FX carry + value + momentum strategy with over 200 years of history
At Quantpedia, we’ve mentioned many times that we love long-term historical analysis because it gives us a unique glimpse into different macro environments and periods of political and economic instability. These long-term studies greatly aid in risk management and also help investors set appropriate expectations about various future outcomes. While historical analysis of stock and bond markets is not uncommon, currency markets are less well researched. Therefore, we would like to investigate a recent paper by Joseph Chen that analyzes FX carry, currency momentum, and currency value strategies over 200 years of history.
Use an expanded data sample spanning over two centuries (!) and Long-term bonds and short-term interest rates as investment vehicles (Chen, 2024 revised edition) investigates the robustness of currency investment strategies. thoughtfully executed Panel regression analysis and portfolio return Some interesting conclusions emerged:
- in FX carry before they They found that, except for the period around World War II, returns on carry trades would have been robust over time regardless of whether short-term interest rates or long-term bonds were used. Dollar carry trading has also been strong except for these periods and the pre- and post-Civil War periods.
- Although there is limited support for a currency momentum effect, using a longer sample we find that a currency reversal effect does not exist. Finally, if we examine currency investing based on the slope of the yield curve, we find that the cross-currency yield curve flattening trade has been surprisingly robust over the years.
- This study revealed some surprising stylized facts. In contrast to Lustig, Stathopoulos, and Verdelhan (2019), long bond carry trades show solid positive returns. There is no evidence that the term structure of currencies with trade risk premiums is declining.
- It seems that the fixed exchange rate system does not eliminate exchange effects such as carry trades. Rather, fixed exchange rate regimes offer higher risk-adjusted returns because exchange rate volatility is lower during these periods.
Overall, we highly recommend this paper as weekend reading for all investors/traders who systematically collect factor premiums in currency markets due to the very long length of the covered sample.
author: joseph chen
title: Currency investments over the last few centuries
Link: https://papers.ssrn.com/sol3/papers.cfm?abstract_id=4692257
Abstract:
The literature on currency investing, such as carry trades, typically bases its analysis on the most recent period since 1983. I use over two centuries of data to analyze the risk and return characteristics of long-term currency investments across 21 currencies. In addition to using short-term interest rates as an investment vehicle, we also explore the use of long-term bonds. The risk premium estimates from these investments provide information about the volatility of the pricing kernel in the reduced-form pricing model. We find that returns on carry trades have been surprisingly strong throughout history and would have been robust to the use of long-term bonds. Additionally, cross-currency yield curve flattening trades generate solid returns with twice the Sharpe ratio of carry trades. These results provide a deeper understanding of the nature and sources of premiums in currency investments and the volatility of pricing kernels.
As always, we present some interesting figures and tables.
Notable quotes from academic research papers:
“In this paper, we build an extended historical dataset of foreign exchange rates, short-term interest rates, and long-term interest rates and design an empirical experiment that overcomes these technical difficulties. U.S. investments that allocate all funds to foreign investments Rather than constructing the dataset from a home perspective, we include investments in the United States as trivial zero excess return investments. Including the option to invest in the home currency makes this study neutral to the reference currency. In addition, by estimating capital gains from holding long-term bonds from changes in bond yields with some assumptions, we can construct long-term bond returns that total over 230 years. We have data covering currencies. We will leverage this dataset to construct currency investment portfolios based on different strategies and perform panel regression analysis to examine their statistical significance.
This historical perspective also provides further insight. Currency momentum effects are likely to be stronger when using longer histories, unless the sample goes far enough back in history. Based on data starting in 1788, currency momentum trading produced positive returns only in the second half of the sample, but not in the entire sample. Panel regression also fails to detect a strong monetary momentum effect. In the case of currency reversals, the effect appears to exist only within samples studied by others. There is no evidence that currency inversion trading has produced abnormal positive returns, except for the most recent period after the collapse of the Bretton Woods agreement. These results suggest that bias in data snooping may be an issue, or that there are underlying mechanisms that change over time that cause these effects.
In my data, the short rate is relatively stable compared to other variables, as indicated by the low volatility (standard deviation). Swiss Franc (CHF) has one of the most stable short interest rates (volatility 1.88%) and also has the lowest average short interest rate (average 2.82%). On the other hand, Portugal Escudo (PTE) has the highest volatility short rate (5.39% volatility) and also has one of the highest average short rates (6.26% average). Generally speaking, the lower the average short rate, the more stable the short rate is, and the correlation between the two is 0.66.
This result is in sharp contrast to Lustig, Stathopoulos, and Verdelhan (2019), who reported modest returns for carry trades using long-term bonds for a modern sample. However, my estimates of long-term bond returns differ from theirs in a number of ways, which may explain the differences. 17. In my analysis, returns on carry trades appear to be robust when using bonds with longer maturities. In the context of the reduced-form model and equation (17), long-term bond carry trades can be explained if low long-term bond interest rates reflect high volatility in the permanent components of a country’s pricing kernel. .
Finally, in Panel C of Table 3, we consider the returns of the signal-weighted carry-trade portfolio. Unlike an equal-weighted portfolio, a signal-weighted portfolio places greater portfolio weight on more extreme yields, similar to regression analysis. The resulting portfolio is qualitatively similar to the equal-weighted portfolio shown in Panel B, but with several notable features. The overall return is slightly higher, but the volatility is higher, resulting in a Sharpe ratio of 0.391 and 0.361, depending on whether short-term interest rates or long-term bonds are used as the investment vehicle. The long-short portfolio is slightly less negatively skewed, and it appears that much of the change is coming from the long-side portfolio rather than the short-side portfolio.
Momentum effect and reversal (value) effect are two additional currency investment strategies that have been studied in the literature and can be easily examined in my expanded sample. Menkhoff, Sarno, Schmeling, and Schrimpf (2012a) and Asness, Moskowitz, and Pedersen (2013) both find strong momentum effects in contemporary samples, where currencies that appreciated the most in the past 12 months tend to maintain high returns. I am reporting. . Menkhoff, Sarno, Schmeling, and Schrimpf (2012a) report that this effect is strongest when the holding period spans the next month. This is consistent with the strategy studied in Asness, Moskowitz, and Pedersen (2013). The latter study also demonstrates a reversal effect, where currencies with lower past long-term returns tend to return to higher returns. Menkhoff, Sarno, Schmeling, and Schrimpf (2017) report similar results based on currency appreciation over the past five years relative to changes in purchasing power. Sometimes called the “value” effect. All of these studies are based on observations in modern samples. In this section, we investigate the robustness of currency momentum and reversal effects over time.
Figure 6 shows the cumulative returns of currency reversals. Currency reversals for short-term interest rates and currency reversals for long-term bonds have only been positive since around 1980, when most prior studies began their data. Although there were periods of relatively flat returns when currencies were not freely floating, the general trend of returns from currency reversals was negative. Table 12 examines the returns of currency reversal portfolios in more detail. Panels A and B show returns using short-term interest rates and long-term bonds, respectively, and find similar results to the figure. ”
Looking for more strategies to read? Sign up for our newsletter or visit our blog or screener.
Want to learn more about Quantpedia Premium Services? Find out how Quantpedia works, our mission, and our premium pricing plans.
Want to learn more about Quantpedia Pro services? Read the description, watch the video, check out the reporting features, and access pricing offers.
Looking for historical data and backtesting platforms? Check out our list of algo trading discounts.
Or follow us:
Facebook groups, Facebook pages, twitterLinkedin, Medium, or Youtube
Share withlinkedintwitterFacebookintroduce to a friend