Chapter 2    (01)

State of the Future Index    (02)

Study initiated and conducted by Theodore J. Gordon    (03)

2.1 State of the Future Index––2001    (06)

1. Executive Summary    (025)

Indexes have been constructed to aggregate many factors into a single number to depict the general state of affairs in a variety of areas. The cost of living index combines the cost of food and other consumer goods in a standard “market basket” to show how prices are changing. The Dow Jones Industrial Average aggregates the price of stocks of selected firms to create a number that quantifies the state of the New York Stock Market. The Millennium Project’s intent is to construct a State of the Future Index (SOFI) that measures the changing state of the future and shows whether conditions, globally or nationally, promise to get better or worse. In general, it explores the questions of whether we are making progress on the global challenges previously discussed.    (026)

It is true that polls could be conducted to obtain public perceptions about the future outlook (e.g., “Do you think things are likely to get better or worse?”), but such surveys are subject to recent news and media pressures, and people answering may discount or not know about recent improvements or threats.    (027)

The SOFI is created by a statistical foundation based on the history of key indicators and forecasts to quantitatively answer whether the future promises to be better or worse. If things seem to be changing, then the SOFI would make it clear how, and would make it possible to identify the factors responsible. If confidence were developed in such an index, it could be used for policy purposes: plans could be evaluated and compared on the basis of their impact on a State of the Future Index.    (028)

The State of the Future Index research was intended to answer five questions and, based on the answers, to construct a state of the future index.    (029)

1.   What variables should be included in a State of the Future Index?  If people say that the future seems promising, what do they mean? That life will be good for themselves and their family; that food, water and shelter will be sufficient; that fear will be absent and life fulfilled. What else should be included? The selection of variables forces a person to answer two key questions: What do I consider an improvement? And how would I know it if it happened?    (030)

2.   How can very different variables be combined? It is necessary to make all the measures included in the SOFI commensurate—that is, expressed in terms that are comparable.    (031)

3.   How can the variables be forecast? Measurement is not enough; since we are dealing with the future, the elements of the SOFI must be forecast. How can this be done?    (032)

4.   How can the variables be weighted? The SOFI elements are not all of equal importance to the future; the SOFI uses the concept of nonlinear weighting in order to balance the significance of the measures that are included. But weighting leads to other problems: different people may see one or the other of the measures as being more or less important, or even of different polarity—that is, some may see an increase in a variable as good while others see it as bad. Since the SOFI is designed to be a globally aggregated measure, it can mask differences among groups or nations: the SOFI could look very positive and yet for some groups or nations, the situation could be worsening. Therefore it is important to recognize that disaggregated SOFI analyses will be essential so that groups or nations can determine—using their own data and weights—how things seem to be changing.    (033)

5.   How can double accounting be avoided? This has to be considered or else one area could be over-represented. For example, should SOFI include both a measure of carbon dioxide concentration and global temperature? They measure different things but are important to consider for the SOFI for the same reason.    (034)

Many issues must be considered when constructing such an index. The future cannot be reduced to a single number. But the ongoing pursuit of this index would help raise and inform a discussion about what constitutes improvement. Combining many variables into a single index number can lead to loss of detail about the forces that move the index. Creating an index requires judgments not only in selecting the variables to include, but also in weighing them to create an aggregate number.    (035)

An index of global conditions can mask variations, for better or worse, among regions, nations, or groups. The apparent precision of an index can easily be mistaken for accuracy. For these reasons, many people interested in tracking social or economic conditions prefer to keep separate and distinct the variables that they consider important. These issues are discussed throughout a short overview in the print version and in more detail in this chapter. Nevertheless, the promise of a State of the Future Index is alluring: it offers the hope of identifying positive and negative changes and points of leverage for policy, as well as achieving some measure of balance in answering questions about the outlook for the future.    (036)

The process by which the SOFI indicators were selected involved the following steps.    (037)

First, the 1999–2000 Global Lookout Panel of the Millennium Project was asked to identify indictors by which the status of 15 global challenges could be measured. These nominated indicators were subsequently evaluated by the panel in terms of their availability and usefulness.    (038)

The results, plus a review of other index studies, were submitted to the Global Lookout Panel in 2001 to collect judgments about potential indicators for the SOFI. The respondents provided judgments about what the best (norm) and worst (dystopic) status was for the indicator in 2011. They also rated the importance of reaching the norm and dystopic state. The criteria for assigning a high weight to a variable were: the number of people affected; the significance of the effect; whether some groups seem to be affected differentially; the time over which the effect will be felt; and whether the effect is reversible.    (039)

Millennium Project staff worked with the variables identified in this questionnaire, obtaining 20 years (where possible) of historical data from the most authoritative sources, and forecasting each variable using a time series approach. The data were scaled by assigning the value of 100 for the most desirable (normative state in 10 years) and 0 for the least desirable values (dystopic state in 10 years).    (040)

It was not possible to make a reasonable judgment about a normative and dystopic status of some variables such as percent of urbanization and UN peacekeeping funding. Yet it was clear they were important to the future. Hence, within the SOFI research, staff will keep track of such variables but not include them in the SOFI calculation per se.    (041)

The variables were weighted in a novel way. All indexes studied thus far have assumed that weights are constant and independent of the value of the variable they modulate. Instead, SOFI assumed that the weights assigned to some indicators should change as the values of the indicators rise and fall. When an indicator reaches a level of satiation, it may not be as important as it used to be. For example, when the level of food intake is below 1500 calories per person, the variable is very important. When it is above 3000, the sense of urgency associated with hunger no longer gives this variable much weight.    (042)

To accommodate this nonlinearity, an S-shaped function was developed that allows the weight of a variable to vary with the value of the variable.    (043)

After considering the double accounting issue and removing redundancies, the SOFI was constructed with the following variables:    (044)

•     Infant mortality rate (deaths per 1,000 live births)    (045)

•     Food availability (calories per capita in low-income countries)    (046)

•     GNP per capita (constant 1995 US dollars)    (047)

•     Share of households with access to safe water    (048)

•     Carbon dioxide emissions, industrial countries (million kilotons)    (049)

•     Annual population addition (million)    (050)

•     Percent unemployed    (051)

•     Literacy rate, adult total    (052)

•     Annual AIDS deaths (millions)    (053)

•     Life expectancy (world)    (054)

•     Number of armed conflicts (those with at least 1,000 deaths per year)    (055)

•     Developing-country debt    (056)

•     Forestlands (million hectares)    (057)

•     Rich-poor gap (ratio of global average income of top 5% to bottom 5%)    (058)

•     Terrorist attacks    (059)

•     Violent crime (per 100,000 population)    (060)

•     Share of world population living in countries that are not free    (061)

•     Secondary school enrollment (% of school age)    (062)

•     Share of population with access to local health care (in 15 most populated countries)    (063)

Figure 2-1 presents the results of this exercise.    (064)

   (065)

The past growth in the SOFI occurred because of improvements in:    (066)

•     GNP per capita    (067)

•     Infant mortality rate    (068)

•     Food availability    (069)

•     Share of households with access to safe water    (070)

•     Literacy rate, adult total    (071)

•     Secondary school enrollment    (072)

•     Share of population with access to local health care    (073)

•     Terrorist attacks    (074)

•     Life expectancy    (075)

The factors that are deteriorating and hence retarding the growth of SOFI are:    (076)

•     Carbon dioxide emissions, industrial countries    (077)

•     Percent unemployed    (078)

•     Forestlands    (079)

•     Rich-poor gap    (080)

•     AIDS deaths    (081)

•     Developing-country debt    (082)

There is a delicate balance between the forces that move the curve up and those that depress it. The signs are positive overall, although relatively small changes in the following will swing the curve:    (083)

•     Developing-country debt    (084)

•     Unemployment    (085)

•     Rich-poor gap    (086)

Finally, the SOFI concept can be adapted at the regional, national, state, local, and group levels. Comparisons can be made between local measures and another SOFI.    (087)

Much remains to be done. Here are some thoughts about an agenda for the further development of a system to track the state of the future:    (088)

1.   Even the crude measures presented here could be helpful to policymaking. Therefore, it is suggested that the SOFI be re-evaluated and computed annually and published, together with a narrative and accompanying charts of the key variables that explain reasons why the future appears to be better or worse than before.    (089)

2.   Indicators that are important to the future but do not fit in the statistical analysis should be tracked and displayed with the SOFI.  A new kind of “dashboard” software that integrates databases and graphics should be considered.  It shows change in indictors much like the dashboard of a car shows changes in speed, temperature, fuel, etc. to help the driver make decisions. SOFI could be the central gauge with key indicators around it, while other indictors that don’t quite fit but that are important could be in the other displays or gauges.    (090)

3.  The questionnaire that appears in Appendix B-3 in the CD-ROM should be repeated periodically.    (091)

4.   Some of the data are weak and new data sources should be explored to improve the coverage and accuracy of the SOFI.    (092)

5.   Better means should be explored for forecasting the variables, including perturbing extrapolations with future developments and cross-impacts among the variables. In addition, for at least some of the variables, agent modeling and multi-equation feedback models should be considered.    (093)

6.   An attempt should be made to derive the weights statistically by correlating public opinion about the future, or some objective measures, with the set of indicators designated for the SOFI.    (094)

7.   Some Global Challenges in Chapter 1 are under-represented, such as:    (095)

• Challenge 5. How can policymaking be made more sensitive to global long-term perspectives?    (096)
• Challenge 9. How can the capacity to decide be improved as the nature of work and institutions changes?    (097)
• Challenge 12. How can transnational organized crime be stopped from becoming more powerful and global sophisticated enterprises?    (098)
• Challenge 15. How can ethical considerations become more routinely incorporated into global decisions?    (099)

Reader feedback is sought for measurable variables to capture these better.    (0100)

8.   This work can help illuminate the different aspirations and beliefs about what is important about the future. SOFI could help focus such discussions in conferences, policy debates, and educational sessions.    (0101)

9.   While the work thus far dealt with a global SOFI, it is possible and important to construct similar indexes for countries, regions, or cities. If the parameters selected for the future index for these areas are the same as those used for a global measure, then direct comparisons could be made.  For example, are things improving for our region as much as for the world as a whole? What set of improvements would change our outlook to be more in line with global prospects?    (0102)

This may be the beginning of an interesting and important avenue of futures research and may stimulate thinking about what constitutes—and how to measure—a good future.    (0103)

2. Some Existing Indicator Activities    (0104)

There are many indices that aggregate economic or social indicators (also called “variables” or “index  components”) to obtain a summary measures that “add up” to and elucidates a topic or issue. For example the Consumer Price Index is an aggregation of representative prices of products and services bought by consumers and that, tracked over time, shows how the overall levels of prices have grown or diminished; that is the inflation rate.     (0105)

Several such indices are summarized in Appendix A and are described briefly here by way of illustration.     (0106)

The UNDP has created a Human Development Index (HDI), comprised of life expectancy, literacy, education and GDP per capita. UNDP describes the HDI as follows:    (0107)

The first Human Development Report (1990) introduced a new way of measuring development — by combining indicators of life expectancy, educational attainment and income into a composite human development index, the HDI. The breakthrough for the HDI was to find a common measuring rod for the socio-economic distance traveled. The HDI sets a minimum and a maximum for each dimension and then shows where each country stands in relation to these scales—expressed as a value between 0 and 1. Since the minimum adult literacy rate is 0% and the maximum is 100%, the literacy component of knowledge for a country where the literacy rate is 75% would be 0.75. Similarly, the minimum for life expectancy is 25 years and the maximum 85 years, so the longevity component for a country where life expectancy is 55 years would be 0.5. For income the minimum is $100 (PPP) and the maximum is $40,000 (PPP). Income above the average world income is adjusted using a progressively higher discount rate. The scores for the three dimensions are then averaged in an overall index. [1]    (0108)

The chart below shows the UNDP HDI components, based on 1995 data.    (0109)

The UNDP has also developed a Human Poverty Index (HPI), and a Gender-related Development Index (GDI).    (0149)

The Environmental Sustainability Index (ESI) [3] is another extensive project is being performed by the Center for International Earth Science Information Network (CIESIN), Columbia University, and Yale Center for Environmental Law and Policy for the World Economic Forum which holds global conferences in Davos. This is “a measure of overall progress towards environmental sustainability.”  This group focused on four aspects of sustainability:    (0150)

Environmental Systems    (0151)

A country is environmentally sustainable to the extent that its vital environmental systems are maintained at healthy levels, and to the extent to which levels are improving rather than deteriorating.    (0152)

Reducing Environmental Stresses    (0153)

A country is environmentally sustainable if the levels of anthropogenic stress are low enough to engender no demonstrable harm to its environmental systems.    (0154)

Reducing Human Vulnerability    (0155)

A country is environmentally sustainable to the extent that people and social systems are not vulnerable (in the way of basic needs such as health and nutrition) to environmental disturbances; becoming less vulnerable is a sign that a society is on a track to greater sustainability.    (0156)

Social and Institutional Capacity    (0157)

A country is environmentally sustainable to the extent that it has in place institutions and underlying social patterns of skills, attitudes and networks that foster effective responses to environmental challenges. [4]    (0158)

Within this framework, a set of 67 variables were chosen to make cross national comparisons and rankings for 122 countries. The researchers say that the ESI (and we observe, all such cross sectional indices) enables the following uses:    (0159)

• identification of issues where national environmental results are above or below expectations;    (0160)
• policy tracking to identify areas of success or failure;    (0161)
• benchmarking of environmental performance;    (0162)
• identification of “best practices”; and    (0163)
• investigation into interactions between environmental and economic performance.    (0164)

The UN Commission on Sustainable Development (UNCSD) also conducts extensive work on indicators of sustainability. [5]  The primary goal of this activity is to develop and track indicators that relate to the accomplishment of the goals of Agenda 21 and, in particular “ to monitor and report on implementation of the agreements at the local, national, regional and international levels.”  They track over 130 indicators in the categories of social, economic, natural resource, and institutional. Appendix A includes some 40 of the indicators. The data published by UNCSD is cross sectional and is not aggregated to form an index.    (0165)

The United States has established an interagency working group that conducts a program of monitoring and analyzing indicators that relate to sustainable development.      (0166)

In June 1993, President Clinton established the President’s Council on Sustainable Development (PCSD) with a mandate to develop recommendations on steps the United States could take to realize sustainable development. The Council presented its initial findings to the President in March 1996 in the document Sustainable America: a New Consensus for Prosperity, Opportunity, and a Healthy Environment for the Future. In this report, the Council noted the importance of monitoring the Nation’s progress toward national sustainability goals. It recommended that the Federal Government intensify its efforts to develop national indicators of progress toward sustainable development in collaboration with nongovernmental organizations and the private sector. In response to this recommendation, the Administration established the U.S. Interagency Working Group on Sustainable Development Indicators (SDI Group) in 1996. [6]    (0167)

Appendix B contains more complete descriptions of these and other indexes.    (0168)

• Human Development Index (UNDP)    (0169)
• OECD Development Indicators (OECD)    (0170)
• Calvert-Henderson Quality of Life Indicators    (0171)
• Bribe Payers Index (Transparency)    (0172)
• Corruption Perception Index (Transparency)    (0173)
• World Times Wealth of Nations Triangle Index (Worldpaper)    (0174)
• Information Society Index (Worldpaper)    (0175)
• Environmental Sustainability Index. (Center for International Earth Science Information Network (CIESIN)    (0176)
• Goal 21 Indicators (The UN Commission on Sustainable Development (UNCSD))    (0177)
• US Sustainability Indicators (U.S. Interagency Working Group on Sustainable Development Indicators (SDI Group))    (0178)

These studies used several hundred variables in total to describe various societal and economic conditions.  A few variables appeared in most of the studies dealing with societal change; these can form a checklist for SOFI.    (0179)

Adult literacy    (0180)

Available resources (arable land/ forest area)    (0181)

Climate measures (temperature/ CO₂ emissions)    (0182)

Consumption measures    (0183)

Educational enrollment or attainment         (0184)

Employment and/or unemployment           (0185)

Energy efficiency    (0186)

Food availability and/or nutrition               (0187)

GDP per capita    (0188)

Health resources    (0189)

Infant mortality rate    (0190)

Life expectancy    (0191)

Measures of crime and corruption    (0192)

Number of armed conflicts    (0193)

Population growth and/or fertility rate    (0194)

R&D expenditures    (0195)

Rich/ poor poverty gap    (0196)

Terrorist incidents    (0197)

Water availability    (0198)

3. Concept of the State of the Future Index    (0199)

The SOFI differs from other indexes in several important respects. All indexes that have been reviewed, are concerned with the present or past, the SOFI is designed to measure the promise of the future. Most existing indexes are cross sectional and are designed to compare countries to countries or various groups of countries at some point in time (usually the most recent possible). SOFI is longitudinal and is designed to track and project change over time.    (0200)

In addition, SOFI is unique in that it has been derived from suggestions of the Millennium Project Global Lookout Panel in 1999-2000 who recommended and rated indicators to measure progress or regress on the 15 global challenges. Then in 2001 a special Global Panel rated indicators for the SOFI in terms of their normative and dystopic levels and priorities. 57 participants from about 15 countries participated in this SOFI panel. The process involved collecting judgments about the variables that might be included in the index, weighting of the variables, and perceptions about the best and worst possibilities for each variable, all of the judgments involved in calculating the index. Further, the process involved feedback that allowed respondents to add to and reassess the judgments which they and others provided in earlier global outlook panels. The questionnaire used in this work appears in Appendix B.    (0201)

A SOFI can be computed for individual countries or groups and used to compare their apparent future outlook to each other as well as to the world as a whole. This application requires the use of country rather than global data, and the use of weights that are appropriate to the country or region. Thus it is quite conceivable that two political groups in a single country, working with the same data set, could produce quite different SOFI’s by weighting the variables according to their views of the importance of the variables and by their views of the best and worst outlook for each variable. Political differences can be quantified in this way.    (0202)

     (0203)

4. Five Important Questions in Designing the SOFI    (0204)

1.   What variables should be included in a State of the Future Index?  If people say that the future seems promising, what do they mean? That life will be good for themselves and their family; that food, water and shelter will be sufficient; that fear will be absent and life fulfilled. What else should be included? The selection of variables forces a person to answer two key questions: What do I consider an improvement? And how would I know it if it happened?    (0205)

2.   How can very different variables be combined? It is necessary to make all the measures included in the SOFI commensurate—that is, expressed in terms that are comparable.    (0206)

3.   How can the variables be forecast? Measurement is not enough; since we are dealing with the future, the elements of the SOFI must be forecast. How can this be done?    (0207)

4.   How can the variables be weighted? The SOFI elements are not all of equal importance to the future; the SOFI uses the concept of nonlinear weighting in order to balance the significance of the measures that are included. But weighting leads to other problems: different people may see one or the other of the measures as being more or less important, or even of different polarity—that is, some may see an increase in a variable as good while others see it as bad. Since the SOFI is designed to be a globally aggregated measure, it can mask differences among groups or nations: the SOFI could look very positive and yet for some groups or nations, the situation could be worsening. Therefore it is important to recognize that disaggregated SOFI analyses will be essential so that groups or nations can determine—using their own data and weights—how things seem to be changing.    (0208)

5.   How can double accounting be avoided? This has to be considered or else one area could be over-represented. For example, should SOFI include both a measure of carbon dioxide concentration and global temperature? They measure different things but are important to consider for the SOFI for the same reason.    (0209)

These five issues will be discussed in turn.    (0210)

4.1 Elements of the Index    (0211)

Identifying what makes one possible future better than another lies at the soul of inspired policy-making. In asking how to measure whether the future is improving, one is reminded of the hierarchy developed by the psychologist Abraham Maslow who proposed that basic needs of an individual such as food and sex must be satisfied before higher goals can gain priority. The highest level in his hierarchy is self-actualization—the fulfillment of one's human potential. A person taking the Maslovian view would select variables that measure progress up the ladder toward self-actualization.    (0212)

There are other historic prescriptions for measuring what’s important to a better future. Sixty years ago Franklin Roosevelt stated the four freedoms as set of guiding principles that could easily have formed the basis for an index [7]:    (0213)

In the future days that we seek to make secure, we look forward to a world founded upon four essential human freedoms.
    (0214)

The first is freedom of speech and expression –everywhere in the world.
    (0215)

The second is freedom of every person to worship God in his own way‑ everywhere in the world.
    (0216)

The third is freedom from want, which, translated into world terms, means economic understandings which will secure to every nation a healthy peacetime life for its inhabitants --everywhere in the world.
    (0217)

The fourth is freedom from fear, which, translated into world terms, means a world-wide reduction of armaments to such a point and in such a thorough fashion that no nation will be in a position to commit an act of physical aggression against any neighbor‑anywhere in the world.
    (0218)

Another prescient is found in the Preamble to the UN Charter [8]:    (0219)

We the peoples of the United Nations determined:    (0220)

to save succeeding generations from the scourge of war…    (0221)

to reaffirm faith in fundamental human rights, in the dignity and worth of the human person, in the equal rights of men and women and of nations large and small, and    (0222)

to establish conditions under which justice and respect for the obligations arising from treaties and other sources of international law can be maintained, and    (0223)

to promote social progress and better standards of life in larger freedom,    (0224)

And for these ends    (0225)

to practice tolerance and live together in peace with one another as good neighbors, and    (0226)

to unite our strength to maintain international peace and security, and    (0227)

to ensure, by the acceptance of principles and the institution of methods, that armed force shall not be used, save in the common interest, and    (0228)

to employ international machinery for the promotion of the economic and social advancement of all peoples,    (0229)

The principles stated in Agenda 21 can also provide guidance to recent thinking about what constitutes a better future [9].    (0230)

Humanity stands at a defining moment in history. We are confronted with a perpetuation of disparities between and within nations, a worsening of poverty, hunger, ill health and literacy, and the continuing deterioration of the ecosystems on which we depend for our well-being. However, integration of environment and development concerns and greater attention to them will lead to the fulfillment of basic needs, improved living standards for all, better protected and managed ecosystems and a safer, more prosperous future. No nation can achieve this on its own; but together we can - in a global partnership for sustainable development.    (0231)

And quoting from the Rio Declaration on Environment and Development those items that bare on the selection of variables that depict an improving or worsening future: [10]    (0232)

• Human beings are… entitled to a healthy and productive life in harmony with nature.    (0233)
• The right to development must be fulfilled so as to equitably meet developmental and environmental needs of present and future generations.    (0234)
• In order to achieve sustainable development, environmental protection shall constitute an integral part of the development process and cannot be considered in isolation from it.    (0235)
• All States and all people shall cooperate in the essential task of eradicating poverty …in order to decrease the disparities in standards of living and better meet the needs of the majority of the people of the world.    (0236)
• States shall cooperate to conserve, protect and restore the health and integrity of the Earth's ecosystem.    (0237)
• States should reduce and eliminate unsustainable patterns of production and consumption and promote appropriate demographic policies.    (0238)
• States should cooperate to …[improve] scientific understanding through exchanges of scientific and technological knowledge.    (0239)
• At the national level, each individual shall have appropriate access to information concerning the environment that is held by public authorities… and the opportunity to participate in decision-making processes.    (0240)
• States should cooperate to promote a supportive and open international economic system that would lead to economic growth and sustainable development in all countries.    (0241)
• Women have a vital role in environmental management and development. Their full participation is therefore essential to achieve sustainable development.    (0242)
• Warfare is inherently destructive of sustainable development. States shall therefore respect international law providing protection for the environment in times of armed conflict and cooperate in its further development, as necessary.    (0243)
• Peace, development and environmental protection are interdependent and indivisible.    (0244)

Finally, the Millennium Project itself can provide the primary starting point in the selection of indicators of a changing future. Over the last seven years, the Millennium Project has concentrated on identifying global change and potential actions that might be taken to improve the future. This work has involved between 1996-2000 more than 700 people who identified and rated almost 300 developments. The developments were distilled into sets of issues and opportunities with a range of views from policy makers about actions to address each. Finally the issues, opportunities and actions were distilled into a set of 15 challenges.    (0245)

Favorable resolution of the 15 challenges would clearly improve the global future. In defining these challenges, “importance” was defined as consisting of these attributes: number of people affected, level of the affect, the need for immediate attention, and the level of reversibility of action or consequences of inaction.    (0246)

The 15 global challenges are:    (0247)

1.   How can sustainable development be achieved for all?    (0248)

2.   How can everyone have sufficient clean water without conflict?    (0249)

3.   How can population growth and resources be brought into balance?    (0250)

4.   How can genuine democracy emerge from authoritarian regimes?    (0251)

5.   How can policymaking be made more sensitive to global long-term perspectives?    (0252)

6.   How can the globalization and convergence of information and communications technologies work for everyone?    (0253)

7.   How can ethical market economies be encouraged to help reduce the gap between rich and poor?    (0254)

8.   How can the threat of new and reemerging diseases and immune microorganisms be reduced?    (0255)

9.   How can the capacity to decide be improved as the nature of work and institutions change?    (0256)

10. How can shared values and new security strategies reduce ethnic conflicts, terrorism, and the use of weapons of mass destruction?    (0257)

11. How can the changing status of woman help improve the human condition?    (0258)

12. How can organized crime be stopped from becoming more powerful and sophisticated global enterprises?    (0259)

13. How can growing energy demand be met safely and efficiently?    (0260)

• How can scientific and technological breakthroughs be accelerated to improve the human condition?    (0261)
• How can ethical considerations become more routinely incorporated into global decisions?    (0262)

Chapter 1 on this CD-ROM discusses in details each of the challenges.    (0263)

The process by which the SOFI indicators were selected involved the following steps:    (0264)

1. In 1999-2000, the panel was asked to identify indictors by which the status of these 15 challenges could be measured. The results, together with other work of the Millennium Project, were published in State of the Future at the Millennium in 2000. [11]  These nominated indicators were evaluated by the 1999-2000 Lookout Panel in terms of their availability and usefulness. Some 90 indicators were judged by the panel to be important enough for further consideration.    (0265)

2. The list of 90 variables was reviewed. Specific measures and sources were identified to match the suggestions of the panel where ever possible. Duplications were removed as were those suggestions that appeared to be overly precise, or ambiguous.    (0266)

3. The set that emerged which was scrutinized against other compendia of variables prepared by OECD, UNCSD, UNDP, CIESIN, and others. This comparison permitted “tuning” the final set for SOFI to assure that essentially all important aspects were covered. This produced a list of 43 variables presented in the table below.    (0267)

4. In 2001, a follow-up questionnaire was constructed to collect additional judgments for the SOFI (Appendix B).. The list of variables from Step 3 above was the starting point for this inquiry; respondents were asked to amend or to add to the list of variables if they felt other variables were better suited than those presented. Respondents were asked to conduct this review on the basis of the following criteria: ideally the indicators had to:    (0268)

• be descriptive of the challenges    (0269)
• be available at a global level annually over the past 20 years    (0270)
• be useful in indicating an improving or worsening future    (0271)

In addition, respondents were asked to provide judgments about the values of the variables under best and worst scenario assumptions and the weight that each variable should be accorded in an index under the best and worst assumptions. The criteria for assigning a high weight to a variable were: the number of people affected; the significance of the effect; whether or not some groups seemed to be affected differentially; the time over which the effect will be felt; and whether or not the effect is reversible. Thus, a respondent  assigning a high weight to a variable implied that changes in the course of the variable would affect almost everyone (or some large groups differentially), deeply, for many decades, and it would be very difficult to change the situation once it occurred.  The respondents used the following scale for weights:    (0272)

10 =  Essential to include in any index that is designed to depict the expected state of the future    (0273)

 8  =  Extremely important    (0274)

 5  =  Enough people or groups are affected by changes in this variable to consider it seriously    (0275)

 3  =  Relatively unimportant.    (0276)

 1  =  Do NOT include it in SOFI    (0277)

The following table lists the 43 variables produced in Step 3 that were evaluated in the questionnaire; the table also presents definitions of the variables. [12]    (0278)

A temperature index was formed by combining the meteorological station measurements over land with sea surface temperatures obtained primarily from satellite measurements (Reynolds and Smith, 1994; Smith, Reynolds, Livesay and Stokes, 1996). Best estimate for absolute global mean for 1951-1980 is  14C = 57.2F, that value was added to the temperature change to obtain an absolute scale (Goddard Institute for Space Studies) 
    (0300)

The applicability of these 43 variables to the 15 challenges and their “coverage” of the originally suggested indicators is shown in Appendix D.     (0457)

The questionnaire requested judgments about the value of the variables in a best and dystopic world ten years in the future, as well as weights to be assigned these variables in an index. To aid the respondents, average values for the variables during the period 1980-2000 were presented (shown shaded). The table below shows the average judgments of the panel and the number of responses. Many responses were received after this calculation; they will be included in the next publication of the research.    (0458)

It is important to point out that in almost all instances, the dystopic weight was judged to be higher than the normative weight suggesting that when things get bad, they are more important than when they are good, as shown on the following chart:    (0812)

     (0813)

We see also that the number of responses we received on each variable tended to be a function of the perceived importance, In the chart below, the number responding is the average number of responses to the four questions asked about each variable: normative value, dystopic value, normative weight, and dystopic weight.    (0814)

   (0815)

From the standpoint of dystopic weighting, the variables scoring above 7.0 in terms of importance to SOFI were:    (0816)

The questionnaire also asked:  “If you were restricted to only 5 variables, which would you include?”  The variables selected by 6 or more respondents (about 1/3 of the respondents) were:    (0896)

As might be expected, most variables made both lists (the bold number indicate the items that made one or the other lists):    (0947)

These variables are not independent; most relate to one another. For example, increasing population growth can result in later unemployment (given constant or increasing productivity and output); lowering food availability and increase infant mortality. Even recognizing the potential for non-independence, there are some variables in the list that may measure the same thing from two perspectives, This is the double accounting possibility (See section 4.5 for a more detailed discussion).  We recognize that within the two set nominated by the best five list and the highest dystopic weights, there are some potential double accounting redundancies:    (01039)

First possible redundancy    (01040)

Food availability Cal/cp Low Income Countries    (01041)

World Grain Production    (01042)

Second possible redundancy    (01043)

CO₂ emissions, industrial (mil kt)    (01044)

Average annual global temperature (Centigrade)    (01045)

Third possible redundancy    (01046)

Annual new HIV cases (millions)    (01047)

Annual AIDS deaths (millions)    (01048)

The following graphs show the cross plots of these variables    (01049)

   (01050)

     (01051)

   (01052)

The relationship between grain production and calories per capita is logical and striking. The relationship between AIDs deaths and new HIV cases is also logical since HIV leads to AIDs, but in the future the relationship may change since there are now drug regimes that greatly extend lifespan for HIV patients who can afford the drugs. With much more scatter, the relationship between CO₂ and atmospheric temperature may also be discernable.    (01053)

So as the variables are chosen to include in SOFI, what should be done with these redundancies? It is more efficient, of course to reduce the number of variables, but not to the point where information is lost. For our purposes here, there were selected:    (01054)

Food availability Cal/cp Low Income Countries    (01055)

CO₂ emissions, industrial (mil kt)    (01056)

Annual AIDS deaths (millions)    (01057)

as being the more direct measures of the phenomena that led to their nomination.    (01058)

Thus the SOFI list contains 19 variables and appears as follows:    (01059)

Finally, the respondents were given a chance to suggest other variables “that, in your judgment, are at the importance level of 8 or above (in either the best or worst plausible worlds) and that you believe should be added to the State of the Future Index.” The suggestions appear in Appendix B and should be considered again in the next iteration of SOFI analysis.    (01099)

4.2 Forecasting    (01100)

Appendix C shows the data collected for the original set of 43 variables. These data have been drawn from numerous sources and cover the period from 1980 to 2000. The data series, in general, are global in scale. Where the data series were not continuous, interpolations were made (as shown by italics); when global data were not obtainable, selected regional or national data series were used to compute the global estimates. Over time, we expect that the data set can be improved by obtaining annual, global data for all series.    (01101)

Since the SOFI is to show future values as well as historical, it is necessary to forecast each of the series. The horizon selected was 10 years, a period half as long into the future as the historical database.    (01102)

The simplest form of forecasting involves fitting the historical data with pre-determined curve shapes; the equations used to fit each series were [14]:    (01103)

1. Linear                       v= m*t + b    (01104)

2. Exponential               ln(v)=m*t + b    (01105)

3. Power function          ln(v)=m*ln(t) + b    (01106)

4. Logarithmic               v= m*ln(t) + b    (01107)

5. Inverse v                   1/v=m*t + b    (01108)

6. Inverse t                    v=m/t + b    (01109)

7. S Shaped                  ln{(v/L)/[1-(v/L)]}= m*t + b    (01110)

where v is the value of the variable, L is the variable’s upper limit when one can be discerned, m is the slope of the fitted curve, and b is the intercept at t=0    (01111)

Data for each of the variables was fitted to this set of curves in order to complete missing historical points and to extrapolate to the year 2010. The curve selected for each variable was based on the statistical “goodness of fit” R^2 criterion, consideration of the shape of the extrapolation, any physical limits that might constrain the curve as well. The table below summarizes the fit information for the set and the number of initial data points.    (01112)

The following figures show the SOFI variables, the forecasts, the normative and dystopic judgments of the respondents and the weights given these values.    (01335)

SOFI Variables

   (01336)

   (01337)

   (01338)

   (01339)

4.3 Scaling    (01340)

Suppose a composite indicator were to include time series measures of population additions (number of people) and economic vitality (GDP per capita). Clearly the numbers representing these two variables could not simply be added; it would be necessary first to express their values in the same dimensions or in non-dimensional form. Making them non-dimensional is called “normalizing” and this could be accomplished in several different ways. One could normalize by:    (01341)

• dividing each number in the series by the largest number appearing in the series    (01342)
• dividing each number by the largest number which could ever reasonably be expected    (01343)
• assigning the lowest number in the series the value of zero and the highest number, 100    (01344)
• assigning the value of 100 for the most desirable and 0 for the least desirable values, neither of which may have yet occurred.    (01345)

We used the fourth approach since it provides the basis for inserting into an otherwise deterministic process, an element of subjectivity about what constitutes a desirable and undesirable future, albeit in quantitative terms. To illustrate the approach which was employed suppose the original data were:    (01346)

Using the fourth approach requires that “best and worst” futures be defined and applied, (as UNDP has done in forming their HDI). In this approach, one is forced to confront the questions: just how good or bad could things be?  For example, in this case just how low could GDP/capita go? In 1983 it was $4,311 and it increased to $5,327 in 1999. Might it return to the previous low values in the future, or exceed the 1999 value? Of course. So, imagine a best and worst case for the next decade. The low estimate might be $3,994 (a return to below the 1980 value) and a plausible high estimate might be $6,068 (an appreciable growth over 1999 [15]). These would be the standards used in the following equation:    (01411)

INDEX (t)= (Ind1 (t) - lowest value of Ind1) *100/(highest value of Ind1 - lowest value of Ind1)    (01412)

where Ind1(t) is the value of variable 1 at time t.    (01413)

Assuming a best case value for population added of 61 million and a worst case of 84 million (in 1989 the population added was 87.4 million) results in the following:    (01414)

As mentioned above, the global panel was asked to provide these estimates and the average of the values they provided were listed in an earlier table.    (01479)

4.4 Weighting    (01480)

After the indicators have been scaled as discussed above, they must be combined. In any index, not all of the indicators are equally important. Therefore it is appropriate to combine them as a weighted sum. The weights can be estimated by individuals or groups, or given appropriate historical data, through regression. If the form of the index is:    (01481)

INDEX (t) = k1*ind1 (t) +k2*ind2 (t) +k3*ind3 (t)……..    (01482)

where the k values are the weights and the ind(n)(t)  are the normalized values of the indicators. When sufficient historical data about the index are available from another source such as polling about future perceptions about the future, then the k’s can be estimated statistically.    (01483)

All indexes of which we are aware have assumed that weights are constant and independent of the value of the variable they modulate. We believe that in the real world, the weights assigned to some indicators should change as the values of the indicators rise and fall. When, for example, an indicator reaches a level of satiation, it is no longer as important as it used to be. Take food intake calories per person: when the level is below 1500 calories, the variable is very important. When it is above 3,000, the sense of urgency associated with hunger no longer gives this variable much weight.     (01484)

There are several relationships between variables and the weights that are assigned to them; these relationships are shown below:    (01485)

Linear Weighting    (01486)

Here the value of the weight changes linearly with the value of the variable. For example:    (01487)

   (01488)

Weight = (dW/dV) * Var  +  W(0)    (01489)

Where dW/dV is the change in the weight with a unit change in the variable and W(0) is the value of the weight at a zero value of the variable. For a constant weight, dW/dV = 0    (01490)

This sort of relationship would be used when the more one has, the less valuable it becomes (dW/dV is negative), or conversely, the more one has, the more one values it (dW/dV is positive). The former situation might be exemplified by a farmer who needs a truck. One truck can make the difference between the success or failure of his farm and therefore weighs this heavily. Then as the farmer becomes more affluent and acquires many vehicles, the addition of one more truck becomes less and less important.    (01491)

The later situation might be exemplified by an animal psychology experiment in which a laboratory mouse is placed in a tube with a reward of cheese at the end. When the mouse is at the entry end of the tube, it exerts a small forward pressure. But as it approaches the cheese, the force grows.     (01492)

S-Shaped Weighting    (01493)

Perhaps a more general way to express the relationship is in the form of an s- shaped curve. For a positive change, as the variable increases, its weight increases.    (01494)

   (01495)

This graph was plotted using the equation:    (01496)

ln{(W/C)/[1-(W/C)]}= dW/dV*V + W(0)    (01497)

Where as before, dW/dV is the change in the weight with a unit change in the variable and determines the slope at the inflection point. W(0) is a value that shapes the initial portion of the curve and C is the upper limit of the weight. In this illustration, dW/dV is 7.0 ,  W(0) is –3.5, and C is 1.0.    (01498)

It is more generally the case that weight diminishes as the desired target is approached. In this case, the value of the weight is higher when the variable has a lower value;  dW/dV is –7.0,  W(0) is 3.5, and C remains 1.0, and the graph is:    (01499)

   (01500)

The chart below shows the effect of changing the value of b while other parameters are held constant.    (01501)

   (01502)

Applying these ideas to the SOFI, the global panel was asked (as noted earlier)  to provide an estimate of the weight of each variable under two assumed scenarios: one for the worst case projected for the variable and the second for the best case. The average weights are repeated table below and the value of dw/dv used to fit the s-shaped weight curve is listed. For all variables, the value of b was taken as 8.00 that produced an intuitively satisfying functional shape.    (01503)

To illustrate how the estimates made by the panel were used to create a non-linear weighting function, the curve below shows the weight for  “Food availability Cal/cp Low Income Countries.”  Assume that the panel’s average weight for the best world (abundant food) was 5.50 and for the worst world (food shortage), 8.39. The s-shaped equation which to fit these values was:    (01726)

Wt=  (Weight in worst scenario) x [(exp(dw/dv) * V + b)/ (1 + exp(dw/dv) * V +b)]    (01727)

where (dw/dv) was taken as –7.35 and b as 8. These choices produced a weight of 8.39 when the value of the variable is a scaled zero, and 5.5 when the scaled value of the variable is one. If the value of the scaled variable were to exceed 1.0, the weight of the variable would drop below 5.5, reaching close to zero at a scaled value of 1.5. The following curve shows the function:    (01728)

   (01729)

In this chart, a variable value (abscissa) of zero represented the worst scenario and 1.0 represented the best.     (01730)

4.5 Double Accounting    (01731)

A statistical correlation matrix was constructed to find which of the selected variables were highly correlated with each other. Limiting the search to cases in which the variables had more than 10 points in common and showed a correlation level above 0.9, it was found that:    (01732)

GDP per capita (constant 1995 $US) correlated with:    (01733)

• Infant Mortality Rate (deaths per 1,000 live births) (-.927)    (01734)
• Food availability Cal/cp Low Income Countries (.945)    (01735)
• Life Expectancy (World) (.922)    (01736)
• School Enrollment, secondary (% school age)         (.934)    (01737)
• Literacy rate, adult total (% of people aged 15 and above) (.974)    (01738)

Infant Mortality Rate (deaths per 1,000 live births) correlated with:    (01739)

• CO₂ emissions, industrial (mil kt) (-.938)    (01740)
• Life Expectancy (World) (-.996)    (01741)
• School Enrollment, secondary (% school age) (-.926)    (01742)
• Literacy rate, adult total (% of people aged 15 and above) (-.959)    (01743)

Food availability Cal/cp Low Income Countries correlated with:    (01744)

• Literacy rate, adult total (% of people aged 15 and above) (.957)    (01745)

CO₂ emissions, industrial (million kilotons) correlated with:    (01746)

• Literacy rate, adult total (% of people aged 15 and above) (.920)    (01747)
• Life Expectancy (World) (.929)    (01748)

Life Expectancy (World) correlated with:    (01749)

• School Enrollment, secondary (% school age) (.908)    (01750)
• Literacy rate, adult total (% of people aged 15 and above) (.953)    (01751)

School Enrollment, secondary (% school age) correlates with:    (01752)

• Literacy rate, adult total (% of people aged 15 and above) (.949)    (01753)

To illustrate the correlation, the literacy correlations are cross-plotted below:    (01754)

   (01755)

   (01756)

Almost all of these correlations seem logical enough; they relate to the effects of affluence and that in turn on school enrollments and literacy. An exception is the correlation between increasing CO₂ emissions and increasing life expectancy, shown below. This is an example of a spurious correlation: both are increasing but for reasons that are not connected. Both CO₂ and life expectancy increase with increasing industrial economic wealth; however, increasing knowledge economy wealth is expected to show a decrease in CO₂ emission.    (01757)

   (01758)

In modeling of social systems an effort is usually made to trace causality rather than simply report statistical correlations which may or may not be spurious. Were the work on SOFI to continue in the future it would be useful to explore the use multi-equation models involving causal analyses, and other advanced modeling techniques that could be used not only to forecast the variables, but to reduce list of variables that truly measure different phenomena and therefore constitute a minimum set.     (01759)

For our purposes, the list of SOFI variables remains as stated earlier.    (01760)

5. An Example of a Global SOFI Analysis    (01761)

Appendix B provides the database, that is, the actual values retrieved from the data sources, interpolations where data were missing, and extrapolations through 2010 using the curves which best fit the historical data. The extrapolations and interpolations are shown in italics. These data were normalized and weighted using a s-shaped function; the judgmental factors required in this analysis were provided by the responses to the global questionnaire. To compute the index, the computed values for each year were divided by the value of the index in 2000. This procedure produced the following graph of the SOFI:    (01762)

   (01763)

By and large the variables that make up the SOFI are relatively stable, but in a number of cases (such as the number of conflicts, and terrorist attacks) the volatility of the historical data and the uncertainty of the forces which cause changes in the variables led to great uncertainty about the forecasts.    (01764)

As mentioned earlier there are several variables that are apparently important but have not been included in the SOFI because it is not clear whether increases in value are, on the whole, good or bad.    (01765)

For example:    (01766)

   (01767)

   (01768)

     (01769)

Now, given the index, it is possible to answer questions such as:    (01770)

Why did the SOFI grow from 1980 to 2000?    (01771)

Simply because many of the variables included in the SOFI improved, for example some of the most important favorable changes occurred for the following variables:    (01772)

• GDP per capita grew    (01773)
• Calories per capita increased    (01774)
• Life expectancy grew    (01775)
• Terrorist Attacks diminished    (01776)
• Literacy grew    (01777)
• Infant Mortality dropped    (01778)
• Access to Fresh Water improved    (01779)
• Access to Health Care improved    (01780)
• School Enrollment Increased    (01781)

But some variables moved in unfavorable directions:    (01782)

• Industrial CO₂ emissions grew    (01783)
• Unemployment moved in a generally unfavorable direction    (01784)
• Forest Lands decreased    (01785)
• Rich Poor Gap got wider                                                                             (01786)
• AIDS Deaths grew    (01787)
• Developing Country Debt increased    (01788)

With the weights that were assigned and the levels of changes experienced against the background of best possible and worst possible scenarios, on balance the favorable changes outweighed the unfavorable in the index.    (01789)

What caused the accelerated growth in SOFI from 1987 to 1990?    (01790)

This is an example of how the changes in SOFI can be traced back to find what caused a change. It appears that several variables were primarily responsible. Here are selected data for the time period:    (01791)

These were extraordinary times: GDP/capita rose by about 5% in constant terms, life expectancy increased 6.5%, access to fresh water by 11.8%, and access to health care by 5.8%.    (01816)

If the data had been mundane and unchanging, as shown in this table:    (01817)

the SOFI would have appeared as:    (01842)

   (01843)

What can be done to increase the growth of SOFI in the next ten years?    (01844)

Modifying four variables in the manner shown below would have a very stimulating affect. The original data are:    (01845)

If policies could modify these variables by allowing neither rich-poor gap and developing country debt to worsen, and by improving levels of unemployment as follows:    (01881)

Then SOFI would be:    (01917)

   (01918)

Certainly other variables could be modified to change the slope of the SOFI in the next ten years, but these illustrate the sensitivity of the curve to changes in the variables.    (01919)

Finally, the history of each variable contains information which can be used to gauge uncertainty intrinsic to the SOFI forecast, namely, the errors that existed between the “best fit” curve and the actual data points. These “residuals” indicate a measure of scatter and one can assume that the residuals of the sort that existed in the past will also surround the extrapolation. In our case, the largest positive and largest negative fit errors were computed for each variable. This value of this range was multiplied by a random number between 0 and 1 and added to the projected value. The solution was run 25 times to form a “fan” showing the range of the SOFI that might be expected, just based on the historical residuals. The figure below illustrates the fan obtained in this way.    (01920)

   (01921)

This chart shows the plausible range of the SOFI. The soaring scenario that carries the SOFI to a value of 1.2 in 2010 is a promising a world of goals reached; the scenarios that stagnate at 1.0 in 2010 shows that a future of achievement may still be an elusive chimera.    (01922)

6. Research Issues    (01923)

There are many factors that could cause the projected SOFI to be different than the projections of Section 5.  Unexpected events may occur that will change the extrapolations of the variables. For example a few of the events that could cause great differences are listed below (these developments were suggested or based on responses from previous Millennium Project Lookout Panels and are only an illustration of what is certainly a much longer list):    (01924)