Exploiting yield-curve dynamics in hiking cycles

In the systematic research section of Alphorum, our quarterly assessment of global fixed-income markets, we focus on trading strategies that can help capture positive returns across the US Treasury yield curve in a rising-rate environment. It follows reports on developed-market sovereigns, corporate credit, emerging-market debt and sustainability. Tomorrow, we will conclude with our overarching perspectives as 2022 gets underway.

Exploitable trends

The hawkish shift in rhetoric from the Federal Reserve has led to intensifying expectations of rising interest rates. The obvious concern in fixed-income markets is the downside potential as prices are marked lower from rising base rates, yet we believe that such a scenario need not condemn active managers to negative performance.

In this piece, we analyse previous rate-hiking cycles and look to unearth exploitable trends in yield-curve dynamics that could offer opportunity in these challenging markets. For our analysis, we use US total return swaps to access the returns of different segments across the US yield curve over the last 30 years, focusing on the two-year, five-year, 10-year and 30-year maturity points.

Performance across the yield curve

Central-bank rate hikes affect different maturity segments of the yield curve in different ways. This analysis, therefore, addresses the three components of the traditional decomposition of moves in yield curves: level, slope and curvature.¹

Clearly, the most obvious and direct impact is levied on the level of yield curve. In the simplest of terms, increasing rates lifts the level of interest rates, causing prices to fall. Figure 1 shows how different segments of the curve performed over the last 30 years, in which we identify four hiking regimes: 1994-1995, 1999-2000, 2004-2006 and 2016-2018.

Figure 1: USD interest-rate swap unfunded total returns by duration segment

Source: Bloomberg bellwether swap indices, LOIM Calculations as at January 2022. Returns calculated using monthly data from July 1992 to Dec 2021. We remove six-month swap returns from the headline indices to calculated ‘unfunded’ returns. ‘Hold’ periods are those deemed neither ‘hiking’ nor ‘cutting’ cycles. All figures are annualised.

Figure 1B shows how performance varies across maturity segments and by regime. Over our sample period, all segments show an average loss across the four hiking regimes, with the higher maturity segments suffering the worst absolute performance along with the largest drawdowns. The reverse is experienced in cutting cycles. This is as expected, with the higher duration segments by definition more sensitive to changing rate levels. However, this is not a fair comparison across segments as the durations and therefore volatility of various segments are very different.Source: Bloomberg bellwether swap indices, LOIM Calculations as at January 2022. Returns calculated using monthly data from July 1992 to Dec 2021. We remove six-month swap returns from the headline indices to calculated ‘unfunded’ returns. ‘Hold’ periods are those deemed neither ‘hiking’ nor ‘cutting’ cycles. All figures are annualised.

To correct this, we calculate returns per unit of duration, as shown in Figure 2. Figure 2A shows that the longest maturity segment performs the best in a rate-hiking cycle and is the weakest in a rate-cutting environment. Economically, this is intuitive, since per unit of risk (duration), the front end of the yield curve is more sensitive to changes in rates, while the longer end of the yield curve is driven by long-term growth and inflation expectations. Hiking rates sees front-end rates affected more per unit of risk directly through the rate-hike channel than long-end rates. What’s more, rate hikes can be seen to be expected to rein in growth and inflation expectations, which should exert some downward pressure on long-end yields.

Figure 2: USD interest-rate swap total returns per unit of duration by maturity segment

Source: Bloomberg bellwether swap indices, LOIM Calculations as at January 2022. Returns calculated using monthly data from July 1992 to December 2021. Returns calculated using monthly data from July 1992 to December 2021. We remove six-month swap returns from the headline indices to calculated ‘unfunded’ returns. ‘Hold’ periods are those deemed neither ‘hiking’ nor ‘cutting’ cycles. All figures are annualised.

Nevertheless, the performance of all rate segments after adjusting for duration remains negative, indicating that the ‘level’ effect of the hiking cycle dominates the yield curve. So, to generate positive performance from risk-free assets in this environment, investors need to exploit relative relationships within the curve.

Slope trades (flattener strategies)

Slope trades can be constructed through two approaches: by over- or under-weighting specific segments of a Treasuries portfolio, or by using a variety of derivatives instruments like swaps or futures. Here we focus on unfunded interest-rate swaps to build our strategies.

The previous section established that longer maturity yields have outperformed shorter maturity yields through recent rate hiking cycles. This would favour a curve ‘flattener’ trade, which profits when short-term yields rise more than longer term yields and the curve ‘flattens’. To achieve this, the trade goes long the longer maturity swap and goes short the shorter maturity swap in a duration-neutral manner. In Figure 3 we analyse the performance of three different, commonly implemented duration-neutral flattener trades which focus on different parts of the yield curve: two-year versus 10-year (2s10s), five-year versus 30-year (5s30s), and 10-year versus 30-year (10s30s).

Figure 3: USD interest-rate swap flattener strategies

Source: Bloomberg bellwether swap indices, LOIM Calculations as at January 2022. Returns calculated using monthly data from July 1992 to December 2021. Returns calculated using monthly data from July 1992 to December 2021. We remove six-month swap returns from the headline indices to calculated ‘unfunded’ returns. ‘Hold’ periods are those deemed neither ‘hiking’ nor ‘cutting’ cycles. All figures are annualised.

Figure 3A shows that these trades have negative performance over the whole period. This is in line with the results established in Figure 2, as the 30-year underperforms the five- and 10-year segments, and the 10-year marginally underperforms the two-year, on a duration-adjusted basis. However, the benefits of the flattener trades in hiking environments are clear to see in Figure 3B, with all trades showing strong Sharpe ratios through the four rate-hiking cycles (as well as in each cycle individually).

Having said that, it’s important to note that this outperformance is not persistent across other central bank regimes, and that it suffers heavily outside of hiking cycles. This is not because curves have steepened more than they have flattened across the whole period – in fact, the steepness of curves is actually mean-reverting and tends to follow rate cycles. Instead, it is a structural feature of yield curves, which tend to be flattest at the long end and are generally steep at the short end. This is because at the short end there tends to be a premium for uncertainty in inflation and real rates, while at the long end there is an excess demand for convexity, especially from long-term liability-driven investors. As a result, yield-curve carry reduces as we increase maturity, and so the flattener strategy runs a negative carry position, which creates a performance drag over time.

Curvature trades (butterfly strategy)

This leads us to the final strategy and the final component of rate-curve dynamics: curvature. Here, the strategy essentially takes two slope positions: one flattener and one steepener (this is colloquially known as a ‘butterfly’). By adding a steepener elsewhere in the curve, the strategy can benefit from the positive carry and reduce the drag that the flattener experiences from its negative carry position. From the results in Figure 3B we can see that, during hiking cycles, it would have made the most sense to have a 2s10s flattener (highest Sharpe ratio) and a 10s30s steepener (lowest Sharpe ratio). Ultimately, the strategy is long the 10-year and short the 2-year and 30-year, with the expectation that the curve will develop greater convexity.

The economic intuition for a butterfly trade comes from overtightening by central banks. Overtightening, such as during the 2017-2018 period, often leads to expectations of a policy error. This tends to depress the belly of the curve as it incorporates more rate cuts in the intermediate period. Such expectations result in an outperformance for the butterfly trade described above.    

Figure 4: USD interest-rate swap slope and butterfly strategies

Source: Bloomberg bellwether swap indices, LOIM Calculations as at January 2022. Returns calculated using monthly data from July 1992 to December 2021. Returns calculated using monthly data from July 1992 to December 2021. We remove six-month swap returns from the headline indices to calculated ‘unfunded’ returns. ‘Hold’ periods are those deemed neither ‘hiking’ nor ‘cutting’ cycles. All figures are annualised.

Indeed, we can see from Figure 4 that the 2s10s30s butterfly strategy improves the Sharpe ratio further across the four periods, with a substantial reduction in volatility compared to a 2s10s flattener alone. Again, this strong performance is achieved across each of the four rate-hiking cycles.

A final caveat is that these results are based only on the four hiking cycles experienced in the past 30 years. Any hiking cycle takes place in unique conditions, so there is no single strategy best suited to all hiking scenarios. However, since a lot of the economic intuition behind the success of these strategies remains intact, we believe that deploying them in a timely manner can help nimble managers outperform in challenging fixed-income conditions.

À l’usage des investisseurs professionnels uniquement
Le présent document a été publié par Lombard Odier Funds (Europe) S.A., société anonyme (SA) de droit luxembourgeois, ayant son siège social sis 291, route d’Arlon, 1150 Luxembourg, agréée et réglementée par la CSSF en tant que Société de gestion au sens de la directive 2009/65/CE, telle que modifiée, et au sens de la directive 2011/61/UE sur les gestionnaires de fonds d’investissement alternatifs (directive GFIA). La Société de gestion a pour objet la création, la promotion, l’administration, la gestion et la commercialisation d’OPCVM luxembourgeois et étrangers, de fonds d’investissement alternatifs (« FIA ») et d’autres fonds réglementés, d’organismes de placement collectif ou d’autres véhicules d’investissement, ainsi que l’offre de services de gestion de portefeuille et de conseil en investissement.
Lombard Odier Investment Managers (« LOIM ») est un nom commercial.
Ce document est fourni à titre d’information uniquement et ne constitue pas une offre ou une recommandation d’acquérir ou de vendre un titre ou un service quelconque. Il n’est pas destiné à être distribué, publié ou utilisé dans une quelconque juridiction où une telle distribution, publication ou utilisation serait illégale. Ce document ne contient pas de recommandations ou de conseils personnalisés et n’est pas destiné à remplacer un quelconque conseil professionnel sur l’investissement dans des produits financiers. Avant de conclure une transaction, l’investisseur doit examiner avec soin si celle-ci est adaptée à sa situation personnelle et, si besoin, obtenir des conseils professionnels indépendants au sujet des risques, ainsi que des conséquences juridiques, réglementaires, financières, fiscales ou comptables. Ce document est la propriété de LOIM et est adressé à son destinataire pour son usage personnel exclusivement. Il ne peut être reproduit (en totalité ou en partie), transmis, modifié ou utilisé dans un autre but sans l’accord écrit préalable de LOIM. Ce document contient les opinions de LOIM, à la date de publication.
Ni ce document ni aucune copie de ce dernier ne peuvent être envoyés, emmenés ou distribués aux États-Unis, dans l’un de leurs territoires, possessions ou zones soumises à leur juridiction, ni à une personne américaine ou dans l’intérêt d’une telle personne. À cet effet, l’expression « Personne américaine » désigne tout citoyen, ressortissant ou résident des États-Unis d’Amérique, toute association organisée ou existant dans tout État, territoire ou possession des États-Unis d’Amérique, toute société organisée en vertu des lois des États-Unis ou d’un État, d’un territoire ou d’une possession des États-Unis, ou toute succession ou trust soumis dont le revenu est imposable aux États-Unis, qu’en soit l’origine.
Source des chiffres : sauf mention contraire, les chiffres sont fournis par LOIM.
Bien que certaines informations aient été obtenues auprès de sources publiques réputées fiables, sans vérification indépendante, nous ne pouvons garantir leur exactitude ni l’exhaustivité de toutes les informations disponibles auprès de sources publiques.
Les avis et opinions sont exprimés à titre indicatif uniquement et ne constituent pas une recommandation de LOIM pour l’achat, la vente ou la détention de quelque titre que ce soit. Les avis et opinions sont donnés en date de cette présentation et sont susceptibles de changer. Ils ne devraient pas être interprétés comme des conseils en investissement.
Aucune partie de ce document ne saurait être (i) copiée, photocopiée ou reproduite sous quelque forme et par quelque moyen que ce soit, ou (ii) distribuée à toute personne autre qu’un employé, cadre, administrateur ou agent autorisé du destinataire sans l’accord préalable de Lombard Odier Funds (Europe) S.A. Au Luxembourg, ce document est utilisé à des fins marketing et a été approuvé par Lombard Odier Funds (Europe) S.A., qui est autorisée et réglementée par la CSSF.
© 2022 Lombard Odier IM. Tous droits réservés.