Broad macroeconomic trends, such as inflation, economic growth, and credit creation are critical factors of shifts in monetary policy. Above-target trends support monetary tightening. Below-target dynamics give grounds for monetary easing. Yet, markets may not fully anticipate policy shifts that follow macro trends, for lack of attention or conviction. In this case, macro trends should predict returns in rates markets. In the past, even a very simple point-in-time macro pressure indicator, an average of excess inflation, economic growth, and private credit trends, has been significantly correlated with subsequent rates receiver returns, both in large and small currency areas. Looking at the gap between real rates and macro trend pressure delivers even higher forward correlation and extraordinary directional accuracy with respect to fixed income returns.
The below post is based on proprietary research of Macrosynergy.
Some links for data access and documentation refer to the J.P. Morgan Markets website and require user credentials. Yet, the links are not necessary for understanding the content of the post.
This post ties in with this site’s summary on information efficiency.
The importance of macro trend pressure
Official mandates and economic theory suggest that economic growth, inflation, and other key macro trends relative to target levels cause monetary policy adjustments. Above-target dynamics support shifts towards monetary tightening, while shortfalls support monetary easing. These policy shifts can take the form of communication or actual market operations. Tightening almost always implies an increase in local-currency interest rates.
In accordance with basic macroeconomic convention, we define macro trend pressure on interest rates as macroeconomic dynamics relative to benchmarks or targets. In particular, we focus on excess inflation, excess economic growth, and excess private credit growth, the most widely used indicators in practice. Positive excesses should bias central banks towards tightening.
If markets did fully trust and rely on macroeconomic information and if there were no frictions and information cost only surprises in macro trends would explain returns. The prevalent state of economic trends should have no predictive power. However, it is more plausible that markets fail to consistently track and predict economic and policy trends due to research costs, trading restrictions, and external effects (view article here). Rational information inattentiveness means that market participants update their information set sporadically, rather than continuously. Such inattentiveness should be commensurate to the costs of acquiring information and costs of re-optimizing investment decisions. It causes sticky expectations and goes some way in explaining price momentum after important relevant data releases. Hence, the working hypothesis of the below analysis is that macro trend pressure on interest rates retains some predictive power.
The measurement of macro trend pressure
To assess the historical predictive power of macro trends for fixed income returns, we must not use standard economic time series. That is because these are not correctly timestamped to their availability in markets and may have been significantly revised over time. Therefore, all data for the below analysis have been taken from the J.P. Morgan Macrosynergy Quantamental System (JPMaQS). These data measure the market information status of the latest instance of a macro-fundamental metric for any given day, not an ex-post (and often revised) actual state of the economy. For example, the rate of private credit growth on January 31, 2010, would be the latest such growth rate in the form in which it was available to the market at this time, even if the observed period was further in the past. And if it is an estimated concept, such as GDP growth trend, it would be the latest estimated based only on data available up to the date. Researchers and investors with access to J.P.Morgan Markets can gain access to the data and replicate the below analysis via API by following the instructions here.
Features and targets for the below analysis have been calculated based on quantamental indicators for 18 currency areas with liquid interest rate swap markets: AUD (Australian dollar), BRL (Brazilian real), CAD (Canadian dollar), CHF (Swiss franc), CLP (Chilean peso), COP (Colombian peso), CZK (Czech Republic koruna), EUR (euro), GBP (British pound), HUF (Hungarian forint), IDR (Indonesian rupiah), ILS (Israeli shekel), INR (Indian rupee), JPY (Japanese yen), KRW (Korean won), MXN (Mexican peso), NOK (Norwegian krone), PLN (Polish zloty), (Swedish krona), THB (Thai baht), TRY (Turkish lira), TWD (Taiwanese dollar), USD (U.S. dollar), ZAR (South African rand).
In particular, we consider three standard sources of macro trend pressure:
Excess inflation: The basis for this indicator is the local headline consumer price index (CPI), adjusted in real-time for seasonal factors and suspicious jumps and spikes. The inflation rate here is measured in accordance with various market conventions: the annualized rate of the past 3 months over the previous 3 months, the annualized rate of the past 6 months over the previous 6 months, and the simple rate over a year ago. Then we take a simple average of these series and subtract the concurrent effective inflation target. View more information on these indicators view the JPMaQS documentation here and here.
Excess economic growth: The basis for this indicator is the intuitive GDP growth trends dataset of JPMaQS (view here). For each point in time, they are the latest estimable GDP growth trend (% over a year ago, 3-month moving average) based on actual national accounts and monthly activity data, using sets of regressions that replicate conventional charting methods in markets. The proxy for an excess growth trend is then the difference between this trend and the median reported GDP growth trend over the previous 5 years.
Excess private credit growth: The basis of this indicator is private bank credit at the end of the latest reported month, % change over a year ago, seasonal, and jump-adjusted (view documentation here). From this rate, we subtract the sum of the effective inflation target and the medium-term real GDP growth trends, i.e. the benchmarks used for excess inflation and excess growth above.
Simple macro trend pressure and IRS receiver returns
History since 2000
For the empirical analysis, we simply add up excess inflation, economic growth, and private credit to arrive at a primitive composite macro trend pressure indicator. This is surely not the optimal or most plausible way to combine the three components, as most central banks would put greater emphasis on inflation than on credit growth. However, simplicity trumps sophistication for the purpose of delivering a proof of concept that is not suspect of data mining. The below analysis is not a recommendation of a trading strategy per se, but merely a documentation of the general predictive power of macro trends.
Composite macro trend pressure has on average been slightly positive for most countries since 2000, due to the prevalence of deflationary. Emerging countries post greater fluctuations in macroeconomic trends and Turkey, in particular, stands out as producing large variations.
We look at the predictive power of real-time information states of macro trend pressure with respect so subsequent 2-year interest rate swap receiver returns. This is done separately for the two largest currency areas, the U.S. and the euro area, and the 16 smaller countries. The separation reflects that the big two tend to dominate global trends and have a great influence on the other markets. The smaller markets’ dependence on the larger markets means that their idiosyncratic macro trends should on average be less influential.
U.S. and the euro area
In the large currency areas, there has been significant predictive power of macro trend pressure on subsequent interest rate swap receiver returns. A sizeable negative correlation of 10% or more prevailed in both countries and across the decades of the sample. The negative correlation has been driven mainly by excess growth and inflation, in accordance with the intuition of the popular Taylor rule.
Accuracy (rate of correctly predicted market return direction) has been very high by the standards of macro trading signals over the past 22 years, at 57-58% at a monthly frequency. The same is true for balanced accuracy, which reflects that the macro trend signal predicted both positive and negative returns well. In fact, the ratio of correctly predicted positive returns was 60% and the ratio of positively predicted negative returns was 54%.
The evolution of macro trend pressure signals has been rather gradual over the past 22 years, in accordance with business and credit cycles. The signal would not always have called for the same direction of positions in the U.S. and the euro area. Since the two markets have usually been strongly positively correlated this means that simple trading strategies based on macro trends would result in large intertemporal shifts in position risk (VaR).
Other currency areas
The correlation between macro trend pressure on rates and subsequent swap receiver returns has also been negative in the 16 smaller currency areas. The Pearson correlation coefficient has been a bit smaller but owing to the abundance of data its conventional significance statistic has been near 100%.
The predictive power of country-specific macro trends should, all other things equal, be less for small countries, as their local markets depend disproportionately on returns in the U.S. and the euro area. Efficient prediction of small countries’ fixed income returns based on macro trends should use both local macro trends and those of the big two countries.
Accuracy has been materially lower when predicting local swap returns with local macro trend pressures alone in the smaller currency areas. This plausibly reflects two effects. First, as mentioned above, returns in smaller countries are disproportionately affected by the macro trends and other return factors of the large currency areas. Thus, U.S. markets strongly influence Australia, but not the other way round, Second, benchmarks for excess inflation and growth are much harder to estimate for some of the smaller countries for lack of data, particularly in the early part of the 2000s.
Even with smaller accuracy and without considering any of the U.S. and euro area signals, local macro pressure in the small countries is an indicator with great potential for boosting the performance of trading rules due to the abundance of signals and the diversification across markets. Sharpe ratios of the most naïve strategies have been between 0.5 and 1.0, versus 0.3 for a simple long-only portfolio, after imputation of reasonable transaction costs.
Rate-pressure gaps and IRS receiver returns
An obvious complement to the macro trend pressure is the concurrent state of the real interest rate, here the real 2-year IRS swap yield. For the calculation see the JPMaQS documentation here. For example, positive macro trend pressure should have a stronger negative effect on rates markets if concurrent real yields are low or negative because the required extent of tightening would be larger. By contrast, macro trend pressure may have little or no impact if the concurrent real yield is already very high by historic standards.
For the analysis below we merely subtract the composite macro trend pressure from the real interest rate, as both are denominated in % annualized. Again, this is brutally simple and can certainly be improved upon, but acceptable for delivering a proof of concept. We call this difference the rate-pressure gap. A large positive gap means that the real rate is high and the macro trend pressure is small or negative. This should bias policy towards easing and be positive for subsequent receiver returns. Over the past 20 years, rate pressure-gaps show both cyclical and longer-term dynamics. Time series for some EM countries start later, due to the limited availability of swap yield data.
U.S. and the euro area
The correlation of rate-pressure gaps and subsequent returns in the two large currency areas has been positive and significantly stronger than the correlation for macro trend pressure alone. The 28% Pearson correlation coefficient is extraordinary by the standards of macro relations. It was 21% in the 2000s and rose to 35% in the 2010s/202s to date. A fitted regression line goes almost exactly through the origin of the scatter plot, indicating that positive real rates predict positive returns as long as economic trends are well balanced.
Also, accuracy and balanced accuracy have been over 60% at a monthly frequency, which is extraordinary for macro trend signals. Moreover, rate-pressure gaps have performed almost equally well as predictors of positive and negative swap returns, with negative and positive precision of 63% and 60%. Above-50% accuracy was recorded in 80% of the calendar years of the sample period.
Other currency areas
The predictive power of local rate-pressure gaps has been significantly lower for the 16 smaller currency areas, with a correlation coefficient of “only” 12%. Again, this difference plausibly reflects the dominant influence of the U.S. and – to a smaller extent – the euro area, as well as differences in data quality. Also, real interest rates in many emerging market countries contain significant risk premia, which may compromise their information value with respect to the monetary policy stance.
Accuracy and balanced accuracy have been 55-56% for the panel and only failed the 50% threshold in Taiwan and Japan. The rates-pressure gap in the small countries would have implied a 55% long bias, unlike in the G2, where it had a slight short bias. This plausibly reflects the risk premia that are priced into local currency yields in EM. A more advanced version of the rates-pressure metric would adjust for the risk premium, i.e. estimate and subtract it, since it is not a reflection of monetary tightness.
