Goldman Runs Reality-Based Simulation, Sees Far Lower 2.5%-3.0% GDP Growth In H2 2011-H1 2012

Tyler Durden's picture

This is about as close as we are going to get an admission from Goldman that growth is about to fall off a cliff. The full announcement will come within the next couple of weeks (assuming Goldman is unsuccessful in its 4th upcoming crude Sell recommendation). "If sustained, the recent oil price increase could thus shave around ½pt off growth for the next two years. Moreover, we showed that further increases from current levels could have increasingly detrimental effects on growth, as “nonlinear” effects might set in as oil prices reach new multi-year highs." And "In particular, the simulations suggest that growth in the second half of this year and the first half of 2012 would only range between 2½% and 3%, well below our forecast of 3½-4%. The implied effect on Q2, however, is small given the lags involved, and the gas price drag weakens notably in the second half of 2012." Must. Not. Cut. Q2. GDP.

From Sven Jari Stehn of Goldman Sachs:

Gasoline and Growth: Gauging the Hit (Stehn)

The sharp rise in energy prices has emerged as the most important downside risk to our US growth outlook. Relative to our forecast—which already assumes a significant increase in energy prices—West Texas Intermediate (WTI) crude oil prices are now up by around 25%. Brent crude oil prices have risen even more sharply.

Our analysis suggests that the crude oil price surge could act as a significant drag on growth. Specifically, our estimates imply that a 10% increase in WTI prices—if sustained—could shave around 0.2 percentage point (pt) off real GDP growth for two years; that is, the level of real GDP would be lower by 0.2% and 0.4% after one and two years, respectively. (For details, see Sven Jari Stehn, “Rising Oil Prices—So Far, Only a Modest Hit to Growth,” US Daily, February 23, 2011.) If sustained, the recent oil price increase could thus shave around ½pt off growth for the next two years. Moreover, we showed that further increases from current levels could have increasingly detrimental effects on growth, as “nonlinear” effects might set in as oil prices reach new multi-year highs. (See “Oil, Growth, Inflation, and the Fed,” US Economics Analyst, March 11, 2011.)

It may, however, be preferable to focus on the growth implications of surging gasoline rather than crude oil prices. First, gasoline prices have a more direct effect on economic activity because by far the most important economic use of crude oil in the US nowadays is as transportation fuel. Second, it is much clearer which gas price to use—namely the US average retail gas price—than which crude oil price to look at. Finally, it currently makes an important difference whether we look at gas or oil prices. Seasonally-adjusted retail prices are now higher than their previous peak in July 2008, having risen significantly more than one would have expected based on the increase in WTI prices. (This is less clear for Brent prices, which have increased more sharply than WTI prices.)

In today’s comment we therefore modify our existing model to analyze the growth effect of gasoline rather than crude oil prices. Specifically, we construct a quarterly model that explains the (annualized) growth rate of real GDP with the quarter-on-quarter percentage change in the average US retail gasoline price. To take into account that GDP growth is related to other factors than just gas prices, we also include as “control” variables the change in our financial conditions index and CPI inflation. As in previous work, we focus on a sample from 1985Q1 to 2011Q1. Once the model is estimated, we trace out the response of GDP growth to a shock in gas prices. (To do so, we need to take a view on the relative speed at which gas prices and growth respond to each other. Specifically, we allow growth to respond to gas prices within the same quarter but assume that gas prices do not react to growth within the same quarter. The results below, however, are robust to ordering growth and gas prices the other way around.)

The exhibit below shows the estimated response of (annualized) real GDP growth to a 10% gas price shock. For ease of comparison with our previous estimates, we also include the estimated response of growth to a crude oil price shock. The results suggest that a 10% gas price shock dampens GDP growth by around 0.3pt for two years—that is, the level of real GDP is 0.3% lower after one year and 0.7% after two years. Consistent with the view that gas prices matter more directly for economic activity, these estimates are larger than for a similar increase in crude oil prices.

What are the potential implications for our growth outlook? Relative to our December forecast—which was for a wholesale gasoline price forecast of $2.63/gallon at a 6-month horizon—gas prices have risen by approximately 20%. To gauge the magnitude of the downside risk to our growth outlook, we consider two scenarios.

First, we assume that the recent spike in gas prices mostly persists, giving back only 40% of its recent increase. (Note that this is the path taken by a “typical” shock in our statistical model.) Thereafter, we assume that the gas price will trend up in line with December’s commodity price forecast, which would be consistent with the current commodity price forecast under the assumption that Libyan oil production returns to the market within the next 12 months (see dashed black line in the exhibit below).

These assumptions would imply a substantial hit to GDP growth relative to our current forecast (see dashed black line in the exhibit below). In particular, the simulations suggest that growth in the second half of this year and the first half of 2012 would only range between 2½% and 3%, well below our forecast of 3½-4%. The implied effect on Q2, however, is small given the lags involved, and the gas price drag weakens notably in the second half of 2012.

Second, we assume that the recent run-up in gas prices is transitory and that prices fully revert to the December forecast path by the end of 2011 (see grey dashed line in the first chart above). These more benign assumptions imply a considerably smaller hit to growth. In particular, the simulations suggest that growth would range between 3-3½% in 2011H2 and 2012H1 (grey dashed line in second chart above). Given the assumption of a substantial drop in gas prices from current levels in this scenario, the implied growth rate exceeds our 3½% growth forecast in the second half of 2012.

Our analysis therefore suggests that the run-up in gas prices poses an important downside risk to growth, concentrated in the second half of this year and the first half of next year. Gauging the quantitative effect depends on whether the spike in gasoline prices is persistent or transitory.