PRECIOUSMETALS：GOLDCYCLESETTOTURNONIMPROVINGUSRECOVERY1206.ppt

December 5,2012GlobalPrecious MetalsCommodities ResearchGold cycle set to turn on improving US recovery,Gold prices range bound in 2012 despite perfect set upGold prices have remained range bound in 2012,despite a steady declinein US real rates and rise in central bank holdings that would ordinarily besupportive.To understand this dislocation we expand our modeling of goldprices to include the impact of the US Federal Reserve easing.We find thatgold prices“look through”easing that does not require Fed balance sheet,Damien Courvalin(212)902-3307 Goldman,Sachs&Co.Jeffrey Currie(212)357-6801 Goldman,Sachs&Co.,expansion like Operation Twist increasing instead on announcements ofeasing through expansion on the Feds balance sheet.Improving US growth outlook offsets further Fed easingOur economists forecast that the US economic recovery will slow early in2013 before reaccelerating in the second half.They also expect additionalexpansion of the Feds balance sheet.Near term,the combination of moreeasing and weaker growth should prove supportive to gold prices.Mediumterm however,the gold outlook is caught between the opposing forces ofmore Fed easing and a gradual increase in US real rates on better USeconomic growth.Our expanded modeling suggests that the improving USgrowth outlook will outweigh further Fed balance sheet expansion and thatthe cycle in gold prices will likely turn in 2013.Risks to our growth outlookremain elevated however,especially given the uncertainty around thefiscal cliff,making calling the peak in gold prices a difficult exercise.Gold cycle likely to turn in 2013;lowering gold price forecastsWe lower our 3-,6-and 12-mo gold price forecasts to$1,825/toz,$1,805/tozand$1,800/toz and introduce a$1,750/toz 2014 forecast.While we seepotential for higher gold prices in early 2013,we see growing downsiderisks.As a result,we find that the risk-reward of holding a long goldposition is diminishing and recommend rolling our long Dec-12 COMEXgold position into a long Apr-13 position and selling a$1,850/toz call tofinance a$1,575/toz put to protect against a decline in gold prices.Since2009,this strategy achieved a better Sharpe ratio than a long gold position.Investors should consider this report as only a single factor in making their investment decision.For Reg AC certificationand other important disclosures,see the Disclosure Appendix,or go to,The Goldman Sachs Group,Inc.,Goldman Sachs Global Economics,Commodities and Strategy Research,2,December 5,2012,Global,Gold cycle set to turn on improving US recoveryAlthough up year-to-date,gold prices have remained range bound since October 2011,gyrating widely between lows near$1,550/toz and highs of$1,800/toz.This trading patternstands in sharp contrast to:(1)the remarkable trend higher in gold prices that occurredover the previous decade,(2)continued support from what we view as the two drivers tohigher gold prices:the steady decline in US real rates to new record lows and increase incentral bank gold holdings,and(3)the US Federal Reserve embarking on additionalquantitative easing through Operation Twist and QE3.The breakdown of these trendsraises important questions for the outlook of gold prices:what has been the impact of QEon gold prices?Has the relationship between gold and real rates ended?Is the recentstrong correlation between gold prices and the US dollar the new normal?And mostimportantly,is the multi-year rally in gold prices ending?To answer these questions,we expand our modeling of USD denominated gold prices onUS real interest rates and physical monetary demand for gold to include the impact of Fedeasing since late 2008.We find that(1)QE announcements matter most for gold priceswhile QE flows have little impact,(2)gold prices respond to easing through expansion onthe Fed balance sheet while looking through non-expansionary easing like Operation Twist,and(3)lower US real rates remain the key driver to higher gold prices when accounting forthe response of gold prices to the various iterations of QE.Looking forward,our economists forecast a slowdown in US economic growth in the firsthalf of 2013 with an acceleration in the second half of the year to trend growth levels.Theyexpect an announcement of further expansion of the Feds balance sheet at the upcomingDecember FOMC meeting as well as more easing than consensus in 2014-15.In the shortterm,the combination of more easing and weaker growth should prove supportive to goldprices although our modeling suggests that these catalysts are to some extent alreadypriced in.Medium term however,the gold outlook is caught between the opposing forcesof more Fed easing and a gradual increase in US real rates on better US economic growth.Our expanded modeling suggests that the improving US growth outlook will outweighfurther Fed balance sheet expansion and that the cycle in gold prices will likely turn in 2013.Risks to our growth outlook remain elevated however,especially given uncertainty aroundthe fiscal cliff,making calling the peak in gold prices a difficult exercise.Consequently,we lower our 3-,6-and 12-mo COMEX gold price forecasts to$1,825/toz,$1,805/toz and$1,800/toz,from$1,840/toz,$1,940/toz and$1,940/toz previously.We expectgold prices to average$1,810/toz in 2013 and introduce a$1,750/toz 2014 average goldprice forecast.Importantly,even under a weaker US recovery than our economists forecast,our modeling still points to only modest upside with gold prices reaching$1,900/toz late in2013.Net,while we see potential for higher gold prices in early 2013,especially should aresolution of the fiscal cliff remain elusive,we see growing downside risks.As a result,wefind that the risk-reward of holding a long gold position is diminishing.We insteadrecommend rolling our long Dec-12 COMEX gold position into a long Apr-13 COMEX goldposition and selling a$1,850/toz call to fully finance a$1,575/toz put to protect from adecline in gold prices.Since 2009,such a strategy has achieved a better Sharpe ratio thanan outright long gold position.Goldman Sachs Global Economics,Commodities and Strategy Research,3,December 5,2012,GlobalGold prices range bound in 2012 despite perfect set upOur forecast for higher gold prices over the past few years has been motivated by thecontinued decline in US real rates on weak US growth and aggressive easing by the USFederal Reserve and steady central bank gold buying.Over the past year however,USDdenominated gold prices have traded in a range-bound pattern between lows near$1,550/toz and highs of$1,800/toz(Exhibit 1).This stands in sharp contrast to 10-year TIPSyields declining by 1%to set new-record lows and central bank gold holdings increasing bya near-record 14 mtoz over this same period.Interestingly,gold prices have insteadexhibited a strong correlation to both the US dollar and the size of the US Federal Reservebalance sheet,putting our modeling of gold prices in doubt(Exhibit 2).We believe that the Feds unconventional easing since late 2008 has been the key catalystbehind these dislocations.While our initial approach was to assume that QE would impactreal rate levels and in turn gold prices,it is now apparent that while each QE iterationhelped push real rates lower,not all QE iterations ended up supporting gold prices.Forexample,the sharp decline in real rates that occurred during Operation Twist was in factaccompanied by a decline in gold prices with the September FOMC nearly setting the highsin gold prices.,Exhibit 1:Gold prices have remained range bounddespite sharply lower US real rates$/toz(left axis);%yield(right axis,inverted),Exhibit 2:exhibiting instead a strong correlation to thesize of the Feds security holdings$/toz(left axis);trillion of US dollars(right axis),2,250,-1.0%,1,850,3.1,2,0001,7501,500,-0.5%0.0%0.5%,1,7001,5501,400,2.92.72.5,1,250,1.0%,1,250,2.3,1,1001.5%,1,000,950,QE1,MBSReinv,QE2,Twist&,QE3,2.1,2.0%,Twist2,750Jan-09,Oct-09,Jul-10,Apr-11,Jan-12,Oct-12,8002009,2010,2011,2012,1.9,COMEX Gold price,10-yr TIPS yield(rhs,inverted),COMEX gold prices,FRB total assets,Source:Federal Reserve Board(FRB),COMEX.,Source:Federal Reserve Board(FRB),COMEX.,Not all QE created equalTo try to understand these interactions,we turn to the CFTCs COMEX gold net speculativelength data,as it has exhibited the most straightforward sensitivity to real rates over thepast decades,rising with declines in real rates.This relationship started breaking down in2009 with net speculative length increasing sharply around the announcement of QE1 andoutperforming the decline in real rates.In turn,net speculative length declined sharply inthe fall of 2011 around the announcement of Operation Twist(Exhibit 3).Understandingthese dislocations is key to our gold outlook as net speculative length is the primary driverof our gold price forecast.Goldman Sachs Global Economics,Commodities and Strategy Research,5,4,December 5,2012,GlobalExhibit 3:COMEX net speculative length declined throughout 2012 despite sharply lowerreal ratesMillion toz(left axis);%(right axis,inverted),454035302520151050-5-10,Net speculative length,US 10 year TIPS yield(right axis,inverted),-1.0%-0.5%0.0%0.5%1.0%1.5%2.0%2.5%3.0%3.5%4.0%4.5%,Source:CFTC,FRB and Goldman Sachs Global ECS Research.We expand our modeling of COMEX net speculative positioning to assess the impact of theFeds QE.Specifically,we regress net speculative length on:(1)US real rates,as proxied by10-year TIPS yields,(2)the stock of announced Treasury,MBS and agency debt purchasesin 10-year equivalent notional,and(3)the actual weekly flow of these purchases.Finally,asgold prices failed to rally when asset purchases did not require Fed balance sheetexpansion(Exhibit 2),we differentiate the stock notionals between instances where the Fedasset purchases were funded with reserve creation rather than sales of shorter-termsecurities(Operation Twist).We present the results of this regression in Exhibits 4 and 5and show the stock announcement path in the Appendix.,Exhibit 4:Stock announcements associated with Fedbalance sheet expansion impact gold net lengthRegression results,Exhibit 5:Net speculative length is well explained by realrates and Fed easingmillion toz3530,Parameter,Estimate,t-Statistic,25,Intercept10-year TIPS yieldStock announcement(ex.Operation Twist)Operation Twist 1&2,29.56-821.740.0049-0.037,25.7-20.93.8-11.1,201510,QE flow,Standard ErrorAdjusted R-square,-0.0064.283%,-0.6,0-5-10-15,Jan-97,Jul-98,Jan-00,Jul-01,Jan-03,Jul-04,Jan-06,Jul-07,Jan-09,Jul-10,Jan-12,Net speculative length,Predicted,Source:FRB,CFTC,Goldman Sachs Global ECS Research estimate.Goldman Sachs Global Economics,Commodities and Strategy Research,Source:FRB,CFTC,Goldman Sachs Global ECS Research.,5,December 5,2012,Global,The key results from this modeling are:COMEX net speculative positioning is better explained by this model than by real ratesalone.In particular,it captures the fall in positioning after Operation Twist.Further,even when accounting for the impact of QE,US real rates remain the key driver forgold positioning,with lower rates supporting net speculative length.Positioning responds to the announced stock of purchases,pricing in the size of thepurchase program at announcement while showing little response to the subsequentflow of purchases.This means that when the flow of Fed purchases is discontinued but the stock of Fed asset holdings is unchanged there should be little effect on goldprices.This sensitivity is in line with our US economists finding that it is the expectedstock of Fed asset purchases and not the flow that matters for bond yields.Stock announcements of asset purchases that require Fed balance sheet expansionpush positioning higher.In turn,stock announcements that do not require balancesheet expansion push positioning lower,with this negative coefficient correcting forthe decline in real rates that occurs around such announcements.This differentiatedresponse stands in sharp contrast with the systematic decline in Treasury yieldsaround QE announcements and suggests that gold prices“look through”easing thatdoes not require potentially inflationary Fed balance sheet expansion.Conceptually,while Operation Twist had an impact on yields through an increase in Fed holdingduration,it did not affect the monetary base as it did not require reserve creation.Although not featured in Exhibit 4,we found little evidence of an impact from the Fedsforward guidance on gold prices.One likely issue is that guidance extensions haveoccurred around announcements on Operation Twist,limiting the ability of ourregression to capture their impact.However,while our economists believe thatcredible forward guidance translates into expectations for further Fed balance sheetexpansion,the corresponding cumulative magnitude remains modest.Under our gold price modeling introduced in March 2009,net speculative positioning andphysical monetary demand for gold are the two key drivers to real gold prices.Havingexpanded our positioning model to better account for both the impact of real rates and thevarious iterations of QE,we reassess the outlook for gold prices in 2013 and 2014.Why is the coefficient assigned to Operation Twist in our regression significantly larger than the coefficientfor other stock announcements?While real rates were mostly unchanged in August-September 2011,net speculative length declined by 13 mtoz.Asthe Twist stock announcement is the only changing variable during that time frame,it captures most of this move.Specifically,our regression implies that the 10-year equivalent$400 bn stock announcement of Operation Twistlowered net speculative length by 15 mtoz,a surprisingly large impact.However,if gold prices indeed look througheasing that does not require Fed balance sheet expansion,then the impact of Operation Twist on gold net lengthneeds to correct for the large moves in real rates and net speculative length that occurred earlier in the summer of2011 in anticipation of easing by the Fed.Specifically,our economists believe that the impact of Operation Twist on Treasury yields was c.20 bp.The 80 bpdecline in 10-year TIPS yields that summer would then represent a 60 bp overshoot,worth 5 mtoz according to ourregression.Further,asset market responses around the announcement of Operation Twist suggest that expectationswere likely for bolder action by the Fed.If we assume for example that expectations were for$400 bn of easing withFed balance sheet expansion,then our regression suggests that net speculative length should have increased by 2mtoz.Combined,correcting for these anticipatory moves in real rates and net speculative positioning suggest anOperation Twist impact of c.7 mtoz,slightly less than half of the models estimate of 15 mtoz.While very simplistic,this exercise suggests that such a larger coefficient is not unreasonable and that the run upand subsequent decline in net speculative length around Operation Twist was likely excessive.Goldman Sachs Global Economics,Commodities and S