Defining the Demographic Window

Introduced by UN demographers during the Population Division’s release of a series of long-range projections (UNPD 2004, 2, 70-73), the original intent of the demographic window was to call attention to the segment of the age-structural transition that tends to measurably favor economic development. In its original formulation, UN demographers assumed that this window occurs when both children (0 to 14 years) and seniors (65 and older)—age groups that are composed principally of dependents—are at a low ebb. However, unlike other measures of dependency, in the UNPD’s formulation, children and seniors are differentially weighted; each senior is assumed to represent twice the per-child burden on economic development. According to this method, the demographic window is deemed open when children comprise less than 30 percent of the total population, and simultaneously, seniors comprise less than 15 percent of the total population.

Figure. Total fertility rate versus median age for all independent states (except the GCC states) in 2015, showing states in sub-Saharan Africa (SSA), and Europe. To enter the demographic window (the intermediate phase of the age-structural transition) appears to require a TFR near 2.8 children or below. Two exceptions, Israel (IS) and Algeria (AL) have experienced a rise in TFR. The position of some other states are shown: Niger (NG), South Sudan (SO), Timor Leste (TL), Nigeria (NI), the USA (US), China (CH), South Korea (RK), and Japan (JP).

Despite its seemingly arbitrary rules and speculative boundaries, in practice the UNPD’s demographic window has done well identifying conditions during which states achieve the World Bank’s upper-middle income classification, a milestone in development. For example, among the World Bank’s 2018 list of upper-middle income states, all are either in the UNPD’s demographic window or have passed through it (World Bank, 2019). The remainder are either significant oil and/or mineral exporters or have a population under 5.0 million—a group that includes numerous island states with exceptional per-capita tourist revenue and/or remittances.

Coupled with the UNPD’s demographic estimates and projections, the analytical usefulness of this simply defined window is obvious. To make the concept compatible with the conventions of age-structural modeling, Cincotta (2017) employed the UNPD (2004) metric to estimate the median ages of its assumed boundaries. From a sample comprising the 40 states that had entered that window since 1950 (excluding states with greater than 15 percent of GDP in oil or mineral rents). On this basis, the UNPD’s demographic window generally began at a median age between 26 and 27 years. Among most of the sample, the window ended at a median age between 39 and 42 years.

Using the lower median age boundary, Cincotta (2017) noted that states are generally unable to enter the demographic window without declining below a total fertility rate (TFR) of 2.8 children per woman (Fig. 4). Nonetheless, among states that have sustained rapid rates of fertility decline, populations have often declined well below a TFR of 2.8 before achieving a median age of 26 years. As these states aged—as relatively large cohorts of children and adolescents were replaced by smaller cohorts—they usually crossed the window’s nominal threshold within a decade.

This analysis noted that for most states, their trip through this demographic window of favorable age structures is likely to be a onetime opportunity—but with exceptions. The length of chronological time spent in the demographic window has occasionally been extended, at least temporarily, by: (a.) a baby boom, such as the post-war WWII rebound in fertility in the United States and in some Western European states; (b.) an influx of youthful immigrants; (c.) sustained rapid population growth among an ethnic minority with significantly higher-than-replacement fertility; or (d.) fertility settling near replacement levels, with very low child mortality.

Attachments

Discussion: Does Demographic Change Set the Pace of Development? (with Jane O’Sullivan, U. Queensland)

Original Essay: Does Demographic Change Set the Pace of Development? 

Comment:  Jane O’Sullivan, U. Queensland.

A good title but a disappointing treatment. In the context of the fertility transition, age structure and population growth rate are confounded. There is little evidence that age structure per se delivers these benefits – it is the slowing of population growth that affects “demand side”. Of course, age structure affects specifically which areas of demand are first affected, but that’s not what matters. The problem is that the “demographic dividend” discourse is used as a means to not have to talk about population growth, but in doing this, it undermines the motivation to get fertility down below replacement-level, which is the urgent need. Indeed, by fuelling the “aging crisis” myth, the DD message actively discourages efforts to complete the transition to low fertility. Also, the role of family planning programs is missing from the flow diagram. Education, income and child survival don’t have strong effects on fertility without explicit behaviour-change programs.    [Note: see Dr. O’Sullivan’s research on population impacts, demographic ageing, food security, and climate change issues]

 

Reply:  Richard Cincotta, Woodrow Wilson Center/Stimson Center

Thank you for your interesting comments, Dr. O’Sullivan. And, welcome to the growing ranks of the disappointed.

The curves featured in this essay were developed as part of the (U.S.) National Intelligence Council’s (NIC) long-range effort (the Global Trends series of publications) solely for the purpose of forecasting changes in political, social, and economic indicators from 2 to 20 years into the future.

For forecasting, the method has worked surprisingly well (you can read some of its forecasts in my other NSB posts). Analysts have used the method’s “eight rules of political demography” (also on NSB) to successfully forecast the rise of a liberal democracy in North Africa two years before the Arab Spring, to identify the remaining clusters of countries most at risk of intra-state conflicts, and to predict declines from liberal democracy among youthful countries.

Along the way, however, its findings have disappointed (or even angered) diplomats, advocates, and political scientists who have deeply-held views of how the world works.

One source of disappointment has to do with population growth. While states with a population under 5 million (particularly small island states) do, indeed, appear to develop politically and socioeconomically more quickly than expected, we have found no additional statistical evidence suggesting that larger population sizes or densities are—so far, at least—a “net impediment” to political and socio-economic development.

For example, population size and density may depress economic productivity by limiting per-capita freshwater supplies, exacerbating pollution, and forcing agriculture into marginally productive land. At the same time, however, population growth drives urbanization, which positively affects per-capita income and speeds other development transitions (including fertility decline).

My own experience with assessing population’s influence leads me to conclude that, in general, increasing human density disrupts and alters natural ecosystem processes (i.e., nature). Unfavorable age structures disrupt development processes (i.e., the state). States are, indeed, affected by ecosystem disruption, but our species has become very good at making and remaking its own highly productive ecosystems—most of which create additional long-term disruption (e.g., climate change, species loss, high nitrogen loading in soils, etc.–a point that you know well from your own research).

As for your critique of Fig. 3: I have indicated (in the diagram) that income and child survival’s effects on fertility are (as you suggest) weak or highly variable. However, I believe that most development analysts would disagree with your assessment of education’s impact on fertility—particularly the impact of women’s educational attainment, which in the past has been statistically strong. Nonetheless (consistent with your assertion), the effect of women’s educational attainment on fertility in tropical sub-Saharan African countries seems, so far, to be weaker than expected (when compared to the Asian and Latin American fertility transitions).

While you insist that getting fertility “down below replacement-level” is an urgent need, our analysis differs somewhat. To achieve a median age of 26 years—the beginning of the demographic window—fertility must decline below 2.8 children per woman. However, continued progress toward a median age of 30 years has generally led to additional fertility decline to near-replacement (close to TFR 2.1) or below-replacement levels.

On population aging: A myth? Perhaps you are referring to the mistaken beliefs/rhetoric of some political leaders of tropical African states who fear an aging population—seemingly unaware that the current UN Medium Fertility Variant scenario projects that such large proportions of elderly, for most tropical African states, will likely show up on the far side of 2100 (probably a century into the future). I believe that, to achieve a median age of 26 years, the leadership of those states must be prepared to dismantle the traditional and religious constraints on women’s lives (much like Habib Bourguiba did shortly following Tunisia’s independence), in addition to supporting quality family planning programs, girls’ education, and elevating women into positions of political power.

However, for the group of very low fertility European and East Asian countries that are rapidly advancing into the post-mature phase of the transition (median age 46+ years), the challenges of population aging are hardly mythical. Does population aging qualify as a crisis for economic and political liberalism? No one knows. However, the most credible research foresees substantial fiscal strains on retirement and healthcare systems (see Lee & Mason’s National Transfer Accounts review), and the admixture of population aging and immigration seems (to me, at least) to be yielding some unfavorable political byproducts.

Thanks again for your comments, Dr. O’Sullivan. // Richard Cincotta

The 4 Dividends: PRB’s Pace Project

See the Population Reference Bureau’s excellent video explaining the “Four Dividends” that countries generally attain following fertility decline as they pass through the demographic window. These four dividends are: (1) child survival, (2) educational attainment, (3) per-capita income, and (4) political stability (measured by 10-year risk of intra-state conflict).

Here are links to obtain the IUSSP Conference paper (authored by Elizabeth Madsen and me) that describes the timing of these changes, in terms of the movement of countries through the age-structural transition. A background paper on the Age-structural Theory of State Behavior is published in the Oxford Research Encyclopedia of Politics.  Some of this information is published in a short essay on the “Eight Rules of Political Demography“, on the New Security Beat.

 

Attachments

Article by Leonid Bershidsky in Bloomberg View

“Democracy in Iran? The demographics say YES” — but the Regime Type says NO

Bloomberg View (plus Bloomberg Business Week) has published Leonid Bershidsky’s excellent article on age-structural theory.  Clearly, Bershidsky has read through and grasped much of the research (thank you, Leonid!).  Bershidsky uses the theory to discuss recent anti-regime demonstrations in Iran and their outcome, and he neatly summarizes the theory’s predictions and points out its strengths. Notably, he also explores some of the theory’s weaknesses dealing with various types of authoritarian regimes that persist despite the societal changes that are associated with a more mature population and passage through the demographic window.

The article can be viewed at the Bloomberg View website, here.

Bershidsky is right — that aspect of the theory remains weak.  To strengthen it, I’ve been using the Authoritarian Regime Data Set (Hadenius, A., J. Teorell, and M. Wahman. 2012. “Authoritarian Regimes Data Set, version 5.0: Codebook.” Lund, Sweden: Department of Political Science, Lund University).  Putting Hadenius et al.’s regime types into “age-structural time” produces the following hypothetical relationships with population age structure (click on image to enlarge it).

 

Sub-Saharan Africa: Looking Toward the Demographic Window

Over the past 25 years, economic and political demographers have focused on documenting the improvements in state capacity and political stability that have been realized in the wake of fertility declines in much of East Asia, Latin America, and most recently in the Maghreb of North Africa (Tunisia, Morocco, Algeria). Nonetheless, foreign affairs, defense and intelligence analysts still seem confused over when and where this demographic dividend should occur—and whether the youthful, low-income states of Sub-Saharan Africa are due to experience the dividend’s economically favorable age structures anytime soon. Because two very different development narratives vie for these analysts’ attention, their confusion is not that surprising.

     In this essay, I discuss the concept of “the demographic window” and compare economists’ perspectives on sub-Saharan Africa to that of political demographers.  I also identify 4 groups of countries in sub-Saharan Africa that have very different schedules for reaching the demographic window (and thus reaching the World Bank’s upper middle income category and other development milestones). For the entire essay, posted in the Woodrow Wilson Center’s New Security Beatsee this page.

Download this New Security Beat essay on Sub-Saharan Africa’s Demographic Window .