Personnel Friction Analysis MAJ Mark Zais, MAJ Jose Ramirez Army G1, Military Strength Analysis & Forecasting

Transcription

1 Personnel Friction Analysis MAJ Mark Zais, MAJ Jose Ramirez Army G1, Military Strength Analysis & Forecasting ABSTRACT In this paper we develop a methodology to define, measure, and predict personnel friction in the US Army. Personnel friction represents the number of Soldiers misaligned by grade, military occupation specialty (MOS), and/or location (brigade level) plus any additional Soldiers unavailable for training or deployment. Previous studies have referred to the concept of friction as the cost of doing business for organizations that have numerous specialties, ranks, and locations. However, it is in the Army s interest to provide a personnel inventory that is truly sufficient for accomplishing its missions while minimizing inefficiency. Quantifying and predicting the components of personnel friction should facilitate the development of policies to reduce the cost of doing business. In this study, our goal is to develop a methodology to analyze personnel friction levels based on data and modeling resources that are available to the Army G-1 and Human Resources Command (HRC). INTRODUCTION The Army is resourced for one face in every space, and expectations in the field are that all authorizations will be filled by skill and not lower than one grade below the authorization. Personnel friction prevents Army units from having the right number and types of Soldiers to meet mission requirements. In order to meet authorizations by grade and skill, faces must exceed spaces. If the expectation is to fill by at least grade and skill, the Army must have enough slack or flexibility in the manning system to compensate for the uncertainty in the manpower demand (e.g., changes in force structure, deployment requirements, etc.) and the manpower supply (e.g., changes in policy, Soldier availability, etc.). The purpose of this paper is to better define and quantify the flexibility required. Specifically, it proposes a methodology for measuring the impact of skill and grade imbalances as well as policy constraints on manning the force by grade and skill. In addition, it proposes a way forward to better calculate the additional flexibility required in time of war. When the US Army meets its overall strength goals, why are commanders still saying I don t have all the people I need to accomplish the mission? If there are enough faces to fill the spaces at a macro level (Army), why do they fail to match at the micro level (units)? Without careful analysis, the quick reaction may be to place sole blame on the inefficiencies of the personnel system. However, acceptance of the status quo encourages us to overlook policy changes that could reduce inefficiencies. Given a system comparable to just-in-time delivery with no inventory on the shelves, one could correctly argue that the Army personnel system produces much better results than should be expected. This doesn t mean that improvements cannot be accomplished. Previous studies have referred to the concept of friction as the cost of doing business for organizations that have numerous specialties, ranks and locations such as the Army. A certain level of misalignment has always been viewed as a byproduct of a dynamic system with many moving parts. Even the staunchest critics of the Army personnel system would acknowledge that the distribution process (meeting the specific needs of both units and individuals) is a dubious and unenviable challenge. Unlike object-oriented supply chain systems, personnel systems do not grant the luxury of extra inventory stock to meet just-in-time needs. We are operating in a personnel system with no slack that is not designed for rapid adjustments or just-in-time delivery. However, as persistent conflict increases the stress on the force and the scale tilts towards demand over supply, there exists a greater need and urgency to maximize effectiveness and readiness. Years of persistent conflict have improved some measures of effectiveness by creating a better trained and more experienced force. Unfortunately, persistent conflict appears to compound personnel friction rather than reduce it. 1

2 The purpose of this study is to develop a methodology to define, measure, and predict personnel friction levels. Once an accurate way of measuring and forecasting personnel friction is accepted, we can begin identifying how components of personnel friction are impacted by structure, demand and policy. More importantly, we can shape future decisions in order to reduce the cost of doing business. In doing so, we can answer a series of questions. What portion of the existing friction level is avoidable? How can avoidable friction be reduced? What can we expect of future friction levels? How much personnel inventory is needed to compensate for unavoidable friction? How are any of these answers affected by proposed structure and policy changes? Personnel friction is defined as the number of personnel misaligned by grade, military occupation specialty (MOS) or location; and/or unavailable for training or deployment at a given point in time. It is a measure of how well personnel are aligning with the force structure so that fighting and supporting units have the Soldiers and capabilities they need to accomplish their missions. Personnel friction can be broken up into three separate categories inventory friction, distribution friction, and readiness friction. Each of these categories is described by a different measure of misalignment or unavailability (Figure 1). However, the three categories of personnel friction are not mutually exclusive and therefore, it is possible for individual Soldiers to contribute to more than one category (Figure 2). As we progress towards measuring and Figure 1 forecasting these categories, a lack of mutual exclusiveness greatly complicates the calculations and demands an approach that accurately describes the overall friction level without overstating the problem. Prior to evaluating how to measure the interaction of the personnel friction categories, we identified the most accurate methods for measuring each individual category. INVENTORY FRICTION DEFINING INVENTORY FRICTION Inventory friction is defined as the number of personnel misaligned by grade or MOS. Therefore, inventory friction is the difference between the personnel inventory (faces) and authorizations (spaces) at the grade and MOS level of detail. What measure do we use to determine the level of inventory friction in the personnel system and how do we account for the fact that there may not be enough people to cover all of the structure? Let s look at a simple example where inventory friction is demonstrated with fictional specialties 1A, 2B, 3C, and 4D. Here, the organization (shown in Table 1) has four different specialties (labeled 1A, 2B, 3C, and Figure 2 4D) and only one grade of E5. Inventory and authorizations both sum to 120, but there are discrepancies in the way that personnel have been distributed by specialty. Since [total strength total authorizations = 0], it may look like the organization is filling all of its personnel requirements, but there is an MOS issue that is not addressed by looking at the 2

3 aggregate numbers alone. The misalignment of MOS 1A (5 less people) and MOS 3C (5 more people) is evident only when comparing authorizations and strengths at the MOS level of detail. This table could have easily shown one MOS and four different grades to show how the grades could also be misaligned. MEASURING INVENTORY FRICTION MOS Grade Authorizations Strength Str-Auths 1A E B E C E D E Total Table 1 In order to calculate the amount of inventory friction, it is necessary to look at the strength deviations (strength authorizations) at the grade and MOS level of detail. Table 1 shows why simple addition ( = 0) of the deviations is not a good measure. Another way to measure is to add the absolute value of the deviations which would give the example in Table 1 a friction level of 10. However, while there are ten positions that are either under-filled or over-filled, only 5 people are actually misaligned. By the definition of friction, the number of misaligned personnel should be the friction level and is is determined by just summing the absolute value of the deviations and dividing by 2, trength-authoriza on. The sum of the deviations is divided by 2 since each mismanaged Soldier is counted twice negative in one MOS and positive in the other. The calculation becomes more complex in situations where aggregate inventory and authorizations do not match. Friction must be defined to exclude the under fill due to authorizations greater than strength in the aggregate (negative operating strength deviation). pera ng trength by grade and Authoriza ons by grade and pera ng trength evia on Table 2 shows an organization that has an operating strength deviation (OSD) of -35 and does not have the ability with the personnel on hand to fill the structure for all of its positions. Nothing can be done to align the personnel inventory with the force structure and yet the sum of the deviations is 45. Table 3 shows an organization with an OSD of +10 and a sum of the deviations of 30. The mismanagement of personnel in Table 3 is more egregious than that in Table 2. The negative OSD in Table 2 causes the sum of the deviations to be higher than the sum in Table 3. Clearly, the operating strength deviation needs to be accounted for when trying to compute the amount of friction in an organization. MOS Grade Authorizations Strength Str-Auths 1A E B E C E D E Total OSD =-35 Table 2 = 45 Friction = 5 MOS Grade Authorizations Strength Str-Auths 1A E B E C E D E Total OSD = +10 Table 3 - = 30 Friction = 10 3

4 - Monthly Inventory Friction = Proof: If the operating strength deviation is negative (OSD <0), then the only misaligned personnel are the ones that are in a positive deviation MOS (see Table 2, MOS 1A). =, when =, when = = = = The same approach can be used to show the. In an ideal case, an organization can be perfectly aligned with no friction, but it still may have a large problem of having too many or not enough personnel. In this case, the above equations would result in a friction of zero. OSD is used to compute the amount of friction, but the size of the organization is not a factor measured according to this definition of inventory friction. The friction equation, by definition, only compares the current strength versus the current authorizations. OSD represents the size of the force and whether the organization is under or over-strength. Historical inventory friction can be represented in whole numbers or as a percentage. If represented as a percentage, we use the force structure document to calculate the number of personnel who will be misaligned with their MOS/grade. FSA*friction% = number of personnel who are misaligned by grade and/or MOS (friction level). Therefore, the historical percentage = fric on level A * 100. The friction number is currently calculated by comparing the applicable force structure documents to the monthly operating strength. Skill level 1 Soldiers are combined together based on the assumption that skill level one personnel have flexibility and exchangeability in positions. In order to conduct these calculations, we use SAS 9.2 to compare personnel data from the Total Army Personnel Database (TAPDB) to the valid structure document for each month (structure documents are normally modified on a quarterly basis). Each month of personnel history in TAPDB contains records for more than 500,000 active Army personnel. Each record contains dozens of data fields with attributes for each Soldier. The equations shown above are applied to months of separate personnel and structure databases. In total, SAS programming is used to sort through roughly 50 million records, merge relevant fields and conduct historical measurements. After the data has been manipulated, SAS programming such as in the example below is used to apply historical calculations. PROC SQL; CREATE TABLE MOS_GRADE_FRICTION AS SELECT DISTINCT YM_DT, GB, MOS, SUM(AUTH) AS AUTH, 4

5 QUIT; SUM(OPER) AS OPER, SUM(OPER-AUTH) AS OSD, ABS(SUM(OPER-AUTH)) AS ABS_DEV FROM WORK.MERGED_DATA1_MOSG1 GROUP BY YM_DT, GB, MOS; DATA WORK.FINAL_MOSG_FRICTION; SET MOS_GRADE_FRICTION; FORMAT MOSG_PERCENT 6.5; IF AUTH NE 0; IF OSD < 0 THEN DO; MOSG_FRICTION = (ABS_DEV+OSD)/2; MOSG_FILL = OPER - MOSG_FRICTION; MOSG_PERCENT = MOSG_FRICTION/AUTH; END; IF OSD >= 0 THEN DO; MOSG_FRICTION = (ABS_DEV-OSD)/2; MOSG_FILL = AUTH - MOSG_FRICTION; MOSG_PERCENT = MOSG_FRICTION/AUTH; KEEP YM_DT MOSG_FRICTION MOSG_FILL MOSG_PERCENT; END; FORECASTING INVENTORY FRICTION Knowing the above mentioned methods of calculating historical inventory friction, we can forecast future levels of inventory friction as long as we have an Army strength forecast by grade and skill. Fortunately we do. The Army G-1 Enlisted Grade and Enlisted Specialty models, in concert with the Officer Forecasting Model, project inventory levels of all grades and skills 84 months into the future. A limitation of any technique used to forecast inventory friction is the great unknown, force structure changes. As a result, any inventory friction forecast is merely the best possible scenario, promised to deteriorate with each structure change in inverse proportion to the amount of lead time the personnel system is given to adjust. Once again SAS programming is used not only to perform the calculations for measuring inventory, but is also used to extract forecasted personnel inventory data (from a database produced by another system of optimization and forecasting models). Presented below is an example of a piece of macro programming code used to collect projected inventory data for a range of future months. %MACRO MANYPROJ(ES_PROJ_VER_NM, PROJ1_YM_DT, PROJ2_YM_DT); %LET START = %SUBSTR("&&YM_&PROJ1_YM_DT",6,2); %LET TABLE = %SUBSTR(&&YM_&PROJ1_YM_DT,1,4); %LET MAX = %SYSFUNC(PUTN(84,Z2.)); %LET MIN = %SYSFUNC(PUTN(1,Z2.)); %LET END = %SUBSTR("&&YM_&PROJ2_YM_DT",6,2); %IF &END LT &START %THEN %DO; %DO I = &START %TO &MAX; %LET J&I = %SYSFUNC(PUTN(&I,Z2.)); %LET K = &&J&I; %IF "&TABLE" NE "CURR" %THEN %DO; %IF &K >= %SUBSTR(&FIRST_P_CURR,5,2) %THEN %DO; %LET TABLE = CURR; %PULL_AUTH_STR_PROJ; DATA WORK.COMBINE1; SET %DO I = %EVAL(&START) %TO %EVAL(&MAX); WORK.AUTH_&&J&I WORK.TRAINEE_&&J&I 5

6 ; %GB TTHS = SUM(THS, TRAINEE); TOTAL_STR = SUM(TTHS, OPER); PROC DATASETS LIBRARY = WORK; DELETE AUTH_&START - AUTH_&MAX TRAINEE_&START - TRAINEE_&MAX / MEMTYPE = DATA; QUIT; %DO I = &MIN %TO &END; %LET J&I = %SYSFUNC(PUTN(&I,Z2.)); %LET K = &&J&I; %PULL_AUTH_STR_PROJ; DATA WORK.COMBINE; SET %DO I = %EVAL(&MIN) %TO %EVAL(&END); WORK.AUTH_&&J&I WORK.TRAINEE_&&J&I WORK.COMBINE1; %GB TTHS = SUM(THS, TRAINEE); TOTAL_STR = SUM(TTHS, OPER); PROC DATASETS LIBRARY = WORK; DELETE AUTH_&MIN - AUTH_&END TRAINEE_&MIN - TRAINEE_&END COMBINE1/ MEMTYPE = DATA; QUIT; %ELSE %IF &END GE &START %THEN %DO; %DO I = &START %TO &END; %LET J&I = %SYSFUNC(PUTN(&I,Z2.)); %LET K = &&J&I; %IF "&TABLE" NE "CURR" %THEN %DO; %IF &K >= %SUBSTR(&FIRST_P_CURR,5,2) %THEN %DO; %LET TABLE = CURR; %PULL_AUTH_STR_PROJ; DATA WORK.COMBINE; SET %DO I = %EVAL(&START) %TO %EVAL(&END); WORK.AUTH_&&J&I WORK.TRAINEE_&&J&I %GB TTHS = SUM(THS, TRAINEE); TOTAL_STR = SUM(TTHS, OPER); PROC DATASETS LIBRARY = WORK; DELETE AUTH_&START - AUTH_&END TRAINEE_&START - TRAINEE_&END / MEMTYPE = DATA; QUIT; 6

7 DISTRIBUTION FRICTION DEFINING DISTRUBTION FRICTION Distribution friction is defined as the number of personnel misaligned by location. When a Soldier is correctly assigned by skill and grade but is assigned to the wrong location (unit), distribution friction exists. In this analysis, location is extended down to the brigade or Distribution Management Sub-Level (DMSL). Therefore, distribution friction between battalion, company or smaller size units is not included. While these smaller unit structures are easier to manage in terms of leveling authorizations with personnel inventory, it is arguable that reducing distribution friction among brigade and higher unit structures is a necessary target for readiness improvement. MEASURING DISTRIBUTION FRICTION Historical distribution friction is calculated using the same method as inventory friction. When measuring inventory and distribution friction together and accounting for inclusiveness, double counting is avoided by disaggregating the strength and authorization data and grouping by grade, skill and location at a given interval of time. - Monthly Distribution Friction = Monthly Inventory & Distribution Friction = - For both inventory and distribution friction, historical calculations require the use of multiple force structure documents. In order to calculate these rates, we must compare past personnel inventory strengths (monthly) with the valid structure document. These documents are referred to as either a Personnel Manning Authorization Document (PMAD) or Updated Authorization Document (UAD) depending on which type is valid at the time. Using the personnel inventory database at monthly intervals since , every individual assignment is mapped against the corresponding structure document and friction levels are calculated. As mentioned in the Inventory Distribution section, SAS programming with SAS 9.2 was used to conduct all of the data collection, data manipulation, and calculations. FORECASTING DISTRIBUTION FRICTION Forecasting distribution friction is more complex than forecasting inventory friction. Army strength forecasts include grade and MOS detail but do not attempt to predict location of assignment. Therefore, alternative methods such as simulation must be applied to predict a future level of distribution friction. By simulating the personnel distribution process and incorporating as many predicted events as possible (i.e., deployment schedules, structure changes, policy changes, etc.) we can produce a reasonable forecast of expected distribution friction. Admittedly, the distribution process is complex and a 1 The DMSL codes used in this analysis were not populated in structure documents prior to Prior to 2004, there was a heavier emphasis on division level structure than brigade size forces. 7

8 degree of personal and human elements cannot be accurately simulated. However, along with valid assumptions, the functions that are tied to published policies and procedures can be integrated into a model that allows us to better plan for the future. Currently, the Army G-1 is engaged in two optimization and simulation modeling efforts (Force Shaping and ARFORGEN 2 BOG:Dwell) that it hopes to integrate for distribution friction forecasting. READINESS FRICTION DEFINING READINESS FRICTION Readiness friction is defined as the number of personnel within the operating strength unavailable for training and/or deployment. It is also the hardest category to capture in terms of quantifiable data and requires unique methods for calculating historical and projected measurements. Unlike the previous two categories, readiness friction is tied to individual availability and effectiveness rather than misalignment with force structure. Sources of unavailability include reasons such as P3/P4 profiles, Medical Evaluation Boards (MEBs), temporary profiles, pregnancies, transients, UCMJ, family care plans, dental categories, and minimum dwell policies. So why would we include readiness friction as a component of personnel friction? The answer lies in the earlier observation that the three friction categories are not mutually exclusive. More often than not, non-deployable Soldiers are discussed in conversation separate from personnel misalignment. In doing so, the Army is guilty of depicting them as additive problems when they are not. Simply put, if a Soldier is misaligned by grade, MOS or location and is at the same time non-deployable, we still have only one individual point of personnel friction. Separate analysis of readiness friction without consideration of overlap with other personnel friction categories is flawed at best and misleading at worst. Table 4 provides a simple example of overlapping friction for Soldiers with the grade of E5 in four MOSs within a single unit. Unit X MOS Grade Authorizations Strength Str - Auths Non-Deployable 1A E B E C E D E Total Table 4 If non-deployable Soldiers are considered independently, one could easily arrive at the conclusion that this unit has five Soldiers contributing to readiness friction. However, three of the non-deployable Soldiers are in MOS 3C which is already accounted for in inventory friction. Thus, even without knowledge of the individual non-deployers, we can fairly conclude that two of the non-deployables are compensated by over-strength (part of the inventory friction calculation). Therefore, the amount of readiness friction in this particular grade, MOS and location combination would be 1+(3-2)+1=3. MEASURING READINESS FRICTION Unfortunately, existing personnel databases do not provide the same fertile ground for readiness friction data mining and historical calculations. Even when non-deployable data is available, it is usually consumed in unit statistics, void of SSN level detail and when accurate, is only recorded at a single discrete point in time, Soldier Readiness Processing (SRP). Fortunately, the previous example in Table 4 demonstrates that SSN detail is not necessary in order to merge readiness databases information with the personnel inventory. Although better data collection procedures and a non-deployable status field in the existing personnel database (TAPDB) need to be implemented in order to collect more usable Soldier 2 Army Force Generation 8

9 identifiable non-deployable data at SRP, raw numbers disaggregated by grade, MOS and location are sufficient. A method of measurement that eliminates the aforementioned data redundancy is shown below. by grade Unless readiness data can be collected at discrete points in time other than SRP, the readiness friction calculations will only capture statistics for units on the verge of deployment. Although a valid argument can be made that readiness implications are less severe farther in advance of deployment, a certain level of predictability and knowledge must be generated earlier in order to shape the force and reduce the impact at deployment. FORECASTING READINESS FRICTION More so than the previous two personnel friction categories, readiness friction is directly influenced by the demand of persistent conflict. Therefore, it is presumed that the best way to forecast readiness friction is by applying methods such as regression analysis which can incorporate the dependency of a variable deployment demand. Crucial to the Army G1 s ability to forecast readiness friction is the availability of predicted ARFORGEN demand scenarios. The mission focus of the Army s manpower community is centered on managing and forecasting personnel inventory in order to meet requirements. Simplified in an economic context, the Army G1 has the modeling capabilities to forecast supply, yet we rely on other organizations 3 within the Army to forecast demand. Given a particular demand scenario with an already known ARFORGEN synchronization concept, we can use the relationship of deployment demand and historical non-deployable rates friction to forecast readiness friction. IMPLICATIONS Once we are able to measure the components of personnel friction, we can evaluate how well the Army is aligning personnel with authorizations. In order to present the contribution of each component, it is important to be able to measure each individually and in aggregate. After collecting reliable readiness friction data, it can be merged with existing historical inventory and distribution data to generate monthly personnel friction. By combing the earlier equations for inventory, distribution and readiness friction, we can calculate the amount of personnel friction at a given point in time using the following equation: Personnel Friction To measure the effects of friction in terms of fill, we must compare the friction percentage to the FSA and operating strength. If the operating strength is greater than the FSA, then the amount of friction is subtracted from the enlisted FSA. Where the opposite is true, or the FSA is greater than the operating strength, the friction amount is subtracted from the 3 Projected demand scenarios are provided to G-1 from G-3/5/7 and Forces Command (FORSCOM). 9

10 Enlisted Personnel NESUG 2010 operating strength. When FSA is smaller than the operating strength, the friction is subtracted from the FSA because, despite having extra assignments, there still is friction and the number of correctly assigned personnel cannot be over the number of authorizations. The friction is subtracted from the lower level because the friction is calculated without respect to the total number of personnel in the organization, just how they are aligned. Figure 3 shows the enlisted operating strength (faces) compared to the force structure authorizations (spaces) and how the inventory friction and distribution friction affect the operating strength. The historical fills show the effect of these two personnel friction components broken down by MOS, MOS and Grade, and MOS, Grade and Location. Accounting for personnel friction shows the real effect of how close the Army is to correctly filling the authorizations, despite having overages in the operating strength. or projection, less detail is available because we don t yet have an accurate forecast of distribution friction. Also, Figure 3 does not show readiness friction MOS & Grade Friction MOS Friction OSD Operating Strength Inventory & Distribution Friction Inventory Friction Fill minus MOS friction Fill minus Inventory (MOS & Grade) Friction Fill minus Inventory & Distribution Friction Force Structure Allowance History Projection Figure 3: Inventory & Distribution Friction As shown in Figure 4, different methods of measurement are required for the different components of friction, and for historical versus projected friction. In order to effectively forecast distribution friction, the Strength and Forecasting Division of the Army G1 is leveraging ongoing internal development of improved optimization and simulation models. These models incorporate a multitude of force shaping mechanisms beyond minimizing operating strength deviation to model effects of manpower policies on human capital and the distribution challenges induced by the demand requirements of persistent conflict and ARFORGEN. Likewise, we are continuing to refine data collection methods in order to quantify the true impact of readiness friction and apply appropriate forecasting techniques. 10

11 Friction Type Historical Forecasted Inventory Distribution Readiness Computed Actuals 1 (SAS Programming) Computed Actuals 1 (SAS Programming) Data Collection / Computed Actuals 1 Merge, query and calculate friction from multiple databases (PMAD/UAD, TAPDB) Figure 4: Measurement and Forecasting Methods Optimization Discrete-Event / Agent-Based Simulation & Optimization Regression Analysis = Complete CONCLUSION Manning the Army in aggregate is much simpler than the Human Resources Command task of meeting unit requirements with no slack in a system that is not designed for just-in-time delivery. Unlike object-oriented supply chain systems, personnel systems do not grant the luxury of extra inventory stock to meet just-in-time needs, because as a unit requirement emerges, there is no guarantee that the next Soldier available is a perfect match in skill and grade. As we operate in a system with no slack and imperfect information, there are several factors that affect the level of personnel friction in the Army. Changing force structure documents, fluctuating deployment demand requirements, seasonality of accessions, training course lengths, Soldiers on stop-loss, TTH, A supportability and loss rates for individual s are just some of the major factors that cause inventory and distribution friction. A longer list of factors contributes to readiness friction. Personnel friction prevents Army units from having the right number and skill of Soldiers to meet mission requirements, which causes readiness degradation. Identifying the components of personnel friction and developing a methodology to measure and predict future friction levels won t solve this problem. However, by understanding the causes and relationships of each of the types of friction, we can determine how much of it can be shaped by variable decisions and policies. It is in the Army s interest to both provide a personnel inventory that is truly sufficient for accomplishing its missions while minimizing any level of inefficiency that is not inherent to a dynamic personnel system with imperfect information. 11

12 REFERENCES Army Regulation , Army Force Stabilization System, dated 15 October Army Regulation , Overseas Service, dated 12 August 2008 Army Regulation , Physical Performance Evaluation System, dated 28 February Fastabend, D., An Appraisal of The Brigade-Based New Army. Parameters, Autumn: Jehle, K Cost of Doing Business: Friction in the Personnel & Structure System. Presented at BAFEX, May Olson, K., et al Friction in Manning, DAPE-PRS. Thie, H.J., Yardley, R.J., Harrell, M.C., Brancato, K Alignment of Department of Defense Manpower, Resources, and Personnel Systems. RAND Corporation Technical Report prepared for Office of the Secretary of Defense. ACKNOWLEDGEMENT The U.S. Government is authorized to reproduce and distribute reprints for governmental purposes notwithstanding any copyright annotation thereon. The views and conclusions contained herein are those of the authors and should not be interpreted as representing the official policies or endorsements, either expressed or implied, of, the Army G1 or the U.S. Government. CONTACT INFORMATION MAJ Mark Zais, MS Army G1, Military Strength Analysis & Forecasting Pentagon, Washington, DC MAJ Jose Ramirez, MS Army G1, Military Strength Analysis & Forecasting Pentagon, Washington, DC