全球智库信息集成服务系统

Two things are coming to rattle the stellar post-crisis performance of emerging markets in 2010 – a tightening cycle in the US that the consensus expects to be limited and to take place late in the year (Wall Street Journal 2009) and, most notably, a steepening of the US Treasury yield curve driven by the combination of a growing risk appetite and the undoing of quantitative easing (Bloomberg 2009 and IMF 2009). These are two elements of the high US rates–strong dollar scenario we characterised in a previous column that is already starting to materialise.

While most emerging markets have undergone structural changes that reduced their exposure to a global tightening (see Levy Yeyati et al. 2009), it is only natural, with spreads in many cases close to historical lows, to expect some headwinds from the shift of the US yield curve. Here, we try to shed light on the link between US rates and emerging market spreads and to quantify the impact of plausible future scenarios for the US yield curve over the financing costs in the emerging world.

Relating emerging market spreads to the US yield curve

In a recent paper (Levy-Yeyati and Williams 2009), we build on a simple model of sovereign spreads to disentangle the different channels through which the US rates curve may affect emerging market spreads and assess their economic importance.

Higher US rates may reflect very different things. Even at a simplified level, we can distinguish at least two supply-side stories and one demand-side story behind an upswing in the US yield curve. On the supply front, monetary tightening may shift up the curve and crowd out capital flows to emerging economies (a channel documented early on by Calvo et al., 1994). Related to the former but more relevant for longer durations, rates may widen due to an increase in the net supply of Treasuries to the market (e.g., if the undoing of quantitative easing and a revived borrowing from high-grade corporates start crowding out the growing financing needs of the US government).

In turn, on the demand front, a decline in the preference for money or money-like instruments (and for riskless assets in general), driven by a renewed appetite for risk and duration, could steepen the curve regardless of the pace of monetary tightening. We proxy the latter by the share of government-only money market funds over total money market funds (to capture the preference for liquidity, see Figure 1) and the VIX index to capture perceived market volatility (to capture aversion).1

Figure 1. Which case are we in?

Source: Bloomberg, Barclays Capital

A first, intuitive way to represent the problem is through the exponential spread-rating link illustrated in Figure 2, which for a given time t could be formally expressed as:

1n (spreadit) =

t + tratingit+ t

Regressing this cross-country equation for different periods (in our case, months), we can extract a series of estimated

t and t (where an increase in the former or a decline in the latter reflect rotation and steepening of the spread-rating curve), and plot them against the Fed funds rate and the short UST yield (Figure 3). The advantages of this exercise (as opposed to a comparison between US yield curve and emerging market spreads) is that it allows as to focus on the effect of US rates above and beyond the ratings up cycle that characterised the emerging world in the 2000s.

Figure 2. Spreads and ratings over the cycle

Note: The highest rating is assigned the number 1 and each notch down is associated with an additional 0.5.
Source: Bloomberg, Barclays Capital

Figure 3. Emerging markets spread-rating: The spread rating curve during the US rate cycle (add cases identified in Figure 2)

Source: IMF, Barclays Capital

While still a crude simplification, the figure highlights two stylised facts. First, a widening of the short end of the US yield curve is associated with a steepening of the emerging markets spread-ratings curve, possibly a reflection of growing risk factors rather than monetary policy. Second, a normal tightening cycle such as the one prior to the recent crisis does not seem to exert much influence on the spread-ratings curve, possibly because, in the absence of financial distress, higher US rates are typically associated with lower risk premiums (in line with the results below).

In order to assess more rigorously the channels through which the Treasury curve influences emerging market spreads, we build on González Rozada and Levy Yeyati´s (2008) model of sovereign spreads. More precisely, to the Treasury rate (proxied here by the yield of Barcap’s 7y10y UST index, 7y10y), the sovereign rating (rating) and the US high yield corporate spread (HY), we add a monetary policy proxy (the federal funds rate, Fed), and two risk aversion proxies: for the demand for liquidity (the government to total money market fund ratio, MM), and the demand for safe assets (the VIX index). Finally, we include the international reserve-to-GDP ratio (reserves), which has been shown to systematically reduce spreads beyond what is captured by ratings (Levy Yeyati, 2008).

Importantly, because we want to identify the effects associated with each of the three channels mentioned above, we proceed step-by-step (Table 1). The first specification tells us that US rates increase less than proportionally (i.e., the curve flattens) as a results of a Fed hike, and that they are affected differently by the preference for liquidity and risk aversion (negatively in the first case, positively in the second).

In turn, the second specifi