While Lee Oliver, Euromoney’s FX correspondent, is on summer break, the weeklyFiX is supplied by guest writers from the industry. Our third contributor is By David Woo, managing director and head of global FX strategy at Barclays Capital.
The correlation between the two, which has risen dramatically since the start of the credit crisis last July, has recently climbed to its highest level in a decade. While this correlation has crucial implications for investors and policymakers alike, there is no consensus on either the drivers behind its recent increase or the sustainability of its current level (the correlation by itself tells us nothing about the direction of causality or the nature of the relationship).
In this article, we aim to contribute to the debate by employing the vector autogression (VAR) methodology to better understand what is driving this important correlation. Our findings are consistent with our long-held view that causality between the USD and oil prices runs in both directions. Moreover, they suggest that the difference in the policy reaction function between the ECB and the Fed to rising energy prices since last year has likely magnified the correlation. These results reinforce our view that a USD recovery will require either an exogenous sharp decline in oil prices or a meaningful shift of the focus of the Fed from growth to inflation.
We believe rising oil prices and falling USD have become mutually reinforcing, forming a vicious circle:
1. Rising oil prices have been pushing the USD lower against the EUR. The fact that the US is a highly energy intensive economy means that higher oil prices lead to a much worse inflation-growth trade-off than for most other economies. The dual mandate of the Fed means lower US real interest rates in the face of rising energy prices, which, in turn, lead to lower USD.
2. Falling USD is helping drive oil prices even higher. Investors increasingly view commodities as a hedge against further USD decline. Oil producers want to be compensated for the lower USD, but USD depreciation increases the purchasing power of non-US oil consumers.
So is there any evidence to support our hypothesis of the existence of a vicious circle? To answer this question, we run a reduced form VAR on three key variables of our circle: EUR/USD, oil prices (two-month WTI futures), and differential in two-year real yields of zero coupon swaps between the eurozone and the US. What we hope to find is evidence that: 1) higher oil prices have led to an increase in the differential of real interest rates between the eurozone and the USD; 2) the increase in interest rate differential drives EUR/USD higher; and 3) an increase in EUR/USD pushes oil prices higher still. We will use daily data and focus on two sample periods: July 2007 to July 2008 and July 2005 to July 2007. Given all our variables are highly non-stationary, the VAR is run on first differences. The Schwartz and Hannan-Quinn information criteria suggest that the optimal lag for the VARs is one day.
Impulse response analysis for the 2007/08 sample period shows that indeed oil prices drive interest rate differentials, interest rate differentials drive EUR/USD, and EUR/US drives oil prices. The directions of the responses are in line with our prior assumptions; the impact is immediate (and starts to diminish generally after the first day) and statistically significant. These results are consistent with the existence of our vicious circle. Moreover, these results indicate that a 1% positive shock to EUR/USD leads to a 1.2% increase in oil prices while a 10% shock to oil prices leads to a 1% increase in EUR/USD. Reinforcing these results is the fact that on days when we have seen big movement in one of these three prices, the other two often move in locked steps. A good example is June 6: following the ECB’s signal that they were considering hiking rates at the July meeting, EUR/USD, oil prices and interest rate differential all rose sharply.
The results from the 2005 to 2007 period make interesting comparison with those of the 2007/08 period. The impulse responses for the main channels are still statistically significant, but the magnitude of the responses is much smaller. In particular, the magnitude of the response of interest rate differential to oil prices is less than half of what it is for the 2007/08 sample, and the response of oil prices to EUR/USD is less than one-third of the more recent sample. In our view, the former result likely reflects the different approaches of the Fed and the ECB to dealing with the credit crisis, which has resulted in a divergence of the way the two central banks have reacted to the oil-driven inflation shock over the past year. This divergence probably contributed to the latter result, which is consistent with the view that the aggressive easing of the Fed has pushed investors into commodities as insurance against inflation and further decline in the USD. One interpretation of our findings is that the Fed easing has played a non-trivial role in the recent surge in oil prices.
The existence of a feedback loop between rising oil prices and falling USD has three key implications: 1) The loop, by perpetuating the co-movements of the two prices, is probably one reason the USD downtrend and oil uptrend have been so strong and stable. This means that investors should think very carefully about taking positions against these trends; 2) The loop could lead to overshooting of the two prices from their fundamental equilibrium levels, suggesting that the correction could be significant when these trends finally break; and 3) A big enough exogenous shock to the system could lead to a reversal of the feedback loop, and the vicious circle could give way to a virtuous circle of strengthening USD and falling oil prices.
To us, an exogenous shock could be an accelerated recoupling of the rest of the world, especially China, with the US slowdown that results in lower global oil consumption and lower oil prices or a policy response to break the oil-USD link. A global equity meltdown that accelerates the recoupling of the rest of the world with the US slowdown leading to demand destruction for oil will probably also do the trick. While there are signs that the Chinese economy is beginning to slow, data do not yet point to a sharp slowdown. Moreover, while the desire of Washington to see a stronger USD has raised the probability of a policy response to limit further USD depreciation, the falling stock market and the approach of the US election may deter policymakers from tackling the vicious circle aggressively. A further correction of the equity markets is probably the best chance the USD has to stage a meaningful recovery.