Cross-country equity futures strategies

Jupyter Notebook

Developing macro strategies for cross-country equity futures trading is challenging due to the diverse and dynamic nature of equity indices and the global integration of corporations. This complexity makes it difficult to align futures prices with country-specific economic factors. Therefore, success in cross-country macro trading often relies on differentiating indicators related to monetary policy and corporate earnings growth in local currency. Additionally, cross-country strategies benefit from a broad and diverse set of countries to generate value consistently.
We tested five simple, thematic, and potentially differentiating macro scores across a panel of 16 developed and emerging markets. Our findings suggest that a straightforward, non-optimized composite score could have added significant value beyond a risk-parity exposure to global equity index futures. Furthermore, a purely relative value equity index futures strategy would have produced respectable long-term returns, complementing passive equity exposure.

The post below is based on Macrosynergy’s proprietary research.
Please quote as “Sueppel, Ralph, ‘Cross-country equity futures strategies,’ Macrosynergy research post, August 2024.”

The attached Jupyter notebook allows for the audit and replication of the research results. The notebook’s operation requires access to J.P. Morgan DataQuery to download data from JPMaQS, a premium service of quantamental indicators. J.P. Morgan offers free trials for institutional clients.
Also, an academic research support program sponsors data sets for relevant projects.

The challenges of cross-country equity index trading

Generally, currency area-specific macroeconomic conditions are valid predictors of cross-country returns of similar assets. This makes point-in-time macro indicators valuable for fixed income and FX allocations (see examples here and here). In principle, macro trends and states also affect the relative performance of local-currency equity indices and related futures by influencing financial conditions and corporate earnings.

However, when compared to fixed income and FX markets, macro-based allocations across country-specific equity index futures, or cross-country relative value trading, face tougher challenges:

• Heterogeneity: Unlike high-grade fixed-income markets, the underlying traded contracts are heterogeneous across countries. Companies and sector weights vary greatly, as do the laws and regulations that affect them. Heterogeneity means differences in sensitivity to local economic conditions, such as local economic growth or financing conditions. A country index dominated by manufacturing companies is much more dependent on external conditions than one dominated by banks or utilities.
• Structural change: Stock indices are evolving. Companies and weights change over time due mainly to market capitalization and listings. This means that, unlike in FX and interest swap markets, the nature of contracts and payoffs is changing, and statistical analyses of the past may sometimes be misleading for the future.
• Globalization: Corporate profitability is closely connected across countries. Most companies operate internationally through branches, subsidiaries, or strategic investments. Also, their financing is international, with many companies preferring debt issuance in large and liquid developed countries. Developments in large countries strongly affect corporate revenues and funding costs in smaller countries, and they do so to varying and unknown degrees.

Simply put, the cross-country heterogeneity of equity indices and the global interconnectedness of traded firms make it hard to align equity index futures with currency-area-specific economic factors. Structural differences across economies further dilute the relation between macro factors and returns. These are similar to heterogeneities that also impair the macro-markets nexus in fixed-income markets:

• Institutional differences: Monetary policy, financial conditions, and country risk respond differently to economic trends across countries. For example, highly inflation-averse central banks tighten policy rates more aggressively than more dovish institutions.
• Statistical differences: Data quality varies across countries regarding reliability and timeliness. Countries such as the U.S., which produce a broad range of timely surveys and labour market indicators, reveal their economic trends earlier than countries with few published indicators and longer lags, such as India.

The loose and uncertain connection between local macroeconomic trends and equity index performance does not mean that there is no cross-country trading value in macro factors. Indeed, these difficulties also heighten the rational inattention of investors. They do, however, imply greater demands on factor development. Generally, there are two ways to translate macro factors into meaningful value generation for cross-currency equity allocation:

• Adapting macro factors through structure features of the index: An example could be using the ratio of local versus international revenues as a modifier of the influence of local and international aggregate demand. This approach will be discussed in another post.
• Focusing on distinctive and differentiating macro factors: While many macroeconomic trends and states affect the equity market, only a few are specific to country performance. For example, broad price and wage inflation are widely seen as detrimental to the outlook for stock markets. Yet, it is often a global phenomenon, and its effect on local conditions can vary greatly. In the section below, we will focus on identifying simple differentiating factors.

International equity index futures for this analysis

An equity index future is a financial derivative contract that stipulates the purchase or sale of an equity index at a predetermined future date at a price specified today. Across countries, these futures are based on major local currency-denominated indices that comprise the stocks of the largest companies listed in a local exchange. The constituents and weights in the index depend mainly on the market capitalization of such companies, the liquidity of common stocks, and financial health. Hence, the composition of indices across sectors and their concentration are not just widely different across countries but also changing.

This post looks at eight developed market local-currency index futures and eight emerging market local-currency index futures. The developed market indices are (in alphabetical order of currency symbol):

• AUD: The S&P/ASX 200 is the primary index for the Australian equity market. It measures the performance of the 200 largest companies by market capitalization listed on the Australian Stock Exchange (ASX).
• CAD: The S&P/TSX 60 Index is designed to represent the large-cap segment of the Canadian equity market. It includes 60 of the largest and most liquid companies listed on the Toronto Stock Exchange (TSX).
• CHF: The Swiss Market Index (SMI) is Switzerland’s premier stock market index, representing the 20 largest and most liquid companies listed on the SIX Swiss Exchange.
• EUR: The EURO STOXX 50 Index is a leading stock market index for the euro area, representing the 50 largest and most liquid blue-chip companies from 11 Eurozone countries.
• GBP: The FTSE 100 Index is a stock market index representing the 100 largest publicly listed companies by market capitalization on the London Stock Exchange (LSE).
• JPY: The Nikkei 225 Stock Average Index is a prominent stock market index in Japan, representing 225 of the largest and most liquid companies listed on the Tokyo Stock Exchange (TSE).
• SEK: The OMX Stockholm 30 (OMXS30) Index is a major stock market index in Sweden, representing the 30 most traded large-cap companies listed on the Nasdaq Stockholm exchange.
• USD: The S&P 500 Composite Index is one of the most widely followed stock market indices in the world, representing 500 of the largest publicly traded companies in the United States. (Note that even though the U.S. is about twice the size of the euro area, the market capitalization of the S&P500 has been roughly ten times that of EURO STOXX.)

The emerging market indices are:

• BRL: The Brazil Bovespa Index, also known as the Ibovespa is the main stock market index in Brazil, representing the performance of the most actively traded companies on the B3, the São Paulo Stock Exchange.
• INR: The CNX Nifty 50 Index is one of the leading stock market indices in India. It represents the performance of the top 50 largest and most liquid companies listed on the National Stock Exchange of India (NSE).
• KRW: The KOSPI 200 Index is a major stock market index in South Korea, representing the performance of the 200 largest and most liquid companies listed on the Korea Exchange (KRX).
• MXN: The Mexico IPC, often referred to as the Bolsa Index, is the main stock market index in Mexico. It tracks the performance of the 35 largest and most liquid companies listed on the Mexican Stock Exchange.
• MYR: The FTSE Bursa Malaysia KLCI (Kuala Lumpur Composite Index) is the main stock market index in Malaysia. It tracks the performance of the 30 largest companies by market capitalization listed on the Bursa Malaysia
• THB: The Bangkok SET 50 Index is a major stock market index in Thailand, representing the 50 largest companies listed on the Stock Exchange of Thailand.
• TWD: The Taiwan Stock Exchange Weighted Index (TAIEX) is the main stock market index in Taiwan. It tracks the performance of all listed companies on the Taiwan Stock Exchange.
• ZAR: The FTSE/JSE Top 40 Index is a key stock market index in South Africa, representing the 40 largest companies listed on the Johannesburg Stock Exchange (JSE) by market capitalization.

Not all liquid emerging market futures have been included in this analysis. Some countries’ policy regimes either feature currency pegs or are too idiosyncratic to allow simple relative value analysis. This group includes China, Hong Kong, Singapore and Turkey. Also, there is a trade-off within the chosen set of 16 indices. Using a broad set of developed and emerging markets provides more statistical power to hypothesis tests. However, including the emerging markets set also aggravates institutional and statistical differences, making it particularly important to focus on comparable and differentiating factors.

Equity index futures returns have been taken from the J.P. Morgan-Macrosynergy Quantamental System (JPMaQS) as per cent change of the front futures prices, assuming rolls (from front to second) on standardized settlement days. The emphasis in the analytical sections below will be on returns in the per cent of risk capital of positions scaled to a 10% (annualized) volatility target based on an exponential moving average of returns with a half-life of 11 days (view documentation).

Macro themes for local equity index futures

The focus of this post is on macro conditions that shape trends and divergences of monetary policy and earnings in the local currency area. Point-in-time macro indicators have been taken from JPMaQS and are suitable for testing relations with subsequent returns and backtesting related trading strategies. We test five macro themes for local equity performance based on the simplest plausible formulation and with no optimization within or across the themes. All macro factors have been formulated such that their hypothesized relation to equity index futures returns is positive,

• Inflation shortfall: This theme suggests that inflation rates above the central bank’s effective target augur for monetary tightening. Conversely, an inflation shortfall, i.e., rates below target, argues for easing monetary conditions, which in turn supports equity performance.
  We use four commonly monitored measures of inflation to measure the shortfall: headline CPI and core CPI as % over a year ago and seasonally—and jump-adjusted headline and core CPI as % of the latest 6 months over the previous 6 months at an annualized rate (view documentation). From these rates, we subtract the effective inflation target of the currency area (view documentation) and take the negative values.
• Labour market slackening: This theme hypothesizes that rising unemployment or underemployment reduces labour cost growth and invites monetary policy support for the economy, while tightening labour markets leads to less support. We represent the theme with two indicators. The first is the negative of the difference between employment growth as % over a year ago in 3-month moving averages (view documentation), and the 5-year trend in work force growth (view documentation). The second is simply the increase in the unemployment rate over a year ago in 3-month moving averages (view documentation).
• Effective currency depreciation: The theme simply states that trade-weighted currency appreciation is bad for corporate earnings growth on local currency terms since prices in trading partners’ markets are most fixed in their currency. The concept is represented by three measures of annual appreciation, all in % over a year ago in a 3-month moving average: Openness-adjusted real effective appreciation (view documentation), openness-adjusted nominal appreciation, and real effective appreciation (view documentation). Openness adjustment means that the (estimated) rate of real appreciation is multiplied by the ratio of merchandise exports and imports to GDP (5-year moving averages), which makes the economic impact of effective exchange rate dynamics comparable across currency areas.
• Ease of local finance: The idea behind this theme is simply that accommodative local financial conditions support demand for equity. The theme is represented by two indicators. The first is the negative of an excess real short-term interest rate, which is calculated as the difference between a real expectations-based short-term interest rate (view documentation) and estimated productivity growth, whereby the difference between a GDP growth trend and a workforce growth trend approximates the latter. The second indicator, excess real equity carry, is based on the difference between (i) the average of expected forward dividend and earnings yield and (ii) the main local-currency real short-term interest rate, in % annualized of notional of the contract (view documentation). Excess here means above 3.5%, a judgment call assuming that for equity carry to be attractive, it needs to excess at least 20% of long-term average index future volatility.
• Terms-of-trade improvement: Improving terms-of-trade, i.e., faster growth of export prices than import prices, augurs for relatively strong corporate earnings in the local economy. We represent the theme by the annual growth rate of two metrics: commodity-based terms-of-trade (view documentation) and “mixed” terms-of-trade (view documentation), whose underlying data vintages include actual broad export and import prices indices and, when these are not available, commodity-based indices. Since broad terms-of-trade are typically published at lower frequencies and with much longer lags, the commodity terms-of-trade are effectively used to predict and interpolate the broader series.

This set of hypothesized factors is, of course, not exhaustive and merely serves as an illustration. Moreover, not all these macro factors are clear cross-country differentiators. Inflation, for example, is very important for equity markets but is often a global rather than local phenomenon, and its influence on monetary policy can vary considerably depending on inflation aversion and data quality.

To analyze the predictive power and economic signal value of these themes, we first normalize the constituent categories of each theme by sequentially applying historical standard deviations across the panel up to the real-time date, and capping normalized values at a maximum of three standard deviations at the positive or negative side. Then, the thematic score is calculated as an equally weighted average of the normalized constituents and renormalized. The six thematic scores are plotted for all 16 developed and emerging currency areas below:

Finally, a total macro score is calculated based on conceptual parity, i.e., as a simple average of the thematic scores, and finally re-normalized. Note that if a theme is missing for a country, as with India’s labour market score, the total score is calculated based on the available themes. There is no optimization involved. The selection of themes is based solely on theoretical plausibility, and weights are based solely on the assumption of equal long-term importance for each concept. The timelines of the aggregate thematic macro scores are plotted below.

The total macro scores are slow-moving trading signals in keeping with the nature of the underlying economic trends. The average duration of long or short positions has been close to a year, albeit signal strength has changed at a higher frequency. Overall, this type of gradual trading signal is very transaction cost-friendly but also requires large data panels for statistical validation, and related PnLs can be very seasonal.

Note that in the below section, the macro scores are applied to all countries except when markets were blacklisted for lack of liquidity or due to trading restrictions.

Simple directional relations between thematic scores and future returns

First, we check if the macro scores display directional predictive power in the full panel of 16 developed and emerging markets. The scatter plots and test statistics below show that both the total macro score and all thematic scores have been positively correlated with subsequent vol-targeted equity index futures returns. These predictive relations hold for subsequent quarterly or monthly futures returns and for both simple and vol-targeted positions.

The total macro score and two of the thematic scores, effective depreciation and financial conditions, have also been highly significant predictors. The significance of the other themes has been below the conventional threshold. Significance measures the probability that the predictive relation has not been accidental. Note that the macro factors with the strongest forward correlations are not necessarily the ones with the greatest significance. This is because, in a panel context, significance relates to intertemporal and country-specific relations, not merely global predictive relations. Thus, inflation shortfalls and labour market slackening are the strongest global factors, but their cross-country predictive power is typically more limited.

Technically speaking, statistical significance here is estimated based on panel regression that uses period specific “random effects”, i.e., the test accounts for the commonalities of variations across countries and, to the extent that features or targets are highly correlated, emphasizes the ability of differences in features to predict differences in country returns (view post here). Thus, the hypothesis that is tested is that the macro score predicts both intertemporal variations in returns and their cross-country differences.

Other statistics confirm the directional predictive power of the macro scores. Thus, monthly accuracy, the ratio of correctly predicted return signs, and balanced accuracy, the average of the ratios of correctly predicted negative and positive returns, have been 54.1% and 53.4%, respectively. Finally, non-parametric predictive correlation (“Kendall correlation”) has been positive and highly significant.

Using the macro score to manage a portfolio of futures returns of the global equity index would have added material economic value. The graph below illustrates naïve PnL, following a standard methodology used in previous research posts. This PnL is calculated for simple monthly rebalancing in accordance with the macro score winsorized at three standard deviations at the end of each month. The end-of-month score is the basis for the positions of the next month under the assumption of a 1-day slippage for trading. The naïve PnL does not consider transaction costs or compounding. For the chart below, the PnL has been scaled to an annualized volatility of 10%.

The green line below shows the naïve PnL of a simple long-only risk parity portfolio, i.e., equal exposure in volatility terms across all 16 equity index futures. The blue line plots the naïve PnL of a portfolio that takes vol-targeted long and short positions in equity index futures based solely on the total macro scores. Finally, the orange line is a combination of the two, i.e., a portfolio that takes positions in accordance with the macro score but at one standard deviation to the scores and thus imposes a long bias.

The risk-parity portfolio would have delivered a long-term Sharpe ratio of 0.6 and a Sortino ratio of over 0.8, with roughly 60% positive correlation with the S&P500 index returns. The pure macro score-based portfolio would have produced a Sharpe ratio of above 0.8 and a Sortino ratio of more than 1.2, with just a 10% S&P500 correlation, resulting from a 54% long bias of the signal. The combined long-biased portfolio would have delivered a 25-year Sharpe ratio of more than 0.9 and a Sortino ratio of nearly 1.3. The stylized PnL arises from directional equity risk premia, exposure timing, and cross-country allocations.

A cross-country equity futures strategy

It is much harder to produce PnL value from cross-country allocations alone. Here, we implement a relative futures strategy that uses relative macro scores to trade one country’s equity future position against a basket of all the others. Positions are all vol-targeted to 10% based on historical daily returns of an exponentially weighted lookback window.

For this pure cross-country strategy, we deploy relative macro factor scores. This means that we subtract global average scores from country-specific scores. For example, the inflation-based signal becomes a relative inflation shortfall of one country versus the others. The relative signals change the dynamics of the signals materially.

Analogous to the directional case, the scatters and statistics below check the correlation between end-of-quarter relative macro scores and subsequent relative futures returns. All thematic and total macro scores have displayed positive predictive power. The predictive relation of the total macro score has been highly significant, albeit its correlation coefficient has been lower than for the directional case. Of the thematic scores, effective currency depreciation and ease of local finance have achieved high statistical significance.

Based on relative global macro scores, we can now estimate a naïve PnL for cross-country equity index futures trading across all 16 developed and emerging equity markets, according to the same rules as above. This cross-country strategy construction never takes long or short exposure to the overall international equity market. It thus forfeits all equity risk premia and potential equity timing benefits. The below graph illustrates its 25-year performance.

The long-term Sharpe ratio of this simple pure relative value strategy has been 0.55 and its Sortino ratio over 0.8. Interestingly, PnL value generation has been roughly comparable to the performance statistics of the long-only risk parity strategy, and the relative value is conceptually and statistically additive. The correlation of the cross-country PnL with S&P500 returns has been virtually zero.

Value generation has been seasonal, as is typical for macro trend-based strategies. However, good seasons have been dominant and peak-to-trough drawdowns have never been catastrophic. Thus, PnL generation has not been concentrated on just a few episodes. The best 5% of all months account for roughly two-thirds of the long-term PnL.

PnL generation through pure cross-country relative positions has been even harder in the developed world alone due to the smaller number, financial integration, and economic similarity of developed countries. The long-term Sharpe ratio of a DM relative value strategy would have been just 0.3 and the Sortino ratio a mere 0.4. The PnL graphic still reveals a long-term upward drift but with large intermittent drawdowns and volatility. Moreover, value generation across developed countries would have been dominated by the terms-of-trade score.

The differences in PnL generation between the broad global and narrow developed market sets illustrate the importance of scope and diversity for relative future strategies. Considering that macro trend influences are pretty subtle, the greater the number of currency areas and the more pronounced the differences in their development, the more opportunities the macro-based strategy has to add value.