorporated SRI compared to 54% for those with more than that. Just over half of the group of those smaller holders were however considering it.

Among those who said that their central bank does not bring SRI into reserve management, the largest portion or 35% was taken by respondents from upper middle income countries. These 20 reserve managers were responsible for $600 billion.

If “no but considering it”, what do you see as major obstacles?

For those considering integrating SRI, the main obstacles are concerns over liquidity/returns and difficulties squaring this with the central bank’s mandate. Twenty-four central banks with reserves worth $1.48 trillion were worried about the impact such a move would have on the performance of their portfolio. Somewhat less, but still more than half, a group of 21 responsible for $1.18 trillion were concerned with governance issues.

Fourteen respondents, just under half the group, selected both these options. The deep-lying nature of the issues was reflected in the number of multiple answers: thirteen respondents, just over 40%, selected three or more options and six reserve managers selected all four. A European reserve manager who was one of these six commented: “We do invest in ‘green’ issues but haven’t changed our risk profile to do so. But, we are in the process of defining all of the above but we should have a policy in 2020 unless the current global environment worsens (or doesn’t improve) – it may be that we have to wait a little longer but will get there.”

If “yes”, does this include:

ESG principles are a mainstay of SRI among reserve managers: three-quarters of respondents include these. These 14 were in the main high income European countries and large holders: as a group they were responsible for $1.9 billion. The other region represented was East Asia and Pacific with four, again large holders. A specific climate focus was mentioned by seven, of which three selected this as their only option.

If “yes”, how do you integrate these into your investment process?

When integrating SRI, reserve managers tend to look to changes in their SAA/investment guidelines: two-thirds said they do this. Just over half use standalone mandates; four respondents checked both boxes. Those who selected making it part of the SAA/investment guidelines tended to be larger holders: these 12 are responsible for $1.6 trillion whereas the ten who selected portfolio mandates are responsible for $890 billion.

Which strategies⁴⁴ Here we follow the NGFS categories, which build on those of Eurosif and the PRI. See NGFS page 12 of https://www.ngfs.net/sites/default/files/medias/documents/ngfs-a-sustainable-and-responsibleinvestment-guide.pdf. Accessed 1 June 2020. do you employ?

The most popular strategy when approaching SRI is negative screening, which was chosen by 71% of the group. Just over one-third said they employ ESG integration, although this was in all but one case combined with negative screening. One reserve manager from Europe employs all five strategies. Reserve managers were split 60:40 with respect to using an external specialist data supplier for SRI. Several comments in the latter group mentioned Bloomberg as a provider, as one reserve manager noted: “We primarily use Bloomberg data field ‘Use of Proceeds’ as a criterion for our dedicated green bond portfolio. We consider examining cooperation with external data providers to enhance our monitoring efforts.” Another central banker said: “As the ESG factors are gaining attention we are considering development of our approach. This may require specialist external data. Research on this topic is ongoing.”

Which of the following best describes your view on signing the “Principles for Responsible Investment”? (Please check one.)

The “Principles for Responsible Investing” remain beyond the scope of most reserve managers with just over half of the reserve managers saying are not considering signing. This group of 32 central banks was responsible for just over $3 trillion in reserves, with a handful of respondents noting that responsible investing or ESG was not a factor in their decision-making: ESG-integration is not a primary factor in investment decisions for the central bank, at this moment, so this investment strategy is not being considered,” said a reserve manager from the Americas. Several comments from this group implied that this situation is likely to change however, as a reserve manager from the Americas explained: “Our central bank is now a member of the Network for Greening the Financial System and our authorities are very interested in greening our international investments.” A reserve manager from Europe is actively considering it this year: “Responsible investment is one of the topics covered by the current review of our reserves management strategy.”

A substantial minority, 19 or 32% of respondents, said they were however considering taking this step; although only one in ten have signed. These 19 were responsible for just over $780 billion and included two holders with more than $100 billion. Ten however hold less than $10 billion underscoring the reserves size is not necessarily an influencing factor. As one reserve manager from Africa explained: “The bank is engaging with fund managers and IFSWF [International Forum of Sovereign Wealth Funds].” Indeed, of the regions, respondents from Sub-Saharan Africa showed the greatest interest in the PRI: five of the nine from this region said they were considering signing. While reserve managers among high income countries and those with larger holdings were somewhat more favourably disposed to the PRI, the difference in percentage terms was not significant compared to overall. The 10% of those who have signed the “Principles for Responsible Investment” were from chiefly from high income and lower middle income countries, with respective shares of 67% and 33%. Most respondents in this group were from Europe and Central Asia.

Reserve managers are increasingly looking to diversify beyond their “traditional” investments, notably into green bonds, emerging market bonds and equities. Covered, green and inflation-linked bonds all enjoy support above 40% and a further 23% of respondents said they were considering adding green bonds to their portfolio. One-third invest in emerging-market bonds and slightly more in ABS (asset-based securities) or MBS (mortgage-backed securities) – but a further third are considering the emerging-market assets. Just under a quarter of respondents invest in equities and a further 26% are considering it either now or in the next 5–10 years. Over half respondents said they invest in corporate bonds (investment grade) and a further 13% said they were considering it. Ninety-three per cent make use of central bank or official sector deposits, compared to three-quarters for commercial deposits.

Green bonds were the most prominent in the second-place category: 23% of central bankers indicated that they are considering investing in this asset class now. This group was mostly from high and upper middle income economies, while the largest number of respondents or 44% were located in Europe and Central Asia. The 30 reserve managers who invest in green bonds were large holders: the average size was just over $100 billion. The 16 considering this asset class were noticeably smaller, however, with an average holding of $25 billion.

In the third category, the most popular choice was that of emerging market bonds (above BBB) as 20% of respondents indicated that they would consider investing in this class in 5–10 years. Twenty-three reserve managers with an average holding of $26 billion already invest in this asset class, a group made up mostly of middle income countries. Waiting on the sidelines are ten reserve managers, with an average holding of $38 billion, mostly high income countries. The 14 considering this asset class over the next 5–10 years have a larger average of $69 billion but this is mainly due to two very large middle income holders: most of this group in fact hold less than $10 billion.

The 17 reserve managers who invest in equities are in the main from high income countries and tend to be larger holders: the average holding was $112 billion. The group did however contain seven central banks with less than $10 billion. The four central banks considering this asset class now are all from high income countries and were on average smaller reserve holders ($63 billion.) The 13 reserve managers who said they would consider investing in equities ove the next 5–10 years were mainly from middle income countries and, again smaller holders on average with $44 billion.

Among income categories, high income country central banks tend to favour corporate bonds (66%), green bonds (67%), covered bonds (55%) and equities (43%). This sector’s exposure to commercial deposits and emerging market bonds (above BBB) at 53% and 21% respectively were noticeably lower than the survey as a whole. Reserve managers from upper middle income countries, in contrast, favour US Agency (90%), inflation linked (55%) and emerging market bonds (above BBB, 50%). Central banks from lower middle and low income countries were more conservative in their investments compared to the survey sample, although around one quarter of lower middle income countries were considering MBS/ABS and green bonds.

Larger reserve holders are in the main, as could be expected, more diversified. Forty-four per cent invest in emerging market bonds rated better than BBB with a further 16% considering this asset class now. Thirty per cent of those with more than $25 billion invest in equities with a further 11% considering such a move.

If your central bank is investing in equities, please say how you access this asset class:

Twenty-four reserve managers responsible for $2.4 trillion in reserves, an average holding of $116 billion, provided data on how they access equity markets. Managing a portfolio of shares through external managers is the most common way with just over 60% respondents employing this strategy. A significant minority use ETFs directly and just under a quarter invest in futures directly. Using external managers for futures or ETFs is the preserve of a minority and only two said they own a portfolio of shares directly.

Thirteen reserve managers said they access shares through external managers. This group was almost all high income countries and was responsible for $1.9 trillion. The high average holding of $147 billion though belies the fact that five of these 13 held less than $10 billion. In their comments, several reserve managers mentioned whether they were adopting passive or active strategies, with a large Asian holder noting “We adopt both strategies.” An interesting twist was provided by a European reserve manager: “We have both active and passive. Active management focuses on ESG integration.” For just under half the group, this was the only route to accessing equity markets: these tended to be smaller holders though the group did include one very large Asian holder. Among the remaining seven there was no clear pattern although four did also access futures via external managers.

Eight reserve managers said they access equities through purchasing ETFs. These central banks were smaller holders than the group using external managers, although they included three very large holders and had an average holding of $90 billion. In their comments five mentioned a passive approach and three (two of which held more than $100 billion) also said they get exposure to equities through external managers. A reserve manager from the Middle East noted that their “External fund management is not in place yet, but policy has been developed.”

The five reserve managers trading futures were very, very large holders and these instruments were always used in association with at least one other, either direct purchase of ETFs of portfolios accessed through external managers. As one reserve manager from the Middle East explained: “The strategy is passive index tracking and we use external managers for the bulk of the investment. For relative marginal short term exposure we do it internally through futures on equity indexes.”

The Chinese renminbi stands out as the reserve manager’s choice of nontraditional reserve and emerging-market currencies. Fifty-five per cent of respondents, responsible for reserves worth $3 trillion, said they had exposure to China’s currency, and a further 16% said they were considering it now. The Australian dollar with a 54% share and the Canadian dollar with 52% were closed behind and the percentages considering them were lower. The next group of currencies, clustered around 30%, were those of Norway, Sweden, Denmark and New Zealand. Eighteen per cent of respondents said they invest in the Singapore dollar and 15% the Korean won. Beyond those two participation was in single digit figures.

The reserve managers who said that they were investing in the Chinese renminbi, Australian dollar and Canadian dollar were mainly from high income and upper middle income economies. Among those who chose the renminbi, 45% and 29% of respondents were from high income and upper middle income economies respectively. High income and upper middle income countries dominated those who invest in the Australian dollar with a respective share of 43% and 35% and those who chose the Canadian dollar with 41% and 38% score respectively. Those considering investing in the renminbi were not the same composition as those investing. High income countries still formed the largest share, 36%, but lower middle income countries were more prominent.

Larger reserve holdings mean on average more diversification across currencies. More than two-thirds (68%) of those holding more than $25 billion said they invest in the renminbi compared to 48% for those holding less than that figure. Interestingly the rank of the currencies was more or less the same across both groups from the renminbi to the Korean won, but with a 15–20 percentage points gap between the proportion of larger holders investing and that of the smaller holders. Beyond the won and the Singapore dollar, however there was more difference and there was noticeably more support for the koruna, ringgit and zloty among larger holders and little interest in the real, rupee and rand.

If investing in the renminbi, please give the amount of your portfolio invested (%).

Thirty-five reserve managers provided details of their renminbi holdings, with most, almost two thirds, reporting holdings of between 1% and 5%. The value of their combined renminbi holdings based on the percentages given was $42 billion, giving a weighted average percentage allocation of 1.83%. Broadly speaking, reserve managers from higher income countries have lower percentage allocations: while 20% of respondents from high income countries said they held 6% or more of their reserves in renminbi, the share of low income countries was 60%.

These 23 reserve managers were responsible for $1.6 trillion and included three very large holders. Removing these three brought holdings down to less than $500 billion and the average holding to $20 billion. This group was mainly made up of high income and upper middle income countries. Just over 40% were from Europe and Central Asia with Latin America and the Caribbean and Sub-Saharan African each contributing 17%.

Five reserve managers said their holdings were less than 1%. This group was responsible for $600 billion but within the group were two very large holders who accounted for over $500 billion of that. A respondent from high income economy indicated the reasons behind their institution’s decision to invest a relatively low percentage in renminbi: “The RMB looked like a star currency some years ago. Now I see a bit of stagnation and one of the reasons may be the fact that China is just learning what it means for the country to open the financial market. The US-China trade talks are just nothing else than an attempt to put RMB and USD into fair competition.” Another respondent from a European country also outlined the issues they face when it comes to investing in the renminbi: “Renminbi has small share in the foreign exchange reserves as we are still in learning phase of Chinese investments. Low level of liquidity and timezone difference (bonds not traded non-Chinese hours) make trading very challenging.” Those investing between 6–10% in renminbi were a different composition to that of the two previously discussed percentage groups. Here respondents came from lower middle and low income countries. A reserve manager from Sub-Saharan Africa commented: “The amount has been increasing steadily in recent years.” Those investing more than 10% of their portfolio were again lower middle income and low income economies. The patterns for large and smaller reserve holders were broadly similar although holders with less than $25 billion did have more of a ‘tail’ with respect to holdings greater than 6%.

The IMF added the renminbi to the SDR in October 2016 with a weight of just below 11%. What percentage of global reserves do you think will be invested in the renminbi by 2020, 2025 and 2030?

Reserve managers see China’s currency settling at between 10–20% of global reserves over the next decade. Just under 30% of respondents see the renminbi accounting for 10–20% of reserves by the end of 2020, a figure which rises to 46% of respondents for 2025 and 58% for 2030. Very few reserve managers see China’s currency amounting to more than 20% of global reserves by the end of this decade and a significant minority, 31%, believe it will still makeup less than 10%.

A respondent from high income country set out what they saw as the two main drivers of this move: “The share of global reserves invested in China is expected to increase with the country’s increasing economic power and the internationalisation of its financial markets. The way is although still long.” While another reserve manager from Latin America offered a similar explanation: “Because of the weight of the Chinese economy and its importance to the global financial system.” Some central bankers envision that the impact of Covid-19 and political tensions surrounding China will exert a drag on the process: “Given the current Covid-19 conditions in the global economy, specifically the continent of Asia, the value of renminbi may decline, which will negatively affect its percentage of global reserve,” commented a respondent from upper middle income country. A respondent from European country noted: “Global FX reserves allocation into renminbi will further decelerate this year as result of the Covid-19 and the geopolitical tensions surrounding China.”

In general, reserve managers from lower middle income countries are the most optimistic for the renminbi’s development as a reserve currency: 78% of this group think China’s currency will account for 10–20% of reserves by 2025 and indeed 2030, although none think it will exceed 20% by the end of the decade. High income respondents were more cautious, especially in the near term: only 32% think it will account for 10–20% by 2025, but this percentage almost doubles to 62% when looking at 2030. Reserve managers from upper middle income countries are the most cautious: 43% of this group think the renminbi will account for less than 10% of reserves by 2030. Broadly speaking lower reserve holders are more optimistic: 65% of those with less than $25 billion think China’s currency will make up 10–20% of reserves in 2030 compared with 40% of those with more than that figure.

What percentage of your reserves do you think will be invested in this currency by end 2020, 2025 and 2030?

Reserve managers are more cautious in their outlook for the renminbi when it comes to their own reserves. Most believe China’s currency will account for less than 10% of their reserves through to the end of the decade, although a significant minority see its share at 11–20% by 2030. Smaller reserve holders tend to be more optimistic: 31% said they think the renminbi will 11–20% of their reserves by 2030 compared to 17% of those with more than $25 billion. Reserve managers from upper middle income countries are the most wary: 11% said 11–20% of their reserves would be in renminbi by 2030; the figures for high income and lower middle income countries were 31% and 40%. Among regions, respondents from Asia-Pacific are the most cautious: all said they thought the renminbi would account for less than 10% of their reserves by 2030. In Latin America and the Caribbean all but one said it would account for less than 10% of their reserves by 2030. A large holder who saw their investment getting to 5% expressed their concerns thus: “As long as the Chinese government keeps implementing capital control measures, and the legal framework is not strengthened, we will be reluctant to add a substantial exposure to the aforementioned. Our exposure to CNY/CNH will be a function of the level of liberalisation that the Chinese markets display.”

Reserve managers from Europe and Central Asia are more sanguine with a quarter of this group seeing their share at 10–20%. One respondent from this group explained their thinking: “In the near term the share of our Chinese exposure is expected to be flat although some increase is very probable in the long run. In addition to the size of investments it could be important adding new asset classes to the current pure government exposure.” In contrast, 44% of Sub-Saharan African respondents thought China’s currency would account for 10–20% of their reserves, as one reserve manager commented: “The amount invested in RMB has been increasing steadily in recent years and we expect this trend to continue.”