February 6, 2017

The mission of Exante Data™ is to provide better answers to key macro questions through superior data and analytics. Over the last several years, one of the key global macro uncertainties has been China. The trajectory of China’s economy is probably the single biggest driver of global growth. And even if Trump’s tweets seem to get all the attention currently, China remains one of the biggest unknowns in the global economic outlook. At the same time, the official Chinese data is not always very informative or timely (some would use stronger language here!). Hence, there is a huge information gap to fill.

Over the last year, we have been working hard on developing a set of sophisticated indicators aimed at tracking key trends in China on a real time or ideally leading basis (ex ante is better than ex post!).

Our focus has been on capital flows, for two main reasons: First, the pulse of the Chinese economy, and sentiment more generally, will quickly be reflected in capital movements (in a way you can regard capital flows as a leading indicator). Second, capital flows, as opposed to GDP movements, are what ultimately move asset prices; and that is what investors should care about, whether they are focused on currency, rates or equities.

We have put all our data and models together in a product called China Flow Analytics™. This product embeds a number of unique proprietary analytics, including real-time estimates of China’s currency intervention and leading indicators of Chinese trade flows.  But the product is not just a collection of statistics. We believe that superior insights come from combining the best possible data and analytics with a human touch, which can involve a more holistic assessment, including qualitative variables. Hence, the idea is to leverage data and analytics to the max, while also providing the broader narrative and context around the data, though the blog and thematic research embedded in China Flow Analytics™.

China Flow Analytics™ has been tested over the last six months with a group of expert users from leading buy-side institutions. Based on their feedback, we have done some final fine-tuning of the analytical content and the web-interface over the last two months. We are very excited to launch it for a bigger group this week. If China is relevant for your portfolio (it is for many!), this product may help you make better alpha decisions and manage risk better

We strongly believe in the analytical content and data integrity of our China Flow Analytics™ and are proud to make this letter available to all who carry risk and generate alpha. Please click here if you would like to learn more about the product and get familiar with its analytics.

Below, we include specific examples of how we have found the insights generated by China Flow Analytics™ helpful over the last year. Given the importance of China to the world economy and investment performance, we are sure China Flow Analytics™ will continue to be a valuable input in alpha generation and risk management in the months and years to come.

EXAMPLES OF INSIGHTS FROM EXANTE DATA CHINA FLOW ANALYTICS™

Over the past year, data and analysis from Exante China Flow Analytics have been extremely useful in catching several trend and trend-breaks. We list some examples below.

• Intervention break in January 2017: In early January, our real-time model detected a break in the onshore spot intervention trend by the PBOC. This break turned out to be crucial for USDCNH dynamics in the rest of the month, as well as the trend in the dollar versus a number of Asian currencies. It very likely played a role global USD dynamics observed so far in 2017. Finally, the lower intervention will have implications for the trend in FX reserves, which had been sharply down since September 2016.
• Currency pegging ahead of SDR inclusion: Going into the SDR inclusion date (October 1, 2016), our models picked up a meaningful increase in FX intervention by the PBOC. Perhaps more interestingly, this FX intervention was happening at almost exactly the same level for about two weeks. Using our real-time data, we are able to look at the price levels at which FX intervention takes place (including at intra-day frequency) and that provides a more accurate picture of the specific price points which intervention is happening, allowing for clearer interpretation.
• Cross border loan payback: In commentary on 11 April 2016, we updated our analysis of cross-border banking flow data, both from a flow and balance sheet perspective. Our conclusion was that, given the cumulative reduction in cross-border loans to China already achieved, the FX outflow from China via this channel was nearing its end, and the process would likely run its course by June. This forecast turned out to be very accurate, and we there has been no meaningful negative flow effect of reduction in loan liabilities since then (see Chart).
• The Exante Data Real-time Intervention model leads official data on intervention:  The key advantage of the real-time model is that it is available much quicker than any information relating to intervention from official sources, which is released with a 5-7 week lag relative to when the intervention actually took place. In addition, data is available with a much greater degree of granularity (10-minute frequency) allowing for much more detailed analysis.The chart below compares a monthly aggregation of the real-time model with intervention numbers computed from official Chinese data (we use a holistic approach using four complementary methods for robustness). While the goal of our modeling is not to optimize the fit relative to any given benchmark, the chart does illustrate that in general the model tracks the trend in the official figures very well, with a correlation of over 93% since 2013.

In general, in a world where the default assumption is to extrapolate from the most recent data, our analytical tools helps in 2 key ways:

• The real-time nature of our models (including the intervention model) ensures that the starting point for extrapolation is correct, at a minimum.
• We anchor our discretionary flow analysis not just in a time-series approach, but also in a fundamental understanding of the agents that ultimately drive the flows (including their balance sheet constraints). That allows us to create limits and boundaries beyond which extrapolation is likely to be wrong.

Our approach is to look at the data holistically to understand the bigger picture. We have built tools to help identify general changes in real time, and we try to understand the changes in flows based on our monitoring of economic/financial/political changes. As the examples above demonstrate, this holistic approach has worked well in the past we are confident this will continue to be the case in the coming months.