Interview: Jonas Stulz

Christopher Jeffery spoke with the head of portfolio management and deputy head of asset management at the Swiss National Bank on March 5, 2024.

The Swiss National Bank (SNB) is one of the most diversified central bank reserve managers in the world. What is the current breakdown of your foreign exchange reserves, and how has that changed versus a year earlier?

The breakdown of our reserves has been relatively stable over the last couple of years. Obviously, there are some fluctuations on the tactical and the implementation levels, but much less for the investment strategy. The current breakdown is 64% in government bonds, a category that also includes cash held with central banks or the Bank for International Settlements. Then we have 11% invested in other bonds – so, that’s investment-grade credit as well as supranationals, sovereigns and agencies, that kind of thing. And then we hold a strategic allocation of 25% in equities.

Why has your overall reserves level fallen in the past few years?

On the balance sheet, after the sharp increase over the last decade [FX reserves hit Sfr 946 billion, or $1.1 trillion, in January 2022], it decreased in 2022 and 2023. This is for two reasons. The primary reason was related to monetary policy – and I should stress that the primary objective of our investment policy is to support monetary policy at all times. As we have reported publicly, there was a period in 2022 and 2023 when the SNB sold foreign exchange [2022: Sfr 22.3 billion; Q1–Q3 2023: Sfr 110.2 billion] and bought back Swiss francs for monetary policy reasons. That led to a decrease in the level of reserves. Then, the second factor, given that we mark-to-market our portfolio of reserves in Swiss francs, was a valuation loss, which accounted for another substantial part of the decrease, particularly in 2022.

During 2022, there was a breakdown in typical correlation between bonds and equities, raising potential questions about the diversification benefits of holding both. What lessons, if any, did the SNB take away from this episode? Were there any changes in the SNB’s assets holdings as a result?

It is important to say that our investment policy is driven by long-term considerations. That speaks for a certain stability over time. Consequently, there was no major change in the strategic asset allocation [SAA] in reaction to the situation of 2022.

In terms of a lesson learned, it was obviously an extraordinary year. We know – and we always communicate – that our financial results can fluctuate a lot, especially given the fact that we mark-to-market the reserves portfolio. But what happened in 2022 was certainly extraordinary on any scale. We – along with most other asset owners – recorded substantial losses in absolute terms. In our case, it was pretty much every kind of risk factor added to the losses: we lost on equities, we lost on fixed income, and we lost on FX risk.

How does the SNB manage its communications about the mark-to-market profits and losses from its FX reserves?

It starts with transparency. I would argue that we are very transparent for a central bank regarding both the breakdown of the reserves portfolio and our financial results. The portfolio breakdown (asset and FX allocation) and the investment performance (both in local currencies as well as in Swiss francs) are published on our website on a quarterly basis. The investment performance is the dominant driver of the SNB’s overall financial results, which are also published on a quarterly basis – this includes the final annual results as well as quarterly intermediate results. We have always highlighted that the SNB’s financial result depends largely on the developments of relevant market prices, and strong fluctuations are therefore to be expected. This has materialised, particularly in 2022.

Investors at the beginning of 2023 were wary that major economies might tip into recession. As a result, some reserve managers adopted a ‘risk-off’ posture (trimmed duration, bought short-term developed debt and put some of their diversification efforts on hold). But, as the US (and others) avoided a downturn and inflation fell quicker than many had expected, given the economic conditions, many switched more to ‘risk-on’. What was the SNB’s experience?

Generally speaking, I think it’s important to distinguish between the different layers of reserve management. What you ask for here are active positions based on a certain view of the business cycle and market environment. At the SNB, such positions can be taken on two levels. Firstly, we have an investment committee that can take discretionary tactical positions for the overall investment portfolio. Secondly, for the fixed income part of the investment portfolio, we have an active management approach at the portfolio-management level. It is important to note, however, that the risk-return profile of the SNB’s investment portfolio is first and foremost determined by the annual investment strategy defined by the Governing Board, whereas active positions of the investment committee and fixed income portfolio management are moderate, and do only have a limited performance impact.

Is there anything you could share in terms of what happened in terms of active portfolio management?

Again, active portfolio management is only applied to fixed income at SNB. Equities, by contrast, are managed passively – ie, we replicate selected equity indices. As to active fixed income portfolio management, we have a very solid track record over the past decade – ie, we have been able to achieve positive excess returns versus the strategic and tactical benchmarks in a consistent manner. This shows that we have, on average, been successful in implementing various active strategies (eg, relative duration and yield curve positions in rates portfolios, as well as sector allocation and title selection positions in credit portfolios). Obviously, the relative excess returns of such active strategies are small compared to the absolute return of the overall portfolio – but, again, over time, we are able to make a financial contribution with active fixed income portfolio management.

In the past, the SNB’s lack of FX hedging – due to the need to be active in the foreign exchange markets for monetary policy purposes – has resulted in the SNB having longer duration than its peer central banks. What is the current duration for your fixed income, and how has it changed?

The duration of the total portfolio currently is slightly below five years. Again, there may be limited fluctuations over time due to tactical and/or implementation considerations. But the strategic duration has been stable in recent years. If you compare our duration to other peer central banks, yes, we tend to have longer duration than other central banks. One reason therefore is that we are fully invested in foreign currency (no FX-hedging) but report our portfolio performance in Swiss francs, which is a ‘safe-haven’ currency. In this constellation, duration has historically been a very good portfolio diversifier. That’s because, usually, in a ‘risk-off’ environment, we would suffer from losses on our FX exposure (due to Swiss franc appreciation) but benefit from gains on duration (due to decreasing global yields). And vice versa in a ‘risk-on’ environment. Obviously, this playbook does not work in every environment, such as in 2022. But, in the long term, we still believe duration is a good diversifier for our portfolio.

How do you assess the interplay between increased government debt, quantitative tightening and bouts of illiquidity in key reserve currency bond markets – particularly the US? Are you concerned about a drying-up of liquidity?

I would not say that we are concerned, but, of course, we constantly monitor liquidity conditions and market functioning in all markets we are invested in. This is particularly true in core government bond markets like the US and Germany, as they are the most liquid assets in our investment universe. At the end of the day, given that our reserves fulfil monetary objectives, we have a high demand for liquidity. This is the same in both directions – when we built up the portfolios up to 2021, and then also during the decrease we have experienced, we must be able to buy and sell securities within our portfolio in decent size. So far, our assets have proven liquid enough for whatever we needed to do. So, we feel still comfortable being invested there. When I look how we trade, for instance, US Treasuries, we can shift around decent trade tickets, say, around $50–100 million in 10-year equivalent, at very acceptable prices.

The SNB split asset management into portfolio management and trading a few years ago. This was done to improve analytical capabilities and trading efficiency. What have been the pros and cons of this approach?

We embarked on that journey of functional separation between portfolio management and trading for fixed income in 2018. Quite frankly, the pros have clearly outweighed the cons. Our goal from having a functional separation was to really focus on core competencies on the analytical side, in terms of portfolio management, and in terms of trading capabilities, in the trading unit. With the old universal model, there were just too many responsibilities for each person. It is important to note that, in our model, the trading unit is much more than just an execution desk. Instead, traders are responsible for security selection, and their profound market knowledge is also used to contribute to the determination of active risk positions in the portfolios. In that sense, at SNB, active fixed income portfolio management is a true collaborative approach including portfolio managers and traders.

With the updated structure, staff could really achieve major improvements on both fronts. On the portfolio management side, we have made our analytical processes more standardised and documented. Among other things, we established a so-called FTV approach (fundamental, technicals, valuation) – a framework that helps to define and evaluate active portfolio positions in a unified manner. And, on the trading side, we introduced ‘Trade value chain in fixed income’, a data-driven approach that integrates all elements of trading – pre-trade analyses, execution and post-trade analyses – into one integrated workflow including feedback loops. We are not done with that, it is not finished, but the journey is continuing in the right direction.

In terms of what we could call cons, we need to be aware that the implementation of the new setup was a big challenge, both organisationally and culture-wise. In particular, we needed to make sure that people – be they in portfolio management or trading – adopt the new mindset and focus on their different core competencies. That’s not something that happens immediately. It needs a lot of management attention and careful change management. It was a big effort for everyone involved, and still is. But it’s worth it!

How do you measure if the portfolio changes and the trade execution are carried out in an optimal manner?

At the end of the day, the success of every active decision – be it from portfolio managers or traders – is reflected in the excess return of the underlying portfolio. Such excess returns are calculated on a daily basis for all our portfolios.

In addition, we apply a variety of measures in order to get a more granular picture. On the portfolio management side, we established a fixed income performance attribution in recent years, which helps us to identify the different sources of excess return in the portfolios. On the trading side, we apply extensive post-trade analysis in order to evaluate transaction costs, execution quality as well as the market impact of our transactions.

What are the pros and cons of the use of direct investments, external mangers, exchange-traded funds (ETFs), futures, options and other derivatives when it comes to major asset classes?

Starting with internal versus external management, we have a very small part of the portfolio that is externally managed (currently roughly 2%, distributed among a number of managers). There are two use cases where we would hire external managers. The first one is where there would be an efficiency gain. Very often, when we introduced new asset classes in the past, we would start with external managers because of such efficiency gains. The second reason would be to have an external benchmark for internal portfolio managers.

In terms of whether to use direct investment, funds, ETFs or derivatives, the usual way is via direct investments in securities: bonds and equities. That is also related to the size of the portfolio. There are no ETFs or funded tools usually available that could digest the size of our flows. So, our investment strategy is, first and foremost, implemented by direct investments. In addition, we make use of derivatives for reserve management purposes: mainly futures on the equity and fixed income side, and interest rate swaps on the fixed income side. Generally speaking, derivatives are very useful for us as they enhance flexibility and provide access to an additional source of liquidity. We use derivatives for two purposes: active overlay positions (both for TAA and for portfolio management) and hedging positions for the day-to-day maintenance of the portfolio (implementing benchmark changes, rebalancings, etc).

Which environmental, social and governance (ESG)-related investment processes does the SNB asset management department engage in, and how is this likely to change over the next three years? If you engage in green investments, what are the best approaches to ensuring sufficient liquidity while avoiding the risk of greenwashing?

In general, the SNB endeavours to spread its investments as widely as possible in order to avoid concentration risks and market impact of transactions. However, the SNB deviates from the principle of full market coverage in two instances. First, we do not invest in shares of global systemically important banks. Second, we want to make sure that our investment policy respects Switzerland’s fundamental standards and values. That is why we have an exclusion policy. Specifically, we do not invest in shares or bonds of companies that seriously violate fundamental human rights, systematically cause severe environmental damage or produce internationally condemned weapons. We use the help of external service providers to identify the companies that fall under the exclusion criteria, and we exclude such companies from all internally and externally managed portfolios.

What happens when it comes to environmental investments?

Again, our exclusion list contains companies that systematically cause severe environmental damage. This category also includes companies whose business model is primarily based on the mining of thermal coal. Apart from that, we are obviously invested in green bonds. However, it is important to note that we do not use green or ESG benchmarks. Instead, we are invested in green bonds as part of our traditional benchmarks.

So, how do you deal with ESG risks in reserve management?

There is always a double materiality in terms of ESG. One element is the risk perspective, which focuses on the financial materiality of ESG risks in our portfolios. The other element is the impact perspective, which focuses on the potential impact that our investments may have on the environment or on society. This distinction helps to understand our approach.

As to the risk perspective, we incorporate that in our decision-making process in active fixed income portfolio management, pretty much as we do with every other source of financial risk. So, we look at ESG risks and integrate them into our portfolio analysis and management processes. But that’s only for the risk perspective. When it comes to the impact perspective, this is not part of our mandate. The Swiss constitutional and legislative authorities have not tasked the SNB with using its investment policy to conduct structural policies, eg, by favouring some economic sectors over others with our investments.

What is the SNB’s assessment when it comes to non-public market assets?

The SNB reviews its investment universe and asset allocation on an annual basis. In that context, we regularly evaluate potential additional asset classes to further diversify our portfolio. We have also looked at private markets in recent years, but we have decided to stay away, as private markets such as real estate, infrastructure or private equity do not fulfil the high standards for safety and liquidity that are required for our reserves portfolio.

How is the SNB factoring in artificial intelligence, if at all, into its monitoring and trading?

Artificial intelligence is a wide field. I would say we in SNB’s asset management do a lot in terms of data integration and making our workflows more data-driven, integrated and more efficient. So, data onboarding and handling is a big issue for us. But when it comes to a more narrow definition of artificial intelligence – say, in terms of the use of large language models, machine learning techniques, and so on – we currently only make use of such techniques to a limited extent.

Personally, I see potential for more both on the portfolio analysis and management side and on the trading side. But, again, that’s music of the future.