Since the start of the year, growth in emerging countries has held up quite well. This is reflected not only in business and household confidence, but also in the confidence of foreign investors in the local bond and stock markets. The tightening of US monetary policy from early 2022 to mid-2023 did have a major negative impact on portfolio investment flows. However, this impact was largely offset by the attractiveness of emerging markets for both private and institutional investors, whether for purely financial reasons (carry trade strategies) or as part of a diversification strategy. The forthcoming monetary easing in the United States should, in this way, should lower local bond yields and, in so doing, curb the rise in interest charges on public debt in a large majority of countries over the recent years.
Since the start of the year, growth in the emerging world has help up quite well, and has even strengthened in several countries. In Q1 2024, the aggregate GDP of our sample of 26 EM countries continued to grow at a rate of just over 1% q/q for the third consecutive quarter. Over a year, the increase stood at 4.7%. Excluding China, the global picture is the same. However, unlike China, the main emerging economies have not benefited, or at least not to the same extent, from the recovery in exports.
Whether it is an explanatory factor in per se mirrors other factors, business, investor and, to a lesser extent, household confidence has improved, supporting domestic demand and limiting the impact of US monetary tightening and local monetary policies on domestic interest rates.
Renewed confidence among households, businesses and foreign financial investors
PORTFOLIO INVESTMENTS NET INFLOWS TO THE MAIN EMERGING MARKETS (USD BN)During Q2, the PMI indices consolidated or remained well above 50, the threshold above which manufacturing activity is considered to be expanding. The only exception is South Africa, where the PMI stagnated.
Household confidence presents a more mixed picture, with a clear improvement until the last few months for Central European countries, and, in the majority of cases, a levelling off since the end of 2023 following a clear upturn in 2022-2023, in line with i) the slowdown in inflation, often coupled with monetary policy easing, ii) the wage catch-up and iii) employment growth. In China, however, household mood has remained low since the start of the property crisis, with the after-effects still being felt (see below). In Argentina, household confidence has plummeted since the start of the year as a result of the shock therapy inflicted on the population by President Milei in an attempt to curb hyperinflation (see below). In South Africa, it is stagnating below the waterline against a backdrop of a near recession. Finally, in Turkey, household confidence is in between, with an improving trend since 2022, but it is still at a historically low level and, above all, has a very erratic profile.
Foreign investor confidence in the local stock and bond markets is holding up well, and has shown surprising resilience/capacity to bounce back after external real or financial shocks over the last four years. This has also been the case over the last six months, despite the heightened geopolitical risks and the revision by financial markets of the timing and the scale of US monetary easing.
According to balance of payments data, net portfolio investment inflows from non-residents (in debt securities and equities) into the main emerging markets (see chart 1) stood at USD 179 bn from November 2023 to May 2024 (USD 95 bn excluding China) and USD 277 bn since the US Federal Reserve (FED) first raised its key rate in March 2022 (USD 338 bn excluding China). However, US monetary tightening has been particularly pronounced (+525 basis points between 17/03/2022 and 27/07/2023).
Could it be that the emerging financial markets have become less sensitive to monetary tightening on the other side of the Atlantic?
US monetary policy and portfolio investment in emerging countries: a very strong causal link remains
RESIDUALS OF EQUATIONS OF PORTFOLIO FLOWS TO EM (IN STANDARD DEVIATIONS OF INVESTMENT FLOWS)According to the IMF, a 100 basis point rise in US long-term bond yields leads to an equivalent rise in domestic interest rates in emerging countries after two years[1]. We can therefore assume that US monetary tightening has had a significant knock-on effect on portfolio investment flows in debt securities by non-resident investors (commonly known as the "flight to quality"). However, this negative impact has clearly been more than offset by other factors.
In order to verify these assumptions, we modelled and simulated the (net) inflows of investments in debt securities by non-residents using a simple linear equation, estimated on a panel of 16 countries observed over the period Q1 2012-Q4 2019, with the dependent variable being investment flows (as a % of GDP) and the explanatory variables being 1) quarter-on-quarter real GDP growth, 2) a measure of the carry trade between government bond yields in local currency and US government bond yields, 3) the FED funds rate and 4) a time trend (the same for all countries)[2].
This equation reproduces investment flows satisfactorily over the recent period (applied to the aggregate variables for the 16 countries, the simulation of the equation gives weak residuals that are not very autocorrelated since 2021 - see chart 2).
According to our equation, the latest US monetary tightening would have generated significant investment outflows since 2022: -3.6% of GDP cumulatively over the period Q1 2022-Q1 2024, or USD -384 bn, compared with observed cumulative flows of USD 51bn. Changes in the carry trade variable suggest that around USD 34 bn of capital inflows would have resulted from purely speculative arbitrage transactions. Finally, the contribution of real GDP growth was practically zero.
Therefore, it was the time trend that essentially offset the outflows resulting from US monetary tightening. This variable, which serves as a control variable in the equation, probably captures investors' diversification strategies in favour of emerging markets, against a backdrop of growing financing requirements for governments and companies on their local bond markets. This is particularly true for institutional investors (sovereign wealth funds and central banks), who have very different motivations than carry traders. The confidence of non-resident investors is also reflected in their portfolio investments in equities. The econometric estimate of investment flows, using the same methodology as the one used for debt-security investment flows, but with a slightly different set of explanatory variables, is of a much lower quality[3]. Over the recent period, the equation significantly overestimates and then underestimates the trends observed (see chart 2). A financial stress variable specific to the COVID crisis would help to reduce these discrepancies[4]. However, beyond the econometric caveats, the impact of US monetary tightening was as strong as for debt investment flows, as the value of the elasticity to the Fed Funds rate is very close in both equations and investment flows were of an equivalent amount at the end of 2019.
In 2024, the expected easing of US monetary policy should generate portfolio investment inflows to emerging markets and therefore, all other things being equal, should lower bond yields, even if this driver is, in principle, less powerful than it was previously[5]. Households and businesses could in turn benefit if this fall is passed on to bank lending rates. Above all, it would help to curb the rise in interest charges on public debt that we have seen in recent years.
Article completed on 9 July 2024.
[1] Fiscal Monitor - April 2024
[2] The sample of countries includes Brazil, Chile, Colombia, Mexico, India, Indonesia, Malaysia, the Republic of Korea, Taiwan, Thailand, Hungary, Poland, the Czech Republic, Turkey, Russia and South Africa. It does not include China because, compared with other countries, portfolio investments are still being highly regulated. The period begins in 2012 only because of lack of data for yields and deliberately stops in Q4 2019, as the data beyond this date are affected by purely exogenous external shocks (COVID crisis and the war in Ukraine). However, extending the time horizon for the sample at the cost of introducing two additional time indicator variables does not improve the accuracy of the estimate. The carry trade variable is the bond yield spread divided by the volatility of the exchange rate against the USD (calculated over a 5-year period). The parameters were estimated by applying the ordinary least squares method to data in deviation from individual mean values (Within estimator) in the simplest dynamic form (i.e., with an endogenous variable with a one-quarter lag). Estimates by instrumental variable (which is in principle recommended for dynamic models) are not necessary given the low autocorrelation of the residuals. GDP growth and the carry trade variable were introduced with a one-quarter lag in order to prevent any simultaneity bias. Only the FED funds rate is introduced into the equation as a contemporary variable. The time variable serves as a control variable. In the end, the parameter estimates give a value of -0.11 for the Fed funds rate, 0.5 for the carry trade variable and 0.05 for real GDP growth. The impact of this last variable is very small.
[3] The carry trade variable is replaced by the government bond yield (in local currency) as a proxy for the discount factor. The overall quality of the estimate is much lower than the estimate for the previous equation; the average residual, measured as the standard deviation of the endogenous variable, is 1.2 compared with 0.7 for debt-security investment flows. The estimate gives a value of -0.14 (statistically significant) for the Fed funds rate, -0.07 for the bond yield (statistically non-significant) and 0.01 for real GDP growth (statistically non-significant). As with the debt flow equation, the time trend coefficient is positive and statistically significant.
[4] However, this is to the detriment of the frugality condition imposed by the finite size of the sample in its temporal and, above all, individual dimensions. This is a second reason why we have restricted the estimate to the pre-Covid period.
[5] The share of government bonds denominated in local currency held by non-resident investors has fallen since (at least) the middle of the decade, from around 25% to 15%.
