Maintaining the flow of philanthropy in this economy

January 4th, 2009 by Al Lewis (alewis)

In this economy, the people who suffer the most are the ones who can least afford it, due to the steep decline in philanthropic human services contributions. There are two types of philanthropic human services contributions:

(1) Maintenance programs – shelters, food banks, etc. These programs are the “social safety net,” for which government funding is often inadequate and must be supplemented to prevent homelessness and malnutrition

(2) Investment-oriented programs – scholarships, parental education, rehabilitation, microfinance, etc. These programs are designed to redirect people’s lives in a much more positive and productive direction than their current trajectories would produce.

This proposal addresses the latter.

In theory, this second category of anti-poverty programs, implemented well, should turn around the lives of large numbers of people.

In practice, despite the paybacks, opportunities in these arenas are almost certainly underfunded for a range of political reasons by the government even in good times, one obvious example being that those who would benefit most from these initiatives neither vote nor contribute money to politicians in amounts proportionate to their numbers. In tough times, these programs suffer even more.

The thesis of this White Paper is that it is possible to magnitudinally increase the scale of anti-poverty investment-oriented programs at no cost to taxpayers by involving the private sector (the emerging “venture philanthropy” industry) via an outcomes-based financing methodology which has never been done before in this field, yet which has proven as a funding mechanism to improve the lives of Medicare beneficiaries with chronic disease, at no cost to taxpayers.

Those “disease management” programs offer perhaps the most analogous situation to this one, where money can be saved by improving people’s lives, but most insurers and other payors (such as Medicare) were originally willing to “pay” for the improvements only out of realized savings. At many points below, an analogy from disease management can be used to explain how that field faced issues in study design, intervention and especially performance measurement, and overcame them.

How Outcomes-Based Community Growth Initiatives differ from current approaches and why the Initiatives would work in theory: A metaphor

Imagine if investment in underserved areas was like planting fruit trees. The object is basically to get the most fruit in total. However, most government-funded initiatives, with the goal of ending up with the most fruit, do so by subsidizing the production of trees in the oft-misplaced hopes that the tree owners — though their decision to plant was based largely on subsidies — will tend those trees until fruit is borne.

The best example is subsidized housing. A major 1980s housing initiative provided very generous tax credits, more than dollar-for-dollar in some cases, to build subsidized housing. Today, many of those housing projects are either in bankruptcy or have reverted to tenement status…but most investors in them have done well at taxpayer expense owing to the tax credits in the early years. This is the equivalent of taxpayers paying someone to plant trees and hoping that the trees will bear fruit. However, since the trees are planted for the wrong reason (tax subsidies), possibly by the wrong people (tax shelter experts), in the wrong places (where building costs are lowest, not where need is greatest), with the wrong incentives (the fruit is an afterthought, with tax savings the primary motivator), they failed.

Outcomes-based Community Growth Initiatives (the “Initiatives”) are exactly the opposite. Investors make all the decisions and risk all the capital – hence they decide which figurative trees to plant, where they will grow best, and how best to nurture them, so that they will bear large amounts of fruit for many years.

So far this is a description of any good investment. Most such investments will be made by investors as they are today, meaning that with no particular incentive to maximize the public good, investments get made in businesses which offer the greatest private returns. The goal of the Initiatives program is to create investment in areas which offer the greatest returns in total, public plus private, thus maximizing total return. To do that, the Initiatives promise to transfer enough of the total realized return to the private sector to ensure that these investments get made by making the risk-adjusted returns competitive with straight private investments.

Here is what would make Outcomes-Based Community Growth Initiatives successful in this respect. Normally when (figuratively speaking) trees are grown, they create jobs along the way, income for the planter-investors, increased value of the land, and a “multiplier effect” of support businesses. That’s capital formation Normally the people who benefit along the way pay taxes, which they otherwise wouldn’t have paid. . Government revenue growth is driven, ultimately, by physical and human capital formation.

Likewise, domestic government spending in many categories – Medicaid, corrections, other social services consumed disproportionately by the underserved – is a direct function of the failure of capital formation in many pockets of the United States and elsewhere in the developed world .

So government would gain from this figurative tree-planting activity in two ways: it would collect more taxes than otherwise, and spend less money than otherwise…but if people don’t plant the trees in the socially optimal places, this government windfall wouldn’t exist or would be much smaller. To spur this tree-planting investment without distorting incentives as in tax shelters to encourage investments with the largest possible total private+public returns, government would contract to share some of this “windfall” – if in does indeed materialize — with Outcomes-based Community Growth Initiatives companies.

For instance, these companies would “do the math” to estimate the windfall and figure that they needed 50% of the windfall, to make a risk-adjusted return over time, keeping in mind that most of the return will be in the “out” years. They would bid for a 50% share over twenty years (or whatever the term of the contract is – a shorter term means a higher share to the Initiatives investor). If that is the best bid and they are the best qualified company and are acceptable to the community in question, they win the opportunity to create that windfall.

Figure A: In order to entice Investment B without subsidies, government agrees to share some of its gains (in dark blue) to make the total expected return to the private investor higher than Investment A
(If “Figure A” is not showing up on your screen please email us and we will send it privately.)

Translating Theory into Practice

First, the program is announced as a demonstration project to make sure there are “takers” on the government side. This should be the easier part because there is no risk. Quite the opposite: the main difference between the Initiatives and conventional funding is that in the former case, the government has to be “sold” as much as the investor, since tax subsidies put the risk of failure mostly on the government.

In this case, though, there is no downside for the government. The government only pays out of increased cash flow if and when it materializes.

Before the interested local governmental entity (in partnership with state and federal, which also share in the windfall) can “sell” the Initiatives investors on doing this, it needs to compile summary statistics on particular communities interested in becoming a part of this project, which can be as small as a ZIP code or as large as a small city. Then, adapting widely accepted biostatistical control techniques to this purpose, it “matches” those communities to nearby communities with similar demographics and socioeconomic issues.

The Analogy to Disease Management – an Introduction

This is approximately how Medicare measures success of disease management in its Medicare Health Support projects, and disease management is an excellent model for how this program might work. The government has contracted to pay these disease management companies only if these programs succeed.

In disease management, hospital days, existing drug compliance, and “health status” measurements, like rates of heart attack and other events, are used to estimate potential for a local population’s improvement potential. Likewise, “socioeconomic status” measures would be used in the Initiatives program to “match” communities for measurement. Examples:
• What percentage of people finish high school and go to college?
• What percentage spend time in jail?
• How many babies are born to teenagers?
• What percentage of babies need neonatal care?
• What is the unemployment rate?

For residents of those communities, estimates are made of future tax revenue from and social services outflows, combining all levels of government.

Companies involved in the Initiatives then determine whether the existing level of socioeconomic status and the potential for improvement will merit the investment. They will no doubt consider their ability to attract business (without artificial tax breaks), the sincerity and degree of interest by both official and unofficial community leaders, and the potential entrepreneurial spirit of the population. The health of local and regional economies matter, but since the Initiatives company’s performance is “matched” to similar areas, there is no possibility of simply riding a rising tide.

For an Initiatves company to be interested, there must be enough potential financial benefits in the windfall to yield a return on that investment, and there must be a way for their investors to capture that return via bidding for a portion of it. It is unlikely that many investors will require “venture-level” returns of 40% a year. Even today, there are individuals and organizations which undertake “I Have A Dream”-type programs and other interventions with no expectation of any return, solely because it is the right thing to do. Many more investors will be happy enough to break even.

For the governments at all levels to be interested in working with the private sector, there needs to be some expectation of upside but it is critical that there be no financial downside for the government—no scenario in which it is possible to finish below break-even, paying out money with no offsetting revenues. The whole point of the exercise is to prevent any argument about trade-offs, budgets and deficits, competing spending priorities, tax increases. Once that happens, at least at the federal level, appropriations are required, the Congressional Budget Office must get involved, and the debate centers around costs, not benefits.

The Example of Disease Management in more detail

A new discipline which grew up under a very similar payment model is disease management. It has become the fastest-growing healthcare service, and the two public DM companies have performed very well over the decade. The idea is that by providing educational and mentoring assistance to people with chronic disease soon after the diagnosis, those patients learn how to take care of themselves better and stay healthy enough to avoid the hospital. No one doubts the value to the patients in improved health and quality of life. The economics, though, are not proven to the satisfaction of all observers because it is rather expensive to talk to everyone with chronic disease when in any given year only a small minority would end up in the hospital.

Nonetheless, most health plans, eighteen state Medicaid programs and many large self-insured employers have increased spending on disease management by 25-30% a year in total for the last seven years. For the first time, the federal government is getting heavily involved, with Medicare committed to allocating close to $100-million this year.

Why, if the economics are not proven, would all these payors, especially including the federal government, be spending all this money? One reason: The financial savings in reduced hospitalizations and emergency room visits are contractually guaranteed, by the disease management vendors, to exceed the costs of the program, usually by a considerable margin. In the Medicare-sponsored programs, as of this writing it is unlikely that significant net savings will be shown. That’s precisely the point: If savings targets are missed, Medicare is held harmless…and yet DM companies are still lining up to try new projects with Medicare.

The disease management industry simply would not have achieved critical mass had it not been for these hold-harmless provisions. As in the case of anti-poverty programs, within most health plans the advocates for improving access by people with chronic disease were originally “outvoted” by the advocates of keeping spending low. Both sides presented good arguments (though the low-spending advocates conceded that health would likely improve, just not cost-effectively so). The debate would have stalemated had it not been for the emergence of disease management vendors willing to “put their money where their mouths were,” by being contractually willing to have their payments tied directly to documented reductions in hospitalizations.

The Relevance of the Disease Management Example

Several comparisons suggest that the financial and humanistic success in disease management can be duplicated or even improved upon in this proposed Initiatives program.

Most importantly, the example demonstrates that where there is an arguably substantial positive return on investment from improving the condition of people, there will be companies willing to make large bets that those improvements can be achieved. The entity ultimately paying the bill for those improvements but reluctant to commit to a fee based on the possibility that the advocates of a substantial return are correct need not risk a nickel.

Second, if anything, improvements in the health status of people with chronic disease in health plans are much harder to achieve and maintain than improvements in overall socioeconomic welfare. People change health plans at the rate of about 20% a year. By itself that makes the improvements 20% harder to achieve since the effort is wasted on someone who leaves the plan after all that investment has been made. In the case of outcomes-based social funding, there is no concept of “leaving” the program. Everyone counts.

Third, the time horizon can be longer for socioeconomic improvements. The companies which emerge to address this opportunity will have longer-term time horizons. They won’t expect the 2:1 first-year return-on-investment which buyers of disease management programs demand.

The key issue in Outcomes-Based Community Growth Initiatives: Performance Measurement

Measurement has until recently been a major issue in disease management and there is no reason to expect anything different here. Even so, controversy around measurement is likely to be mitigated by two factors. First, there have been dramatic actuarial learnings in disease management measurement which can be applied. The disease management field has now determined the following:

(1) Multiple measurement techniques need to be used to confirm one another. Oftentimes there is a financial measurement or actual dollars saved or earned by the various parties, backed by “plausibility indicators. The plausibility indicators must confirm the financial measurement for the vendor to have “hit their numbers.” Example from disease management: If a heart disease management program is showing savings, it stands to reason that the actual rate of heart attacks and angina (both easy to track) will have declined vs. baseline, and that filling of certain prescriptions (also easy to track) will have increased. Those unit-based approaches, all relatively easy to track, can directionally and magnitudinally confirm the financial results. In this situation, the “plausibility indicators” such as graduation rates, incarceration rates, and teenage birth rates, are even more closely tied to the financial results and likely to correlate quite closely

(2) There are exogenous variables too, like new drugs or procedures. In a prospective controlled trial such as this one and such as Medicare’s Health Support program, those don’t affect measurement because there is no a priori reason to think they will be adopted faster in one population than another. The analogous situation here is that macroeconomic trends, such as unemployment, globalization, sector shifts, or new technologies should affect both populations equally absent this intervention.

Second, the most accurately measurable study design is not a pre-post but rather a prospective trial in which two matched populations are tracked over the same time interval going forward. That is rarely possible in disease management (though Medicare does it) because a health plan is unlikely to deny disease management to some people, for public relations reasons. However, in this situation there are likely to be more candidates than money, so an accurate, prospective study design is quite feasible.

Third, one doesn’t measure just financial outcomes. For instance, in diabetes, one also measures amputation rates, a major long-term cost which can be avoided by good blood sugar control. Blood sugar control is measured by a specific blood value which predicts future rates of amputation and other complications. This gets measured too because neither party wants to wait to see what the amputation rate is. Sometimes two variables confound each other. For instance, improving that blood value in a population is more difficult if one also excels at getting more people to care enough to be tested. So an improvement in both the blood value and the testing rate might get a special bonus.

Likewise, in the measurement of an Initiatives program, one measures intermediate outcomes like teen pregnancies and dropout rates to predict actual future cash flows, which will lag the underlying socioeconomic improvements.

Does one pay based on these intermediate indicators or wait until actual tax revenues and social spending show an improvement? For political reasons, only payment based on actual tax revenues is likely to be acceptable, in order to maintain unimpeachable budget-neutrality. But a secondary market for these revenue streams will inevitably develop probably even before the first revenue-sharing payment is made, and the early valuation of future revenue streams will be based on changes in intermediate indicators. Intermediate indicators are also important because they provide the basis for performance reviews. For instance, in Medicare’s disease management program, a contract is being awarded to compare the different projects around the country well before the reconciliation, using intermediate indicators to determine which are succeeding and failing, and use the best practices from the former to help the latter.

Questions and Answers

Q: Who would fund this? There are no venture philanthropy companies today.

A: There has never been a situation in American history where a major economic opportunity got ignored for long. These companies—and there already are philanthropists who style themselves “venture philanthropists” (the name is not new to this White Paper)—will start up just as disease management companies did when faced with similar opportunity. One disease management company, Healthways, was the most successful stock on Wall Street over a six-year period, an observation which will not be lost on potential “venture philanthropy” investors.

Q: Do you measure people or places?

This is a perfect example of what disease management has learned—there is no single right answer. (In their case, one such dilemma is measurement of everybody or just people who’ve been in the health plan for a period of time.) Primarily, you measure people. There is always movement out and movement in to communities. Paradoxically, if an intervention succeeds, movement out could increase. This almost certainly indicates success, especially if one moves into an owned home as opposed to another rental. Clearly one also wants to improve the actual community too—but not by “gentrifying” it and causing current residents to flee to other depressed areas. This measurement could be done based on a comparison of per-capita income of the zip code to which relocation take place vs. that of the original community. People are succeeding if they are moving to a higher-income location, failing if they are moving to a lower-income one. Both outcomes can be measured.

Q: Variables needing to be tracked can get multiplied if there are multiple venture philanthropists in a market. How do you determine who gets credit?

One other learning from disease management: Avoid multiple vendors—even doing different diseases—within the same payor. Instead use one vendor and if the vendor isn’t expert in all areas, they can subcontract. There should be one contract, one point of contact, and one “macro” set of measurement. In a prime-sub arrangement, the contractors can split things up for themselves. To be sure, in venture philanthropy just as in disease management there will be organizations which excel at specific interventions but lack expertise in others. They should be encouraged to partner to create a single contracting entity.

Also, these issues can’t be addressed in a vacuum. One can’t reduce crime without improving schools. Yet students might stay away from schools if the crime rate in the school is too high. One could imagine that many other such interrelationships exist. Just as disease management started with single-disease vendors and moved to a multiple-disease approach, one would expect venture philanthropists to assemble an intervention to address many issues.

Q: It sounds like there are a lot of variables to measure.

Yes. In addition to wanting to ensure the economics, satisfaction/happiness and other subjective indicators are equally important. In disease management, it’s not just about the savings – people have to feel good about the program. The same would, of course, be true here. In the contract, some quality and community satisfaction variables would also be tracked, and there would be penalties and possible loss of contract if satisfaction is low.

Q: Can’t some of these outcomes be “gamed,” like dropout rates? Haven’t private-sector education companies attempted to “game” dropout rates because they were being compensated on graduation rates?

The discipline of the private sector provides the key difference. The venture philanthropists would be “prime contractors” in that situation, rather than the education companies. It would be in their interest to prevent “gaming” because recall that they will want to “sell out” to investors with longer-term time horizons. If those investors think the outcomes are gamed, they will pay much less for achieving them. The distinction: the ultimate outcomes are mostly relatively ungamable financial cash flows, though for very good political and social policy reasons some arguably gamable intermediate outcomes will be included in the evaluation.

Q: Talk more about the importance of satisfaction of the people in the program.

In disease management, satisfaction is generally high enough that it goes without saying, but nonetheless it is an important variable for measurement and empowerment in this case—it can’t be taken for granted. Communities are going to want the right to terminate a contract with a “venture philanthropist” for failure to perform. Tracking satisfaction is one way to trigger that.

The right to terminate is a perfectly legitimate contractual expectation but the question is, how low does satisfaction have to be? What are the remedies? Is there a “recall” process? It is entirely up to a community wishing to attract a venture philanthropist, how much power they want to retain for themselves. The more power they wish to retain and the easier they wish to make it to sever ties, the less likely they will be to attract a partner, and/or that partner is going to demand a greater share-of-savings. (The analogy in disease management is, not-for-cause termination clauses. The easier it is for a payor to walk away, the higher the fee and the lower the net savings guarantee.) On the other hand, the community may decide it’s worth it. For instance, there is some dissatisfaction with Edison as the contractor for schools in many communities, but those tend to be long-term contracts which are difficult to terminate. Other things equal, long-term contracts reduce the incentive of Edison to respond to those criticisms. On the other hand, Edison claims that its results will take years to be realized, so they wouldn’t enter a market with a short-term contract.

Q: Speaking of Edison, would “venture philanthropists” overhaul school systems and bring in contractors like Edison whose track record is controversial?

In the selection process and RFP for venture philanthropists, that is one excellent question to ask…how exactly do they intend to get their results? In disease management in the old days, vendors came in and tried to start their own physician network and tell the doctors how to practice. It wasn’t popular. Nowadays, DM vendors work within the system, only recommending de-credentialing a physician from the network on rare occasion. Likewise, one would expect the more popular approach would be an “I Have a Dream Foundation” approach. This is far less controversial than privatized school districts, and its outcomes are well-tracked and very positive in the large majority of cases. On the other hand, there may be communities which want their school system overhauled or possibly privatized.

Q: Is there truly enough economic opportunity here to interest investors who want venture-level returns?

A: One could easily argue on paper by looking at the results of “I Have A Dream”-type programs that just keeping kids in school eventually pays back multiples of the original investment. But the actual level of return is definitely in the realm of the unknown. In venture capital, by definition, the level of returns is always an unknown. For instance, in the case of biotechnology, often there is no expectation of revenue for a decade from inception. Yet biotech has attracted much venture capital over the years. So fear of the unknown should not be a deterrent here either. One positive difference between this and biotech or other investments: Some venture investors, owing to the socially beneficial nature of this investment, would settle for lower returns than their usual “hurdle rate.” Some might even be happy breaking even, capturing only enough of the savings to fund future programs. Keep in mind that “venture philanthropists” today already exist, and don’t require any financial return. It stands to reason that at any given level of expected return there will be takers in the venture community, the higher the return the more takers there will be.

Q: Aren’t some of the benefits very long-term investments that venture philanthropists won’t want to delay their returns for? For instance, even if an intervention increases the college graduation rate, the differential in income taxes vs. earnings from a high school graduate doesn’t max out until people retire. Who would wait that long?

A: One option would be to actuarially estimate the present value of that income tax stream and for the government to pay back the VCs out of the “present value” once certain milestones have been achieved. However, to make this proposal truly inclusive by getting support from the entire political spectrum including those who oppose any new government commitments, it may be critical to show that there is not one iota of additional government spending required to implement it (other than the relatively trivial expenses of results measurement and contract procurement). That would prevent the government from paying out on milestones, before offsetting cash flow is received. Under that constraint, it is quite likely that, just as happens in almost any other situation where investor time horizons differ, the streams of savings are “securitized” and initial venture investors are paid off by investors who want what almost amounts to an annuity.

A secondary market would likely be more inefficient than simply allowing the government to settle up based on intermediate outcomes, but might be necessary to attract more political support and to prevent the Initiatives program from needing an official allocation.

Q: What if these various projected “streams” change over time? For instance, projections of differential earnings for a college vs. high school graduate might vary over time. The government-venture philanthropist contract could lock in one rate and have it be wrong. Or, tax rates themselves could rise and fall, making the value of a dollar of future earnings lower or higher.

That is true. It is a risk applied to future earnings and as such one or the other party to the contract will be willing to accept that risk in exchange for a better division of the savings. For instance, supposed a college graduate earns twice as much today as a high school graduate. One offer to the government might be to “lock in” the difference and therefore have future cash flow depend only on the variables it is trying to influence. Another offer might be willing to place a bet that tax rates and earnings differentials are going to rise, and this might be a more attractive offer. Likewise, the government may be ideologically so tied to the concept that there can be no chance of a net outflow of funds here that they would only accept proposals in which these “uncontrollable” risks were assumed by the Initiatives contractor. This would probably result in the contractor, now assuming more risk, to propose a different split. And some of the strategy would depend on how the contractor felt that the securities markets would react to a securitized offering of a stream of income with more variability associated with it.

In any event, uncertainty of future trends is not a reason not to do the program. It might factor into the government’s contracting philosophy and strategy but as long as the government can be 100% assured that it can’t lose money all other measurement issues are second-order issues, not deal-killers. Likewise, in disease management, some buyers today prefer programs where they have 100% certainty that “hard” dollar savings at least match program costs than for programs which have 80% certainty of 2:1 returns but where there is no guarantee of at least a 1:1.

Q: How is this different from government tax incentives, “enterprise zones,” low-income housing tax shelters and other initiatives to involve the private sector which have been tried from time to time?

A: In one fundamental way: The investor’s incentives and the government’s projected outcome are joined at the hip. There is no way the investor gets paid unless and until the results are achieved. Other programs were the opposite. There is a saying “What’s managed is what’s measured,” which was very true here. The private sector worked to maximize the tax benefit rather than to achieve the desired outcome. In low-income housing, for instance, the tax benefit was paid out over the first years, rather than tied to the occupancy rate or other success factors. Basically the “outcome” was to get the units built — which is laudable but is not a true “outcome”– It has been quite common in the case of low-income housing for the investors to have already made money on the front-loaded tax benefits when the units themselves fall into disrepair and go bankrupt.

Q: If this is such a good idea how come nobody has ever thought of it?

A: One could say the same thing about any idea, including disease management, which has now grown at 25% a year for eight years. The best-performing stock over the last six years is a disease management company.

To conceive this one needs a background in outcomes-based contracting and measurement in large populations (which describes disease management), development of new industries, financial markets, venture capital, and biostatistics. The author has all five.

Q: What is the major reason this wouldn’t work?

A: To create the financial benefits pool, all three levels of government, or at least state and federal, must actuarially work together. There is not enough financial benefit just in, for instance, increased federal income tax receipts, to create a savings pool large enough to attract many investors who spend a lot of money to achieve a successful outcome. Since some philanthropists will do projects such as these with no return at all, it does stand to reason that more investors will be attracted even if only a municipality wants to do this. To attract the most investors, though, all three levels of government would need to participate in the pool.

Q: Why is the focus on urban poverty? Rural poverty is in some ways even worse because there is less of a local “safety net.”

A: There may be opportunities in rural areas as well but part of the upside is educating and training people for jobs, which are generally more plentiful in urban areas. Just as disease management started in the “obvious” disease of heart failure and has subsequently migrated to other diseases, the suggestion would be to start where the opportunity is most likely to be realized.

Q: This paper makes references to aligning incentives and bringing people together, but it doesn’t appear that the middle class benefits. What would be the benefit to the middle class?

The government (particularly if the federal government participates) could offer dollar-for-dollar middle class tax cuts to be financed with its share of the savings. The important thing is to build consensus by having everyone share in the benefits.

Share This Idea:
  • Digg
  • StumbleUpon
  • del.icio.us
  • Facebook
  • Mixx
  • Google

Email

2 Responses to “Maintaining the flow of philanthropy in this economy”

  1. kim Says:

    this fits with investors’ unwillingness to take any risk these days. I hope you are right in that these things are cost-effective. One recommendation: It doesn’t bother me because I am one of the policy wonks who eats this stuff up but you need to post shorter or you’ll lose your casual audience

  2. harrymellen Says:

    This was a bit too sophisticated for a blog, I gotta agree with Kim. it took me two reads. when I finally got it, it sounded like a great idea but no way your average idiot politician is going to take the time to understand this.

Leave a Reply