1. Introduction
I would like to thank the Bank of Albania, the IMF, and the Swiss Embassy for your kind invitation to speak at this forum. I was quite interested in the various discussions on the macroeconomic vulnerabilities of highly euroized economies in the context of negative euro area interest rates. These discussions provide a good background for my presentation on reserve management in the region.
In order to address the issue of negative euro area interest rates on the return of reserves, rather than discussing investment strategies designed to enhance the income on reserve management in this environment, which I suspect most central banks in the region have already considered, I would like to focus on the risk that these strategies entail given the external vulnerabilities that highly euroized economies are exposed to in a rising interest rate scenario. The reason for taking this approach is to emphasize that, first and foremost, reserves serve an insurance purpose, and as managers of insurance, central banks need to ensure that the level of risk-taking in the management of these funds is consistent with the size and nature of their external vulnerabilities.
I would like to begin my presentation with a description of the current market consensus on the trend of interest rates in the euro area, then discuss specific reserve adequacy indicators for highly euroized economies as the basis for defining the risk tolerance with which reserves are managed and, lastly, the risk in a rising yield environment of investment strategies designed to enhance returns in a negative rate environment.
2. Market Consensus on the Trend of Interest Rates in the Euro Area
As we all know, the prevailing forecast is that euro yields will be positive in the next three years. This is not necessarily a World Bank view, but the view of the consensus of the market. In accordance with the latest European Commission Economic Forecast (Figure 1), for the first time in a decade the economies of the EU member states without exception are expected to grow, bar any political surprise from another member state. After the French elections last week, President Draghi announced that he might be considering ending the ECB’s tapering plans sooner rather than later. However, after Brexit, political uncertainty is likely to persist and the focus may now shift toward the approaching Italian elections next year. But what drives this consensus?
Private consumption has been the engine of the recovery (Figure 2). Private consumption has been supported by robust employment growth and increases in compensation, while public consumption is expected to continue providing stable support (Figure 2)
Investment growth has recovered, but remains subdued notwithstanding substantial policy support and the improvement of financing conditions. Brexit, uncertainty on the outcome of the elections in France, and the ongoing process of corporate deleveraging may explain the disappointing growth of investment. But going forward, there is a reasonable expectation that investors’ confidence should pick up as the economy continues to grow and financing costs remain low (Figure 3).
The contribution of net exports to growth is expected to be neutral. Exports have gradually increased, supported by the recovery in global demand, particularly in the US and emerging markets, but imports have increased at the same pace given the recovery of private consumption (Figure 4).
Inflation is expected to be below the 2 percent ECB target and inflation swap rates at the three-year forward three-year ahead horizon price an average inflation of 1.4 percent (Figure 5)
Given the above, market consensus predicts that, barring a political surprise like Brexit, the gradual economic recovery will induce the ECB to announce in Q4 the end of its QE tapering plan for 2018. In response to this, the yield curve is expected to be less negative next year (Figure 6). In three years, the concensus expects that the yield curve will be positive, while short-term rates remain close to zero (Figure 7).
Expected changes in German Government yield curve over a one-year horizon
Macroeconomic scenario based on Oxford Economics Market Consensus based on Bloomberg.Expected changes in German Government yield curve over a one-year horizon
Macroeconomic scenario based on Oxford Economics Market Consensus based on Bloomberg.Expected changes in German Government yield curve over a one-year horizon
Macroeconomic scenario based on Oxford Economics Market Consensus based on Bloomberg.
Expected changes in German Governmen yield curve over a three horizon
Macroeconomic scenario based on Oxford Economics Market Consensus based on Bloomberg.Expected changes in German Governmen yield curve over a three horizon
Macroeconomic scenario based on Oxford Economics Market Consensus based on Bloomberg.Expected changes in German Governmen yield curve over a three horizon
Macroeconomic scenario based on Oxford Economics Market Consensus based on Bloomberg.Given the above, what would be the risk of strategies to enhance returns in a negative interest rate environment in a rising yield consensus scenario? Before addressing this question, I would like to first assess the risk taking capacity of a central bank to manage its reserves on the basis of an assessment of a country’s external vulnerabilities.
3. Reserve Adequacy and Risk Tolerance
What could the impact be in the region, and indeed in all developing countries, of higher interest rates in the euro area and in the US? During our Reserve Advisory and Management Program (RAMP) Executive Forum last week, President Trichet warned that low interest rates have pushed the boundaries of risk for international investors looking for returns. He foresaw that the next global crisis could very well come from emerging markets, both from the significant increase of capital flows to emerging markets in a low interest rate environment and unsustainable debt to GDP ratios. Furthermore, Mohamed El-Arian warned that global risks have increased considerably, characterizing the future as a camel with two humps, meaning two fat tails, and central banks will need to manage the reserves in an environment where abnormality is the norm. The fundamental question then becomes whether developing countries have enough reserves to adequately hedge a potential reversal of capital flows, harder access to refinance debt rollovers, global shocks, and, in the case of highly euroized or dollarized economies, risks to the stability of their financial systems vulnerable to significant currency mismatches, as we have already observed in some Western Balkan countries.
In the case of the Balkan region, it is difficult to generalize a single measure of reserve adequacy given that each country has different external vulnerabilities, exchange rates regimes, and EU membership status (Figure 8). In this presentation, I will mainly focus on countries with a floating exchange rate arrangement.
I shall also use as a measure of reserve adequacy, the IMF’s assessment of reserve adequacy model (ARA) comprising the following indicators with the weights shown in Figure 9:
Exports: to hedge the risk of a potential loss of export income as a result of a drop in external demand or a shock in the terms of trade.
Short-term debt (STD): to hedge debt rollover risk.
Other liabilities: to hedge the risk of non-resident capital outflows as a result of the liquidation of domestic equity holdings and medium- to long-term debt holdings in domestic currency.
Broad money: to hedge the risk of resident capital flight.
The ARA model is a considerable improvement over the Greenspan-Guidotti rule for reserve adequacy, based on short-term debt plus a measure of current account volatility. I also believe that the ARA model is more relevant to developing countries because it incorporates the importance of non-resident capital flows, as capital markets have become more open, and the broad money measure incorporates financial stability concerns, especially relevant to highly euroized and dollarized economies.
In accordance with the ARA model, as observed in Figure 10, most countries in the Western Balkan region with floating rate regimes hold reserves within the target band defined by the model.
Reserve adequacy metrics for Albania, Bosnia and Herzegovina, Macedonia, Serbia, and Croatia
Reserve adequacy metrics for Albania, Bosnia and Herzegovina, Macedonia, Serbia, and Croatia
Reserve adequacy metrics for Albania, Bosnia and Herzegovina, Macedonia, Serbia, and Croatia
The fact that reserves in the region are within the adequacy range used by the ARA model has profound implications for the risk tolerance with which they are managed. But before delving into the specifics of the region, the following global generalizations could be relevant:
In countries with flexible exchange rate regimes, accumulating reserves within an adequate range is necessary to preserve the credibility of monetary policy, but the tolerance to liquidity, market, and credit risk is usually higher, as intervention is designed to only smoothen excessive exchange rate volatility.
In countries with managed currency arrangements, not only maintaining reserves within an adequate range is highly relevant, but also their tolerance to liquidity, market, and credit risks is typically lower.
The insurance motive for holding reserves to manage fat tail events, for example, disruptions in international markets and political events, affects both.
In the case of highly euroized or dollarized economies, even those with inflation-targeting regimes, the vulnerabilities captured by the ARA model provide some insights into their specific vulnerabilities that could be used in setting the tranching structure of the portfolio comprising: (1) a highly liquid low-risk Working Capital Tranche for day-to-day intervention; (2) a Liquidity Tranche invested in liquid assets with a short-term capital preservation objective; and (3) an Investment Tranche invested in less liquid assets with a longer than one year capital preservation objective. In this respect, I would like to make the following observations:
Particularly after the financial crisis, portfolio flows have increased significantly, increasing the exposure of the region to non-resident capital outflows and potentially requiring higher low-risk liquidity buffers. Non-resident capital flows are pro-cyclical, raising vulnerabilities and amplifying external shocks, of which portfolio investments represent the more volatile component. As observed in Figure 11, before the financial crisis capital flows reached the equivalent of 25 percent of GDP, predominately because of FDIs and bank lending consistent with the expansion of foreign bank networks in the region. After the financial crisis, capital inflows collapsed below 10 percent of GDP, but portfolio debt flows increased.
As observed in Figure 12, the region has high debt-to-GDP ratios, and at least half of this debt is denominated in a foreign currency. However, all of the countries in the region meet the yardstick for holding enough liquid reserves to cover at least short-term debt amortizations. In this respect, many central banks have established coordinating committees with the Ministry of Finance to ensure that debt repayments are smoothened over time.
Indicators of the soundness of the financial system (Figure 13), as well as the robustness of macro-prudential safeguards such as the existence of regulatory deposits, can provide useful information on the allocation of this metric to the Liquidity and Investment tranche. Although the Western Balkan banks appear well capitalized, hold high liquidity-to-short-term-liabilities ratios, and maintain a high level of regulatory deposits in the central bank, the region has suffered bouts of high levels of non-performing loans, currency mismatches, and contagion from parent companies abroad.
The exchange rate competitiveness indicator may not be as relevant to central banks in countries with floating rate regimes in the Balkan region. However, an assessment of the degree of depreciation or appreciation of the exchange rate relative to the currencies of its main trading partners can provide useful information on the allocation of funds to the Liquidity and Investment tranches.
Western Balkan-resident capital inflows, % of GDP
Western Balkan-resident capital inflows, % of GDP
Western Balkan-resident capital inflows, % of GDP
Considering the above, the risk bearing capacity of central banks in the region may be tempered by the fact that the level of reserves is in most cases within the ARA reserve adequacy metrics. Furthermore, the nature of the vulnerabilities of each country in the region may also limit the risk-taking capacity of central banks to manage their reserves within the reserve adequacy range, in particular because of financial stability concerns given the high level of euroization.
4. The Risk of Investment Strategies to Enhance Return in a Negative Rate Environment.
In the current negative interest rate environment, the need for income generation and the limited risk-bearing capacity of central banks in the region poses a challenging problem. On the one hand, reserve managers may argue that negative income generation weakens central bank balance sheets and, at the extreme, may indirectly hamper the accomplishment of their fundamental objectives unless the central bank charter, like the one governing the central bank of Albania, obliges the government to finance the central bank with the issuance of interest-bearing domestic bonds while this situation persists. On the other hand, the managers of monetary policy may argue that increasing the risk tolerance beyond a level that is consistent with the size and nature of their contingent liabilities may expose the central bank to the realization of significant losses if risky investments are required to be liquidated before maturity to cover a potential need for intervention.
One way to resolve this conflict is by measuring the potential risk to future returns of different investment strategies designed to enhance returns in a negative rate environment – duration, credit spreads, and currency – should the consensus view of the market occurs.
In the case of duration, the steepness of the yield curve provides an opportunity for reducing the impact of negative yields by simply increasing the duration of the portfolio. To assess this risk, we analyze the impact on returns of portfolios with increasing durations if yields of German Government bonds, increase as forecasted by the consensus over a one- and three-year investment horizon, as shown in Figure 14.
Consensus forecasts
Macroeconomic scenario based on Oxford Economics Market Consensus based on Bloomberg.Consensus forecasts
Macroeconomic scenario based on Oxford Economics Market Consensus based on Bloomberg.Consensus forecasts
Macroeconomic scenario based on Oxford Economics Market Consensus based on Bloomberg.The results are not surprising. Duration strategies do improve short-term returns given the steepness of the curve, but will forgo future returns if there is a normalization of the yield curve. Over a one-year investment horizon (Figure 15), the C-VaR looks progressively worse even though yield increases are not significant, mainly because the portfolio does not benefit from positive carry. Over a three-year investment horizon (Figure 16), the losses worsen considerably for the same reason. The question then becomes whether it is preferable to accept negative rates now by investing in short duration securities, rather than face bigger losses in the future if the yield curve normalizes.
Duration risk over a one-year investment horizon (annualized)
World Bank Asset Allocation Workbench simulation resultsDuration risk over a one-year investment horizon (annualized)
World Bank Asset Allocation Workbench simulation resultsDuration risk over a one-year investment horizon (annualized)
World Bank Asset Allocation Workbench simulation results
Duration risk over a three-year investment horizon (annualized)
World Bank Asset Allocation Workbench simulation resultsDuration risk over a three-year investment horizon (annualized)
World Bank Asset Allocation Workbench simulation resultsDuration risk over a three-year investment horizon (annualized)
World Bank Asset Allocation Workbench simulation resultsIn the case of spread strategies, as shown in Figure 18, the consensus predicts a slight increase in spreads. Since this strategy benefits from positive carry in a magnitude that depends on the credits chosen, the returns of the portfolio benefit both in the current negative rate environment and in the future (Figure 19). However, investing in credit spreads exposes the portfolio to the non-systemic risk of investing in a “bad apple,” which can be managed through diversification, but may not address reputational risk. To illustrate this point, some central banks, including the central bank of Colombia, invested in a well-diversified portfolio of corporates before the financial crisis. During the financial crisis, a default occurred worth 10 basis points of the value of the portfolio and, although the loss had been compensated by the additional returns accrued over the whole portfolio, to this day the decision has been questioned by the public, and the central bank has undergone several investigations.
Eastern European spread over two year German Government bond
Based on projected spreads of Eastern European countries that are part of the EU but not in the eurozone.Eastern European spread over two year German Government bond
Based on projected spreads of Eastern European countries that are part of the EU but not in the eurozone.Eastern European spread over two year German Government bond
Based on projected spreads of Eastern European countries that are part of the EU but not in the eurozone.
Spread risk over a one and three-year invesment horizon using two-year bonds
World Bank Asset Allocation Workbench simulation resultsSpread risk over a one and three-year invesment horizon using two-year bonds
World Bank Asset Allocation Workbench simulation resultsSpread risk over a one and three-year invesment horizon using two-year bonds
World Bank Asset Allocation Workbench simulation results
Return and risk of investing in non-Euro currencies
Return and risk of investing in non-Euro currencies
Return and risk of investing in non-Euro currencies
An alternative used by central banks to reduce the impact of negative returns on their reserve portfolios is to invest in currencies that are not necessarily their natural numeraire to benefit from positive carry. This is a more complex strategy to assess from a balance sheet perspective. In most central banks, accrued income benefits the income statement, but exchange rate changes relative to the local currency are usually registered in equity in a currency stabilization account in order to avoid monetization. In other words, foreign exchange strategies can improve income, while, at the same time, deteriorate the capital of the central bank, and, as such, they should be treated with care.
In most central banks, the currency composition of core reserves (defined as reserves within the reserve adequacy range) seeks to preserve their value relative to the macro vulnerabilities of the country. As such, in setting the currency composition, central banks consider the denomination of foreign debt, trade, and the currency of intervention, among other variables. In highly euroized or dollarized economies, the currency composition is heavily influenced by such, currencies. By following an asset-liability management approach for the selection of the currency composition of a reserves portfolio, the impact on equity of changes in the value of these currencies relative to the local currency can be explained to the public and usually, over the long term, follows a countercyclical behavior. When the local currency appreciates relative to these currencies, the probability of intervention is lower and the capital of the central bank accumulates unrealized losses. In the opposite case, when the local currency depreciates, the probability of intervention is higher and the central bank can potentially register realized gains. This may not be the case when a central bank invests core reserves in currencies with a positive correlation to the local currency, for example, by investing in high-yielding commodity currencies.
As can be observed in Figure 19, investing in Australian dollars generates higher cumulative income relative to investing in German Government bonds, as well as provides diversification benefits, but exposes the portfolio to a much higher volatility against the predominant reserve management numeraire in the region.
However, central banks that hold reserves beyond precautionary needs, and are confident that these reserves will not be used for intervention over investment horizons longer than one year, have looked at different rationales for minimizing the cost of accumulating reserves beyond an “insurance” purpose. Some of the objectives used by central banks are:
Hedge the balance sheet from non-euro-denominated external debt contracted by the central bank, such as IMF loans.
Hedge public debt contracted in other currencies, as part of a strategy to reduce inefficiencies in the consolidated balance sheet of the central bank and the government.
Protect the balance sheet of the central bank and increase income by investing in high-yielding currencies with a positive correlation to the local currency.
5. Conclusions
I would like to conclude my presentation by reemphasizing that when it comes to core reserves, as defined by their reserve adequacy target, it is of fundamental importance that they are managed with a risk tolerance that is consistent with the contingent macro risks that they are designed to cover. Furthermore, the analysis of the specific external vulnerabilities of each country using the ARA model, or other reserve adequacy methodologies, is fundamental in defining the tolerance to market, credit, and liquidity risks with which reserves should be invested. Not to do so can expose central banks to the realization of significant losses at a time when it needs reserves the most. Income considerations, while important, are secondary in this case. However, in the case of central banks that have either excess reserves over core needs, or feel that they have sufficient short-term liquidity to invest a portion of reserves over longer investment horizons, they can and do invest the reserves with a higher risk tolerance and different rationales in order to reduce the cost of accumulating reserves beyond short-term contingent needs.
This presentation has been prepared by the Treasury of IBRD (TRE) for working purposes for the clients participating in the RAMP program to guide them in understanding certain concepts underlying investment management. It does not represent, and shall not be interpreted as, specific advice or recommendation as to any particular matter covered herein, nor as an indication of market standard in a particular area. Nothing contained in the presentation constitutes or shall be construed as a rpresentation or warranty by IBRD.
