Click for next page ( 113


The National Academies | 500 Fifth St. N.W. | Washington, D.C. 20001
Copyright © National Academy of Sciences. All rights reserved.
Terms of Use and Privacy Statement



Below are the first 10 and last 10 pages of uncorrected machine-read text (when available) of this chapter, followed by the top 30 algorithmically extracted key phrases from the chapter as a whole.
Intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text on the opening pages of each chapter. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.

Do not use for reproduction, copying, pasting, or reading; exclusively for search engines.

OCR for page 112
G The Smog and Levee of Joint and Service-SpecKic Vising n this chapter me discuss research regarding two questions. First is Mere a minimum level of advert/sing necessary for a cost-e~ective recru1Ung program' even ~ that advert/sing ~ not necessary to achieve conte~or~eous enUs~ent conbact goad? Pastoral Men Me recruits big climate is good and recruits are plenties' m1Utary planers tend to cut advert/sing budgets Derby contributing to a reducUon in "awareness" capital and propensity Even. ~~ may possibly set up a boom or bust cycle' in which propensity faUs' recru1Ung becomes more Dracula and Men advert/sing funds have to be restored or even increased beyond what they would have been gout Me 1nhial cut to stimulate propen- ~ty and enUstmenL lo our knowledge' this Sue has not been Budged extensively' and therefore the best research designs may not be 1mmed1- ately clear) Is chapter discusses Me poss~11hy of both econometric and experimental designs to address the quest/on. Me second quest/on concerns the proper levels of joint and Service- spec1Ac advert/sing. Certain types of advert/sing themes' such as generic themes designed to increase overaU propensHy' may be best done as a joint program' Shoe advert/sing themes Maturing specific benefits of m1~- 1Col~erg and ~~ko (2003) studied aggregate enbstments into the four Sedges over the and esh~ate larger business cycle gag., unemployment) effects for the Any and the Navy than are generally estimated in me studies summarized in Chapter 5. hey recom- mend an early warning system that wiU help the Department of De~nse (DoD) predict turning points in recruiting and undertake policy actions that wig reduce the cyclically of ~cmU~g. 112

OCR for page 112
fOINT AND SERVICE-SPECIFIC ADVERTISING 113 tary service are best done in the Service programs. We do not currently know much about what levels of Service and joint advertising will most efficiently sustain different levels of propensity and enlistment. Answer- ing this question is most likely to require a combination of research designs, both econometric and experimental. MINIMUM LEVEL OF ADVERTISING In the discussion in Chapter 5 of the optimal levels of recruiting resources to achieve a given set of recruiting goals, the optimal amount of advertising, like the optimal amount of any other recruiting resource, was the amount that permits goal achievement at lowest cost. Is there some minimum level of advertising that is above the level of advertising that minimizes the total cost of achieving current recruiting goals, but may be optimal in a longer term perspective? For example, consider a case in which, because of a poor economy and high unemployment rates, or a significant reduction in the current demand for new recruits, or both, current recruiting goals can be achieved even if recruiting resources, including advertising, are drastically reduced. Is there a reason to main- tain resources, particularly advertising, above the minimum level to achieve current goals? If the answer is yes, it must be because greater advertising expendi- tures in one period will make recruiting goals in a future period less costly to achieve. That is, there must be dynamic aspects to some types of recruiting resources such that current expenditures on those resources affect both current and future recruiting. As we discussed in Chapter 5 and as reported in the literature (Hogan, Dali, Mackin, and Mackie, 1996; Dertouzos and Garber, 2003), current advertising expenditures contribute to a stock of information and awareness capital. This stock changes over time, as the existing stock of information decays but is replenished by new advertising expenditures. Hence, under this basic concept of a flow of advertising contributing to a stock of awareness or information capital, one can anticipate that the effect of reducing the flow of new advertising on recruiting will, at first, be small. But as the stock continues to decay without replenishment, the adverse effect on recruiting can grow. This is illustrated in Figure 6-1, which assumes a depreciation rate of 10 percent. The stock of information and awareness capital has been built up to a notional value of 100 units in month 1. However, from that month forward, new advertising, which contributes to the stock, is cut to zero. Some level of new advertising is necessary simply to compensate for the depreciation of the existing stock. The effect on the ability to recruit is, at first, modest, because the stock of capital by month 3 is about 80 percent of the original stock. However,

OCR for page 112
114 120 100 80 o Oh 60 40 20 EVALUATING MILITARY ADVERTISING AND RECRUITING O- + Stock ~ New Adv 1 2 3 4 5 6 7 8 9 10 11 12 13 14 Months FIGURE 6-1 Stock of information and awareness capital at 10 percent monthly depreciation (notional). with no new advertising and continued depreciation, the stock would quickly be depleted to less than 30 percent of its original value. Now, assume that new advertising, equivalent to a contribution to the stock of 15 units per month, is provided. In this case, the stock of information, initially at 100, rises eventually to a new equilibrium of 150, where depreciation (10 percent of 150) equals new advertising of 15 units. At this point, new advertising exactly offsets the depreciation, maintain- ing a stock of information and awareness capital of about 150. When new advertising exceeds depreciation, the stock grows. This is shown in Fig- ure 6-2. This simple example suggests how one may reduce advertising resources and in the short term experience no significant reduction in recruits. Moreover, assume that, initially, the optimal level (i.e., cost- minimizing level) of the stock of advertising is 100. Then, because the economy sours and the demand for recruits declines, the new optimal level of the stock is 50. That is, significantly less awareness is necessary to recruit the desired numbers. In this case, it would make sense to reduce new advertising until the new equilibrium stock is approached, then . . . Increase it again. However, if there is reason to believe that external economic condi- tions are likely to improve significantly, or that the demand for new recruits is to increase, it may be imprudent to permit the stock of aware- ness capital decline to this level. This would be of greater concern if the marginal cost of increasing awareness capital in a particular period was

OCR for page 112
fOINT AND SERVICE-SPECIFIC ADVERTISING 150 115 100 ~ New Adv 50~.................................................................................................................................................................................................................................................... O 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 3 5 7 9 11 13 15 Months FIGURE 6-2 Stock of information and awareness capital (notional). increasing. That is, let A be the augmentation to the stock of awareness capital in a given time period. Then, let C(A) be the cost of increasing awareness capital by A in a given period (month). We have: DC(A) / DA, O C(A) / DA , O If the marginal cost of increasing the stock of awareness capital increases with the magnitude of the desired increase in capital, then it may be less costly to begin increasing the stock of information capital in periods earlier than is needed for current recruiting. To abstract from uncertainty over the demand for recruits, let us assume that the time path of recruit quotas for both highly qualified recruits (QH) and less-well-qualified recruits (QL) is known and that it fluctuates over time. One way to express the issue is whether there is a case for which T T ~Cmin(QH t,QL t,Pt,Et) / (1+ r) > ~COpt(QH t,QLt,pt,Et) / (1+ r) t=1 t=1

OCR for page 112
116 EVALUATING MILITARY ADVERTISING AND RECRUITING where C(....) is the cost function for producing highly qualified and less qualified recruits, respectively, in period t, and r is the discount rate. The variable P is the vector of recruiting resource prices in the period and E is a vector representing the state of the economy. The notation min repre- sents the cost-minimizing levels of inputs to exactly produce the supply of recruits to meet goals in period t, while opt is the level of resources that minimizes, not the costs in a specific period, but the present value of costs over periods 1 through T. when discounted at interest rate r. In the second case, potential supply may be greater than necessary to meet current goals. If so, it is optimal because it lowers the costs of meeting subsequent goals. If the above inequality holds as a general case, then there is a level of resources above the level necessary to meet current goals that mini- mizes the present value of costs over the longer term. How can we test the hypothesis that the inequality, defined above, may hold? We are likely to be able to do so only indirectly. If advertising is the reason the inequality holds, it is necessary that: (1) advertising expenditures have effects over time and (2) the marginal cost of increas- ing the stock of information or awareness capital in a given time period is . . ncreasmg. If advertising affects the stock of information capital only in the period in which it is produced, then the stock of capital and the flow of advertis- ing in a given period are equal, except for any external factors affecting the stock. In a sense, the depreciation rate of the inherited stock is 100 percent. The only reason to increase advertising in a given period would be to affect current enlistment supply, and there would be no reason not to reduce advertising if not needed in the current period. Furthermore, if advertising adds to the effective stock of information at a constant mar- ginal cost, it would be less costly, in a present value sense, to wait until the advertising is needed before increasing it, and to do it all in one period. Only if marginal costs are rising, and advertising effects are dynamic, would it be optimal (i.e., cost minimizing) to incur advertising costs cur- rently, not because they are needed currently but because they affect future recruits. To determine if there is a minimal level of advertising expenditure that is above that necessary to meet current goals, an econometric model must be able to capture the effects of advertising over time and allow for nonlinear effects of advertising on recruiting in a given time period.2 2Earlier we presented the notion that advertising contributed to a stock of information or awareness capital, and that this stock then affected recruiting. This is a useful concept for exposition purposes. However, the nonlinear effect, in practice, is between advertising in a given period and enlistments in that and subsequent periods. There is no necessity for the concept of the stock.

OCR for page 112
fOINT AND SERVICE-SPECIFIC ADVERTISING 117 Dertouzos and Garber (2003) estimate a supply curve that includes advertis- ing. As discussed in Chapter 5, this specification permitted both dynamic effects of advertising that is, it allowed advertising in one period to affect recruiting not only in the same period but in subsequent periods- and nonlinear effects within a given time period. Advertising in each period, both current and lagged, enters as a nonlinear S-shaped or logistic function. The functional form provides ranges for small marginal effects of advertising, then increasingly larger marginal effects, followed again by increasingly small marginal effects, as if there were a saturation level for advertising. If the unit price of advertising is constant, this would correspond to areas of high but decreasing marginal costs, followed by an inflection point and increasing marginal costs. Hence, this specification has the ability to capture the two features of advertising that would be required if there is a minimum level of advertising that is above the cost- minimizing level when only current recruiting goals are considered. The findings reported in Dertouzos and Garber (2003) and those from the literature reviewed briefly in Chapter 4 are that although advertising has dynamic effects, the lag structure suggests that most effects are real- ized within about two to three months of the advertising expenditure. These data suggest that current advertising may affect recruiting one, two, or three months in the future, but it does not affect recruiting six months or a year in the future. If the data on which these findings are based are strong and compelling, it would mean that a compelling case could not be made for maintaining advertising expenditures in a given period, if they are not needed to achieve current goals, solely because it is desirable to increase recruit supply six months or a year in the future. However, there is not a large body of evidence on the issue, and most authors who do provide some evidence also note that the data from which the estimates are derived are weak, and that more research is necessary. Hence, because of the paucity and limitations of evidence available to date, we conclude that the issue remains open and subject to further research. In our view it would be a serious mistake to view the available research as sufficient grounds for drawing conclusions one way or another about the effects of current advertising on future outcomes. Furthermore, the functional form used by Dertouzos and Garber (2003), while allowing for nonlinear effects of advertising a necessary feature for establishing some minimal level of advertising constrains the nonlinearity to follow a particular form. Because of this, it imposes the nature of the nonlinearity, rather than having a very flexible functional form that would allow the data to determine the nature of the nonlinearity. Hence, while the S-shaped functional form for advertising is plausible, it cannot be used by itself to test the hypothesis that advertising has non-

OCR for page 112
118 EVALUATING MILITARY ADVERTISING AND RECRUITING linear effects on recruiting.3 Instead, a more flexible functional form, one that permits lagged effects of advertising but also nonlinear effects of advertising within a period and does not constrain the precise nonlinear relationship, should be applied. The data necessary to estimate more flexible functional forms are typically quite demanding. Because the functional form itself places fewer constraints on the nature of the relationship between the supply of recruits, recruiting resources, and other factors affecting recruiting, the variation in the data must be sufficient to support distinguishing among alternative relationships. It is probably not the case that the most recent data over the l990s and early 2000s are sufficient to estimate these more demanding relationships. Dertouzos and Garber (2003), for example, were not able to estimate a Service-specific enlistment supply function from data over the l990s, even though they placed constraints on the nature of the functional relationships. A controlled experiment using a quasi-experimental design to gener- ate the data necessary to estimate supply relationships that incorporate the potential for both dynamic effects of advertising and nonlinear rela- tionships is likely to be impractical. The reason is that the effects to be estimated are not simple impact effects of an intervention, but a more complicated set of relationships regarding the dynamic, nonlinear struc- ture of advertising effects. To generate sufficient data to estimate such relationships would require that the experiment be continued over an extended period of time at least three or four years. In an operational environment in which there are changing supply conditions and real recruiting demands to be met, this is unlikely. Instead, we propose a focused effort to maintain data, especially advertising data, in a systematic and careful way, for the purposes of estimating a supply curve that incorporates the potential both for dynamic and nonlinear advertising effects. This is consistent with our recommen- dation in Chapter 5 for better and more thorough data collection and estimation using more flexible functional forms. Finally, it is important to note that most recruiting advertising has been targeted directly on recruit-age youth. Advertising designed to influ- ence parents, others who counsel youth, and youth who are not yet old enough to enlist may have effects that extend over time. Tests for the dynamic effects of advertising should attempt to distinguish between 3It could be used in conjunction with other constrained forms to test the hypothesis. In that case, one would specify several alternative forms, including a linear form, and deter- mine which form explained the observed data the best, and which could provide the best out-of-sample prediction.

OCR for page 112
fOINT AND SERVICE-SPECIFIC ADVERTISING 119 advertising targeted at recruit-age cohorts and advertising targeted at influencers or those who are not yet of recruiting age, if possible. ADVERTISING TARGETED AT THE LONGER TERM Historically, advertising for recruiting has focused almost exclusively on recruit-age cohorts of youth. That is, it has been focused on providing 17-21-year-olds with information about opportunities in the Services and on convincing recruit-age youth that enlisting is an option that they should seriously consider. This being the case, it is perhaps not surprising that the literature finds that current advertising expenditures affect recruiting only in the near term two or three months in the future and do not have effects on the market in the longer term six months or a year or more in the future. An alternative type of advertising would target audi- ences that have, perhaps, a longer time horizon for affecting the recruiting market. This would include targeting youth earlier in the career decision process before they can actually enlist, say at ages 14 through 16 and targeting parents and other adult authority figures who may influence the decisions of youth both immediately and in the future.4 This type of advertising has not been systematically employed by the Services or the Department of Defense. Arguably, it is a possible role for joint advertising. There is a solid logical argument for this type of adver- tising, based on analyses of survey and other data (see National Research Council, 2002, for further discussion). However, because it has not been tested in the recruiting market, little is known about its effects. The first step toward understanding its potential effectiveness and how it may influence the recruiting market over time would be to develop advertising of this nature and test it in the recruiting market. Because we anticipate that this type of advertising has the potential for affecting the market over longer periods of time, the experiment should be designed to devote advertising expenditures of this nature in the recruiting market for an extended period of time at least two years. Ideally, the quasi-experimental design should attempt to include variation over time and cross-sectionally in the advertising test. However, in an operational environment, a simple pre-post design may be both 4warner, Simon, and Payne (2002) provide evidence that military advertising affects pro- pensity to enlist. In an analysis of the Youth Attitude Tracking Study propensity over the 1988-2000 time period, they found a youths propensity to enlist to be positively related to total military advertising expenditures per youth in the youths state of residence in the 12 months prior to the survey. But data limitations prevented determining whether the impact of advertising on propensity varies by media or by whether it is joint or Service-specific advertising.

OCR for page 112
120 EVALUATING MILITARY ADVERTISING AND RECRUITING more workable and may serve to deliver the advertising in a more effi- cient way.5 If the experimental period is of sufficient duration, it would be useful to vary the level of advertising over time, rather than keeping it at a single level "dose." In addition to econometric analyses of effects, one would also include more qualitative methods of evaluation, including surveys and focus groups. THE LEVELS OF SERVICE-SPECIFIC AND JOINT ADVERTISING In fiscal year (FY) 1989, the last large recruiting year before the draw- down, the Services spent $100 million on enlisted advertising compared with a joint level of $20 million. In that year, almost the entire joint budget was for TV advertising. By FY 2000, joint advertising had dwindled to less than $5 million despite Service advertising of $250 million (Office of the Assistant Secretary of Defense [Force Management Policy], 2000~. The scaled-back joint program has been either direct mail or magazine adver- tising. The joint program raises four fundamental questions: What should be the content of joint advertising, and should it differ from the content of Service-specific advertising? To whom should it be targeted? What media should the joint program use? Do the efficient levels of joint and Service-specific advertising depend on the scale of the recruiting effort? As discussed below, the answers to these questions are interrelated. Past research is of little use in answering them. The first two questions have barely been addressed in previous research (Carroll, 1987~. Several econometric studies have tried to answer the latter two questions. Here the research strategy has been to estimate the responsiveness of enlist- ments to Service-specific and joint advertising in different media and then infer the cost-effectiveness of the various forms of advertising in generat- ing enlistments. Evidence from such studies about the relative effective- ness of joint advertising is, at best, mixed (see Table 5-2~. Hogan et al. (1996) found that joint TV and joint direct mailing were as effective as Navy advertising in the same media. But estimates by Dertouzos (1989), Warner (1991), and Warner, Simon, and Payne (2001) are less kind to the . . Font program. 5Excluding geographic areas to induce cross-sectional variation in advertising is difficult and likely to be cumbersome to recruiting operations. Moreover, it may preclude some efficient ways of delivering advertising, such as national magazines and network television.

OCR for page 112
fOINT AND SERVICE-SPECIFIC ADVERTISING 121 Dertouzos and Garber (2003) question the validity of past estimates of the effects of advertising (summarized in Chapter 5~. They argue that enlistments follow an S-shaped relationship with advertising and that the form of the relationship varies by media. If their analysis is correct, then we actually know very little about the effectiveness of joint advertising (and therefore the answers to the above questions). In fact, the insignifi- cance of the joint program in past econometric studies may have resulted from its being below the minimum efficient scale (MES) in Figure 5-1. Despite criticizing past studies on theoretical grounds, estimates pro- vided in Dertouzos and Garber (2003) do not help answer the above ques- tions. Their updated analysis of the Ad Mix Test data does not distinguish joint advertising from Army advertising, their analysis of the 1993-1997 recruiting experience is conducted at the DoD level, and the joint pro- gram is added to Service-specific advertising to derive measures of total DoD advertising. Since past work is of little use in answering the four fundamental questions above, this section now explores a series of ques- tions that need to be addressed in order to adequately answer these fun- damental questions. It is apparent from this discussion that an experiment is needed. Advertising Content If the joint program advertises in the same media as the Service pro- grams with the same message content, it is just an add-on to Service- specific advertising. That is not necessarily bad. Joint advertising that mirrors the Service programs essentially expands the scale of a common program and thereby just moves enlistments up the S-shaped curve por- trayed in Figure 5-1. If the unit cost of joint and Service-specific advertis- ing are the same, joint advertising that mimics the message and content of Service advertising would be a perfect substitute for Service advertising. Such advertising would be pointless since an equivalent expansion of the Service programs would yield the same result. In fact, due to the over- head costs of ad development and effects of program scale on advertising rates negotiated with ad agencies, newspapers, and other media, it is unlikely that the joint program could perfectly replicate the Service pro- grams at the same cost. In order to be useful, joint advertising must do something that Service-specific advertising does not. One way for joint advertising to distinguish itself is in the content of its messages. The committee's previous report (National Research Coun- cil, 2003, p. 227-230) discussed the message strategies in current military advertising. Service-specific advertising appears to be occupationally ori- ented, career-oriented, or oriented toward appealing to youths' desires for adventure, challenge, self-development, and self-actualization. The

OCR for page 112
122 EVALUATING MILITARY ADVERTISING AND RECRUITING report labeled these "package-oriented" advertising and went on to note that only Marine advertising seems directed toward the more noble virtues of patriotism, self-sacrifice, and service to country. (This seems to have changed with recent Service advertising in the wake of September 11, 2001, and Operation Iraqi Freedom.) Two recent DoD study groups the Defense Science Board Task Force on Human Resources Strategy (DSB) and the DoD Quality of Life Panel (also called the Jeremiah panel) expressed a belief that DoD needs to do more to engage the American public about the importance of public service.6 More values-oriented advertising that stresses the virtues of patriotism, self-sacrifice, and service to country would serve that pur- pose. Of course, the ultimate goal of advertising is to make youth more inclined to serve (i.e., increase propensity), so that more of them walk into recruiting offices on their own and more respond positively when they are contacted by military recruiters. One major research question, there- fore, is whether the message content of military advertising in fact affects propensity and whether a move toward more values-oriented advertising would increase it. The other question is whether a move toward values- oriented advertising is more efficiently accomplished in a joint program or whether redirected Service programs would have the same impact at the same cost. An advantage of loading all advertising into Service advertising was alluded to above: namely, that advertising unit costs are lower in pro- grams of larger scale. This advantage probably disappears once programs reach some minimal scale. In fact, Navy and Army TV advertising seem to generate about the same impressions per dollar of spending despite the much larger scale of the Army TV program.7 A potential advantage of having both joint and Service-specific advertising is that there are more competition and more innovation in the development and delivery of advertising the more firms are involved in the program.8 6The Dss report is available in The Defense science Board Task Force on Human Resources strategy, Office of the Under secretary of Defense for Acquisition, Technology, and Logistics, Washington DC 20301-3140, February 2000. The Quality of Life panel was commissioned by secretary of Defense Donald Rumsfeld and was chaired by Admiral David Jeremiah tRety. It did not publish a formal report, but an informal report and briefing of its findings and recommendations are available from the Undersecretary of Defense for Personnel and Research. Unpublished calculations using the data supplied by PEP Research. sin fact, there is some evidence from weapons programs that the cost-reducing competi- tion from a second supplier outweighs the economies of scale that are lost when procure- ment is split between two firms.

OCR for page 112
fOINT AND SERVICE-SPECIFIC ADVERTISING The Target Audience 123 Current Service advertising is directed at youth themselves. The committee's previous report expressed the belief that some advertising needs to be directed toward civilian influencers of youth: parents, school teachers and counselors, and others (for the role of family in influencing behavior, see, for example, Sewell and Hauser, 1972; Otto, 2000; Mortimer and Finch, 1996~. Such advertising, if effective, would build up the sup- port base from which youth make early career decisions upon leaving high school. Changing the attitudes of adult influencers would not happen overnight and would require a sustained effort over time. One purpose of targeting adult influencers would be to counteract the decline in the adult veteran population, which one study (Warner et al., 2001) found to be related to the decline in enlistment in the 1990s. An advertising campaign aimed at adult influencers would be consis- tent with the DSB recommendation to engage the American public about the value of public service.9 It is an empirical question whether such advertising, if done properly, would have more impact on youth propen- sity and the ultimate goal, enlistment than advertising directed at the youth themselves. Advertising aimed at adult influencers is likely to be more effectively accomplished in the context of a generic (joint) program than via Service- specific advertising whose purpose is to channel recruits to particular services. But that is an empirically testable hypothesis. The Media The Services advertise in a number of media: TV, radio, magazines, newspapers, direct mail, billboards, etc. In recent years about 60 percent of Army and Navy advertising expenditures have been for TV. Until recently, most Marine Corps advertising was TV and most Air Force advertising was non-TV. According to the Dertouzos-Garber model, an organization with a low advertising budget should spend the bulk of its budget on non-TV media. Non-TV advertising generates many more im- pressions per dollar than TV advertising (according to PEP data) and (theoretically) has a lower MES. But, theoretically, such advertising reaches its saturation point at a relatively low level of expenditure. TV advertising becomes cost-effective only at higher budget levels due to its higher MES, but it remains cost-effective much longer due to the fact that the saturation point is not reached until a fairly high level of expenditure. 9 The committee's previous report also expressed the belief that more advertising should be values-oriented (National Research Council, 2003, p. 233-234~.

OCR for page 112
124 EVALUATING MILITARY ADVERTISING AND RECRUITING These considerations suggest that the decision to eliminate joint TV adver- tising during the latter half of the 1990s as the joint program was scaled back made sense. However, devoting almost the whole 1989 joint budget of $20 million to TV advertising probably did not make sense. Factors that would have to be considered upon any rebuilding of the joint program, or during experiments with advertising, are the message content and the target audience. Ads that appeal to patriotism, for exam- ple, hope to evoke a strong emotional response from viewers. TV or radio ads may be more likely to evoke such responses than magazine or news- paper ads. Messages that are targeted to youth in one medium may be more productively targeted to adults via other media. For example, if adults are more likely to read newspapers and less likely to watch MTV, ads aimed at adult influencers might be more effective in newspapers than on MTV. Overall Scale of Recruiting For several reasons alluded to above, the optimal levels of joint and Service-specific advertising are likely to depend on the overall scale of the advertising effort. In particular, joint advertising is likely to be more cost- effective in the context of a large recruiting effort than a small recruiting effort. This section discusses another reason why this might be the case. Studies find a strong link between propensity to enlist, as measured from surveys such as the Youth Attitude Tracking Study, and actual enlist- ment (e.g., Warner et al., 2002~. Given the youth propensity to enlist, the lower the military's demand for recruits, the greater the fraction of enlist- ments that will come from youth who are already positively inclined to enlist without military advertising to attract them. In such an environ- ment, advertising can be targeted at those with a propensity and can be largely informational and aimed at informing them about career opportu- nities in particular Services and the like. It need not attempt to persuade youth that military service is a good thing; they are already so inclined. But the larger the demand for recruits, the more effort the military must make to attract youth with no propensity to enlist. The issue is not "high" versus "low" demand on some absolute scale, but demand rela- tive to propensity. Holding constant the recruiting goal, any change in propensity arising from factors other than military advertising (e.g., a lower unemployment rate) can influence the degree to which advertising needs to build propensity in the youth population rather than channel those who have already decided to join to a specific Service. How a change in the scale of the overall recruiting effort affects the levels of joint and Service advertising depends in part on whether the two sources of advertising differ in the degrees to which they build propen-

OCR for page 112
fOINT AND SERVICE-SPECIFIC ADVERTISING 125 sity. Much Service advertising is aimed at inducing youth already with a propensity to enlist to join a particular Service and may not affect overall propensity in the youth population. To the extent that joint advertising can affect the propensity base, it should play a larger role in the overall advertising program when the larger the recruiting mission is compared with the base of youth with a propensity to enlist. As an example, joint advertising might not have been very productive in 1992, a very easy recruiting year, but it could have been very productive in the late 1990s, when propensity was lower and recruiting was much more difficult. CONCLUSIONS AND RECOMMENDATIONS We have addressed the question of whether there is a minimum level of advertising necessary for a cost-effective recruiting program, even if that advertising is not necessary to achieve current enlistment contract goals. Historically, when the recruiting climate is good and recruits are plentiful, military planners tend to cut advertising budgets, thereby con- tributing to a reduction in awareness capital and propensity levels. This may possibly set up a boom or bust cycle, in which propensity falls, recruiting becomes more difficult, and advertising funds have to be restored. We presented a model that describes the conditions under which it would be cost-effective to advertise in the interests of future enlistment supply, and we reviewed research to date that speaks to the issue. While extant research suggests that advertising may have effects only for a short period of time, the data available to previous researchers are limited for several reasons. First, they do not permit examining both lagged effects and nonlinear effects within a time period. Second, they focus on adver- tising aimed at youth at the point of the enlistment decision and do not permit examining possible supplemental advertising approaches, such as those aimed a youth several years prior to an enlistment decision or those aimed at adult influencers, such as parents. As a result, research to date does not permit a definitive answer to the question of the cost-effectiveness of advertising above and beyond that which is necessary to achieve current recruiting goals. We recommend a focused effort to maintain advertising data in a systematic way for pur- poses of estimating a supply curve that incorporates the potential for both time-lagged and nonlinear advertising effects. We further recommend a program of research, incorporating quasi-experimental methods, to examine advertising effects over an extended period of time. We then turn to the optimal levels of joint and Service advertising. It is our opinion that certain types of advertising themes, such as generic themes designed to increase overall propensity, are best done as a joint program, while advertising themes featuring specific benefits of military

OCR for page 112
126 EVALUATING MILITARY ADVERTISING AND RECRUITING service are best done in the Service programs. What we do not know is whether there is an optimal level of joint and Service advertising or, more specifically, what advertising fund level should be allocated to joint pro- grams. We note that issues of scale play a role in addressing this question, as certain types of advertising (e.g., television) do not appear to have a constant effect across levels of expenditure. The larger the recruiting effort and the larger the budget, the greater the potential value of a multifaceted campaign, with some resources targeted toward providing information about specific Services to those already with a propensity to enlist and others targeted toward increasing propensity among those currently with- out it. We recommend a program of research aimed at examining the effects and cost-effectiveness of information-oriented versus values-oriented advertising in joint and Service advertising programs.