Daria Parkhomenko:
Hello everyone and welcome back to another episode of the RBC Emerging Markets Report. My name is Daria Parkhomenko and today I am joined by my colleagues Alvin Tan and Daniel Rico and we will dive into two short topics: EM Central Bank flows and the vulnerability of EM FX to investor flows.
Daria Parkhomenko:
Hi Alvin, You recently wrote a piece about the accumulation of EM Asia reserves. Can you please tell us about the model and what is the difference between the fundamental & non-fundamental drivers of reserve accumulation?
Alvin Tan:
In general, larger trade surpluses will tend to generate higher reserves, all else being equal on the other hand reserves may also grow if a central bank decides to intervene in a market to sell its currency and accumulate US dollars. This latter case would be a non-fundamental cause of reserves accumulation. I wrote a piece to try to estimate if Asian reserve accumulation in recent months was much more than expected, given shifts in the trade balance and other factors, and also how this might differ from country to country. For this purpose I used a model to estimate how much reserve accumulation had been driven by the trade balance, USD valuation and the value of likely reserve investments in US Treasury bonds.
The model I used is a Vector Autoregression (VAR) model for time series data, basically where every variable in the country-specific model is being simultaneously estimated by the other variables, and their respective lagged values in individual simultaneous regressions. From the model, I extracted the predicted monthly values of two variables for the May through December 2020 period, namely the trade balance and the reserves accumulation, both measured in US dollars. The sum of these predicted values were then compared to the actual realised values.
Daria Parkhomenko:
What do the results show? Are there certain countries that have been more active than others when it comes to intervention activity in Asia?
Alvin Tan:
The findings were that the trade balances generally did not differ much from the realized values, with the ratio between the total actual trade balance and the total predicted trade balance mostly close to 1 for the countries in the sample.
On the other hand, the ratio of the total actual reserves accumulation between May and December 2020 and the predicted total reserves accumulation were much higher than 1 for most countries in the sample. In other words, the model did a decent job predicting the trade balance, but a really poor job predicting the reserves accumulation. It basically under predicted the reserve accumulation in most cases. This suggested that a significant amount of the increase in reserves over the May through December period was driven by non-fundamental factors, especially in India, South Korea, Singapore and Taiwan.
This implied there was significant intervention activity by these central banks and this also means one should temper the still generally positive FX outlook for these currencies, as it indicates determined official resistance against currency appreciation.
Daria Parkhomenko:
Given the sizeable implied intervention activity that you just mentioned, do you think the phenomenon of USD recycling is back as a factor in FX markets Alvin?
Alvin Tan:
While it may be important to distinguish between fundamental-driven reserves growth and non-fundamental growth, reserves are still reserves in the end. The pace of annualised Asian reserves growth between May 2020 through February 2021 reached almost US$600bn. That is a rate that is almost comparable to the previous period of high reserves growth in 2009 through mid-2014, when it reached an annualised pace of $615 billion a year among the Asian countries. That said, the oil exporters are not yet a factor in the reserve accumulation story, so we are only speaking of the Asian central banks at his point. Thus, overall reserve accumulation globally is still lagging the pace that was seen a decade ago.
Daria Parkhomenko:
Thank you Alvin. In contrast to some of the EM central banks that have been accumulating reserves, we have seen large US ETFs investing in Emerging Markets, both debt and equity at a total of about 10bn US dollars as of February YTD. But despite the strong inflows, EMFX was down on average 2.5% year to date as of February on a total return basis. Daniel, you recently looked at the asymmetric response of EM FX to investor flows. Can you tell us more about it?
Daniel Rico:
Hi, Dasha. Investor flows and Central Bank flows are different sides of the same coin in FX. While the 2021 consensus view was for a weak dollar and a large rebound in EMFX, we have actually experienced the contrary. Central Banks as Alvin highlighted have been accumulating reserves effectively adding a limit to where some of this currencies can move. A good example of this is Chile where the Central Bank announced a reserves accumulation program once the currency rallied below the 700 level. In the last 10y, EM ETF flows have accounted for billions of dollars into Emerging Markets local currency debt and equity. However, with the rise of China as an asset class and low share of foreign holdings of Chinese assets we have seen an uneven distribution of flows this time around. With investors going overweight Asia vs underweight Latam.
Daria Parkhomenko:
The recent increase in US yields, has challenged the weak dollar consensus view and the price action over the past few weeks suggests that sudden changes to the FED’s communication or any of their programs could trigger a large taper tantrum. How do you think flows will play out in such a scenario?
Daniel Rico:
Inflows in EM markets are important in creating a wealth effect in EM economies as local financial assets appreciate. Intuition suggest that inflows should results into EMFX appreciation but history shows that EMFX appreciated only 50% of the times during periods of net inflows, while depreciated 70% of the times during outflows. Clearly the reaction function is quite asymmetric given the market structure of EMFX and when investors want to leave, the door is much smaller. On average a large outflow could have 4x larger impact than inflows of the same magnitude. But not all EMs are the same and according to our analysis EM Asia is much less sensitive to this change while Latam and South Africa are quite sensible. This, explains in part why we are seeing an uneven distribution of flows to EM Asia and an underweight position in Latam.
Daria Parkhomenko:
Thank you Daniel. And thank you everyone for listening. Have a wonderful rest of your day.