The long-run demographic consequences of the economic crisis
The Greek economy went off the rails in 2009 with a large fiscal deficit, an external current account deficit, a drop in productivity, and a jump in public debt. The ensuing crisis burst onto the international scene in 2010 with the beginning of a series of economic programs with partner countries, European institutions, and the IMF, to try and steady the listing economic ship. Eight years later we can summarize that national income collapsed by about a quarter, the unemployment rate peaked at 27.5 percent, and the country went through significant political and social turmoil. In more recent years, with stops and starts, the economy has begun to level off and stabilize. We are now in the “second 5-year period” (2016-2020) of this crisis decade, and some interesting data from the first 5-year period are starting to settle down as well, so that we can begin a reflection about the deeper influence of the economic crisis on Greek society. (Data are understandably preliminary in the heat of the present; it takes about 5 years for measurement to solidify.)
If one wants to understand the basics of any macroeconomy, one has to start by studying population and demographic data. The macroeconomy is anchored in the size and dynamics of the population. There is also an important feedback effect from economic management to the size of the population and its demographic developments. When an economy is well managed and the fortunes of its citizens are developing adequately, population dynamics tend to be stable and families have the financial confidence to have children. In societies that are poorly governed or have come under strong economic stress, population dynamics can be interrupted and families can have a tendency to leave and/or have fewer children. There are, of course, many other social, cultural, and personal influences on family planning and demographic developments, but economic confidence is one of the fundamentals that play a role. This “Note for Discussion” presents data from ELSTAT submitted to EUROSTAT on population and demographic developments that offer insight as to the circular, or inter-active, link between economics and demographics. This note offers a hypothesis of what may be going on below the surface. The reader may agree or disagree; all discussion and comments are welcome.
Figure 1 shows the population (POP) of Greece since 1960, appended by the latest projections from the Greek authorities, as submitted by ELSTAT to Eurostat. The data are used as presented and no filtering has been applied. If the data are incorrect, the diagnosis or hypothesis of what is going on may also be incorrect. It goes to show how important it is to present accurate data without worrying whether these data may be “pleasant” or “unpleasant.”
The reader will see two lines in Figure 1 on population: an upper dashed line, which shows the demographic projections available in 2011 based on information and tendencies that were known at that time. These projections carried through 2060 (a horizon of some 50 years). The lower solid line shows the vintage of projections that I downloaded in 2018, and these carry through 2080 (a horizon of some 60 years). It is at once evident that ELSTAT/EUROSTAT have updated the projections considerably, based on data that have come in for the period 2010-2015: the projections for population and population growth have been significantly revised downward, beginning with the onset of the economic crisis in 2009. By 2060, the population estimate is now some 2.4 million persons less than were projected in the 2011 data. This is critical for government policy in managing the economy, and we can dig deeper why this is the case.
Notice in Figure 1 that the new population line has a pronounced kink in 2011; before this date the population was growing slowly and projected to transition to a gradual decline as aging set in for future years. In the new data, 2011 was a peak and the population has started to decline steadily thereafter. This coincides with the economic recession and the stress this put on families and society more broadly—Greeks started to look for their fortunes outside of the country and outward migrated to other European countries and farther afield—causing the domestic population to decline. It is possible that data will be further revised in a next iteration, but such a pronounced kink in population growth is not often seen in other countries. It is one indicator that the recession had a substantial impact on Greece.
A smaller population in future means that the economy will also be smaller, as compared with the expectations at the beginning of this decade. With fewer persons, there will be less consumption and less investment and a lower absolute level of employment. The potential for the size of the Greek economy (the level of overall activity) is hereby reduced. Since Greece is coming into this projection with high debt, the smaller economy means that debt per capita (per person in Greece) is higher than was projected before, and this adds to the challenge how to manage over time the reduction in this debt. This underscores the sense that many Greek persons have that the crisis is past its most impactful immediate point (the recession), but the aftermath is far from resolved—the economy will require careful data-based management going forward to manage the challenges from high debt and declining population dynamics.
Let us dig one level deeper and see what happens to the population of working age—those between 15-64 years of age. ELSTAT/EUROSTAT report on this subgroup of the population as well. Figure 2 shows the dynamics of the “working-age population” (WAP) as reported by these institutions.
Here we can see that the estimates of the historical working-age population are slightly higher in the 2018 projections than in the 2011 projections, but the difference is small. However, the sharp downturn now also shows up in the working-age population in the new data projections (the lower line in the future through 2080). By 2060, the working-age population is now some 1.2 million persons smaller than it was in the older projections. Again, these are significant numbers that affect what we should expect the economy, in terms of its potential, to be able to deliver as Greece moves forward into the future.
There is yet more information in these data than meets the eye. To see this, we have to look at growth rates, rather than levels of population and working-age population. This is shown in Figures 3 and 4 which present time series of the annual percent change of the two variables of interest. Figure 3 shows the annual growth rate in the population and its projections, and a smoothed line (to take out the blips up and down that separate noise from the underlying signal), for the 2011 projection, and also for the 2018 projection. Figure 3 shows that the growth rate of the population took a hit around 2010-2011, and, interestingly, it stays lower for some 30-40 years compared to the previous projections. The convergence of growth rates is slow and takes time to unfold.
Figure 4 shows that the working-age population growth rate took an even bigger hit than the top-line population dynamics. It collapsed around 2010-2011 also, and stays below previous estimates until around 2050—some 30 years from now—before the effect of the economic crisis wanes from the projections. These data raise two questions: what is the impact of this demographic dynamic on the economy and why does it take so long for the data to re-converge?
To explore an answer to the first question, assume that all people of working-age find a job in the economy and produce at the “average level and growth rate”—in other words, assume that they all have average “productivity.” This would mean that the potential growth rate of the economy is the product of growth in the working-age population and average labor productivity. Looking at many economies, the reader will find that a labor productivity growth rate of around 1.5 percent a year is possible—therefore, let us assume for now that average Greek labor can deliver the same productivity growth rate in the long run, on average. Then we need to assess the growth rate in the number of persons in the working-age population to get a sense of the growth in available labor. This growth rate, based on the data above, has shifted down from around -0.5 percent on average in the next 50 years in the 2011 projections to around -1.0 percent in the newer projections, until the WAP growth rates converge after 2050. This means that this “primitive” estimate of average potential growth of the Greek economy will now be around 0.5 percent a year over the next decades, from an estimate of around 1.0 percent per year in the projections from 2011.
Bottom line: the loss of population dynamics and especially working-age population dynamics (adults have left the country more than dependents (<15; >65 years) as may have been expected) that are suggested by the new projections and that seem to have been induced by the economic crisis, can have long-term consequences and may have lowered Greek potential growth by about 0.5 percentage points.
As to the second question, why does it take so long for growth rates to converge again, this may be interpreted as having to do with Greek fertility and generational recovery. Greek people tend to marry somewhat late, around 25-35 years of age, and have children somewhat later than in some other countries. This means that a “generation” can be defined as spanning a time period of around 30 years (using round numbers). If the crisis hit the current generation of child bearing Greeks, with a number of them having now moved to other countries, especially of working-age, which also often overlaps with child-bearing age, then the generational gap that has emerged with the economic crisis and the outmigration of potential child-bearing persons, will be visible for some 30 years in the population statistics, until the population finds a new “steady state” when the next generation of the reduced number of children born during the crisis itself matures into the next child-bearing cohort and stabilizes the demographic trends. Conceptually, this is some 30 years from now, and the projections shown above do indeed bear this out. If out-migration reverses, or fertility goes up, then the current projections could be too cautious; a next and future iteration of the population projections will therefore be interesting.
Bottom line: barring changes in migration and fertility from current assumptions, it may take a full generation (some 30 years) for the Greek population to stabilize from the impact of the economic shock and the influence this had on demographic developments. If one looks at the numbers above, it is indeed around 2050, some 30 years after the crisis, that the population dynamics converge on the previous estimates that did not anticipate the effects of the economic shock in 2010-2015.
Some further remarks and caveats:
* The statistics may be revised and assumptions and projections may be updated. Given its importance, it would be helpful if ELSTAT could produce on an annual basis a review and update of long-run demographic data and projections. The government is well taken to pay close attention to and comment to the public on these fundamental data.
* This note uses “baseline” projections, which are the projections that seem most likely given available information. To get a sense of risk around the baseline, both for upside potential and downside risk, the government may wish to run scenarios with slightly different assumptions around this baseline, to get a range of estimates.
* Projections are always to some extent “linear.” That means, as mentioned, that they do not assume major changes in behavior by the population going forward. They are “relative extensions of recent trends.” But perhaps as the economy finds a new wind, some that have left will come back to Greece and thus boost the working-age population back up. If so, the newer estimates provided above would be too cautious and one would converge somewhat to the previous projections. Again, sound government policy and good communication to boost confidence that it is worth investing and working in Greece are essential in this respect.
* This note stops, for now, at decomposing the population into working-age population and dependency. For economic analysis, this is not enough. The next step would be to extract from the working-age population a signal for the labor force (this depends on participation), employment (this depends on investment and labor costs), and unemployment. Even if the working-age population does not grow, the economy could still grow by increasing participation and reducing unemployment (increasing employment). There is ample room to boost these variables in Greece, and, again, sound government management plays a vital role in this respect. The next “note for discussion” will attempt to decompose the working-age population into the labor force, employment and unemployment to see what more detail we can learn.
Bob Traa is an independent economist. This is the first of a series of articles he will be contributing to Kathimerini.
