XIV. Payments Made Under the New Payment Systems
Payments made to providers under the new payment systems provide a direct measure of the extent to which TTW participants under these systems have achieved the earnings levels that trigger payments. Outcome payments are of particular interest because they are made only when the participant receives no DI or SSI payment as a consequence of earnings; in essence, the participant has exited the rolls, at least temporarily, and is on a path that can lead to a formal exit due to work.
We use the payment data to assess the extent to which earnings and benefit outcomes for early participants will improve after 2003. In the previous chapter, we concluded that any impact of TTW on earnings and benefit payments in the first two years of operation (that is, through 2003) were too small to detect, given historical variation in these two outcomes and underlying unmeasured differences between the experiences of beneficiaries in Phase 1 states and those in the other states. As we pointed out, however, impacts on these outcomes for early participants could be delayed. For those who assigned their Ticket in the first two years (that is, by the end of 2003), the payment data allow us to develop an upper bound for the impact of TTW on a closely related outcome for the same period: the number of participants who at least temporarily go off cash benefits due to work. We can then assess how that upper bound is likely to increase after 2003, as the experience of those who assigned their Ticket late in the first two years of TTW catches up to the experience of those who assigned their Tickets earlier. That upper bound is also an upper bound for program exits among beneficiaries who assign their Ticket.
Of the participants we have observed the longest (those who assigned their Ticket in the first half of 2002), 14.5 percent generated at least one payment from the new TTW payment systems by July 2005, including 15.8 percent who assigned their Ticket under the milestone-outcome system and 9.0 percent of who assigned their Ticket under the outcome-only system. Although the latter were less likely than the former to generate at least one payment, they were more likely to generate payments over a sustained period. In the first cohort, 75 percent of the participants who generated payments under the outcome-only system had generated at least 12 payments by July 2005, compared with just one-third of those generating payments under the milestone-outcome system. In fact, only about half of those generating payments under this system generated more than four payments—the maximum number of milestone payments.
The Ticket Act set a benchmark of increasing permanent exits due to work by at least half a percentage point. The findings from the analysis of the payment data imply that impacts through the end of 2003 were well below that benchmark. Further, we have to conclude that the program’s impact on participant exits will not reach the Ticket Act benchmark unless participation increases to well above the level reached in Phase 1 states by the end of 2004 or unless TTW somehow induced a large number of exits that are not reflected in the outcome payment data.
When interpreting these findings, it is useful to keep in mind the challenges of using payment data to measure earnings and benefit activity. One challenge is that payments are observed only if a payment is actually made, and a payment is often made long after the “earnings month” (that is, the month in which the participant generated the earnings that triggered payment). Another challenge is that the available data cover only the approximately 10 percent of participants who assign their Ticket under one of the new payment systems, reflecting major differences between the traditional payment system and the new systems. We discuss these and other methodological issues in Section A of this chapter. Section B presents payment statistics on Tickets assigned in the first two years of TTW. These Tickets had been assigned long enough so that by July 2005, we could both estimate how many would be likely to ever generate payments and begin to see how the number of payments would be likely to increase with time. In Section C, we narrow our focus to payments generated by earnings during the first two years of TTW for participants in Phase 1 states only and thereby measure the level of participant earnings activity for that period in those states—the period and the states that are the focus of the impact estimates in Chapter XII. Section D discusses the implications of the findings for TTW’s impact on program exits due to work.
A. Methodological Issues
1. The Payment Process
The analysis in this chapter exploits the fact that payments under the two new payment systems are indicative of earnings attained by participants. The connection between earnings and payments is, however, inexact. Earnings generate payments only if the provider files a payment claim and then only after the completion of the payment process. Although providers have a strong incentive to file claims for months in which beneficiaries achieve the required earnings levels, participant earnings might not be reported to the provider quickly, if at all. Even after the provider files the claim with the Program Manager, several months may elapse before the Program Manager and SSA process the payment, especially in the case of a Ticket’s initial claim. As reported in Chapter VIII, the median “payment lag” (that is, the duration from the earnings month to the payment month) exceeded eight months for first payments in the period we are examining. As a result, payments made as of any date can substantially understate the number of months in which participants achieved enough earnings to generate payments.
We do not have any data on the extent to which providers have not filed claims for which they might be eligible. Although interviewed providers have indicated that obtaining earnings documentation is problematic, they have not suggested that failure to obtain earnings data has prevented them from filing significant numbers of payment claims. Our working assumption is that the number of payments will be a reasonably accurate reflection of the number of months for which earnings were sufficient to generate payments once enough time has passed for those payments to be made.
Given the substantial payment lag, we limit the analysis to Tickets assigned by December 2003, but examine payments made through July 2005.1 Thus, we observe payments made from at least 19 to as many as 41 months after assignment (counting the assignment month). Even with the expected lags in payment processing, it seems likely that enough time has passed to observe which beneficiaries who assigned their Ticket to an EN (or to an SVRA acting as an EN) during the first two years will have generated at least one payment.
In addition, we divide these early participants into four assignment cohorts, according to the six-month period in which they assigned their Tickets (Exhibit XIV.1). By comparing payment statistics from the first cohort with each of the three later cohorts, we are able to assess the extent to which payment statistics change with time, as well as the extent to which the experiences of the three later cohorts are similar to those of the first cohort. We also classify participants by payment system and “payment title” (that is, DI or SSI). The latter class is of interest because payments for DI beneficiaries are higher than for those who receive only SSI. Participants who receive both DI and SSI (that is, concurrent beneficiaries) are in the DI category for payment purposes.
|Cohort||Month of First Assignment||Number of Participants|
|Total||Payment System and Payment Title|
|Source: March 2004 Ticket Research File merged to Ticket payment data through July 2005.|
The payment analysis in Section C considers all payments made for Tickets assigned under a new payment system by the end of 2003. Some of these payments are for earnings months after 2003. Their utility for assessing the impact estimates of the previous chapter is limited because the latter are based on Phase 1 states and consider outcomes in 2002 and 2003 only. Hence, Section D presents statistics for payments on Tickets assigned by just Phase 1 participants, and includes only payments for earnings months in 2002 and 2003.
2. Traditional Payments
The payment data analyzed here cover only the approximately 10 percent of Tickets assigned under the two new payment systems. A comparable analysis is not possible for traditional payments because of fundamental differences in the payment systems themselves. Viable claims for payments under the two new systems can be made as soon as the provider can document that the participant has achieved earnings above a specified level, potentially while the participant is still receiving services. Viable claims for payments under the traditional system can be made only after the participant has achieved sufficient earnings over a nine-month period and only after the SVRA has formally closed the case.
The findings reported in Sections B and C refer only to beneficiaries who assigned their Ticket under one of the new payment systems. Section D extends the inferences drawn from outcome payments to participants under the traditional payment system, under the assumption that those in the traditional payment system achieve months of zero benefits due to earnings no more frequently than participants under the new payment systems. Although we cannot verify this assumption directly, it is consistent with the survey data on participant earnings reported in Chapter VI. Specifically, during the survey month, 18.6 percent of participants receiving services from ENs (all under one of the new payment systems) had earnings above SGA, while only 8.9 percent of participants receiving services from SVRAs (the vast majority of whom were served under the traditional payment system) had earnings above SGA.2
B. Payments for Tickets Assigned in the First Two Years
By December 2003, beneficiaries had assigned 7,438 Tickets under one of the new payment systems, representing 11.5 percent of the 27,346 Tickets assigned as of that month. Of all Tickets assigned under a new payment system, 76.0 percent (5,654) were assigned under the milestone-outcome system. By July 2005, 11.4 percent of participants who assigned a Ticket under the two new systems (849) had generated at least one payment. The mean number of payments for participants generating payments was 6.4, and the mean of total payments for participants generating payments was $2,262. The mean payment for all beneficiaries who had assigned Tickets under the new payment systems (that is, including those with no payments) was $258. SSA payments on behalf of all of these beneficiaries totaled $1.9 million.3
1. Statistics for the First Assignment Cohort
Payment statistics for the first assignment cohort appear in Exhibit XIV.2. Of the 1,011 beneficiaries in this cohort, 14.5 percent generated at least one payment. For that group, the median number of months from Ticket assignment to first payment was just over 12, the mean number of payments for Tickets with payments was 9.7, the mean total payment for Tickets with payments was $2,800, and the mean total payment for all Tickets was $407. Although the percent of assignments in this cohort with at least one payment is not likely to increase substantially in the future, the other statistics are likely to increase as additional payments are made.
|Payment Title and System||Number Assigned||Payments by July 2005|
|Number with Payments||Percent with Payments||Tickets with Payments||Mean Total Payments for All Assigned Tickets|
|Mean Number of Payments||Mean Total Payment Amount|
|Source: March 2004 Ticket Research File merged to Ticket payment data through July 2005.
Note: The first assignment cohort includes beneficiaries who assigned their Tickets under the new payment systems between February and June 2002.
Some other important features of payments are also apparent from the findings for this first assignment cohort. Those who assigned their Tickets under the milestone-outcome system were substantially more likely to generate payments than those who assigned their Tickets under the outcome-only system. The likely reason is that a beneficiary must earn enough to be ineligible for benefit payments before a payment under the outcome-only system can be made, whereas payments under the milestone-outcome system are normally made before benefits are zero. The same reasoning is also likely to explain why the mean duration from assignment to first payment for participants generating payments is much longer for Tickets assigned under the outcome-only system than for Tickets assigned under the milestone-outcome system (19 months versus 14 months).4 We also found that DI beneficiaries who assigned their Ticket are more likely to generate payments than are SSI-only beneficiaries, perhaps reflecting differences in marketable skills, in the amount a participant must earn before benefits are reduced to zero, or in provider incentives created by the different payment rates.
Of participants in the first assignment cohort who generated payments, those who assigned their Ticket under the outcome-only system generated more payments and larger payment amounts, on average, than participants who assigned their Ticket under the milestone-outcome system. When averaged over all participants who assigned their Ticket (that is, including those with zero payments), however, those who assigned their Ticket under the outcome-only system were less likely to generate any payment and tended to generate lower payment amounts than did those who assigned their Ticket under the milestone-outcome system. The payment differential could change in the future if Tickets assigned under the outcome-only system continue to generate more payments than those assigned under the milestone-outcome system. This scenario seems likely because the maximum number of milestone payments is four, and some assignments that have already generated milestone payments are not likely to generate outcome payments.
Mean payment amounts for SSI-only beneficiaries who generated payments are lower than for DI beneficiaries. For the first assignment cohort, the percentage of SSI-only beneficiaries who assigned their Ticket under the outcome-only system that generated at least one payment was higher than the corresponding percentage of DI beneficiaries (14.6 versus 7.5 percent). The finding is surprising because the Section 1619(a) program, which automatically applies to SSI beneficiaries, has the effect of increasing the earnings threshold at which benefits for SSI recipients fall to zero to an amount above the threshold for DI (that is, the SGA level) unless the beneficiary has substantial other income. This finding is not replicated in later assignment cohorts, however. For instance, for the second cohort (those assigning a Ticket from July through December 2002), none of the 49 SSI-only beneficiaries who assigned their Ticket under the outcome-only system generated a payment by July 2005, as compared with 5.5 percent of the corresponding 235 DI beneficiaries.
2. Later Assignment Cohorts
To compare payments generated by later assignment cohorts with payments generated by the first assignment cohort, it is necessary to consider the interval from the assignment month to the payment month. Hence, this section presents statistics on the share of Tickets generating at least one payment by month since first assignment for each of the four cohorts. Also considered are the distributions of the number of payments generated and total amounts paid, by cohort.
a. Percentage Generating Payments
Of the beneficiaries in the first assignment cohort under the milestone-outcome system, the percentage generating at least one payment rose fairly rapidly over the first 12 months after assignment and continued to rise, albeit more slowly (Exhibit XIV.3). New first payments largely end after month 30 (2.5 years), although one first payment for the first assignment cohort was received 38 months after assignment.
First payment percentages for later cohorts are remarkably similar to those for the first assignment cohort during the months observed. Reasonable extrapolation from the trends shown in Exhibit XIV.3 suggests that the percentage of first payments for the later cohorts will eventually be slightly lower than it was for the first cohort, perhaps by one or two percentage points.
We also observed a noteworthy change in the first-payment experience by month 13 between the first cohort and later cohorts. The share of beneficiaries in the last three cohorts who received first payments was higher in each of these months than the share of beneficiaries in the first cohort. The apparent reason is a decline in processing times as SSA, the Program Manager, and providers gained experience in processing payment claims (see Chapter XIII).
The first-payment experience of participants who assigned their Ticket under the outcome-only system differs from that of participants who assigned their Ticket under the milestone-outcome system. As we saw earlier, the share of beneficiaries in the first assignment cohort with outcome-only assignments generating payments by July 2005 was smaller than the corresponding share for milestone-outcome assignments (9.0 percent versus 15.8 percent). However, as shown in Exhibit XIV.4, the percentage with first payments among those with outcome-only assignments in the first cohort continues to rise at the end of the observation period.
The difference between the first-payment experience under the two new systems reflects the difference between the two systems themselves. It is easier to generate first payments under the milestone-outcome system because milestone payments do not require earnings at a level that would reduce benefits to zero. Furthermore, processing time for first payments under this system should be shorter than for first payments under the outcome-only system because the latter require SSA to verify that benefits have been reduced to zero. As a result, essentially no payments were made under the outcome-only system in the first five months after assignment, whereas some payments were made under the milestone-outcome system during that period. Moreover, while the percentage of first payments under the milestone-outcome system began to level off at about 15 months after assignment, the same percentage of first payments made under the outcome-only system continued to rise through 24 months before leveling off.
Although first-payment experience improved under the milestone-outcome system from the first to the later cohorts, we see no substantial evidence of improvement under the outcome-only system. In fact, in months 6 through 18, the percentage of beneficiaries with first payments among the first cohort was somewhat higher than it was among beneficiaries in the later cohorts. In addition, the percentage in the second cohort with first payments was notably lower than it was for all other cohorts over the entire 30-month observation period. We do not have an explanation for the difference.
b. Number of Payments Generated
The number of payments generated by participants who generated at least one payment is an indicator of the extent to which participants are sustaining their earnings at a high level over a long period. Furthermore, four payments represent an especially important benchmark for Tickets assigned under the milestone-outcome payment system because the number of payments cannot exceed four unless benefits are reduced to zero due to earnings (that is, unless at least one outcome payment is generated).5
Exhibit XIV.5 provides information on the number of payments generated through July 2005 for beneficiaries who generated at least one payment under the milestone-outcome and outcome-only systems, respectively, again by assignment cohort. For each cohort, the exhibit shows the percentage of Ticket assignments with payments generating at least the number of payments indicated on the horizontal axis. The percentage of beneficiaries achieving each number among later cohorts is generally lower than for earlier cohorts because the later cohorts’ Tickets have been assigned for a shorter period. For the earliest cohort, we do not expect to observe much change in the reported figures as additional months pass because we have already observed each member of the cohort for at least 36 months since assignment; if, over that period, cohort members have generated fewer than 12 payments, they will probably not generate many more payments in the future. At the other extreme, we have observed some members of the latest cohort for only 19 months since they assigned their Ticket. Hence, many members with fewer than 12 payments to date may generate more payments in the future.
The distributions differ markedly for the two payment systems. A large majority of participants generating payments under the outcome-only system generated payments over a sustained period, but only a minority of participants generating payments under the milestone-outcome system generated payments over a sustained period. For the first and second assignment cohorts, 75 percent of participants who generated at least one payment under the outcome-only system generated at least 12 payments by the end of our observation period. The same is true of only one-third of those in the first cohort who generated at least one payment under the milestone-outcome system, with only half generating more than four payments; the corresponding percentages for the second cohort are lower.
|Source: March 2004 Ticket Research File merged to Ticket payment data through July 2005.
Note: Based on assignments made by December 2003 and payments through July 2005.
c. Payment Amounts
Payment amounts are of interest because they represent both SSA’s programmatic (that is, non-administrative) expenses for TTW and provider revenue. Exhibit XIV.6 shows the distribution of total payment amounts for Tickets with payments assigned by December 2003, by assignment cohort, payment system, and payment Title. The exhibit does not show the distribution for SSI outcome-only payments because so few Ticket generated payments in that group. Each graph shows the percentage of Tickets with payments generating at least the amount indicated on the horizontal axis.
The distributions of payment amounts reflect the distributions for the number of payments shown in Exhibit XIV.5. Whereas the median total payment amount for DI milestone-outcome assignments in the first cohort is just over $2,000, the median total payment amount for DI outcome-only is over $4,500, with almost 75 percent of the Tickets assigned under this system generating $4,000 or more. The median for SSI milestone-outcome assignments in the first cohort is about $1,500, reflecting the lower SSI payment schedule. As time passes, we expect that the distributions of the later cohorts will look more like those of the first cohort, and that a larger share of the beneficiaries assigning Tickets for all cohorts will achieve high payment levels, especially those DI beneficiaries assigning Tickets under the outcome-only system.
C. Payments for Earnings in the First Two Years in Phase 1 States
This section focuses on payments for participants in Phase 1 states only and addresses only payments for earnings months in the first two years of TTW—the same states and years covered by the impact analysis described in the previous chapter. Enough time has passed since December 2003 to ensure that nearly all payments that will be paid for earnings months during the two years have already been paid.
A total of 475 Phase 1 participants generated payments for an earnings month in the first two years (Exhibit XIV.7). SSA made payments for a total of 2,277 months during this period, with payments totaling $728,000. Assignments under the milestone-outcome system accounted for 89.9 percent of all assignments associated with payments, 86.3 percent of all payments, and 87.6 percent of the total amount paid. Assignments from DI beneficiaries accounted for 78.3 percent of all assignments associated with payments, 71.3 percent of all payments, and 83.3 percent of the total paid.
Outcome payments, which represent months with zero benefits, were made on 239 Tickets. The number of outcome payments totaled 1,564, or an average of 6.5 payments for those Tickets where the beneficiary generated an outcome payment. Thus, the 239 participants received no benefit payments due to earnings for an average of a little over half a year during the two-year period. The total amount paid in outcome payments was $375,000, an average of $238 per payment, or $1,570 per participant with at least one outcome payment.
D. Inferences About Impacts on Exits Due to Work
The payment data can be used, along with participation data from Chapter III, to (1) develop an upper bound for TTW’s impact on at least temporary exits due to work by participants in Phase 1 states during the program’s first two years and (2) assess the extent to which impacts are likely to be larger in later years.
|Payment System||Payment Title||Total|
|Tickets with Payments|
|With Outcome Payments||121||70||191|
|Total with Outcome Payments||161||78||239|
|Number of Payments|
|Total Payment Amounts (thousands)|
Source: March 2004 Ticket Research File merged to Ticket payment data through July 2005.
Note: Based on payments made and reported as of July 2005.
As mentioned, outcome payments were made on behalf of 239 Phase 1 participants for earnings achieved by the end of 2003, representing 7.6 percent of all Phase 1 participants under the new payment systems as of December 2003. All of these participants generated sufficient earnings to be ineligible for benefit payments for at least one month. That is, they had at least temporarily exited from cash benefits. In addition to having earnings above SGA, DI beneficiaries completed both the nine-month trial work period and the three-month grace period, and they had entered the extended period of eligibility. For at least one month, SSI recipients had earnings above the Section 1619A threshold amount, which is generally above SGA.
If the same proportion of participants under the traditional payment system exited cash benefits by December 2003 for at least one month due to work, then the total number of participants achieving that standard would be nine times as large—or 2,043—and would represent 0.078 percent of the more than 2.5 million Ticket-eligible beneficiaries in Phase 1 states as of December 2003. We interpret this figure as an upper bound for TTW’s impact on exits of at least month due to work by Phase I participants through the end of 2003. For two reasons, however, the true impact is likely smaller. First, a comparison of NBS data on monthly earnings in 2003 by participants with Tickets assigned to SVRAs and ENs suggests that participants under the traditional payment system likely achieved at least one zero-benefit month at a rate substantially lower than for participants under the new payment system (see Section A of this Chapter and Chapter VI). Second, a substantial number of the same participants might have exited due to work for at least one month by the end of 2003 in the absence of TTW.
For two reasons, impacts on exits due to work for at least one month will certainly be larger in later years than in the first two years. First, some of those who participated in the first two years but did not earn enough to generate at least one outcome payment during that period will generate at least one outcome payment after 2003. Not enough time has passed to determine exactly how many will generate a payment, but a reasonable projection can be made from the experience of the first assignment cohort through July 2005. By that date, 9.4 percent of the first assignment cohort had generated at least one outcome payment. We think that such a figure is a reasonable upper bound for the percentage of all participants who will eventually generate at least one outcome payment. It is important to note that the cohort analysis of the percentage of assignments with first payments (Exhibit XIV.3 above) suggests that few, if any, first payments will be made for members of the first assignment cohort after July 2005 and that first payments for subsequent cohorts are on a slightly lower track than for the first cohort. If 9.4 percent of all participants as of December 2003 (that is, under all payment systems) had exited for at least one month due to work, the number exiting TTW would represent 0.096 percent of Ticket-eligible beneficiaries.
The second reason that the impact on exits should be expected to increase is that the Ticket participation rate in Phase 1 states continued to increase after December 2003 (see Chapter III). The participation rate increased by 34 percent from December 2003 to December 2004 (from 1.03 to 1.38 percent). If we assume no increase in participation after December 2004, and if we further assume that the share of all participants who exit for at least one month eventually reaches our upper bound estimate of 9.4 percent, then the percentage of all eligible beneficiaries who participate and exit for at least one month will eventually reach 0.13 percent.
This upper bound projection for at least temporary exits from cash benefits is well below the 0.5 percent benchmark for permanent exits due to work that appears in the Ticket Act itself:
“If only an additional one-half of one percent of the current Social Security Disability Insurance and Supplemental Security Income recipients were to cease receiving benefits as a result of employment, the savings to the Social Security Trust Funds and to the Treasury in cash assistance would total $3,500,000,000 over the work life of such individuals, far exceeding the cost of providing incentives and services needed to assist them in entering work and achieving financial independence to the best of their abilities.” 42 USC 1320b-19, Section 2(b)(12).
Even if the percentage of participants who achieve at least one month of no benefits due to work reaches our upper-bound estimate of 9.4 percent, the TTW participation rate would have to increase to 5.3 percent—almost four times the December 2004 value in Phase 1 states—for the number of participants exiting for at least one month to reach 0.5 percent.
For reasons discussed above, we think that the actual impact on at least temporary exits from cash benefits by participants is well below our upper-bound estimates. The impact on permanent exits would be lower still. Although the early statistics on the number of outcome payments for participants who generate at least one outcome payment offer some encouragement, we have to expect that some participants who stop receiving cash benefits due to work will return to the rolls after a short period, and perhaps many will return after a few years.
For two reasons, it is at least possible that the impact of TTW on exits due to work is larger than indicated by the upper-bound estimate for participants because outcome payments do not capture all the instances where a participant stops receiving cash benefits because of work. First, providers might not file claims for some payments that they are eligible to receive. As discussed, we think that the number of such cases is likely to be small, but we have no empirical evidence on this point. Second, beneficiaries can stop getting cash benefits without participating in TTW, and it is possible that TTW induced a significant number of nonparticipating beneficiaries to work enough to reduce their benefits to zero. SSA’s efforts to reduce the post-entitlement workload backlog and improve the process for reporting and validating earnings might have resulted in termination for some. SSA’s efforts to provide benefit counseling might also have affected the decisions of some nonparticipants. More broadly, the beginnings of a shift at SSA toward a culture that is more supportive of beneficiary efforts to search for and retain work could be having a positive impact on exits by nonparticipants. Even if the number of such exits is large, however, it might be a mistake to attribute them to TTW. Although TTW might have been the driving force behind these other changes, presumably many, if not all of the changes, could have been implemented without TTW.
We have to conclude, however, that as it is currently configured, the TTW program’s impact on participant exits will not reach the Ticket act benchmark unless participation increases above the level reached in Phase 1 states by the end of 2004, there is a surprising large change in beneficiary behavior, or TTW somehow induced a large number of exits that are not reflected in the outcome payment data.
1 The bulk of payments made appear in the administrative files shortly after the payments are made, but a few do not. For instance, the July 2005 extract used here includes data for 32 payments made in 2004 that were not in a February 2005 extract. That number represents just 1.1 percent of all payments made in 2004 (based on the July 2005 extract). (back)
2 The percentages reported here were obtained by multiplying the percentage with earnings, as reported in the table (30.6 percent for ENs and 32.7 percent for SVRAs) by the corresponding percentages with earnings above SGA for those with earnings (60.6 and 27.1, respectively). (back)
3 In Chapter VIII we reported that SSA had paid $2.5 million under the new payment systems through July 2005. The figure reported here is lower because it excludes payments made on behalf of beneficiaries who assigned their Ticket after December 2003. (back)
5Some beneficiaries who assigned their Ticket under the milestone-outcome system generated outcome payments before generating four milestone-outcome payments. Hence, not all payments for the large share of assignments with four or fewer payments under the milestone-outcome systems are milestone payments. (back)