Wright Petition Affidavit of Douglas a Dawson 2003
Download original document:
This text is machine-read, and may contain errors. Check the original document to verify accuracy.
Before the FEDERAL COMMUNICATIONS COMMISSION Washington, D.C. 20554 In the Matter of: Martha Wright, Dorothy Wade, Annette Wade, Ethel Peoples, Mattie Lucas, Laurie Nelson, Winston Bliss, Sheila Taylor, Gaffney & Schember, M. Elizabeth Kent, Katharine Goray, Ulandis Forte, Charles Wade, Earl Peoples, Darrell Nelson, Melvin Taylor, Jackie Lucas, Peter Bliss, David Hernandez, Lisa Hernandez and Vendella F. Oura Petition for Rulemaking or, in the Alternative, Petition to Address Referral Issues In Pending Rulemaking ) ) ) ) ) ) ) ) ) ) ) ) ) ) AFFIDAVIT OF DOUGLAS A. DAWSON STATE OF MARYLAND COUNTY OF PRINCE GEORGES: ss Douglas A. Dawson, being duly sworn, deposes and says: I. INTRODUCTION 1. My name is Douglas A. Dawson, and I am the President of CCG Consulting, Inc. ("CCG"), located at 6811 Kenilworth Ave., Suite 300, Riverdale, Maryland, 20737. CCG is a general telephone consulting firm. CCG works for over 250 communications companies, which include local exchange carriers ("LECs"), competitive LECs ("CLECs"), cable TV providers, electric utilities, wireless providers, paging companies, mumcipalities and other governments and interexchange carriers ("IXCs"). 2. I submit this affidavit in support of the above-captioned petition to have the Federal Communications Commission ("Commission" or "FCC") address certain issues involving prison inmate calling services referred to the Commission by the United States District Court for the District of Columbia in Wright, et al. v. Corrections Corporation of America, et al. (" Wright").' I have specific experience and expertise relevant to the issues in this proceeding, which involves the provisioning oflong distance calling for prison inmates. I have assisted in the launch of over 50 long distance companies in my career. In that role, I have done virtually everything associated with creating or running long distance businesses. I am also familiar with all regulatory aspects oflong distance, including the development of rates and costs and the preparation and filing of tariffs. I have helped numerous companies select switching hardware for long distance service, and I know the capabilities and technical specifications of such hardware. I have negotiated numerous wholesale long distance service agreements between facilities-based IXCs such as Sprint, Frontier, Qwest and WorldCom, and resale carriers, and I understand the underlying long distance networks and issues associated with using them. I have had extensive experience with, and, consequently, have an in-depth understanding of, the capabilities and configurations of the network switching systems that lie at the heart of all telephone systems. I also have helped numerous companies with the provisioning of ancillary long distance products such as calling cards, operator services, pre-paid cards, international toll and Internet telephony. My CV, including prior testimony, is appended as Exhibit I. 11_ PURPOSE OF THIS TESTIMONY 3. In this affidavit, I have been asked to examine whether competition would work in the prison calling environment. Because the Wright case focuses largely on inmate calling at three specific prisons operated by the Corrections Corporation of America CCCA'') - the Central Arizona Detention Center ("CADC'') in Florence, Arizona, the Torrence County Detention Facility ('TCDF") in Estancia, New Mexico, and the Northeast Ohio Correction Center ("NOCC") in Youngstown, Ohio - during a period when inmate calling services were , CA No. 00-293 (GK) (D.D.C.). 2 provided there by Evercom Systems, Inc. oflrving, Texas ("Evercom"), I will use data relating to those facilities and Evercom to illustrate the points I want to make. 2 Evercom's inmate calling services to those prisons are typical, with regard to the rates and the methods used to bill long distance calls by prisoners, of most prison inmate calling services. The issue of inmate service competition is a generic question, and the conclusions drawn in this analysis would apply to all prison calling systems. eCA and Evercom controlled, and, in the case of the CADC and TeDF, still control, inmate calling on a monopoly basis from those three prisons and have permitted only a limited set of very expensive options for making long distance calls. I will analyze how competition could be brought to bear in inmate calling and demonstrate how it could lower inmate calling rates. 4. For the reasons set forth in this affidavit and based on my extensive background in the telecommunications field, I conclude that there are competitive alternatives to the monopoly environment found in these prisons. I will demonstrate a way that any prison system could allow open access to competition and still meet all of the security and other penological requirements of the prisons. 5. In brief, in this affidavit, I will: a) describe the history and development of telephone systems - both generally as well as specifically for prison systems; b) discuss the various penological requirements that must be satisfied by a prison calling system; c) discuss specifically the current payment methods that are used with prison calling systems; d) demonstrate that there are no justifications for prison administrators not to allow debit card or debit account calling or for inmate service providers not to offer debit card or debit account calling; and e) demonstrate the feasibility and reasonableness of opening inmate calling services to competition, so that inmates have a choice of carriers. On information and belief, Evercom is still providing inmate calling services to the CADC and TCDF. 2 3 III. PRISON TELEPHONE SYSTEMS 6. Since I will be discussing specific details of the various telephone systems used in prisons, such as debit systems and collect call systems, I will first discuss telephone systems generally and describe how they work. I will then discuss the specific attributes of the prison systems that relate to this proceeding. 7. Historically, all telephone systems in the U.S. began with operator assisted calling. Every call required an operator to complete a call using the large plug panels that we have all seen in movies. Even today, it is still possible to use a live operator to complete a call. In the late I 930s and into the I 940s, local switches were developed that allowed some automation in completing local calls; that is, a caller could complete some calls without using a, live operator, as long as the called party was connected to the same local switch. However, all long distance calls, or even calls to other switches in the same city, still required live operators. Beginning in the 1940s and into the 1950s, automated switches were introduced that allowed for the automatic switching of calls between local switches, and this allowed for the long distance network in place today, where dialing" I" plus the long distance number allows a call~r to directly dial long distance calls without the intervention of an operator. The early local and long distance switches were electromechanical. They worked by creating a mechanical connection between the called and calling party, much as operators had done mechanically before that. These electromechanical switches were not very sophisticated, and they could not perform very many functions beyond connecting calls. 8. In the late 1960s, computer technology Was introduced into telephone networks. With the advent of computers, a new set of telephone services, referred to as vertical features, was developed. Vertical features are computerized functions that provide callers more sophisticated services than simply the completion of calls, such as call waiting, call forwarding, call hold and speed dialing. These features relied on the new computer core of the switch to perform logical processes. With these new switches, the old electromechanical portions of the 4 switch used for basic call completion were replaced with computerized hardware. During this same period, the hardware that was used by the remaining operators was also computerized, and terminals that automated many of the operator's tasks replaced the old manual plug panels. However, even with computer assistance, collect and other similar calls still required live operators in order to be completed. 9. The next big breakthrough in telephone switching systems cmne in the early 1980s and was referred to as Signaling System 7 ("SS7"). SS7 is a technology that provides a second electrical path in the telephone network. The original path, referred to as the voice path, is where the electrical voice signal is sent across the network to complete calls. This new second signal, the SS7 signal, uses a different frequency and allows the switching system to communicate and perform tasks without disrupting the voice path. For example, the SS7 signal is the mechanism used to transmit the telephone number of the calling party and is what enables a new service like caller!D, which allows a called party to see the caller's phone number. The new telephone products that were enabled by SS7 were referred to generically as "CLASS" (Custom Local Area Signaling Services) features. The SS7 system allowed for many of the features present in the prison telephone systems in place today. For example, SS7 allows for prison officials to monitor the numbers that prisoners dial. Many of the new CLASS Features using SS7 required computerized databases, and these were introduced into the network in the early 1980s along with SS7. 10. The next technology breakthrough that is relevant to this case is the introduction of dial pulse recognition. With dial pulse recognition, any caller with a touchtone phone is able to give feedback to questions asked by a mechanized recording. For example, in the prison system, a mechanized recording may say "You have a call from prisoner X. To accept this call dial 5." The technology needed to do this on an automated basis was created in the late 19805. This was a significant technological breakthrough in that, for the first time, collect calls and other similar types of calls could be completed without utilizing a live operator. This 5 technology relied on two technologies to be implemented. First, a phone company needed to update each subscriber line card so that a given subscriber could dial using a touch-tone phone. This required significant capital outlay and was usually done as part of updating and replacing the entire switch. Second, the phone company had to update the switch core itself to be able to recognize dial pUlses. 11. There are recent technological changes that also impact prison telephone systems. The most recent breakthrough is voice recognition. Voice recognition just entered the market in a useable format in the late 1990s. Voice recognition technology allows the phone system to elicit responses from customers verbally without requiring them to dial digits, as is needed with dial pulse recognition. For example, a customer may be asked to answer "yes" or "no" to a question, and the voice recognition software is set to recognize one of these two answers. This technology is now widely used in the marketplace in various collect calling systems. Today, technology has taken another leap forward, and there are now switching systems that can recognize a person by his or her voice print using voice recognition software, thus eliminating the need for PIN numbers or the use of dial pulse recognition. 12. There is one additional technology that has evolved over time that is key to prison telephone systems, and that is recording technologies that make it possible to record and monitor calls. For most of the history outlined above, no widespread technology was available to record and monitor calls on an automated basis. It has been possible for a very long time to monitor calls by having a person tap into the calls and listen to them. The ability to record calls and to later listen to them, as prison officials require, is now a key penological requirement. The first hardware that could record calls on a wide-scale basis was available in the early 19708. This consisted of little more than a bank of tape recorders that could allow for the simultaneous recording of many calls. Such a system required a massive storage of computer tapes, and it was not easy in such a system to pinpoint or retrieve a specific call from a specific inmate. Newer recording technology is available that solves such problems. Modem recording systems 6 use computer drum storage, much as is done for the storage of data on a commercial company's local area network. Such storage is done digitally, and a digital record is made of each call, thus making it easy to later retrieve specific recorded calls. The size and cost of the storage devices that can be used for such a purpose have drastically decreased over time, and the cost continues to decline as digital storage techniques improve year after year, with a seeming doubling in storage capacity per dollar every 18 months or so. 13. Because of the need to satisfy penological requirements, there are unique features of prison calling systems that, in combination, differentiate them from other types of telephone systems. For many years, prison systems were at the cutting edge of technology, as prisons tried to meet their requirements with the latest available technologies. However, with the advent of modem switching technologies, technology has finally caught up to the penological requirements, and there are now many different switching platforms that can be modified to meet the requirements of prison systems. 14. A prison calling system is comprised of four basic components. First is the switching platform referred to above. This is essentially a piece of hardware that allows for the dialing and completion of calls along with a core computer logic system that allows for the creation of specific features and functions that, taken together, are unique to prison calling requirements. The second requirement for a prison telephone system is a recording storage system that allows for the easy monitoring, recording and retrieval of prisoner calls as needed. The ideal prison recording system records calls automatically and also allows authorities to easily listen to calls later. Third, the prison telephone system requires a master control system that allows the authorities to quickly intervene and modify prison calling patterns as needed. Such a master control system is basically a terminal with an easy interface into the switching system software, where authorities can make quick changes to such functions as the list of numbers that a specific prisoner is permitted to call. All modem switching systems have such control interfaces. The last component of a prison telephone system is the software 7 programming that enables the features that are unique to the prison system. For example, a feature allowing a called party to request to be automatically removed from a prisoner's calling list is unique to the prison system. Such a feature is created by specific software developed by a prison switch vendor to meet this specific requirement. IS. Prison telephone systems have evolved over the years in response to two trends. First, such systems have evolved to introduce new functions and features in response to the availability of new technology, as outlined above. To illustrate, consider the example of one specific penological requirement: that prison telephone systems allow prison administrators to restrict prisoners to a relatively short list of pre-approved telephone numbers that they may cal\. This particular requirement was not feasible until the late 1960s, when similar features were introduced into commercial telephone switching systems. As switches became more like computers, it became technically possible to devise a system that could limit prisoner calls to specific numbers. Thus, each separate penological requirement for prison telephoneswitching systems has only been made possible, and thus really created, in response to changes in technology. In summary, technology has expanded the ability to provide more functions with a switch, and the basic requirements for prison switching systems have constantly evolved to exploit these technical capabilities. 16. The second trend that affected the development of prison switching systems was the expansion of prisoner calling rights. For a long time, prisoners were allowed to make very few calls. However, as prisoners won greater calling rights, prison telephone systems were developed to respond to these expanded calling rights while meeting penological requirements. As prisoners called more, the penological requirements for the prison systems have grown to meet the evolving challenges presented by prisoners. 17 _ For many years, all prison inmate calls were collect calls. This was largely due to the fact that only a live operator could satisfy the basic penological requirement that prisoners could not make calls to those who did not wish to talk to them. There was no other way 8 -- ------------ - -- --------------- historically to automate this function, and thus the intervention of a live operator and the use of collect calling was necessary to ensure against the harassment of witnesses and other similar abuses. Live operators are no longer needed to meet this requirement. With easily programmable switches, very complex features can be introduced today, and if a switching requirement can be imagined, it probably can be programmed. 18. The three prison facilities under examination in this proceeding -- the CADC, the TCDF and the NOCC -- have used or now use Evercom's telephone calling systems and services for inmate calling. Evercom specializes in prison caJling systems and services. According to Evercom's year-end 2000 10-K Report ("IO-K Report''), it served almost 2000 prisons in the United States as of December 31, 2000. 3 Evercom refers to its product as CAM (Inmate Call Access Management).4 The Evercom CAM system can meet all of the penological requirements described in this affidavit. Note that Evercom is not the only provider of prison telephone systems. There are several other prison switch providers, but Evercom is the predominant supplier of prison calling systems in the U.S. marketplace today. IV. PENOLOGICAL REQUIREMENTS OF PRISON CALLING SYSTEMS 19. The foJlowing description of the penological requirements of prison inmate telephone systems is derived from various documents gathered from the manufacturers of such systems. Additionally, these requirements are usually specified in great detail in the various periodic Requests for Proposal ("RFPs") issued by the prison administrators when they are seeking a new telephone service provider. For example, these requirements are specified in detail by the Federal Bureau of Prisons ("BOP") in its 1997 Request for Proposal for its inmate Evercom, Inc., SEC Fonn IO-K, Part II, Item 7, at "Overview" (filed June 1,2001 for the fiscal year ended December 31,2000) ("IO-K Report"). The relevant portions of the 10-K Report are attached hereto as Exhibit 2. 3 4 ld. at Part J, Item 1, "Systems." 9 telephone system, relevant portions of which are attached hereto as Exhibit 3 ("BOP RFP',).' I also understand from a technical perspective how all of these penological requirements can be made to work in a prison calling system. These penological requirements for a prison calling system can be broken down into the major categories listed below. Different prisons have selected different subsets of these requirements, but overall, most prison systems are designed to fulfill the same basic list of penological requirements, which are: • Number Control • Personal Allowed Numbers ("PAN") • Individual Phone and Phone Group Definitions • Voice Prompts • Personal Identification Numbers ("PIN") • Monitoring • Recording and Playback • Reporting • Calling as a Commodity 20. Number Control consists of those telephone features, such as blocking, unblocking, validation and the defining of telephone numbers, that allow the prison to control the telephone calls that can be placed by prisoners. With number control, prisons can satisfy various penological requirements. One almost universal use of number control is the prohibition against inmate calls to certain types of numbers, such as 800 or other toll-free numbers or 900 numbers. This stops prisoners from re-originating calls. It is possible, when dialing 800 or other toll-free access numbers that terminate to a non-prison telephone switch, to connect with call systems that allow the caller to get an additional dial tone and then re-originate the call to another number. The blocking of 800 and 900 calls greatly reduces the chances of 'Federal Bureau of Prisons, Request for Proposal, June 2,1997 ("BOP RFP"). 10 call re-origination. In a modem switch, numerous types of blocking can be performed. Universal blocking rules block certain categories of calls for all inmates, such as not allowing any prisoner to call an 800 number. Individual blocking rules can also be applied, allowing certain categories of calls to be blocked for certain prisoners. Blocking can be made very specific. For example, a prison can prevent calls to an individual number, and many prison systems allow outsiders to elect not to receive calls from prisoners. 21. A related feature to blocking is Personal Allowed Numbers ("PAN"). PAN is a penological requirement that enables prison administrators to restrict inmate calling to a preapproved list of telephone numbers. A PAN system thus prevents harassing calls and fraudulent telephone schemes involving calls to non-approved numbers. Any attempt to dial a number not on a PAN list is blocked by the switch. 22. Another important set of penological tools is Individual Phone and Phone Group Definitions. This means that prisons can control calling in any manner they choose. For example, they can limit the duration of calls. They can track the time used by a given prisoner and cap his total usage at some fixed ceiling amount per day. The prison can restrict the hours of phone usage, either universally or by prisoner. Phone Group Definitions give prison administrators control over the basic functioning ofthe phone system. 23. Voice Prompts is a series of functions that allow the prison to control how prisoners can place and use calls. For example, voice prompts can be used to warn prisoners that a call will soon be terminated if it is running too long. One penological use of voice prompts is the use of a pre-recorded announcement to let a called party know the name of the inmate making the call. Voice prompts also allow the called party to accept or reject the call before the prisoner comes on the line. The announcements now provided by voice prompts were historically provided by live operators, but these functions have been replaced today with a mechanized and computerized series of recordings designed to meet every possible and allowable type of call. 11 24. Another penological concern is that each inmate should have a unique Personal Identification Number ("PIN") that must be used in order to initiate calls. PINs ensure that inmates are identified and tracked individually. Every call can be tracked and traced to an individual inmate. The use of PINs also enables administrators to provide different telephone privileges to each inmate. The prison can place restrictions on any aspect of calling, from who can be called to how long calls last, by having all calls use the PIN system for access. The use of PINs is widespread in the telephone industry outside of prisons. PINs are used routinely for credit card calls, debit card calls, pre-paid card calls, international callback calls, within the PBXs of many large companies and in many other applications. PIN verification works by using a lookup table. In the prison example, the lookup table is a very simple one that consists of just one PIN for each prisoner. If the prisoner attempts to use a PIN that is not in the table, a call cannot be completed, and, usually, the prison is notified of the fraudulent attempt. 25. Modern prison telephone systems also require Monitoring. Monitoring allows prison officials to listen to calls on a real-time basis. Prisons routinely monitor inmate calls to make certain that no crimes are being committed or that people are not being harassed. A monitoring system allows the prison administrators to listen at any time to specific prisoners or to choose calls at random to monitor. Many prison telephone systems include camera surveillance of telephones along with voice monitoring. This allows the prison officials to see who is making the call while listening to the conversation. 26. Another requirement of modem prison telephone calling systems is Recording and Playback. This allows prison officials to listen to calls that were made in the past. For example, should a prison administrator discover a case of telephone fraud, the administrator can listen to phone calls made by the same prisoner in the past. The recording of calls is done by separate hardware that is not an integrated part of the switching system. Modem telephone recording systems usually use drum storage devices to capture and store calls, and the number of calls and the length of retention of recorded calls is limited only by the size of the storage 12 system chosen. Such storage devices can be programmed to allow for instant retrieval of recorded messages by the authorities, much as is done by voice mail systems widely in use. In order to control the costs, most recording systems also allow the calls to be moved from drum storage to more permanent media for long-term retention. 27. Another penological requirement is Reporting, which allows the prison officials to create rules for calling and then to report any violations. For example, a system might record instances when a prisoner does not know his PIN on the first try. This will help identify any prisoner who is fishing for valid PINs by trial and error. The same sort of system can be used to track sequence calling by an inmate, that is, in calling numbers that are close to each other numerically. Such calling patterns are often associated with attempts at fraud. Reporting can , also show when prisoners try to call people whose numbers are blocked for them, such as witnesses and judges. Modem reporting systems have become quite sophisticated in response to the demands placed upon the telephone system by prisoners. 28. A final penological requirement is one that is not directly related to the phone system hardware. Prisons prefer to have an inmate calling system that does not create a commodity, and thus is not subject to coercion or extortion among prisoners." Typically, any system that involves funds or a commodity that can be used by prisoners can be subject to these types of abuses. No calling system - be it collect only or a debit system - can completely eliminate such problems in a prison. The ideal system will have stringent enough rules to make calling reasonably unattractive as a commodity. For example. closely scrutinizing the pre-approved list of telephone numbers that each prisoner is allowed to call greatly reduces the attractiveness of another prisoner's account, particularly if such scrutiny is combined with blocking that precludes the re-origination of calls. " This issue is not unique to a prison's telephone system. inasmuch as inmates routinely maintain commissary accounts for the purchase of sundry items. 13 29. These penological requirements, taken together, are unique to a prison calling system.' Many of these features are used individually elsewhere in the telephony world, but only the prison systems brings all of these unique attributes together as a package. There is a definite incremental cost of providing these features. These are costs that should be recoverable by the provider of the prison calling system. V. COST ISSUES 30. Historically prison inmate calling required collect calls using Jive operators. Only a live operator could make sure that prisoners were limited to the types of calling that the prison authorities allowed. But with today's technology, there is no longer any reason to use only collect calling for prison calls. For example, the Evercom system in the three sample prisons it serves or has served allows for at least two types of calling. First, it offers an automated collect call. meaning that the called party pays for the call. Second, it offers a debit product, meaning that the call is pre-paid before being placed. 31. As described above. collect calling systems historically required live operators. Ascertaining whether the called party was willing to accept charges for a call required a live operator because there was no technology available to automate such a function. Today, the vast majority of commercial collect calls are performed entirely by computers and do not require a live operator. There are a number of automated collect call products available to the general public such as 1-800-COLLECT and 1-800-CALLATT. To a large degree, except for the extra layer of penological functions, these commercial collect systems operate much like the prison collect system. To place a prison collect call, a prisoner must first dial a desired number. The prison system then maintains complete control of the call. Typically, it mutes out the prisoner so that he cannot hear the called party being queried by the automated prompts. The computerized system connects to the desired number, and when the called party answers, a , The requirements discussed above are also reflected in the portions of the BOP RFP attached hereto as Exhibit 3. 14 voice prompt will ask whether the called party wishes to accept the charges for a call from the prisoner. Because the prisoner is muted, the system uses a recording of the prisoner's name to announce the request. The called party is given instructions on how to accept the call ifhe or she wishes to pay for it. In some newer systems, the called party can accept the call by verbally saying "yes," using voice recognition software that recognizes simple words. In most prison systems, the called party will be asked to dial a digit on the phone, for example, "Dial 5 if you want to accept charges for this call." When the system receives an affirmation that the call will be paid for, the prisoner is taken off of mute, and the call is completed. 32. The network process required for completing a prison pre-paid debit call is almost identical to the processing of a collect call. In a debit system, a prisoner will also dial the desired number. The system will then put the prisoner on hold until it determines that there are enough funds available to pay for the desired call. Once it has been determined that sufficient funds exist, the call is completed. A debit platfonn is virtually identical to a collect system. The debit system requires the same major components -- a switching platform, a storage device with a voice mail-like system, a master control system and unique software_ The only real difference between the prison collect call product and the pre-paid debit product is who pays for the calls and hence how payment is made. 33. This is a very, important distinction and something that has been brought about by the convergence of technology. For most of the history of the industry, collect calls were very different from other types of calls. They required unique equipment and the use of live operators. As such, collect calls were billed under unique rate structures. However, the unique nature of collect calling has now disappeared. As can be seen in these prison systems, there is no practical difference between a prison debit call and a prison collect call, except for the decision of who is going to pay and how payment will be made. Moreover, because, as discussed below, debit calling eliminates the significant amounts of uncollected revenues that service providers experience with collect calls, debit calls ought to be the preferred prison 15 calling methodology. Both debit and collect calls meet all of the same penological requirements and use the same equipment. From a network perspective, the only difference is a very minor one related to call routing in the case of debit calling in order to verify that there are existing funds for the call -- a change that does not add cost to the call processing. Because debit card calling meets all of the same penological requirements as collect calling, there is no justification for restricting inmates to collect calling. All prisons thus should be required to allow debit calls. Such calls are less expensive for the providers, by definition, and should thus cost less for prisoners and families of prisoners. 34. Some prisons have not allowed debit calling, typically, on the grounds that the administrators do not want the extra administrative burdens of handling the cash for the debit payments.s Prison administrators claim that creating an additional source of prisoner funds might generate an additional possibility of extortion among prisoners. However, there are many options for establishing a debit calling system that can overcome these objections. For example, the federal prison system has had a debit product for prisoners for many years. One way to avoid having an extortable commodity is to have a debit system where the called part.ies (the families) control the funds. In such a system, a family member would purchase a debit account under his or her own name and control. A prisoner would be allowed to call this family member as long as there were funds in the pre-paid account. Removing the cash from prisoner control will remove most of the penological concern and eliminate any additional administrative costs for the prison in handling debit accounts. As will be demonstrated below, the collect calls initiated from the sample prison systems are quite expensive. At the end of the day, it is the families and acquaintances of the prisoners who pay for collect calls. Given a choice, many of these called parties would much rather establish a personal debit fund if the calls could be cheaper. Upon information and belief, one of the CCA facilities involved in the Wright case, the Northfork Correctional Facility located in Sayre, Oklahoma, did not allow inmates to make debit card or debit account calls; they were provided only the option of collect calling. 8 16 35. A debit system that allows families to pay for calls instead of having the prisoners pay would not increase costs or administrative burdens for the prison. In most contracts between prisons and providers that I have seen, the carrier usually absorbs all of the costs of running the prison telephone system, including the switch and the software. In this case, of course, Evercom also bills everyone who accepts collect calls. As long as the service provider is responsible for the cost of maintaining external family debit systems, there should be no additional cost or burdens for the prisons. 36. In the telephone industry, revenues that are billed but not collected from customers are classified as uncollectibles. A significant number of people who accept collect calls from prisoners subsequently refuse or are unable to pay for the calls. The underlying prison calling provider must absorb the lost revenues from any calls that are not collected. The uncollectible rate for inmate collect calls can be very high. According to its year-end 2000 lOoK Report, Evercom states that it has always had high uncollectible revenues from inmate collect calling.' However, Evercom should experience very little, ifany, uncollectibles from debit calls. In a debit system, the calls are pre-paid, and when a caU is placed, the service provider can instantly collect from the debit card account. Accordingly, uncollectibles in a debit system should be virtually zero. A debit system would also allow the service provider to collect the cash from calls in advance -- at least thirty days earlier than with conect calling -- which is a big plus for any telecom provider. 37. Prison administrators have argued that debit calling does not offer as many penological safeguards as collect calling. In particular, they point to the penological requirement that telephone privileges not become a commodity. They suggest that allowing prisoner debit accounts can create a currency or credit that can be sold or extorted. The federal system, however, which allows debit calling, has taken several steps to reduce the possibility 'Evercom's IO-K Report, which is attached hereto as Exhibit 2, states, in Part I, Item 1, at "Federal Regulation," that "[bJad debt is substantially higher in the irunate telephone industry than in other segments of the telecommunications industry." 17 that debit calling might result in the creation of a commodity. The Federal BOP has very strict rules concerning the ability of prisoners' families to replenish the funds in a debit account. They restrict such debit fund payments to a small list of outside parties that includes lawyers and direct family members. Other penological tools also help to reduce the possibility of creating a commodity. For example, strictly limiting the calling for each prisoner to a preapproved list of telephone numbers greatly reduces the attractiveness of any other inmate's account, particularly if this technique is combined with the inability to re-originate calls. It should also be kept in mind that a collect calling system can be abused as much as a debit calling system. Whatever value can be extorted from another inmate's debit account could also be extorted from his collect calling PIN. If implemented properly, as has been done in many prisons, there is no specific advantage to collect calling over a debit system. 38. In summary, a debit card system can meet all of the same penological requirements as a collect system. The only real difference between the two systems is who pays for calls and how they pay. In a properly designed debit system, there is no additional burden for prison officials. There also does not have to be an additional source of funds available to prisoners that can be extorted. The only real difference between a well-designed debit system and a collect system is how the prisoners or the families of prisoners pay for calls. There is therefore no penological justification for limiting inmates to collect calling services, rather than providing a choice between collect and debit calling. VI. COMPETITION IN PRISON CALLING 39. Many prison inmates and families of prisoners, including the petitioners in this proceeding, have asked for the introduction of competition into inmate calling services. In every other segment of the telephone industry, competition has very effectively lowered the cost of long distance calling. The cost of calling has tumbled everywhere over the last few decades except within prisons like the ones in the referral case. This affidavit will demonstrate that it would be economically and technologically feasible to introduce competition into prison inmate 18 calling services, consistently with all legitimate security and other penological requirements, thereby allowing for more options for families and ultimately resulting in lower rates. As other observers have noted, the penological justifications for exclusive inmate calling service arrangements are factually unsubstantiated and pretextual. lO 40. The best way to get competition into inmate calling services, and thereby benefit prisoners' families or other telephone service bill payers receiving calls from prisoners, Vy'ould be to allow inmates to choose among different IXCs; in effect, to create an equal access multicarrier platform for each prison calling system. One possible mechanism for such a system will be discussed in more detail below. One question that is routinely asked by family members is why the prisons do not allow the use of commercial calling products, such as 1-800-COLLECT or commercial debit cards. As described above, these commercial products allow the reorigination of calls. Prison administrators claim that the prison system needs to maintain control of the call from beginning to end for security reasons and that if a prisoner were allowed to use a commercial calling platform that allows the re-origination of calls, many of the penological safeguards discussed above would be bypassed, thereby making abuses possible. Leaving aside the merits of such claims and the potential use oftechniques to maintain control over reoriginated calls, it would be feasible to allow multiple IXCs to offer services to any given prison facility, and thereby bring the benefits of competition to prison inmate calling, while meeting all of these objections to the use of standard commercial calling products. 41. Following is one such mechanism that could be used to allow multiple carriers to compete within a prison calling system. There may be other mechanisms that will work, but the goal of this example is to demonstrate that competition is technologically and economically feasible, consistent with all ofthe security and other penological concerns discussed above. The 10 (0 See Justin Carver, An Efficiency Analysis of Contracts for the Provision o/Telephone Services Prisons, 54 Fed. Comm. LJ. 391, 394 (2002) ("Carver"). A copy of this article is attached as Exhibit 4 hereto. 19 primary reason to provide for multiple carriers is to allow choice, thereby creating competition and the resultant lower rates. The FCC has spent considerable effort in the last twenty years to ensure that consumers everywhere have choice, and the presumption has always been that choice is beneficial. The evolution to more choices for long distance and local calling has led to lower prices, creative new products and overall greater satisfaction among telephone subscribers in the U.S. However, the families of prisoners in the CCA and other prison systems are the last group of telephone consumers in the U.S. who are still being denied choice. 42. One way to allow competition in prison inmate long distance calling services would be to authorize a multi-carrier platform provided by an underlying service provider in each prison that would supply the prison telephone system hardware and software. This underlying. provider would supply the switch and software, the phones, the management control system and any other required components of the prison calling system. The various carriers offering competitive long distance services to the inmates would interconnect with the underlying carrier's prison telephone system. The underlying service provider could be compensated for providing the prison telephone system by a charge imposed on the interconnecting cOJ:'lpetitive carriers, based on the costs of installing and operating the prison system. This charge would compensate the underlying carrier for the switch, software, maintenance and operating costs for providing the system, but would not include the cost of providing the long distance transmission. The underlying provider could recover its costs through a per minute charge levied against all long distance calls placed from the prison and carried by one of the competitive service providers. As discussed below, these costs would range from 4.4 to 5.9 cents per minute. 43. In order to implement a long distance multi-carrier choice through a prison telephone system switch, each competitive carrier should be required, at its own cost, to provide long distance transport facilities to the prison switch. These facilities would typically consist of T-J t~nks (a digital transmission link with a capacity of 1.544Mbps, enough for 24 simultaneous voice conversations), that go from the prison switch to the IXC's point of presence 20 ("POP"). Each IXC also would be required to pay the underlying carrier for the fixed cost per minute of providing the prison system. Each IXC would then be free to compete on price and service to get the prison calling business. Each IXC would be free to charge any rate it chose as long as it agreed to first pay the underlying provider to interconnect with the prison system. In such a competitive system, the underlying provider could also be allowed to offer a competitive long distance product along with the other competitive carriers, as long as it also covered its basic per minute system fee on an imputed basis. 44. In this way, prisoners, or the prisoners' families, would be able to select the carrier of choice from a menu of available interconnecting carriers. Today, the prisoners get a prompt in most prisons to choose between debit calls and collect calls. In the competitive environment, they would get an additional prompt asking them to select a carrier for whichever type of call they elected to use. Prisoners could also be allowed to "choose" a carrier on a more permanent basis in order to avoid going through the carrier selection screen for each call. The competitive carriers would be free to market directly to the people who actually pay for the long distance calls made by prisoners -- in most cases, the families. Families could elect to purchase calling products from the competitive carriers offering the best deals. Since there is such a large volume of calls made from prisons, a number of different carriers could be expected to compete for the business from each prison. There is little doubt that such side-by-side competition among mUltiple IXCs would lead to much lower long distance rates than those in place in these prisons today. 45. It is important to note that even in such a multiple-provider system, all of the penological requirements discussed above would continue to be met. The software in the prison telephone system switch would continue to provide all of the necessary security functions, just as it does today. Adding a choice of carrier to the calling process would not affect or modify any of the penological safeguards built into today's systems. Prisoners would still place calls under the complete control of the prison phone system. This system would maintain control of the entire 21 call using all of the rules and safeguards in place today. A call would only be completed after it could be ascertained that the prisoner was not making unauthorized calls and that the carrier was being paid for the call. Because the long distance provider carrying the call would be interconnected at the prison system switch, control over the entire call could be maintained, just as it is today. At the end of each call, the underlying service provider would assess the system fee to the lXC carrying the call. The IXC that handled the call would then charge the inmate's debit account for the call, including the underlying system fee. 46. There have been other proposals in the past that have suggested ways to offer competitive calling in prisons. Some of them involve handing off inmate calls to another network not directly interconnected with the prison telephone system. The proposal set forth in this affidavit would require that the underlying carrier process a call up to the point where the call was handed off to an IXC for completion. That hand-off would take place at the switch exclusively serving, and under the administrative control of, the prison. There would be a requirement that calls remain under the control of the initial switch for the entire duration ofthe call. Competitive carriers would be prohibited from transferring any inmate calls to o\her IXCs or to any carriers other than the terminating LEC serving the called party. The interconnecting carriers thus would be in the business of completing long distance calls, but, because they would take the calls at the prison system switch and deliver them to terminating LECs, they would not have the ability to bypass any of the penological requirements of each prison, which would be implemented and enforced by the underlying switch provider, just as Evercom enforces those requirements today. 47. As demonstrated above, this competitive proposal would be technically feasible and would not be a major burden for carriers, it would safeguard the rights of consumers, and it would maintain all oftoday's penological safeguards. It would also attract numerous additional competitive IXCs to compete for long distance inmate calling service. Most IXCs would view a 22 ------------- ...~.~.-~.-.. prison system, with its many concentrated minutes, to be a premium opportunity to be pursued. If we build a competitive environment, the carriers will come. 48. Implementing such a competitive system would cause a fundamental change in the way that the underlying provider does business. Allowing multiple carriers to compete would require some hardware and software changes to the prison calling systems. While these changes are relatively minor, there would be some small incremental start-up costs in implementing competition. In the past, the FCC has not hesitated to impose requirements that increase carriers' short run costs when such changes were necessary to facilitate competition. There are numerous examples of FCC orders that have required carriers to expend money for capital and software. In recent years, we have seen orders requiring the provision of "LIDB" (line information database) functions," payphone call tracking 12 and others. As will be demonstrated below, the capital required to implement a competitive solution is too insignificant to be a barrier to change, especially given that the underlying telephone system provider would be able to recover the complete cost of providing the prison calling system from each call, including a reasonable profit. 49. There is no question that introducing competition into the prison calling system is in the public interest. Regulatory bodies have often assumed that exclusive inmate calling service arrangements were required in order to meet legitimate security and other penological See generally Policies and Rules Concerning Local Exchange Carrier Validation and Billing Informationfor Joint Use Calling Cards, 7 FCC Rcd 3528 (1992) (subsequent history omitted) (requiring LECs to provide non-discriminatory access to the validation and screening infonnation located in the LECs' line information database so that IXCs can accept and complete calling card calls). \I 12 See implementation of the Pay Telephone Reclassification and Compensation Provisions of the Telecommunications Act of I 996, I I FCC Rcd 20541, 20588, 20590-91 (1996) (subsequent history omitted) (requiring IXCs to track calls they receive from payphones in order to ensure fair compensation for each payphone call, despite the IXCs' claims that implementing tracking mechanisms would require significant expenditures of capital). 23 requirements, and this assumption has contributed to past rulings that have upheld the current prison inmate calling regime. In the past, that assumption might have been valid. At this point, however, as explained above, it is clear that competitive long distance inmate calling services are perfectly compatible with security, antifraud and other penological requirements. Given that it is typically non-inmates -- families and attorneys -- that ultimately pay for inmate long distance calls, it must be concluded that these consumers deserve the same rights to choice as do all other callers. Moreover, lowering the cost of prison inmate calling would bring about penological benefits, such as improving family relations for prisoners and improving the chance of successful rehabilitation and integration into the community after the sentence is completed. Finally, as demonstrated below, the competitive system envisioned here would be economically feasible. VII. THE COST OF PRISON INMATE CALLING 50. This section will explore the potential cost of providing the competitive prison system described above. The goal in this section is not to specifically identify the precise costs of providing inmate long distance calling services. Rather, this section is intended to examine whether such a system would be economically feasible by analyzing the potential range of costs, particularly the costs of the underlying system that would be used by all of the competitive IXCs.13 As will be demonstrated, even the most conservative estimate of the cost of implementing this proposal is so reasonable that any objections to it based on cost burdens could not be valid. Several different sources have been reviewed in analyzing the costs that would be incurred by the underlying system provider, including Evercom's public financial data. Evercom is a useful source of data, not only because it is the primary provider involved in the referral case, but also because it is one of the largest prison inmate cal\ing service providers in the country. The cost of providing the long distance segment of the service will also be discussed, but only as a comparison with other estimates filed with the FCC by inmate service providers. The primary focus of this analysis will be the costs of providing the underlying telephone system. The costs of the long distance segment "wash out" of any economic feasibility analysis because competition in the provision of the long distance segment of the inmate service will quickly reduce the rates charged by the competitive long distance carriers to the most efficient cost. 13 24 Cost data provided to the FCC in filings by inmate telephone calling service providers also provide confirmation of the conclusions reached below. 51. The following calculations are intended to quantify a range of rates that would need to be charged by the underlying system provider under the proposal. These rates are intended to be profitable for the underlying system provider; thus, the rates include a profit margin in addition to costs. Because Evercom is the primary provider in the prisons under examination, the first set of calculations is based upon Evercom's costs as an example of how such costs might be calculated. The costs for other experienced providers should be similar. 52. Based on my knowledge of the industry, financial reports from Evercom,14 and evidence about Evercom's and other inmate service providers' costs from the public record in other cases, the basic components of prison system costs are defined below. The costs of a prison calling system include the hardware that makes up the prison phone system, maintenance, billing, administration and sales, uncollectibles, and the cost of providing long distance transmission and local termination. 53. The hardware in a prison calling system consists of the switch, the reco~ding system, the monitoring interface and the cost of telephones in those cases where the phones are not provided by the prison. The cost of switching hardware has dropped tremendously over the past few years. There are two primary types of switches that can be purchased -- carrier class switches and enterprise switches." A carrier class switch must be able to interface with the Evercom's December 31, 2000 Independent Auditor's Report by Deloitte & Touche LLP is an attachment to the I O-K Report, relevant portions of which are attached as Exhibit 2 hereto. I have also reviewed Evercom's 10-Q Report for the quarter ended September 30, 2001. Evercom's lOoK Report for 2000 is the most recent SEC report covering a full year, however. Because the September 30, 2001 10-Q Report covers only one quarter and shows little change from the data in the 10-K Report relevant to this analysis, this affidavit relies on the more complete 10-K Report. 14 " As used in this discussion, the term "enterprise switch" has a different meaning from the way that term is used in the FCC's Triennial Review Order. See Review of the Section 251 Unbundling Obligations of Incumbent Local Exchange Carriers, Report and Order and Order on 25 larger public switched telephone network and is generally used only by LECs. The switches required for prisons are enterprise switches, and are somewhat analogous to the large PBXs used by many businesses. Because enterprise switches are smaller and simpler than carrier class switches, they are far less expensive. A switch has several major components -- the line side connections, the trunk side connections, the operating software and a user interface. The line side connection is the hardware that interfaces with the telephone sets that use the switch. From the line side perspective, prison switches are relatively small switches. According to data included in the June 2, 1997 BOP RFP, the average federal prison has just under 44 telephone sets. 16 The trunk side connection is the interface to the public telephone network. As described elsewhere in this paper, these switches today only require only a handful ofT-} connections to the public switched telephone network -- making these relatively small switches. The most costly feature on a prison switch is the specific software that allows the switch to meet the various penological requirements listed earlier. If one were to develop such a switch for only one prison, such software would be quite expensive. However, most prison providers supply switching to many prisons, thus lowering the cost of this software on a per location basis. Evercom supplies switches to about 2,000 prisons, and thus its software cost is spread over many locations and is relatively inexpensive per switch. 54. The cost of switching has dropped drastically over the last few years. As an example, a small Class 5 carrier grade switch that can handle 5,000 lines would have cost $2 million - $3 million just a few years ago. In the last several months, such switches have been available from every major switch manufacturer -- Lucent, Nortel and Siemens, plus a number of the new soft switch manufacturers -- for under $600,000, due in part to the collapsing of the Remand and Further Notice of Proposed Rulemaking, CC Docket No. 01-338, FCC 03-36 (Aug. 21, 2003). There, "enterprise switch," see id at ~ 428 n. \335, refers not to a type of switch but to any carrier class switch used by a CLEC to serve large business customers. Id at ~'\l419-22. Here, it refers to the type of switch used by large non-carrier entities. 16 See BOP RFP, Exhibit J-I, attached hereto as Exhibit 5 (3850 telephones in 88 prisons). 26 ., telecom sector and resulting overcapacity. Even as far back as 1999, this Commission calculated that carrier class switches cost less than $500,000. 17 Switch costs have fallen considerably since then, and especially since the collapse of the high-tech bubble. Recently, observers have found carrier switches advertised for as little as $100,000.'8 As noted above, enterprise switches are far simpler and less expensive than carrier switches. Based on my recent experience in pricing switches for clients, a conservative current estimate for an enterprise switch with the features needed for a prison telephone system, including monitoring and recording equipment, would be approximately $350,000. 55. Moreover, there is a trend in the switching world that is going to lower the cost of switching even further in the near future, and this innovation is particularly relevant to prison calling systems. There are a number of new switches in the market referred to as soft switches. A soft switch is a switching device that separates the various switching functions into separate components. The major components of a soft switch are referred to as the call processor, the media gateway, the signaling gateway and the feature server. The call processor is the same as the core of the older switches and is the device that actually switches and routes calls. The media gateway is a device that allows for the interface to various other switching platforms. There is no real analog to the media gateway in older switches -- they were proprietary and self-contained. The signaling gateway allows the switch to interface with the SS7 network and thus use advanced features such as caller !D. Finally, the call feature server is the device that contains the unique systems and programs that operate the various features on the switch. The feature server in a soft switch would contain al1 of the unique penological features that distinguish prison 11 See Implementation o/the Local Competition Provisions o/the Telecommunications Act 0/ 1996, Third Report and Order and Fourth Further Notice of Proposed Rulemaking, 15 FCC Red 3696, 3812-13 (l999). J8 "State Regulators Courted by ILECs and IXCs on UNE-P Role," Communications Daily, April 28,2003, at 2 (comment attributed to Link Hoewing, Verizon Assistant Vice President-Internet), attached hereto as Exhibit 6. 27 switching systems from other systems. The availability of soft switches is relevant because they will allow a further large reduction in the cost of providing prison calling. With a soft switch platform, a prison provider could serve many prisons from one switching platform. For example, it would need only one feature server and one signaling gateway at some central site in the U.S. At each prison, it would need only the call processor. Such a distributed network would probably represent an additional 50 percent reduction over today's cost of switching, and this distributed architecture is ideally suited for applications like prison calling that require services at many different locations. Thus, any costs quoted in this paper can be expected to further decrease over time as technology takes yet another leap forward. 56. Service providers like Evercom are often required to provide the telephone sets as part of providing service to a prison. The phones used by prisons are more expensive than the average phones used by most business and residential users. Prison phones are more like payphone sets, in that they are built to stand up to heavy use. There are a vast number of types of payphones available in the marketplace. Payphones vary in the functions they must perform and in the ruggedness of the environment for which they are constructed. Prison paypbones can be of the "dumb" variety, i. e., they do not need to be able to perform such functions as coin counting. "Smart" telephones that process coins cost more than dumb phones that do not. Additionally, a prison phone does not need any of the advanced features often seen on payphones today, such as a scanner that can read in calling card information from a caller. Prison payphones can be of the most basic type, in that they require a keypad on which to dial the desired numbers, and, in the case of the competition proposal presented in this affidavit, to choose the desired IXC, but very little else. The payphone industry is very competitive, and there are a large number of manufacturers and thus a wide range of prices. Based on recent market research, there are 28 --_._---------------- pay phones that would work in a prison environment that are available from as low as $280 up to $550, with an average prison payphone price of$400.'9 57. In order to translate that per-phone set estimate into an estimated pay phone equipment cost for a typical prison, it is necessary to examine prison inmate telephone data. The data in the BOP RFP shows that the federal prison system has one telephone for every 25 prisoners.'· Using a subset of the federal data, an attachment to a report from the Virginia State Corporation Commission" shows a ratio of I telephone per 26 inmates. 22 The three sample CCA prisons served by Evercom have an average capacity of 1,743 prisoners'3 Using that sample, applying a ratio of one phone per 25 inmates yields an average of 70 telephones per prison. Applying the $400 average payphone cost to the estimate of 70 phones per sample prison yields a total payphone cost of $28,000 per prison. Adding that cost to the $350,000 switch estimate above results in an average total equipment cost per prison of $378,000. 19 Attached as Exhibit 7 are sample advertisements from payphone websites for equipment that would be suitable for inmate services showing prices as low as $149. An estimate of $400 for an inmate telephone set is also consistent with the Commission's estimate of$225 for a coinless payphone for general use in the Third Report and Order, and Order on Reconsideration of the Second Report and Order, Implementation of the Pay Telephone Reclassification and Compensation Provisions of the Telecommunications Act of 1996,14 FCC Red 2545, 2622, 2634 n.404 (1999), affd sub nom. American Pub. Communications Council v. FCC, 215 F.3d 51 (D.C. CiT. 2000). See BOP RFP, Exhibit J-l, attached hereto as Exhibit 5. The total number offederal prisoners shown in this chart is 97,579, and the total number of phones is 3,850, or 25.35 prisoners per phone. 2. Div. ofCommuns., Virginia State Corp. Comm'n, Report on Rates Charged to Recipients of Inmate Long Distance Calls (2000) ("Virginia Inmate Report"), attached hereto. as Exhibit 8. 21 Analysis of the Federal Bureau of Prisons Inmate Telephone System and Applicability to the California Department of Corrections, Executive Summary at 1 ("CDC Report") (attached to Virginia Inmate Report) (see Exhibit 8). 22 23 The three prisons are as follows: Central Arizona Detention Center - 2,304, Torrence County Detention Facility - 910, and Northeast Ohio Correction Center - 2,016. See Correctional Corporation of America web site, at http://www.correctionscorp.com!map.html. 29 58. In deriving an estimate of total operating costs, the cost of the switch and the telephones is reflected as depreciation expense. Evercom's audited financial statements show that it uses straight-line depreciation and that it uses depreciation lives of between 3.5 years and 7.5 years for telephone system equipment. 2' In the cost calculation set forth below, a depreciation life of 5.5 years is used, which is in the middle of Evercom's range of depreciation lives. This depreciation life also aligns very well with the typical length of a typical inmate service provider contract with a prison system, which is approximately five years,2S and with data filed by the Inmate Calling Services Providers Coalition ("Coalition"),2" of which Evercom is a member.27 59. Another major cost of providing service is maintenance expense. Maintenance expense includes spare parts, repairs and the personnel required to answer customer questions and keep the systems working. Most companies budget maintenance as a percentage of equipment costs. This ratio can be used for Evercom by taking the maintenance expense figure set forth in its 10-K Report. There, Evercom states that its maintenance expense has been steady See 10-K Report at Part II, Item 8, Notes to Consolidated Financial Statements, Note I, "Property and Equipment," attached hereto as Exhibit 2. 24 " See Carver, 54 Fed. Comm. LJ. at 395 n.20, attached as Exhibit 4 hereto. 2" The Coalition uses a depreciation life of five years in its calculations of equipment costs. See, e.g., Don J. Wood et al., "Inmate Phone Local Call Cost Study" D.3.3 (May 24, 2002) ("Inmate Cost Study") (attached to Comments of the Inmate Calling Service Providers Coalition, Implementation of the Pay Telephone Reclassification and Compensation Provisions of the Telecommunications Act of1996, CC Docket No. 96-128 (May 24, 2002) ("2002 Coalition Comments"». The relevant portions of the 2002 Coalition Comments are attached hereto as Exhibit 9. See ex parte letter from Robert F. Aldrich, Counsel to the Inmate Calling Service Providers Coalition, to Magalie Roman Salas, Secretary, FCC, at attachment captioned "Independent Inmate Phone Service Providers (as of May, 2000)" (May 9, 2000), the relevant portions of which are attached as Exhibit 10 hereto. 27 30 and varies little over time. 28 The amount of maintenance expense equates to approximately 13.2 percent of equipment costS.'9 This is the ratio used for estimating maintenance expense in the operating cost calculations set forth below. 60. Another cost that prison providers face is billing costs. Most inmate calling service providers do not have direct billing relationships with the family members of prisoners, or others receiving collect calls from prisoners, across the U.S. Instead, the inmate service providers typically pay a third party, usually the Regional Bell Operating Company or other LEC serving the party paying for an inmate call, to bill such parties for them. In its 10-K Report, Evercom states that billing costs paid to third parties vary between 2 percent and 3 percent of the revenues billed. 3• Accordingly, in the cost calculations set forth below, a figure equivalent to 2.5 percent of revenues is used to estimate billing costs, It should be noted that significant billing costs apply only to collect calling. The only billing cost required for a debit call is the cost of electronically extracting revenues from the pre-paid debit account, an insignificant expense per transaction. Given that Evercom provides both collect and debit calling to inmates, its reported billing costs represent an average for both types of services. Thus, its actual billing costs for collect calling only are probably much higher than 2.S percent of collect calling revenues. 61. In order to derive an estimate of billing costs, it is necessary to compute a composite average per-minute revenue amount to which the 2.5 percent ratio can be applied. The ratio of collect to debit calling varies from prison to prison, although there is still more collect calling than debit calling. For simplicity, it is assumed that, once this competitive 28 10-K Report at Part II, Item 7, "Field Operations and Maintenance," attached hereto as Exhibit 2. Evercom's 10-K Report shows maintenance expense of $6.67 million (in Part II, Item 6) and total equipment costs of $50.39 (in Part II, Item 8, Note 4 to Consolidated Financial Statements, "Property and Equipment"), a ratio of 13.2 percent ($6.67M1 $50.39M). See Exhibit 2. 29 3°Id. at Part I, Item I, "Billing Arrangements." 31 proposal is adopted, it would be reasonable to expect that half of the calls will be debit calls and half will be collect. Using Evercom's tariffed rates during a portion of the period it was providing service to the CADC, TCDF and NOCC -- a debit card rate of $0.65 per minute and a collect calling rate of $0.59 per minute plus a $3.95 per collect call charge -- and assuming a tenminute call, the composite calling rate charged to inmates would be $0.82 per minute in the cost calculation below. 'l If prisons were to switch to debit calling only for inmate calls, billing costs would essentially disappear. 62. Another major cost for inmate service providers offering collect calling is the cost of uncollectibles, as mentioned previously. Evercom does not show uncollectibles as a separate item in the financial statements in its 10-K Report. Evercom does state in the 10-K Report, however, that although inmate prepaid calling services have minimal uncollectible expenses,32 called parties' failure to pay for inmate collect calls place unique demands on this sector of the industry.'3 Data provided by the Coalition in an ex parle letter filed in April 2000 with an During the period from September 14,1999 to the detariffing of Evercom's rates on June 27, 2000, Evercom's standard tariffed debit card service rate, which applied to its Inmate-only Debit Account Service, was $0.65 per minute. See Evercom Systems, Inc. Tariff FCC No.1, Section 3.4.1 (effective Sept. 14, 1999), and FCC Public Notice, Tariff Transmittal Public Reference Log (June 29, 2000), and its standard tariffed rate for interstate, interexchange operator assisted inmate calls, including collect calls, was $0.59 per minute plus a $3.95 service charge. See Evercom Systems, Inc. Tariff FCC No. I, Section 3.5 (effective Sept. 14, 1999). For a tenminute collect call, that comes to $0.99 per minute for collect calls. The average 0£$0.65 and $0.99 is $0.82 per minute. The relevant portions of Evercom's Tariff No. I are attached as Exhibit 11 hereto. 31 It should be noted that in the event that this competitive proposal is adopted, actual inmate rates will be far lower than they have been in the recent past. The 82 cent rate is used here purely as a conservative estimate. As demonstrated below, the cost of billing drops out in deriving the cost of providing the underlying inmate telephone system. 32 IO-K Report at Part I, Item I, "Products and Services" ("Prepaid Services''), attached hereto as Exhibit 2. 33Id. at Part I, Item I, "Industry Overview." 32 attached analysis of the cost of providing a 12-minute local inmate collect call ("Coalition Cost Analysis"), show a typical uncollectibles rate for inmate collect calling of 14 percent ofrevenues, and, in some cases, over 23 percent. 34 Accordingly, the cost calculation below uses a 15 percent uncollectibles rate to apply to collect calling. Because there are virtually no uncollectibles from debit calls, for which revenue is collected directly from prepaid accounts, however, the overall uncollectibles rate must be adjusted to take into account a mix of collect and debit calling. Using the assumption discussed above that half of the calls will be debit calls and half will be collect, the composite uncollectible rate would be 7.5 percent of total revenue, and that rate is applied to an assumed composite calling rate of 82 cents per minute in the cost calculation below.ls 63. One of the largest costs incurred by inmate calling service providers is the category of "Administration, General and Sales" expenses. On Evercom's financial statements, this includes a broad category of costs. In addition to the cost of the salespeople who sell to prisons and related expenses, it includes the following types of costs: executive salaries, board of director expenses, accounting, legal, human resources, computer networks, insurance, the cost of running corporate headquarters and other overhead costs. In Evercom's case, for 2000, these costs were roughly 2.6 times greater than maintenance costs. 36 Accordingly, a ratio of2.6 times Ex parte letter from Jacob S. Farber, Counsel to the Inmate Calling Service Providers Coalition, to Magalie Roman Salas, Secretary, FCC, at attachment, "Inmate Service Fee - 12 Minute Local Call Cost Analysis" (April 6, 2000) (uncollectibles rate for inmate collect calls of 14 percent) ("Coalition Cost Analysis',), attached hereto as Exhibit 12. See also, 2002 Coalition Comments at 3-4; Inmate Cost Study at Workpapers labelled Input C, Input G, Input H, Input N, Input 0 and Input P (showing inmate collect uncollectibles rate of over 23 percent), and Input Q (showing uncollectibles rate of over 19 percent), attached hereto as Exhibit 9. 34 3S As explained below, the cost of uncollectibles, like billing costs, drops out in deriving the cost of providing the underlying inmate telephone system, since the underlying system operator recovers its costs through rates charged to the competitive interconnected long distance carriers terminating each call, obviating any billing or uncollectibles costs. 10-K Report at Part I1,ltem 6 (showing maintenance costs of $6.7 million and seJling, general and administrative costs of$17.7 million). See Exhibit 2. 36 33 .----.~~---------- maintenance costs is used to represent an allocation for administration, general and sales expenses in the cost calculation below. It should also be noted that the Coalition has represented overhead expenses to be just slightly less than 2.S times maintenance expenses in their filings with the FCC. J7 64. Another cost of providing long distance inmate calling service is the cost incurred in the transmission and termination of the calls, i.e., the cost oflong distance transport to the called party's local calling area and the cost of terminating each long distance minute at the final destination. In the competitive scheme described here, this cost would be borne by the competitive interconnecting carriers, rather than the underlying inmate telephone system provider. For long distance transport, carriers typically use T-Is or larger circuits. In this case, such circuits would begin at each prison switch and reach to the nearest POP on the interconnecting carrier's toll network. Using the average of70 telephones in each prison, as discussed above, a service provider would need approximately three T -I s for transport to its long distance network. Because a single T-I has 24 voice channels available, three T -1 s would allow for 72 simultaneous calls. Based on my recent experience, an average T-I circuit costs around $400 per month. T-I costs vary drastically across the U.S. by market, but $400 is a conservatively high estimate ofthe composite monthly cost ofT-Is across the country. Accordingly, an annual transport cost of$14,400 is used in the cost calculation below?8 65. Evercom, like most inmate service providers and other IXCs, does not own a nationwide long distance network. It therefore has to pay a wholesale IXC to carry each long distance inmate call to the recipient's local calling area and to arrange for local terminating 31 See Coalition Cost Analysis (showing overhead -- $0.224 per call-- equal to 2.49 times maintenance -- $0.09 per call -- for an inmate local collect call), attached hereto as Exhibit 12. Four hundred dollars per month for an average T-l circuit is a rate that would be available only to a carrier purchasing a fairly large volume of capacity. That rate multiplied by three circuits mUltiplied by 12 months equals $14,400. 38 34 access to the recipient. Such wholesale long distance contracts are routine for long distance resellers like Evercom, which typically use the underlying network of one or more large IXCs, such as AT&T, MCI (formerly known as WorldCom) or Sprint.'· A carrier would have to pay no more than 2.5 cents per minute to get long distance calls terminated through one of these facilities-based IXCs. As far back as 1996, the Coalition estimated that its members' long distance transmission cost was approximately 2.5 cents per minute.'· Since then, long distance wholesale costs have declined drastically, and wholesale long distance terminating rates, including terminating access charges paid to the terminating LEC, are now as low as 1.8 cents per minute for large volume users." Accordingly, the 2.5 cents per minute rate will be used as a conservative estimate oflong distance transmission plus termination in the cost calculation below, although these costs are certainly lower today. 66. Finally, it is necessary to estimate the volume oflong distance usage from the average prison. Based on available data, a low and a high estimate of calling volume can be derived in order to develop a range of possible per-minute costs. This exercise also shows that costs decrease with increased calling volume. The low estimate assumes that each pri~oner averages one hour of calling per week, and the high estimate assumes that each prisoner averages It should be noted that, although resellers obtain facilities from other carriers, a reseller carrying an inmate call would nevertheless be fully capable of retaining complete control over the entire transmission of the call. 39 See Comments oflnmate Calling Services Providers Coalition at 8 n.14, Implementation of the Pay Telephone Reclassification and Compensation Provisions of the Telecommunications Act of 1996, CC Docket No. 96-128 (July I, 1996) (" 1996 Coalition Comments''), attached hereto as Exhibit 13. AmeriTel Pay Phones, Inc. and InVision Telecom, Inc., see id. at 1 n.l, were predecessors to Evercom. See IO-K Report at Part I, Item I, "General," attached hereto as Exhibit 2. 40 " Of that 1.8 cents per minute, only .71 cents per minute was accounted for by local terminating access charges as of June 2003. See Industry Analysis and Technology Division, Wirelinc Competition Bureau, Federal Communications Commission, Trends in Telephone Service at Table 1.2 (August 2003). The relevant portions of the FCC's report are attached hereto as Exhibit 14. 35 1.5 hours of calling per week. The low estimate is derived from a report prepared by the California Department of Corrections concerning the BOP inmate telephone system, which estimated that BOP inmates average 242 minutes of calling per month (approximately one hour per week)." The higher estimate is derived from the BOP RFP discussed above, which indicates that the average federal prisoner makes 1.4 hours of long distance calls per week." 67. There is one additional cost of prison calling that is not included in these cost figures. Many prison systems charge a commission to inmate service providers as a cost of doing business in the prison. As this Commission reiterated in the Inmate Payphone NP RM, location rents (i.e., commissions) are not legitimate costs of providing service; rather, they are an element of profit." Additionally, not all prisons systems charge commissions. For these reaSons, commissions have been excluded from these cost calculations. Commissions have also been removed from the comparable costs figures cited from other FCC filings discussed throughout this affidavit. It should be noted that, although commissions are not a legitimate expense of inmate calling services, as a practical matter, they nevertheless inflate the rates charged by Evercom and other service providers. According to the Coalition Cost Analysis, com~issions amount to 30 percent of the total cost of inmate calls, including all profit." If that is true, commissions add another 43 percent (i.e., 30% /70%), to total costs before commissions, which must be presumed to exert a commensurate upward pressure on calling rates. " CDC Report, Executive Summary at I (attached to Virginia Inmate Report) (attached hereto as Exhibit 8). Exhibit J-2 of the BOP RFP shows an average of 4,991 minutes per year ofte1ephone usage per inmate, of which 749 minutes are local calls, for an average of 4,242 long distance minutes per year, which is slightly under 1.4 hours per week oflong distance calling. Exhibit J-2 is attached hereto as Exhibit 15. 43 Order on Remand & Notice of Proposed Rulemaking, Implementation of the Pay Telephone Reclassification and Compensation Provisions of the Telecommunications Act of 1996,17 FCC Rcd 3248, 3255 & n.49 (2002) ("Inmate Payphone NPRM,). 44 4j See Coalition Cost Analysis, attached hereto as Exhibit 12. 36 68. Following is a calculation of the total cost per minute of running a prison calling system using all of the assumptions and inputs discussed above. Because costs vary by call volume, one can easily postulate that costs also vary by prison size, with larger prisons having lower per minute costs. As noted above, the three sample CCA prisons currently or previously served by Evercom have an average population of 1,743 prisoners. The cost calculation is set forth in two columns, with the first column showing low prisoner calling at one hour per prisoner per week and the second column showing 1.5 hours of calling per prisoner per week. Each entry will first be calculated on an annualized basis, rounded off to the nearest thousands of dollars, and then divided by the low and high call volume estimates in order to derive low and high perminute costs.46 Estimate of Evercom Costs Low Estimate High Estimate Average Number of Prisoners Average Calling Per Prisoner Per Week 1,743 1.0 hr 1,743 1.5 hr Calling Hours Per Week Annual Minutes 1,743 5,438K 2,615 8,157K $ 136K $ 14K $ 150K $ 204K $ 14K $ 218K Operating Costs Wholesale Long Distance and Termination" Transport Total Long Distance Costs Because of the unavoidable inefficiencies of serving extremely small facilities, this analysis may not apply to locally-administered jails and other low-capacity prison facilities. 46 The estimated wholesale cost oflong distance transmission and termination of2.5 cents per minute was multiplied by the low and high annual estimated minutes to derive low and high annualized totals. 47 37 Depreciation'8 Maintenance'9 Billing'O Uncollectibles'l Administration & Sales s2 $ $ $ $ $ Total Expenses Total Cost per Minute 69. 69K SOK 111K 334K J30K $ 69K $ SOK $ 167K $ S02K $ 130K $ 844K $I,136K $ 0.155 $ 0.139 This demonstrates that the total cost of providing long distance inmate calling service, before profit and taxes, is somewhere between 13.9 cents and 15.5 cents per minute. This is far below the revenues providers like Evercom collect for interstate calling, as discussed above. 70. From these total cost estimates, it is then possible to break out the cost of providing just the underlying inmate telephone system by eliminating the long distance and other '8 As described in paragraph 58, depreciation is based on an average useful life for all equipment of 5.5 years. The equipment costs are as follows: Hardware Switch Telephones Total Hardware $ 3S0K $ 28K $ 378K ($378,000/5.5 = $69,000). As described in paragraph 59, maintenance is estimated at 13.2 percent of the total equipment costs. 49 50 Billing costs are estimated to be 2.5 percent of billed revenues per paragraph 60. These amounts were calculated by assuming that average billing is 82 cents per minute for the assumed minutes multiplied by 2.5 percent. 'I As explained in paragraph 62, uncollectibles are calculated by taking 7.5 percent of total revenues, based on a composite revenue estimate of 82 cents per minute. As explained in paragraph 63, general, administration and sales expenses are estimated by multiplying maintenance expenses by 2.6. ,2 38 costs that could be avoided by a finn acting solely as the provider of the underlying system. For example, the actual cost of providing the long distance transmission -- both the network costs per minute and the transport -- would become the responsibility of each competing interconnecting lXC. Also, under the system described here, because the underlying system provider would bill its per-minute charge to the competitive interconnected IXC tenninating each call, the underlying system provider would have no billing or uncollectibles costs. Moreover, because most calls would become prepaid debit calls under a competitive system, the cost of billing and uncollectibles would largely disappear in any event." Eliminating the avoided costs, the costs of providing the underlying inmate telephone system for long distance service is as follows: Low Estimate High Estimate Total Expenses (from above) $ 844K $1,136K Less Avoided Costs Less Long Distance Costs Less Billing Less Uncollectibles Total Underlying System Costs $ 150K $ IIIK $ 334K $ 249K $ $ $ $ Underlying System Cost per Minute $ 0.046 $ 0.031 218K 167K 502K 249K This demonstrates a range of costs for the underlying system provider of 3.1 cents to 4.6 cents per minute. Note that the cost per minute decreases with a greater calling volume. 71. There are two possible categories of costs to add to these figures. First, it is reasonable to allow the underlying system provider to make a profit. In the wholesale long distance business, a reasonable profit for most carriers, after all costs, is roughly one cent per " It should be noted that in the BOP inmate telephone system, 92 percent ofthe long distance calls are prepaid debit calls, and the rest are collect. Virginia Inmate Report at 14, attached hereto as Exhibit 8. 39 '4 minute. This estimate of profit compares well with the profit estimated by the Coalition and thus is a reasonable profit component. Along with profit comes the need to recognize the cost of income taxes. Evercom is a relatively young company and, as such, it has yet to pay any significant income taxes." For other providers, however, and, eventually, for Evercom, there would be income taxes to be recovered. While taxes for most providers are theoretically as much as 40 percent (when using the maximum possible tax rate), most telecommunications carriers pay less than a full tax rate because of various tax loopholes and write-offs. A tax level of 25 percent is typical for the industry over the long run. Accordingly, assuming profit of one cent per minute, income taxes might eventually be around $0.0025 per minute, or $0.003 per minute, rounded off to the nearest tenth of a cent. Adding $0.013 per minute for income taxes and profit, the reasonable rate for providing the underlying inmate telephone system is calculated to be between $0.044 and $0.059 per minute. 72. These calculated costs are comparable to the costs of providing inmate calling services as reflected in the Coalition Cost Analysis, which is attached hereto as Exhibit 12. That analysis shows a total cost, less commissions, of$1.508 for a 12-minute local call, or $0.126 per minute." The basic costs for providing local inmate collect calls are very similar to the costs of providing long distance inmate collect calls. The difference between the two categories, from a cost perspective, is the difference between the cost of transport and termination of the long distance call and the local service charge for carrying the local call to the public telephone "See Coalition Cost Analysis (showing profit of8.2 cents on a 12 minute local inmate collect call), attached hereto as Exhibit 12. II See, e.g., 10-K Report at Part II, Item 6 (income taxes for 2000 -- $553,000 -- slightly above one quarter of one percent of total operating expenses of$218,804,000), attached hereto as Exhibit 2. S6 As explained above, commission payments to prisons are not a legitimate expense. The commissions cost of $0.647 for a 12 minute local inmate call has therefore been removed from the Coalition's total cost estimate of$2.155 in the Coalition Cost Analysis, attached hereto as Exhibit 12. 40 network. In the Coalition Cost Analysis, the Coalition indicates that the LEC service charges for carrying a 12-minute local inmate collect call to the public telephone network are $0.243, or $0.020 per minute.51 In order to use the Coalition's data in an apples-to-apples comparison with the long distance inmate service cost calculations presented in this affidavit, the cost of the long distance transmission and termination plus the cost of transport to the long distance carrier must be substituted for the Coalition's local service charges. In the long distance cost calculations presented above, the costs oflong distance transport and termination equate to about $0.027 per minute." Substituting that figure for the Coalition's local service charge in its cost analysis yields the following: Adjusted Coalition Costs Long Distance Costs (substituted for local costs) Billing & Validation Maintenance Depreciation Overheads Profit Uncollectibles Total Cost 12-Minute Call I-Minute Call $0.324 $0.027 $0.350 $0.090 $0.110 $0.224 $0.082 $0.410 $1.590 $0.029 $0.008 $0.009 $0.019 $0.007 $0.034 $0.133 It should be noted that the Coalition's adjusted cost of$0.133 per minute is even less than the lower estimate of the cost of inmate calling presented above, or $0.139 per minute, which does not include profit or taxes. It must be assumed that, in light of the inmate calling service providers' interest in higher rates, the Coalition data does not understate the cost of providing 51 Coalition Cost Analysis, attached hereto as Exhibit 12. ,. Dividing the "low estimate" long distance costs of $150,000 by the low annual traffic estimate of 5,438,000 minutes yields a per-minute cost of $0.02758. Dividing the "high estimate" long distance costs of $218,000 by the high annual traffic estimate of 8, 157,000 minutes yields a perminute cost of $0.02672, for an overall estimate of slightly over $0.027 per minute. 41 inmate telephone service. Because the adjusted Coalition-based data results in a lower cost estimate than the low estimate calculated above from Evercom data, the cost of providing the underlying inmate telephone system is likely to be at the low end of the range of costs calculated above, if not even lower. Moreover, the cost analysis presented here allows the underlying provider a greater profit per minute than is claimed in the Coalition figures',further confirming that the cost estimates presented here might overstate, but certainly do not understate, the costs of inmate calling. 73. In order to compare the adjusted Coalition estimate to the estimated cost of providing the underlying inmate telephone system presented above, it is necessary to remove the avoided costs of long distance, billing and the uncollectibles to arrive at the cost of the underlying system, as follows: 12-minute Call Per Minute Adjusted Coalition Total Costs $1.590 $0.133 Less Avoided Costs Long Distance Costs Billing & Validation Uncollectibles Underlying System Costs $0.324 $0.350 $0.410 $0.506 $0.027 $0.029 $0.034 $0.043 The adjusted Coalition data demonstrates a cost of $0.506 for a 12-minute call, or $0.043 per minute. This is even lower than the low end of the estimates of the cost of providing the underlying system presented above, which range between $0.044 and $0.059 per minute, thereby confirming the conservative nature of the cost calculations presented here. 74. Finally, As explained previously, the estimates of the total cost of providing inmate long distance calling service presented above -- $0.139 to $0.155 per minute before profit and taxes -- are a composite of debit and collect calling costs. As also explained above, billing costs and uncollectibles virtually disappear in the case of debit account or debit card calling. 42 _ _._--------------------.. Because bilJing costs and uncollectibles account for such a large portion of the total cost of providing inmate long distance calling service, debit calling could be provided much more cheaply than collect calling. Removing billing and uncollectibles costs from the composite total cost estimates reduces them by over six cents per minute, which is a tremendous proportion of the total cost of providing inmate long distance debit and collect services. Thus, long distance inmate debit calling could be provided at much lower rates than long distance inmate collect calling service. 75. Taken together, the analysis presented here and the comparison with the Coalition's data demonstrate that there exists a reasonable range of rates at which an inmate telephone system provider could operate an inmate calling system, make a reasonable profit and still leave room for multiple interconnecting long distance carriers to compete for inmate long distance calling. The range of estimates reflects the economies of scale in providing prison inmate calling and the different possible methods of calculating costs. These estimates demonstrate that a competitive prison inmate calling system of the type described in this affidavit is technologically and economically feasible and would result in much more affordable calling for prisoners. Moreover, as explained in Part VI above, such a system would meet all legitimate security, anti-fraud and other penological goals. DOUGLAS A. DAWSON STATE OF (flarylo.lIJ CITY OF th.U, r dM~ Sworn to before me this ;;'9~ day of October, 2003. ~IiA!J~) SHERRI N. SPENCE NOTARY PUBLIC STATf Of MARYlAND My Commi"ion expire. Augu.t 14, 2007 43 .. -~ - ~.-------------------