Memrandum DATE: February 20, 2015 TO: FROM: SUBJECT: Mike Reininger, AAF Ali Sule, AAF Al Racciatti Review f Ecnmic Analysis f AAF by Jhn N. Friedman As yu have requested, Luis Berger has reviewed the paper prepared by Jhn N. Friedman f Brwn University entitled An Ecnmic Analysis f All Abard Flrida (February 2015). While we agree that a tp-dwn aggregate calculatin f ridership and fare revenue ptential can ften be a valuable exercise in understanding and cnfirming mre detailed ridership estimates, we believe that Prf. Friedman lacked sme key infrmatin regarding the travel market which cntributes t a substantial underestimate f ridership and revenue ptential in his cnclusins. Belw is a discussin f several key aspects f his reprt as cmpared t the investment grade ridership and revenue study we cnducted fr AAF ver the past several years. This mem des nt speak t the CapEx, OpEx and financing assumptins in the Friedman reprt which are beynd the scpe f ur ridership and revenue wrk fr AAF. Size f the Addressable Market A key shrtcming f the Friedman reprt is the lack f infrmatin abut the magnitude and extent f the addressable market fr intercity travel in Central and Sutheast Flrida. An imprtant first step in the preparatin f a passenger rail frecast is an understanding f the existing number f travelers between the cities served, the rigin destinatin patterns fr thse trips with respect t statin lcatins, the purpse f the trip, and the existing mde f travel used. Friedman Reprt: The reprt makes n reference t specific travel patterns within and between Central and Sutheast Flrida. Instead the reprt assumes a level fr AAF ridership (nt the verall market) based n a cmparisn t ridership n Amtrak s Nrtheast Crridr services, scaling dwn t the ppulatin f the cities t be served by AAF. This verall ridership assumptin is then distributed t statin pairs withut reference t data n travel between the cities served by the statins. AAF Frecast: The AAF frecast was prepared with a detailed data set f trips n the radway netwrk fr the entire state f Flrida derived frm an assemblage f reginal travel mdels develped by Metrplitan Planning Organizatins frm surveys and traffic cunts. The final mdel was separated int lng distance travel (Central t Sutheast Flrida) and shrt-distance mdel (trips within Sutheast Flrida). Each mdel had several hundred travel analysis znes and represented trips and travel time/cst between thse znes. Luis Berger supplemented this data with research n the size f the air travel market, lng-distance and cmmuter rail market, and bus market between the cities served referencing federal, state, and lcal gvernment surces. Because the size f the market is the basis fr the mde chice analysis (see belw), 48 Wall Street, 16 th Flr New Yrk NY 10005 USA Tel +1.212.612.7963 luisberger.cm
we emplyed several methds t cnfirm the detailed data n trips and rigin/destinatin patterns, including: An rigin/destinatin survey f 8,000 travelers in the regin implemented by Luis Berger specifically fr this study. The survey cnfirmed the verall magnitude f trips between the cities served, the current mde f travel, trip purpse, and party size. An analysis f traffic cunt data frm the Flrida Turnpike Enterprise and Flrida DOT. This cnfirmed the verall number f aut travelers between Orland and Sutheast Flrida currently n the radways adjusting fr thrugh trips and intermediate destinatins. A review f previus studies was cnducted by the study team. The verall travel market estimate we develped clsely crrespnds t market research cnducted previusly fr the Flrida Statewide Rail Plan and previus frecast studies fr high-speed rail prpsal between Central and Sutheast Flrida. Cnclusin: Passenger rail ridership is directly related t price, frequency f service and quality f service in additin t the verall size f the market and the extent f cmpetitive alternatives. A tp-dwn estimate based reference t anther crridr cannt take these fundamental attributes int accunt. AAF service will have a prfile f service and pricing that will differ substantially frm Amtrak in the Nrtheast Crridr. By assuming a value fr AAF ridership, Friedman cannt accunt fr these differences r benchmark his estimate t the bserved patterns f travel between the fur cities served by AAF. Friedman s analysis als makes n reference t the verall size f the travel market in the Nrtheast Crridr and Amtrak s share f that market based n the key service attributes. The AAF frecast was based n a detailed dataset f trips cnfirmed independently thrugh survey wrk, traffic cunts, and reference t previus studies. Mde Chice Evaluatin As nted abve, the Friedman reprt assumes a level f ridership fr AAF. The AAF Frecast is based n mde chice mdel which cnsiders traveler preferences fr the time and cst f travel and calculates the share f verall travel market AAF is likely t capture given these preferences and the attributes f cmpeting mdes. Friedman Reprt: The reprt assumes AAF ridership thrugh a cmparisn t Amtrak Nrtheast Crridr service scaled t the size f the cities served by AAF. One tp-dwn estimate fr all travel between the fur cities served was assumed. AAF Frecast: The AAF frecast was prepared with a detailed data set f trips by all mdes f travel with a disaggregate representatin f rigin/destinatin patterns and travel time and cst (between several hundred travel analysis znes) fr each mde f travel. The frecast used the fllwing methds. A stated preference survey f 1,800 travelers in the regin was implemented by Luis Berger specifically fr this study. The survey cnfirmed cllected infrmatin n a recent reference trip between the cities served and presented a set f 8 chice experiments t gauge traveler chice f mde with the intrductin f a new mde f travel (AAF), given a range f travel time and cst scenaris. Discrete chice analysis was used t derive parameters fr the mde chice mdel including invehicle travel time, travel cst, access/egress travel time/cst, and frequency f service. The analysis was cnducted separately fr Business and Nn-Business trip purpses and lng-distance luisberger.cm
and shrt-distance travel. The implied value f time resulting frm the mdel was benchmarked t U.S. DOT guidelines and ther surveys fr the regin and cmpared favrably t thse benchmarks. Mdels were calibrated t base year (2010) travel cnditins fr each city pair, trip purpse, and mde f travel. Future frecasts were develped by estimating future year travel vlumes by mde and applying the mde chice mdels. This resulted in an estimate f future travel diverted t AAF frm the traditinal mdes f travel. The addressable market used in the estimates was adjusted dwnward t accunt fr travelers that needed t make intermediate stps by aut (en-rute captives) and was further limited t accunt fr reasnable catchment area arund statins (20-30 minutes access time fr Shrt Distance and 30-40 minute access time fr lng distance). Within the catchment areas the mde chice mdel accunted fr actual travel and access time by each mde f travel. The resulting market capture fr AAF as estimated by the mdels ranged frm 5% t 10% f the ttal market fr lng-distance travel (Central Sutheast Flrida) depending n distance, and 0.5% t 4% fr shrt-distance travel (within Sutheast Flrida) depending n distance. Capture rates are calculated n a zne-by-zne basis reflecting the actual time and mney cst f travel and access/egress (including taxi/rental car, parking) fr each mde. These capture rates are reasnable given knwn values fr ther U.S. and internatinal rail services, and previus frecasts fr the Flrida market. Cnclusin: The aggregate assumptin fr ridership in the Friedman Reprt excludes infrmatin n ttal market size and the capture ptential f a new service based n a disaggregate understanding f travel time and cst preferences and attributes by mde in the markets t be served. Withut this detailed infrmatin, Friedman substantially underestimates the ridership ptential f AAF. Induced Ridership Friedman s methd f deriving an aggregate ridership estimate by scaling frm ther rail crridrs des nt allw fr the estimatin f hw the additin f a new mde f travel will encurage travelers t make trips they may nt therwise have made. This ptential induced travel was cnfirmed thrugh respnses t the AAF stated preference survey. Friedman Reprt: Makes n allwance fr induced travel. AAF Frecast: The AAF frecast estimates induced travel based n the verall imprvement in the generalized cst f travel with the additin f a new mde (AAF) with the assumed fares, travel time, and frequency f services. This methd is standard practice fr passenger rail frecasting and the resulting rates f induced travel (10% t 20% based n distance) cmpare favrably t benchmarks in the literature. Cnclusin: The Friedman estimate des nt allw fr a reasnable level f induced ridership. Expansin t Additinal Markets AAF management has utlined identified a variety f travel markets nt currently adequately served by existing mdes f transprtatin. These markets include internatinal travelers wh may want t expand a trip t Suth Flrida t the resrts in Central Flrida, special event attendees, resrt visitrs, university students, and ther types f travelers, that culd be encuraged t luisberger.cm
make trips that wuld nt therwise be made, thrugh targeted marketing effrts by AAF and arrangements with travel arrangers, resrt peratrs, and special events venues. AAF management has assembled a detailed plan fr these marketing effrts. Luis Berger has reviewed that plan and independently cnfirmed the market size assumptins and examined the justificatin fr the capture rates. Friedman Reprt: Makes n reference t the ptential fr expansin f the market thrugh targeted initiatives. AAF Frecast: The AAF frecast includes an enumeratin f these ptential markets fr new travelers and cmments n the assumptins fr market capture t AAF. Cnclusin: The Friedman estimate des nt allw fr a reasnable level f market expansin thrugh targeted marketing initiatives. Travel Market Grwth The Friedman reprt assumes a level f ridership that appears t apply t an initial year f service and des nt appear t grw int the future. The revenue estimates, therefre, appear t be based n a static picture f annual ridership and revenue. Friedman Reprt: Assumes a static annual level f ridership and fare revenue. AAF Frecast: The AAF reprt makes allwance fr the underlying grwth in the verall travel market given reasnable frecast prjectins fr grwth in ppulatin, emplyment, visitatin, and ther surces f intercity travel. The frecast fr travel market grwth was prepared separately fr each traditinal mde f travel in the ttal market (aut, air, rail, and bus) in bth the shrt-distance and lng-distance markets. The grwth frecasts were benchmarked against ther reliable surces f sciecnmic and travel market infrmatin including the reginal planning agencies, the Flrida Bureau fr Ecnmic and Business Research, Flrida Department f Labr, and third-party cmmercial ecnmic frecasting firms. Overall, intercity travel has grwn substantially in the last 10 years with the Flrida Turnpike mainline lng-distance traffic grwing at 3.6% per year and intercity air travel grwing at 3.2% per year. The turnpike and aviatin authrities anticipate that this grwth will cntinue int the future with lng-range frecasts shwing 3.2% grwth in turnpike traffic and 3.0% grwth in intercity airline trips. The AAF frecast based n a disaggregate view f grwth in each market segment yields grwth rates f 1.5% t 3% per year depending n the distance, well within the demnstrated recrd f travel grwth fr the regin. Cnclusin: In excluding cnsideratin f travel market grwth, the Friedman reprt s estimate f financial feasibility substantially underestimates future ridership and revenue ptential. Willingness t Pay fr Travel Time Savings / Fare Estimatin Friedman assumes value f time based n US DOT guidelines in a tp-dwn calculatin based n a simplified view f travel time and cst t calculate the applicable fares a single fare level assumed per mile traveled. The AAF study was based n a detailed fare ptimizatin prcedure utilizing the mde chice mdels t estimate the revenue maximizing fare fr each city pair based n value f time by trip purpse. Friedman Reprt: Uses U.S. DOT guidelines fr Value f Time assumptins applied t a simplified aggregate view f travel time savings ptential. AAF Frecast: The AAF frecast calculates the willingness t pay fr travel time savings based n results luisberger.cm
f a stated-preference survey f actual travelers making trips between the cities t be served by AAF. The calculatin accunts fr preferences by trip purpse and length f trip and accunts fr preferences in invehicle travel time, cst, access/egress time/cst, and frequency f service. The mde chice mdels use this infrmatin t calculate the prbability f AAF ridership with disaggregate infrmatin n travel time and cst by lcatin reflecting detailed knwledge f rigin/destinatin patterns and the precise lcatin f the statins. This disaggregate view allws fr the evaluatin f the revenue maximizing fare fr each city pair fr business and nn-business trip purpses accunting fr the rigin-destinatin patterns in each market, the travel times and csts in statin access and egress, and the actual travel times and csts f cmpeting mdes. These fares are substantially higher than Friedman s estimates and vary by distance and trip purpse. The AAF frecast included a detailed review f AAF fares in reference t Amtrak fares fr cmparable distances and levels f service in the Nrtheast Crridr and fund the estimated AAF fares t be generally lwer than the Amtrak benchmarks cnsistent with the AAF estimates f verall demand and willingness t pay in the Central/Sutheast Flrida market. Cnclusin: The tp-dwn assumptins fr value f time and fare levels d nt accunt fr market variatins in willingness t pay revealed in the AAF survey. The tp-dwn methds substantially underestimates the ptential fare levels fr the city pairs given AAF s plan f service. By separating ut the calculatin f estimated ridership and applicable fares, the Friedman reprt ignres the interplay between ridership and revenue based n fare elasticity. It is difficult t directly cmpare the estimates prduced by Friedman with the AAF frecast because f this fundamental difference in methdlgy. Ancillary Revenue The Friedman reprt assumes that all ancillary revenues are scalable with ridership. Luis Berger reviewed the ancillary fare estimates prepared by AAF and determined them t be reasnable. The estimates include items that are nt directly related t AAF ridership and items that have been sized apprpriately t the anticipated levels f ridership and revenue. Friedman Reprt: Assumes direct crrespndence between ancillary revenues and ridership which is static ver time given the methds emplyed. AAF Frecast: The AAF management estimate f ancillary revenues is based n a bttm-up view based n market research and reference t cmparables. The management view accunts fr grwth ver time fr each line items based n the cmpnents driving demand. Cnclusin: The ptential f ancillary revenue is underestimated based n the lack f infrmatin related t ridership and elements f ancillary revenue that are nt expected t vary with ridership. General Cmment n Methdlgy Althugh the Friedman Reprt addresses sme f the basic elements underpinning ridership demand and fare revenue n AAF in high-level verview, the reprt was nt based n the cmprehensive level f infrmatin r detailed methds that are custmary fr an investment-grade level analysis. The AAF frecast was develped in keeping with industry standards as a reference fr AAF management and ptential investrs and has been subject t thrugh peer review by ther cnsultants and infrastructure investment prfessinal as planning fr the railrad has advanced and due diligence by investrs and rating agencies has been undertaken. The AAF frecast study included sensitivity testing that allws readers t assess the uncertainty and variability underlying the frecast inputs and has been subject t stress-case scenari testing during rating agency review. luisberger.cm