J.P. 摩根-美股-铁路运输业-美国铁路运输：美国铁路PSR潜力分析-2019.1.7-21页.pdf

North America Equity Research07 January 2019Equity Ratings and Price TargetsMkt CapRatingPrice TargetCompanyTicker($mn)Price($)CurPrevCurEnd DatePrevEnd DateCSXCSX US53,622.6662.79OWn/c86.00Dec-19n/cn/cNorfolk SouthernNSC US41,877.45150.53OWn/c203.00Dec-19n/cn/cUnion PacificUNP US102,088.60137.79Nn/c169.00Dec-19n/cn/cSource:Company data,Bloomberg,J.P.Morgan estimates.n/c=no change.All prices as of 04 Jan 19.RailroadsU.S.Railroad PSR Potential Analysis-Feedback,Pushback,and FAQs on Upgrade of NSC to OWAirfreight&Surface TransportationBrian P.Ossenbeck,CFA AC(1-212)622-Bloomberg JPMA OSSENBECK J.P.Morgan Securities LLCCaleb B Hogan(1-212)622-J.P.Morgan Securities LLCSanket P Parab(91-22)6157-J.P.Morgan India Private LimitedLacey-Ann Wisdom(1-212)622-9566lacey-J.P.Morgan Securities LLCSee page 16 for analyst certification and important disclosures,including non-US analyst disclosures.J.P.Morgan does and seeks to do business with companies covered in its research reports.As a result,investors should be aware that the firm may have a conflict of interest that could affect the objectivity of this report.Investors should consider this report as only a singlefactor in making their investment decision.The centerpiece of our 2019 Outlook and upgrade of NSC to Overweight was adifferentiated analysis of key performance indicators,rather than simple operating ratio benchmarking,to identify potential productivity gains at each U.S.rail network.In this note we address frequent questions and feedback on the analysis,including our preference for NSC over UNP,which appears further out of consensus than we initially anticipated.Investors appear more confident that UNP has the stronger“PSR put”based on the perception management is more committed and under greater pressure to drive productivity gains than Norfolk.This view is widely accepted,in our opinion,but we believe the operating efficiency and competitiveness of BNSF is still being underestimated,especially in a slower growth environment(see pages 3-4).Conversely,we believe the$600mm+of specific productivity opportunities at NS quantified in our analysis are not well understood;see pages 7-8 for key figures supporting our thesis.Our work also confirmed the strength of Norfolks intermodal franchise,which bodes well for CSX if it can narrow the gap in loading efficiency and track density as part of the current intermodal restructuring.Expectations and valuation favor NSC,in our view,but a potential truckload capacity squeeze in 2020 from the last phase of the ELD mandate is also a key factor in our preference for the Eastern rails versus UNP.NS has$600mm+of specific potential productivity gains.Our analysis shows NS is lagging peers across nearly every aspect of headcount efficiency even prior to CSXs operating turnaround in 2018.Key areas for cost savings include:back office consolidation where NS is the least efficient U.S.rail(+$100mm),closing the 10%fuel efficiency gap with CSX(+$120mm),and T&E which is overstaffed relative to car miles per day(+$400mm)after leading the East during most of the last decade.NS also lags U.S.peers in maintenance efficiency but operates more productive switching yards than CSX based on our estimates.UP already runs T&E efficiently and faces more competition.After sluggish performance over the last two years,UP has room for improvement and startedmaking strides in 4Q18.However,labor productivity has yet to increase and presentsa challenge for UP since we estimate it already runs an efficient T&E workforce.The biggest potential gains appear to be in locomotive utilization where UP“overpowered”the network while consolidating routes should increase lane density(see pages 5-6).Volume growth can also boost this key metric although BNSF poses a significant risk after lowering OR each year since 2014,closing the revenue per train hour gap,and taking share from UP during periods of slower growth.2North America Equity Research07 January 2019Brian P.Ossenbeck,CFA(1-212)622-Choose your own PSR adventure comparing key drivers,risks at UP and NS.Our framework for assessing potential operating gains and regional competition factored heavily into our decision to upgrade NSC and not UNP.Truckload exposure estimated by length of haul is another important consideration given our outlook for the last ELD squeeze will benefit Eastern rails more than the West in 2020.Our recent ratings changes were more of a relative exercise,but we will continue evaluating the progress of UP 2020 as well as the other considerations in Figure 1.Figure 1:Comparison of key drivers,risk factors,and expectations related to pursuing PSR at UP and NSSource:J.P.Morgan estimatesNSC is pricing in the least amount of PSR potential.We estimate NSC is pricing in 60-65%of the potential operating efficiencies identified in our analysis,which mirrors CPs prior cost saving targets in its failed M&A bid during 2015.The most frequent debate on the stock centers on whether or not the next strategic plan will set aggressive enough targets to satisfy investors.The majority we have spoken with think NS will fall short of expectations but the mitigating factor appears to be consensus estimates at 64%for OR in 2020.UNP reflects 70%of managements 2020 guidance,which mirrors consensus,although we are mindful this target represents a similar pace of efficiency gains as CN during its initial PSR transition.CSX should not be overlooked even after a record run.Our PSR potential analysis also provided a few extra reasons to remain positive on CSX when looking at the network and overall track density.Specifically,the potential upside from intermodal upgrades in 2019 was not considered during the initial 2020 guidance.We expect the redesign currently underway can help CSX narrow the significant intermodal loadings gap with NS while helping create track density and potentially increasing speeds in high density corridors,which are 10mph slower than NS(see page 9).CSXs goal to significantly improve switching productivity by 2020 also appears achievable considering Norfolk is already 20%more efficient.CategoryCompetitionRevenue driversEnd market riskCapex needsExpectations-STExpectations-LTTruckload exposureNear term opportunitiesPotential catalystsPSR experienceOperating track recordInteresting intangiblesLingering questionsOther observationssub-60%OR matching CSX operating improvementSignificant,estimated 550 mile average length of haulHQ move to Atlanta could reduce back office headcountPlaybook for a successful investor day provided by CSXNone,management open to adding in the futurePoor as the only U.S.rail to not improve OR in 2018Management widely viewed as under the most pressureWhy has pricing lagged TL rates?Will BNSF follow with PSR?KPIs do not appear completely volume dependentFebruary 11,2019 investor dayOne publicly announced hire now SVP of TransportationPoor after several bouts of service issues in the last yearOperating playbooks provided by CPs M&A and CSXs PSRCan PSR principles mirror the CSX success story?UP 2020 momentum drives upside to 2019 estimatesUnion PacificNorfolk SouthernBNSF a persistent issue,amplified when volumes declineCSX expected to take share in a softer freight marketGulf Coast chemical plants,West Coast port shareFrac sand falling,US$ag headwind,Mexico is 10%Export coal tailwinds fading while TL capacity increasesIntermodal conversions,low coal stockpilesShedding locomotives,historically a slow volume growerMixed,concerns on rushed UP 2020 roll-out55%OR,with some debateLimited,estimated 950 mile average length of haulImproving operating efficiencies builds credibilityLocomotive rebuild program underwayMixed,concerns on export coal exposure3North America Equity Research07 January 2019Brian P.Ossenbeck,CFA(1-212)622-Illustrated U.S.Rail PSR Potential AnalysisBNSF:More Competitive and Efficient than RecognizedOperating long corridors with consistent volume supports efficienciesWestern rails historically run their highest traffic corridors 20%above than the eastern rails and BNSF is best in class despite a 51%intermodal volume mix.Notably,BNSF average density per track mile has varied only+/-5%since 2009 while UP was at+/-10%,driven primarily by volume growth differentials.Figure 2:UP density fell in 2017 after remaining constant since 2009Source:J.P.Morgan estimates,Company ReportsFigure 3:BNSF density has remained stable and best in classSource:J.P.Morgan estimates,Company ReportsHistory suggests BNSF grabs volume in a downturnWe monitor the weekly share of originations in each region for trends across key categories and in the aggregate.BNSF has consistently taken share after the service issues in 2014 has now reached a level below the financial crisis.BNSF reported volume growth(two year stack)has also surpassed UP for the vast majority of the last five years despite running more carloads and containers in absolute terms.Figure 4:UP has consistently lost market share vs.BNSF in the West%share of total originationsSource:AAR,J.P.Morgan estimates,Company ReportsFigure 5:BNSF has outgrown UP in all but three quarters since 1Q13Source:AAR,J.P.Morgan estimates,Company Reports41%43%45%47%49%Jan-08 Jan-09 Jan-10 Jan-11 Jan-12 Jan-13 Jan-14 Jan-15 Jan-16 Jan-17 Jan-18UP Share of Total Originations vs BNSF4North America Equity Research07 January 2019Brian P.Ossenbeck,CFA(1-212)622-BNSF has recently caught up with UP on RTM/train hourTrain lengths are increasingly less relevant as they reach the upper bounds and especially as the focus shifts toward moving cars and not trains.Moreover,they lack context around utilization of track time and freight mix.We prefer revenue ton miles per train hour to capture this additional information and on this metric UP has seen a steady decline since 2009 while BNSF continues to show improvement.Figure 6:BNSF has closed the gap with UP after network issues flared up in 2014Source:J.P.Morgan estimates,Company ReportsA lower yielding mix requires BNSF to operate with a better cost structureWe attribute BNSFs operating efficiency primarily to locomotive utilization,consistent lane density,and productivity of administrative and maintenance employees.However,this lower cost structure is required to profitably serve a mix that yields fewer cents per RTM because of a heavy intermodal exposure.Figure 7:BNSF leads on cost efficiency but carries a lower value mixSource:J.P.Morgan estimates,Company Reports5North America Equity Research07 January 2019Brian P.Ossenbeck,CFA(1-212)622-UP:T&E Already Efficient,Upside in Locomotive UtilizationNetwork appears“overpowered”especially compared to BNSFUP is near the bottom of the group based on GTM per horsepower day and 20%below BNSF.Through late November 2018,UP had made progress against the initial 5-10%targeted gain on this metric(+2%)and continues to shed locomotives.Figure 8:UPs utilization of locomotive horsepower lags BNSF significantlySource:J.P.Morgan estimates,Company ReportsT&E employee already running near best in class based on car milesWe estimate UP already runs an efficient T&E workforce when comparing car miles per employee as show in the following figure.A similar conclusion is evident when analyzing GTM or RTM per T&E employee.Considering the level of efficiency UP has already attained,it is not surprising that the network has not improved daily car miles per FTE as disclosed in the November 2018 operating update.Figure 9:UPs T&E workforce is already operating near best in class productivitySource:J.P.Morgan estimates,Company Reports6North America Equity Research07 January 2019Brian P.Ossenbeck,CFA(1-212)622-Daily car miles rebounded slightly in 2017 and improved in late 2018Car miles per day fell across the industry as inefficiencies crept in with cars online growing at a faster pace than volume.UP has improved its daily car miles per full time employee,which builds on the recovery from the trough in 2016.Figure 10:Average car miles per day in 2016 were equivalent with 2009 levelsSource:J.P.Morgan estimates,Company ReportsUP should improve after sluggish performance and an easy comp for UP 2020Until the recent recovery off the lows in 4Q18,the UP network had experienced a gradual and steady decline in fluidity and throughput.The basis for comparison of the UP 2020 plan is the 7 day average ending Oct 24,which provides a relatively easy comp for improvement,including operating inventory,which is already-7%.Figure 11:UP average train speed is down 12%from the 2016 peakSource:J.P.Morgan estimates,Company ReportsFigure 12:Cars online declining but still 3%above the 2016 averageSource:J.P.Morgan estimates,Company Reports7North America Equity Research07 January 2019Brian P.Ossenbeck,CFA(1-212)622-NS:Specific and Quantifiable Efficiency OpportunitiesBack office and maintenance productivity significantly lag U.S.peersWe estimate Norfolk could generate$100mm of savings if it matched CSXs back office efficiency using the average comp and benefit per employee this could prove conservative given exec and admin compensation is likely above the company average.Norfolks maintenance employees also appear less productive than peers,which represents potential upside to our current productivity estimates.Figure 13:NS trails peers significantly on back office efficiencySource:J.P.Morgan estimates,Company ReportsFigure 14:Maintenance employees are also less productiveSource:J.P.Morgan estimates,Company ReportsNS generates more car miles per day than CSX but has lagged in efficiencyWe estimate Norfolk is ahead of CSX in average car miles per day in absolute terms after improving significantly off the 2009 trough.However,NS appears to get these results with significantly more T&E employees than CSX after slipping from the lead in the East in 2016.As shown in the figure below,NS had been at parity or above CSX in car miles per T&E employee over the last decade based on our estimates.Closing the gap with CSX on this metric would generate$400mm of lower labor expense using a conservative benchmark to 2017 efficiency-prior to the significant gains recognized by CSX as the PSR transition gained momentum.Figure 15:Norfolk leads CSX in average car miles per daySource:J.P.Morgan estimates,Company ReportsFigure 16:but lags significantly when adjusting for T&E headcountSource:J.P.Morgan estimates,Company Reports8North America Equity Research07 January 2019Brian P.Ossenbeck,CFA(1-212)622-CSX continues to lead NS on fuel efficiency and aims to widen its lead in 2020Norfolk appears to utilize its locomotive horsepower per GTM better than CSX but the network lags when comparing overall fuel efficiency.We estimate the 10%gap represents a$120mm opportunity at current fuel prices,with further upside if Norfolk can keep pace with the additional 10%gain CSX is targeting by 2020.Figure 17:Norfolks fuel efficiency has historically lagged CSX even pre-PSRSource:AAR,J.P.Morgan estimates,Company Reports9North America Equity Research07 January 2019Brian P