The Volokh Conspiracy - More Details on the Financial Cost of DADT:

I've now gone through the recent University of California study of "Don't Ask, Don't Tell," which estimates the costs of the policy in its first ten years (1994 through 2003). The commission that produced the report includes several experts in military and national security policy. Notable among them are former Secretary of Defense William Perry under President Clinton; former Assistant Secretary of Defense Lawrence Korb under President Reagan; retired Admiral John D. Hutson; Professors Donald Campbell and Kathleen Campbell of the U.S. Military Academy at West Point; and Professors Frank Barrett and Mark Eitelberg of the Naval Postgraduate School. Experts in economics, cost accounting, management control systems, and other fields assisted the commission. The Williams Project of the UCLA Law School, which studies gay legal issues and works for gay equality, loaned out the time of Dr. Gary Gates, who provided extensive statistical and conceptual analysis as Senior Project Consultant. While some of those who produced the study may personally oppose DADT, the study itself cannot be dismissed on this basis. It is a serious effort to weigh some of the financial consequences of DADT.

The report breaks down the financial cost of firing service members for homosexuality under DADT into four discrete categories: (1) recruiting costs for enlisted service members; (2) training costs for enlisted service members; (3) training costs for officers; and (4) separation travel costs. Let's take a look at each of these:

(1) Recruiting costs for enlisted service members fired for homosexuality (1994-2003): $79.2 million

The military spends a lot of money to recruit. Some of this money is spent to recruit service members who are eventually fired for homosexuality. In a February 2005 report tellingly entitled "Financial Costs and Loss of Critical Skills Due to DOD's Homosexual Conduct Policy Cannot Be Completely Estimated," the congressional General Accounting Office estimated these costs attributable to DADT at $95.4 million.

The UC Commission believes this overstates the actual cost:

The critical value for estimating this cost, we would argue, is not how much the military spent to replace service members fired for homosexuality. Rather, the appropriate consideration is how much value the military lost as a result of each homosexual discharge. For example, in [an] extreme hypothetical situation [], in which the service member served for almost 30 years in uniform prior to discharge, we suggest that the military barely lost any value from the premature discharge for homosexuality.

To correct this type of error, the UC Commission took the GAO cost for enlisted recruiting of DADT-discharged service members ($95.4 million) as a lodestar and subtracted from that an estimate of the value of this cost the military recovered from the service members' time in service, as follows:

To determine the military's monthly return on investment, we divided the average cost of recruiting each enlisted service member ($10,193) by the number of months during which the military could have recovered its investment in that individual's recruiting. . . . For each enlisted service member, we credited the military with a monthly return on its investment in recruiting for each month served, except for those months spent in initial and mid-career training. The cost of enlisted recruiting was determined by GAO to be $95,393,000. Total recovery on investment . . . is calculated as $16,113,715. The total spent on recruiting, $95,393,000, minus the recovery on investment, $16,113,715 yields a total of $79,279,285.

(2) Training costs for enlisted service members fired for homosexuality (1994-2003): $252.3 million

Once the military recruits a person for service, it invests even more heavily in both basic and initial skills training. The GAO estimated the cost of training recruits fired for homosexuality to be $95.1 million.

This is almost certainly a large underestimate of the cost of training these recruits, for a couple of reasons. First, the GAO number does not reflect training costs for Marines discharged for homosexuality (since the Marines apparently did not provide GAO with training estimates). Second, even the GAO's training cost figures for the other services are substantially lower than the GAO's own previous estimates of training costs and estimates available in other public sources. The UC Commission corrected the GAO figures by relying on the Defense Department's and GAO's own previous figures for both basic and initial skills training for each of the branches. After that, the UC Commission again credited the military for recovering at least a part of this cost through the member's service before discharge. Here's the calculation:

Spending on enlisted training, prior to any recovery of costs, is $331,866,779. Total recovery on investment . . . is calculated as $79,492,728. The total spent on training, $331,866,779, minus the recovery on investment, $79,492,728, yields a total cost to the military of $252,374,051.

(3) Training costs for officers (1994-2003): $17.7 million

The GAO report did not factor the cost of training officers into its report. In the period 1994-2003, 137 officers were discharged for homosexuality.

To quantify the losses associated with firing officers for homosexuality, we estimated the cost of training to commission as well as post-commission training. Then, as was the case with our estimates of recruiting and enlisted training costs, we reduced our estimates by crediting the military with any recovered value on its initial investment in officer training for those officers who served after the completion of their training. Unlike enlisted service members, however, in the case of officers we did not include mid-career training costs in our estimates.

The UC Commission then estimated the cost of training for officers who go through one of five different routes: service academies (like West Point), ROTC, Officer Candidate School, direct appointment, and other paths. Since the UC Commission was unable to get cost estimates for the latter two, it assumed these costs to be zero. This approach yielded these numbers:

Spending on officer training, prior to any recovery of costs, is $27,553,701, of which $15,752,353 is for pre-commission training, and $11,801,348 is for post-commission training. Total recovery on investment . . . is calculated as $9,781,631. The total spent on training, $27,553,701, minus the recovery on investment, $9,781,631, yields a total loss to the military of $17,772,070.

(4) Separation travel costs (1994-2003): $14.3 million

Recruiting and training costs are front-end: they occur at the beginning of a military career. There are also costs associated with separation from the military, the back-end of service. These "out-processing" costs are numerous and are also investments the military must make when it discharges a member. One such cost is travel expense. Using the Army's own lower-range estimates for such travel costs, and deducting for recovery of costs through time served, the UC Commission found as follows:

Spending on enlisted and officer separation travel, prior to any recovery of costs, is $16,633,308 and $638,381, respectively. Total recovery on investment . . . is calculated as $2,926,816. The total spent on separation travel, $17,271,689 minus the recovery on investment, $2,926,816, yields a total of $14,344,873.

Putting all these numbers together, we arrive at a total cost of $363, 770, 279 to implement DADT during its first ten years.

How accurate is this number? I am not an economist and can't vouch for the inputs the UC Commission used to calculate costs. Several factors, however, suggest that the UC Commission estimate -- while an improvement over the GAO figure -- is still a substantial underestimate of the financial cost of expelling gay service members. The Commission itself points to five ways in which its analysis may underestimate costs:

First, we were unable to obtain reliable data for some costs that were omitted from GAO's original report. For example, we were unable to obtain reliable data for the costs of discharge review boards, security clearances, out-processing costs, investigations into service members' sexual orientation, re-enlistment bonuses, and officer recruiting.

...

Second, as noted above, our use of the training costs for a surface warfare officer as a proxy for the cost of training all officers reflects a conservative assumption that probably reduced our overall cost estimate. The cost to train a surface warfare officer is $92,924, while the cost to train one jet pilot (T-45 line) is $1,439,754. The list of officers fired for homosexuality includes physicians, pilots, dentists, and other individuals with highly technical training.

Third, many gays and lesbians do not re-enlist after fulfilling their service obligations because they are unwilling to continue to conceal their identity. According to a new survey of 445 gay, lesbian, bisexual and transgendered veterans, 19.6 percent of respondents left the armed forces "voluntarily because they could not be open about being LGBT while in the military." . . . [T]he military may be losing some of its investment in recruiting and training individuals who would remain in uniform if the ban were repealed.

Fourth, we assumed that the benefits of a service member to the Defense Department accrue evenly over the cost recovery period. . . . This is a conservative assumption given that, as is the case in most industries, service members' value to the military increases with experience.

...

Fifth, we did not include the costs of marriage benefits for gays and lesbians who get married to opposite-sex individuals to avoid military scrutiny of their sexual orientation, and who then file claims for military benefits for their spouses. According to the new survey of 445 gay, lesbian, bisexual and transgendered veterans mentioned above, 18 percent of respondents (80 individuals) got married to avoid military scrutiny of their sexual orientation.

Of these, the first (additional separation costs) and the third (premature loss of non-discharged gay personnel) seem most likely to add considerably to the real cost of DADT.

There are a few more reasons, in addition to these, that the UC Commission may be underestimating the cost of DADT. At just about every turn, the Commission used very conservative estimates of costs. For example, the UC Commission used a low-ball estimate of the number of service members fired for homosexuality in the ten-year period, putting the number at 9,359 enlisted, active-duty members. But this number, as the Commission notes (fn. 14), does not include some members of the Coast Guard and reserve forces. Including them would add about another 300 discharged under DADT. Further, the UC Commission's estimate of training costs, though more reasonable than GAO's, may still be too low. The Commission cites a "senior level military operations research analyst" who estimates that actual training costs are much higher than even the Commission accepts as the basis for its calculations (p. 13). Since enlisted training costs are by far the largest fraction of the overall cost of DADT, under the UC Commission's own calculation, even small error in per capita costs could have a big effect on the final calculation.

On the other hand, there are a few ways in which the UC Commission may have overestimated the total cost of DADT. First, as the Commission acknowledges, it did not calculate the cost of paying same-sex partner benefits if the ban is lifted. Experience in other countries so far shows this cost is very low. Potentially more significantly, the UC Commission does not attempt to calculate the cost of recruiting, training, and retaining service members who might disdain military service if the ban is lifted. Nobody can know what this cost would be, though it would surely be greater than zero. I doubt it would be a large cost, since open homosexuals would be such a tiny portion of the military services, and whatever the initial cost it could be expected to dwindle as military culture adjusted. The experience of other countries does not indicate any recruitment and retention difficulties attributable to allowing service by openly gay service members. But perhaps the experience of other countries cannot be fully extrapolated to the United States. Finally, as a reader and former service member helpfully noted in a private message to me, the Commission assumes that gay service members discharged under DADT would have served just as long as straight service members if not for DADT. The assumption seems reasonable, and the Commission makes no attempt to defend it. But I do not really know. The difference between the expected length of service and actual service forms the basis for some of the Commission's cost estimates. If, contrary to the Commission's assumption, service members expelled under DADT would, on average, serve shorter periods of time than their peers even in the absence of DADT then the "loss" the military suffers because of early discharge would be correspondingly smaller.

Finally, whatever the accuracy of the numbers in the UC Commission report, the financial costs alone do not resolve the debate over DADT, just as financial cost does not resolve a debate over any policy that might be worthwhile. The military exists to deter wars and, when that fails, to win them. If allowing homosexuals to serve openly would likely hinder that mission to any substantial degree, then the financial cost of expelling them would have to be just one more item in the Defense Department's budget. Nor do "financial costs" alone measure all the costs -- in broken careers, broken lives, broken families, in the dignitary loss to gay Americans in general and to those who want to serve their country with integrity in particular -- of having a policy that expels American service members simply because we learn they're gay.

But we cannot have a debate over the overall costs and benefits of DADT without at least a reasonable baseline estimate of the financial costs. For that reason, the UC Commission has made what appears to be a useful and important contribution. Military policy leaders in Congress should take a close look at the study, since Congress is the body that must repeal or amend DADT.

I am interested in readers' views on the specifics of this study. I am especially interested in hearing from readers who have expertise in economics, accounting, and military-cost analysis.

	[Dale Carpenter, February 16, 2006 at 11:13pm] Trackbacks More Details on the Financial Cost of DADT: I've now gone through the recent University of California study of "Don't Ask, Don't Tell," which estimates the costs of the policy in its first ten years (1994 through 2003). The commission that produced the report includes several experts in military and national security policy. Notable among them are former Secretary of Defense William Perry under President Clinton; former Assistant Secretary of Defense Lawrence Korb under President Reagan; retired Admiral John D. Hutson; Professors Donald Campbell and Kathleen Campbell of the U.S. Military Academy at West Point; and Professors Frank Barrett and Mark Eitelberg of the Naval Postgraduate School. Experts in economics, cost accounting, management control systems, and other fields assisted the commission. The Williams Project of the UCLA Law School, which studies gay legal issues and works for gay equality, loaned out the time of Dr. Gary Gates, who provided extensive statistical and conceptual analysis as Senior Project Consultant. While some of those who produced the study may personally oppose DADT, the study itself cannot be dismissed on this basis. It is a serious effort to weigh some of the financial consequences of DADT. The report breaks down the financial cost of firing service members for homosexuality under DADT into four discrete categories: (1) recruiting costs for enlisted service members; (2) training costs for enlisted service members; (3) training costs for officers; and (4) separation travel costs. Let's take a look at each of these: (1) Recruiting costs for enlisted service members fired for homosexuality (1994-2003): $79.2 million The military spends a lot of money to recruit. Some of this money is spent to recruit service members who are eventually fired for homosexuality. In a February 2005 report tellingly entitled "Financial Costs and Loss of Critical Skills Due to DOD's Homosexual Conduct Policy Cannot Be Completely Estimated," the congressional General Accounting Office estimated these costs attributable to DADT at $95.4 million. The UC Commission believes this overstates the actual cost: The critical value for estimating this cost, we would argue, is not how much the military spent to replace service members fired for homosexuality. Rather, the appropriate consideration is how much value the military lost as a result of each homosexual discharge. For example, in [an] extreme hypothetical situation [], in which the service member served for almost 30 years in uniform prior to discharge, we suggest that the military barely lost any value from the premature discharge for homosexuality. To correct this type of error, the UC Commission took the GAO cost for enlisted recruiting of DADT-discharged service members ($95.4 million) as a lodestar and subtracted from that an estimate of the value of this cost the military recovered from the service members' time in service, as follows: To determine the military's monthly return on investment, we divided the average cost of recruiting each enlisted service member ($10,193) by the number of months during which the military could have recovered its investment in that individual's recruiting. . . . For each enlisted service member, we credited the military with a monthly return on its investment in recruiting for each month served, except for those months spent in initial and mid-career training. The cost of enlisted recruiting was determined by GAO to be $95,393,000. Total recovery on investment . . . is calculated as $16,113,715. The total spent on recruiting, $95,393,000, minus the recovery on investment, $16,113,715 yields a total of $79,279,285. (2) Training costs for enlisted service members fired for homosexuality (1994-2003): $252.3 million Once the military recruits a person for service, it invests even more heavily in both basic and initial skills training. The GAO estimated the cost of training recruits fired for homosexuality to be $95.1 million. This is almost certainly a large underestimate of the cost of training these recruits, for a couple of reasons. First, the GAO number does not reflect training costs for Marines discharged for homosexuality (since the Marines apparently did not provide GAO with training estimates). Second, even the GAO's training cost figures for the other services are substantially lower than the GAO's own previous estimates of training costs and estimates available in other public sources. The UC Commission corrected the GAO figures by relying on the Defense Department's and GAO's own previous figures for both basic and initial skills training for each of the branches. After that, the UC Commission again credited the military for recovering at least a part of this cost through the member's service before discharge. Here's the calculation: Spending on enlisted training, prior to any recovery of costs, is $331,866,779. Total recovery on investment . . . is calculated as $79,492,728. The total spent on training, $331,866,779, minus the recovery on investment, $79,492,728, yields a total cost to the military of $252,374,051. (3) Training costs for officers (1994-2003): $17.7 million The GAO report did not factor the cost of training officers into its report. In the period 1994-2003, 137 officers were discharged for homosexuality. To quantify the losses associated with firing officers for homosexuality, we estimated the cost of training to commission as well as post-commission training. Then, as was the case with our estimates of recruiting and enlisted training costs, we reduced our estimates by crediting the military with any recovered value on its initial investment in officer training for those officers who served after the completion of their training. Unlike enlisted service members, however, in the case of officers we did not include mid-career training costs in our estimates. The UC Commission then estimated the cost of training for officers who go through one of five different routes: service academies (like West Point), ROTC, Officer Candidate School, direct appointment, and other paths. Since the UC Commission was unable to get cost estimates for the latter two, it assumed these costs to be zero. This approach yielded these numbers: Spending on officer training, prior to any recovery of costs, is $27,553,701, of which $15,752,353 is for pre-commission training, and $11,801,348 is for post-commission training. Total recovery on investment . . . is calculated as $9,781,631. The total spent on training, $27,553,701, minus the recovery on investment, $9,781,631, yields a total loss to the military of $17,772,070. (4) Separation travel costs (1994-2003): $14.3 million Recruiting and training costs are front-end: they occur at the beginning of a military career. There are also costs associated with separation from the military, the back-end of service. These "out-processing" costs are numerous and are also investments the military must make when it discharges a member. One such cost is travel expense. Using the Army's own lower-range estimates for such travel costs, and deducting for recovery of costs through time served, the UC Commission found as follows: Spending on enlisted and officer separation travel, prior to any recovery of costs, is $16,633,308 and $638,381, respectively. Total recovery on investment . . . is calculated as $2,926,816. The total spent on separation travel, $17,271,689 minus the recovery on investment, $2,926,816, yields a total of $14,344,873. Putting all these numbers together, we arrive at a total cost of $363, 770, 279 to implement DADT during its first ten years. How accurate is this number? I am not an economist and can't vouch for the inputs the UC Commission used to calculate costs. Several factors, however, suggest that the UC Commission estimate -- while an improvement over the GAO figure -- is still a substantial underestimate of the financial cost of expelling gay service members. The Commission itself points to five ways in which its analysis may underestimate costs: First, we were unable to obtain reliable data for some costs that were omitted from GAO's original report. For example, we were unable to obtain reliable data for the costs of discharge review boards, security clearances, out-processing costs, investigations into service members' sexual orientation, re-enlistment bonuses, and officer recruiting. ... Second, as noted above, our use of the training costs for a surface warfare officer as a proxy for the cost of training all officers reflects a conservative assumption that probably reduced our overall cost estimate. The cost to train a surface warfare officer is $92,924, while the cost to train one jet pilot (T-45 line) is $1,439,754. The list of officers fired for homosexuality includes physicians, pilots, dentists, and other individuals with highly technical training. Third, many gays and lesbians do not re-enlist after fulfilling their service obligations because they are unwilling to continue to conceal their identity. According to a new survey of 445 gay, lesbian, bisexual and transgendered veterans, 19.6 percent of respondents left the armed forces "voluntarily because they could not be open about being LGBT while in the military." . . . [T]he military may be losing some of its investment in recruiting and training individuals who would remain in uniform if the ban were repealed. Fourth, we assumed that the benefits of a service member to the Defense Department accrue evenly over the cost recovery period. . . . This is a conservative assumption given that, as is the case in most industries, service members' value to the military increases with experience. ... Fifth, we did not include the costs of marriage benefits for gays and lesbians who get married to opposite-sex individuals to avoid military scrutiny of their sexual orientation, and who then file claims for military benefits for their spouses. According to the new survey of 445 gay, lesbian, bisexual and transgendered veterans mentioned above, 18 percent of respondents (80 individuals) got married to avoid military scrutiny of their sexual orientation. Of these, the first (additional separation costs) and the third (premature loss of non-discharged gay personnel) seem most likely to add considerably to the real cost of DADT. There are a few more reasons, in addition to these, that the UC Commission may be underestimating the cost of DADT. At just about every turn, the Commission used very conservative estimates of costs. For example, the UC Commission used a low-ball estimate of the number of service members fired for homosexuality in the ten-year period, putting the number at 9,359 enlisted, active-duty members. But this number, as the Commission notes (fn. 14), does not include some members of the Coast Guard and reserve forces. Including them would add about another 300 discharged under DADT. Further, the UC Commission's estimate of training costs, though more reasonable than GAO's, may still be too low. The Commission cites a "senior level military operations research analyst" who estimates that actual training costs are much higher than even the Commission accepts as the basis for its calculations (p. 13). Since enlisted training costs are by far the largest fraction of the overall cost of DADT, under the UC Commission's own calculation, even small error in per capita costs could have a big effect on the final calculation. On the other hand, there are a few ways in which the UC Commission may have overestimated the total cost of DADT. First, as the Commission acknowledges, it did not calculate the cost of paying same-sex partner benefits if the ban is lifted. Experience in other countries so far shows this cost is very low. Potentially more significantly, the UC Commission does not attempt to calculate the cost of recruiting, training, and retaining service members who might disdain military service if the ban is lifted. Nobody can know what this cost would be, though it would surely be greater than zero. I doubt it would be a large cost, since open homosexuals would be such a tiny portion of the military services, and whatever the initial cost it could be expected to dwindle as military culture adjusted. The experience of other countries does not indicate any recruitment and retention difficulties attributable to allowing service by openly gay service members. But perhaps the experience of other countries cannot be fully extrapolated to the United States. Finally, as a reader and former service member helpfully noted in a private message to me, the Commission assumes that gay service members discharged under DADT would have served just as long as straight service members if not for DADT. The assumption seems reasonable, and the Commission makes no attempt to defend it. But I do not really know. The difference between the expected length of service and actual service forms the basis for some of the Commission's cost estimates. If, contrary to the Commission's assumption, service members expelled under DADT would, on average, serve shorter periods of time than their peers even in the absence of DADT then the "loss" the military suffers because of early discharge would be correspondingly smaller. Finally, whatever the accuracy of the numbers in the UC Commission report, the financial costs alone do not resolve the debate over DADT, just as financial cost does not resolve a debate over any policy that might be worthwhile. The military exists to deter wars and, when that fails, to win them. If allowing homosexuals to serve openly would likely hinder that mission to any substantial degree, then the financial cost of expelling them would have to be just one more item in the Defense Department's budget. Nor do "financial costs" alone measure all the costs -- in broken careers, broken lives, broken families, in the dignitary loss to gay Americans in general and to those who want to serve their country with integrity in particular -- of having a policy that expels American service members simply because we learn they're gay. But we cannot have a debate over the overall costs and benefits of DADT without at least a reasonable baseline estimate of the financial costs. For that reason, the UC Commission has made what appears to be a useful and important contribution. Military policy leaders in Congress should take a close look at the study, since Congress is the body that must repeal or amend DADT. I am interested in readers' views on the specifics of this study. I am especially interested in hearing from readers who have expertise in economics, accounting, and military-cost analysis.