Reproductive Behavior of the Greek Population in the Interwar Period: Regional Variations and Impact of Refugee Integration

Abstract

The 1920s are often described by demographers as a period of “population depression” across Europe. By the end of the decade, many Western and Northern European countries were approaching or had already fallen below replacement level fertility for the first time in recorded history. Greece provides a distinct contrast to the Western European trend, recording volatile but generally ascending fertility rates during the 1920s. This rise in fertility is often attributed to the arrival of over 1.2 million Greek refugees from Asia Minor, which presumably exhibited more prolific reproductive behavior than the native population. This paper examines the reproductive behavior of the Greek population by geographical department and refugee status (refugees and natives), employing established methods of historical demography to overcome interwar data limitations. The findings reveal that the observed rise in fertility was driven both by an increase in marital fecundity and by a shift in nuptiality. The shift in marriage patterns bolstered the Total Fertility Rates (TFRs), while the Total Marital Fertility Rates (TMFRs) followed a proportionally lower increase. Moreover, the regional analysis reveals a finding previously unaddressed in the relevant literature: in the early 1920s, fertility in certain Greek regions had already dropped below replacement level, even though the national average remained above it. These spatial disparities underscore that interwar Greece was a complex demographic mosaic of cultural and social identities, particularly across territories recently annexed into the modern Greek state.

1. Introduction

Sub-replacement fertility is often cited as a characteristic of the second demographic transition which occurred in the “Western” countries from the 1970s onwards (Van de Kaa 1987). However, sub-replacement fertility was recorded for the first time in a number of European countries in the interwar period (1918–1939), a fact that appears to have been forgotten or ignored by the contemporary literature. A reason for this omission is the notion that generation replacement is secured if every woman of reproductive age bears to life, on average, 2.1 children. This may be the case for developed countries with low mortality and life expectancy at birth, reaching 80 years. Yet, it was not the case with Europe in the 1920s and the 1930s, where mortality was much higher than in the last decades of the 20th century (when the “magic” number of 2.1 was established), and life expectancy (e₀) was a little higher than 50 or 55 years in the best case (Sardon 1991; Van Bavel 2010; Vallin and Meslé 2001).

Therefore, in the early 1920s, replacement fertility meant Total Fertility Rate (TFR) values above 2.5 and in some countries, like Greece and Spain, above 3.0 children per woman. During the 1920s, as European countries were pacing into the peaceful interlude of the interwar period, mortality improved, life expectancy was rising and, consequently, the TFR value necessary to ensure population replacement became lower. For Greece, this value dropped from 3.15 in 1920 to 2.9 in 1928 (Sardon 1991, p. 17). It is in that era (in the 1920s) that countries like England and Wales, Sweden and Germany recorded fertility below replacement level. In fact, out of the Council of Europe countries, only Greece, Italy, Spain, Portugal, Ireland, Iceland and the Netherlands had a fertility level high enough to ensure replacement in the 1920s. France recorded below replacement fertility even before WWI. England and Wales fell below replacement level in 1922, coinciding with prolonged post-WWI economic stagnation, while Germany followed in 1923, Sweden in 1925, Austria and Switzerland in the mid-1920s and Denmark in 1927 (ibid., p. 22).

The 1920s marked a crucial point in the history of the modern Greek state. The defeat of the Greek army in Anatolia, which followed the campaign in Asia Minor (1919–1922), brought 1.2 million refugees to the Greek state that was economically devastated by the continuous wars, and unprepared to receive and house them in decent conditions. Many of these refugees died in the hastily erected camps without basic sanitary conditions, while some emigrated to foreign countries (Rapp 1924, p. 408). The 1920 census enumerated 5,021,790 people within the borders, later finalized by the Treaty of Lausanne in 1923 (Statistique Générale de la Grèce 1928, κζ’). By the 1928 census, the population had reached 6,184,645, including 1,067,518 refugees who had arrived after September 1922 (Statistique Générale de la Grèce 1935, η’). The actual cumulative number of refugees was higher, as significant contingents had fled to Greece during earlier periods of hostility—namely the Balkan Wars (1912–1913), World War I (1914–1918), and the Asia Minor Campaign (1919–1922). This earlier refugee influx was omitted from the 1928 census. Nonetheless, the impact on fertility indices remained substantial. As shown in

Table 1. Age structure of female refugees and their relative share in the total female population in 1928. Source: (Statistique Générale de la Grèce 1935, Vol. II) (author’s calculations).

Ages	Female Refugees	% Distribution	Native Females	% Distribution	Female Refugees as % of Total Females
0–14	143,366	26%	822,333	32%	15%
15–49	321,566	57%	1,287,780	50%	20%
50–64	63,669	11%	272,484	11%	19%
65+	31,539	6%	176,193	7%	15%
Total	560,140	100%	2,558,790	100%	18%

, refugee women comprised 20% of the total female population of childbearing age (15–49). While the refugee population overall was not younger than the native population, it exhibited a higher proportion of women in their peak reproductive years, thereby generating a structurally higher potential for aggregate fecundity.

This study examines the reproductive behavior of the Greek population during the 1920s, disaggregated by geographical department and ethnic group (refugees and natives). To test the hypothesis of differential reproductive behavior between refugees and natives, this paper investigates the structural drivers of interwar fertility—specifically evaluating whether regional and sub-population differentials were governed by earlier nuptiality timing ($I_m$) or marital fertility control ($I_g$). By employing established methods of historical demography, this study derives fertility indices indirectly to overcome significant data limitations. The analysis is guided by the following research questions:

• What were the levels and trends of fertility in the Greek population during the 1920s?
• What distinct geographical variations existed, and how can they be explained?
• How did the reproductive behavior of refugees differ from that of the native population?
• How did Greece, or any of its regions, record sub-replacement fertility at any point during the interwar era?

To provide context for these questions, the following section offers a review of the existing literature.

2. Literature Review

The primary challenge in studying Greek fertility is the lack of reliable, continuous data series prior to the 1950s. Although the modern Greek state was established in 1830, the first published statistics of vital events refer to the period 1860–1891, and they are of poor quality due to the under-registration of births (and to a lesser degree that of deaths). Political instability (1861–1863) and extended operational lapses in the statistical system (1886–1920 and 1940–1949) created major gaps where no vital statistics were published at all.

Even into the 20th century, registration remained incomplete in the first half of that century, with the notable exception of the 1930s when birth registration achieved near-universal coverage (Valaoras 1960; Siampos and Valaoras 1969). Furthermore, the absence of interwar statistics on births by maternal age precludes the direct calculation of the TFR or the Net Reproduction Rate (NRR).

Furthermore, under-enumeration of the population in the census was common in Greece in the 1920s. The 1920 and 1928 censuses were characterized by extremely skewed sex ratios at certain age groups. The main pattern was a lack of males in the ages 15–44 and a plenitude of males in the ages 10–14 and 45–59. This is an indication that males were trying to evade military service in an era that Greece was at war or immediately after it. There was also a dearth of females in the ages below ten years of age (Gavalas and Baltas 2022).

Despite these data constraints, Greek demographers have made significant efforts to reconstruct the nation’s demographic history, including reproductive behavior during the interwar period (Valaoras 1960; Siampos 1973; Hionidou 2016; Gavalas 2001, 2002; Zafeiris and Kaklamani 2019; Gavalas and Baltas 2021, 2022). Most of these works rely on aggregate statistics. Valaoras employed demographic methods (model life tables and population projection and inverse projection techniques) to estimate the vital events of the population over a century (1860–1960) and calculated the Crude Birth Rate (CBR) at 31.4‰ in 1922 and 32.4‰ in 1927 (Valaoras 1960, p. 132). By contrast, calculations based solely on raw published statistics yield a CBR of 21.5‰ for 1922 and 28.5‰ for 1928 (Gavalas 2001, p. 71). This discrepancy underscores the magnitude of the registration gap. Furthermore, while Valaoras’ estimates show the CBR marginally increasing between 1922 and 1927, the substantial increase found in official statistics likely reflects improvements in registration of births during the 1920s.

Early research based on aggregate data identified substantial regional variations in fertility. By the turn of the 20th century, the capital area and the Ionian Islands had already begun to deviate from natural fertility patterns (Siampos and Valaoras 1969). More recent family reconstitution studies on two Cycladic islands provide clear evidence that fertility control was established there by the 1930s (Hionidou 2016; Gavalas 2002). Gavalas and Baltas (2021) attempted to indirectly derive period total and marital fertility indices for the 1920s using historical demographic methods; however, that study was limited to the Aegean Islands and to Greece as a whole. To date, there remains a gap in the literature regarding regional variations in Greek reproductive behavior during the 1920s, specifically through the lens of TFR and TMFR.

3. Material and Methods

In demographic analysis, TFR is among the most widely applied indicators of fertility patterns. This metric reflects the average number of children a woman would bear across her childbearing years (from 15 to 49 years of age), assuming she experiences the age-specific fertility rates recorded in a given year. Computing this index requires two datasets: live births categorized by maternal age, and the female population by age group. The mathematical expression of this index is:

$$\mathrm{T}\mathrm{F}\mathrm{R}=5\ast {\displaystyle \sum _{x=15}^{45}}{}_{5}f_x$$

(1)

where ₅f_x = age-specific fertility rate from age χ to age χ + 5, which is calculated as:

$${}_{5}f_x=\frac{{}_{5}B_x}{{}_{5}W_x}$$

(2)

where: ₅Β_x = births from women aged χ to χ + 5

₅W_x =women aged χ to χ + 5

A related measure, TMFR, uses the same underlying logic but restricts its scope to legitimate births and married women. Its calculation requires data on legitimate births distributed by maternal age, alongside the number of married women in each reproductive age group (Newell 1990).

$$\mathrm{T}\mathrm{M}\mathrm{F}\mathrm{R}=5\ast {\displaystyle \sum _{x=15}^{45}}{}_{5}mf_x$$

(3)

where ₅mf_x = age-specific marital fertility rate from age χ to age χ + 5, which is calculated as:

₅mf_x = (₅B^L_x/₅MW_x) = legitimate births from women aged χ to χ + 5

₅MW_x = married women aged χ to χ + 5

As far as Greece is concerned, data on births by age of mother exist only from 1956 onwards. For earlier periods, no such published data existed. For the 1920s, population broken down by age, sex and marital status is provided by the 1920 and 1928 census returns. Data on births by legal status of the child (but not by age of mother) are also provided by published vital statistics from 1921 onwards. Therefore, an approximation of TFR and TMFR has been attempted in this paper by modifying some indices used in historical demography. Our estimations rely specifically on Coale’s indices (or the Princeton fertility indices). Ansley Coale originally devised these tools for the Princeton Fertility Project, where they were used to track and record fertility decline at the county level across Europe during the demographic transition (Coale and Watkins 1986). Of these indices, two—Coale’s overall fertility index (I_g) and his marital fertility index (I_g)—were computed and subsequently converted into the more widely recognized and intuitively interpretable measures: the TFR and TMFR.

In practice, Coale’s fertility indices were designed to address a specific data limitation: the unavailability of maternal age at childbirth. For the overall fertility index ($I_f$), the sole data requirement is the total number of births in the population under study (B) and the age structure of the females in the reproductive age span (w_i).

$$I_f={\displaystyle \frac{\mathrm{B}}{\sum {\mathrm{w}}_{\mathrm{i}}·{\mathrm{F}}_{\mathrm{i}}}}$$

The term F_i in the above formula represents the age-specific fertility rates of married female Hutterites1 in the 1920s, which is the highest fertility schedule ever recorded in human populations (Coale 1971).

The Coale’s index of marital fertility (I_g) considers married women only (m_i) and legal births only (B_L)

$$I_g={\displaystyle \frac{B_L}{\sum m_i·F_i}}$$

In short, Coale’s indices express a population’s observed fertility as a proportion of what it would be if women followed the maximum fertility rates of 1920s’ Hutterites. In other words, these indices measure the deviation of the examined population from the maximum fertility level, which the population would have attained under certain conditions. Coale published the marital fertility schedule of Hutterite women during the 1920s (ibid., p. 208). According to this schedule, a Hutterite woman married at age 15 could expect to bear 12.44 children on average, provided she survived to age 49. The value 12.44 is the TMFR of married Hutterites in the 1920s and corresponds to Coale’s index of marital fertility I_g = 1. Furthermore, driven by religious beliefs, nearly every woman in the Hutterite community got married at a young age and there was no childbearing out of marriage. As a result, their marital fertility (TMFR) did not differ substantially from their total fertility (TFR). For this reason, a TFR of 12.4 was assigned to the female Hutterites of the 1920s, which corresponds to a Coale’s index of overall fertility I_f = 1.

Following this logic, Coale’s fertility indices were converted into the more commonly used and readily interpretable TFR and TMFR using the following formulas:

TFR_exp = I_f* TFR_Hut

TMFR_exp = I_g* TMFR_Hut

The subscript exp denotes the expected rate within the population under study, whereas Hut refers to the relevant index recorded for the 1920s’ Hutterite population. In modern populations characterized by birth control rather than natural fertility, marital fertility is greatly defined by the choices of the couples on the number of children they want, and to a lesser extent by the marriage pattern. However, in the 1920s the Greeks had only just begun to depart from the natural fertility schedule (Siampos and Valaoras 1969, p. 606; Gavalas 2001; Gavalas and Baltas 2022). Consequently, marriage patterns remained a major factor in shaping their TFR. While the marriage patterns in Greece—analyzed by geographical region and refugee status in a previous study (Gavalas 2025)—provided valuable insights, metrics such as permanent celibacy rates and mean age at first marriage do not always correlate directly with fertility levels. Of greater significance is the proportion of women married within each specific age group. A robust synthetic measure for this parameter is Coale’s index of marriage (I_m).

This index, symbolized by I_m, takes values from 0 to 1 and shows how nuptiality deviated from being universal. In fact, Coale’s index of marriage (I_m) is calculated as the ratio of two hypothetical birth totals based on a standard maximum fertility schedule (that of the Hutterites during the 1920s):

• The numerator is the number of births that would occur if only married women in the population followed this maximum schedule.
• The denominator is the number of births that would occur if all women of reproductive age—regardless of marital status—followed the same schedule.

This ratio *(1)/(2), thus measures the extent to which actual fertility in a natural fertility population falls short of its potential under universal marriage. The formula for the index of marriage is typically expressed as:

$$I_m={\displaystyle \frac{\sum m_i\ast H_i}{\sum w_i\ast H_i}}$$

where:

m_i = number of married women in age group i.

w_i = total number of women in age group i.

H_i = age-specific fertility rates of the Hutterites.

To derive the aforementioned indices, data were extracted from the 1920 and 1928 census returns (Statistique Générale de la Grèce 1928, 1935) and from published vital statistics on births by legal status (Statistique Générale de la Grèce 1924, 1931). Data on births for 1921 have been treated for under-registration by applying a correction weight of 1.24. This was derived by contrasting the birth estimates of Valaoras (1960, p. 132) for the quinquennial 1920–24 with the published statistics of the same period (Statistique Générale de la Grèce 1924, 1928, 1931, 1935). The discrepancy between the two reveals that under-registration of births in the early 1920s amounted to 24%. For data on births for 1928, no corrective action was taken. This is attributable to the fact that, by 1928, the share of municipalities failing to submit vital statistics to the central government had declined from 13.1% in 1921 to 0% (NSSG 1956, p. XV). While birth registration required external correction weights (as detailed above), census counts were accepted as the official baseline population. Under-enumeration in interwar censuses predominantly affected males of military age and young girls under the age of ten, rather than women in their peak reproductive years (15–49). Consequently, the female population utilized in our denominators is that of the census returns.

The analysis was conducted at the geographical department level, based on the administrative boundaries of the Greek state as they existed in 1928. Eastern Thrace, although part of the Greek territory during the 1920 census, was ceded to Turkey in 1923 (Treaty of Peace with Turkey 1923) and is therefore excluded from this analysis. The Dodecanese, which were annexed in 1947, are also excluded from the analysis. In total 10 geographical departments comprised the Greek State in 1928.

4. Results

4.1. Reproductive Behavior of the General Population

Under the mortality regime of 1921 in Greece, replacement fertility meant a TFR of at least 3.15 children per woman (Sardon 1991). What is noteworthy in Figure 1a is that 3 out of 10 geographical departments of Greece recorded sub-replacement fertility. Those were the Ionian islands (TFR = 2.82), Crete (3.05) and marginally Peloponnese (3.13). On the other end of the spectrum, the departments with the highest recorded TFR were the Cyclades and the newly annexed Thrace, both of which recorded a period total fertility above four children per woman. The Dodecanese Islands, annexed to Greece in 1947, likely had higher 1920s’ fertility rates than the other study areas. However, due to a lack of pre-annexation data, this inference relies primarily on general post-1947 fertility trends (Gavalas and Baltas 2021, p. 7).

In 1928 the situation changed dramatically (Figure 1b). Greece’s TFR rose from 3.8 (in 1921) to 4.3. There was no geographical department with sub-replacement fertility and none with TFR below 3.25. The highest TFRs were recorded again in Northern Greece, with Thrace remaining at the top with 5.1 children per woman. In 1928, the Cyclades remained among the geographical departments with high fertility (more than four children per woman) as they were in 1921.

Figure 2a, however, paints a different picture regarding TMFR. Geographical departments that presented a TFR higher than the national average in 1921 did not necessarily exhibit the highest TMFRs in the same year. For instance, all departments north of Central Greece and Euboea—specifically Thrace, Macedonia, Thessaly, and Epirus—recorded TMFRs below the national average (Figure 2a), despite their TFRs being above average (Figure 1a). On the other hand, Crete’s low TMFR was in accordance with its low TFR, while the Aegean and the Ionian islands exhibited higher TMFRs than expected.

In a parallel evolution with TFR, TMFR also increased at the national level from 6.4 in 1921 to 7.4 children per married woman in 1928 (Figure 2a,b). There was only one geographical department (Epirus) with a TMFR below 6.5 in 1928, in contrast to 1921, when all mainland Greece northern to Central Greece and Euboea had recorded TMFRs below 6.5. In six out of ten geographical departments in 1928, marital fertility was above seven children per woman (while only two geographical departments recorded such high marital fertility in 1921).

This discrepancy between TFRs and TMFRs is explained by variations in regional marriage patterns. Northern Greece customarily favored early and universal marriage for women (Figure 3a,b); the mean female age at first marriage in 1921 was lower than 22 years of age while permanent celibacy was below 3% (Gavalas 2025).

Consequently, although period TFR in Northern Greece was above average in 1921 due to the marriage pattern, women in that region actually bore fewer children within marriage, ranging from 5.6 to 6.2 compared with the national average of 6.4. On the other hand, Southern Greece and the Aegean Islands exhibited lower nuptiality, characterized by higher ages at first marriage and significant rates of permanent celibacy, particularly among island populations (Gavalas 2025). This explains the divergence between their high TMFRs and below average TFRs: although a smaller proportion of the total female population was married in Southern Greece and the Aegean islands, married women in those areas were more prolific than the national average. Crete remains a notable exception to this trend, recording low values for both TFR and TMFR.

In a nutshell, the spectacular increase in TFRs between 1921 and 1928 was partly due to elevated nuptiality. This is obvious when one compares Figure 3a and Figure 3b. In 1921 five out of ten geographical departments exhibited an index of marriage below 0.5 (Figure 3a). In 1928 only three of the geographical departments had an I_m below 0.5, and even this was marginally lower than 0.5. In 1928 the lowest value was 0.48, while in 1921 the lowest I_m was 0.44. However, the even greater increase in marital fertility between 1921 and 1928 (from 6.4 to 7.4) was not due to elevated nuptiality but due to a series of historical circumstances, like the return of the soldiers from the war, which made the married population more prolific and which (circumstances) are accounted for in the Discussion section.

Another issue that arises is whether there were urban–rural differentials in reproductive behavior. The primary geographical unit of analysis in this study is the geographical department, which does not allow us to get into detailed urban–rural differentials. However, a provincial-level analysis of Greece’s two major urban centers, Athens and Thessaloniki, has been attempted (

Table 2. Indicative urban–rural differentials in reproductive behavior in 1928. Source: Indirectly estimated fertility indices (TFR and TMFR) based on Coale’s $I_f$, $I_g$, and $I_m$. Primary data drawn from Statistique Générale de la Grèce (1935, Vol. II); author’s calculations.

Region Type	Province	TFR	TMFR	Im
Main urban	Attica	3.08	8.51	0.35
	Thessaloniki	3.55	6.20	0.56
Neighboring non-urban	Megara	4.46	7.75	0.57
	Lagkadas	5.21	7.45	0.70

). Comparing the urban provinces of Attica and Thessaloniki with their neighboring non-urban provinces, Megara and Lagkadas, reveals that urbanization depressed overall fertility, with urban areas recording lower rates (Attica: 3.08, Thessaloniki: 3.55) than their rural counterparts (Megara: 4.46, Lagkadas: 5.21). The primary driver for this depressed urban fertility seemed to be lower nuptiality ($I_m$ of 0.35 in Attica vs. 0.57 in Megara; 0.56 in Thessaloniki vs. 0.70 in Lagkadas). Nevertheless, high marital fertility persisted in Athens. The TMFR of Athenian women remained exceptionally high at 8.51 children per married woman, surpassing Megara’s 7.75.

4.2. Reproductive Behavior of the Refugees

Although the 1928 census enumerated the refugees—detailing population by refugee status and demographic characteristics—the published vital statistics of the era do not disaggregate births by the mother’s refugee status. Consequently, it is impossible to calculate distinct fertility indices for native and refugee populations. However, insights into their respective reproductive behavior can be inferred by comparing refugee marriage patterns with those of the native population.

Generally, women who migrated from Asia Minor following the 1922 catastrophe tended to marry at a younger age than their native Greek counterparts. The Singulate Mean Age at Marriage (SMAM) for native Greek women was 24.2 compared with 22 for the refugee women. (Gavalas 2025, p. 32). Figure 4 shows that, in the primary reproductive age groups (15–29), the percentage of married female refugees was indeed higher than that of natives. After the age of 30, the trend reverses, as the percentage of married refugees falls below that of the native population. This decline does not necessarily suggest a higher rate of lifelong celibacy among refugee women in the ages above 30. It may be that the refugee population had higher mortality than the natives. Scholars have documented tragically high mortality rates within refugee settlements, which were hastily erected by a Greek state unprepared for an influx of approximately 1.2 million people. This loss of life was primarily driven by a lack of sanitary infrastructure and severely harsh living conditions (Rapp 1924, p. 408). Consequently, the lower percentage of “currently married” refugees over age 30 may reflect a population where the proportion of widows was significantly higher than among natives, because marriages were frequently dissolved by the death of a spouse. This hypothesis is supported when one compares the percentages of widows in each age group between natives and refugees. As illustrated in Figure 5, refugees exhibited spectacularly greater percentages of widowed women than the natives.

In demographic terms, a marriage at age 20 contributes more to total fertility than a marriage at age 30 or 35, especially in populations that had just started to deviate from natural fertility. Since the refugees “out-marry” the natives during the most biologically fertile years, they would produce more children per capita even if their marriage rates lagged behind in later life. This educated hypothesis is strengthened by the Coale’s index of marriage I_m (

Table 3. Coale’s index of marriage in 1928 by geographical department. Source: census returns (Statistique Générale de la Grèce 1928, Vol. I; 1935, Vol. II) (own elaboration).

		Im	Im
	Geographical Department	Refugees	Natives	Difference	% Refugees *
1	Central Greece and Euboea	0.49	0.53	−0.03	17.8%
2	Thessaly	0.59	0.57	0.01	5.4%
3	Ionian islands	0.44	0.49	−0.05	1.3%
4	Macedonia	0.67	0.65	0.02	38.1%
5	Epirus	0.53	0.62	−0.10	2.4%
6	Crete	0.54	0.49	0.05	8.5%
7	Cyclades	0.37	0.57	−0.21	3.5%
8	Aegean islands	0.47	0.49	−0.02	16.9%
9	Thrace	0.64	0.72	−0.08	31.8%
10	Peloponnese	0.56	0.49	0.07	2.4%
	Greece	0.60	0.56	0.04	17.2%

* These percentages are based on the refugee population that arrived in Greece after the Asia Minor Disaster of September 1922 and was recorded in the 1928 census. In reality, the total number of refugees was higher, as a significant number of people had already fled to Greece during earlier periods of hostility and persecution, namely the Balkan Wars (1912–1913), World War I (1914–1918), and the Asia Minor Campaign (1919–1922). This earlier refugee influx was omitted from the 1928 census. Nonetheless, their impact on fertility indices remained substantial.

). As mentioned in the methodology section, I_m measures the extent to which marriage contributes to a population’s fertility by comparing the actual number of births to the number that would occur if all women of reproductive age were married. Although in six out of ten geographical departments, the I_m of the refugees was slightly lower than that of natives, the overall exposure (for the whole of Greece) was greater for the refugees (0.6) than for the natives (0.56). This difference of 0.04 confirms that, on a national scale, the refugee population had a higher overall “marital potential” for childbearing. Therefore, even if the percentage of currently married refugees “lagged behind” in later life (due to the high number of widows), the higher I_m attests that the intensity of marriage during peak reproductive years more than compensated for the losses in later age groups.

5. Discussion and Conclusions

The demographic landscape of interwar Greece presents a striking counter-narrative to the “population depression” that characterized much of Western Europe during the 1920s. While many nations saw fertility fall below replacement levels, Greece experienced a rise in its TFR from 3.15 in 1921 to 4.3 in 1928, a spectacular increase of 36.5%. This growth was driven by a dual force: a shift in nuptiality—where women married more frequently and at younger ages—and a general increase in marital fertility. By 1928, the national TMFR had risen to 7.4 children per married woman, up from 6.4 in 1921, a notable increase of 15.6%.

This “rejuvenation” of Greek demographics is probably associated with certain historical circumstances, which took place in the 1920s. The first of these events was the arrival of 1.2 million ethnic Greeks from Asia Minor. This movement—part of a population exchange already underway following the 1922 Asia Minor Catastrophe—was formalized by a Greco-Turkish agreement in 1923. Two additional factors likely contributed to the fertility rise in the 1920s: the homecoming of soldiers following the series of wars fought between 1912 and 1922, and the halt in overseas emigration after the United States imposed its first numerical immigration quota in 1921 (Gavalas 2001, p. 93). These historical conditions, which shaped the identity of modern Greece, erased the sub-replacement fertility previously seen in regions like the Ionian Islands, Peloponnese, and Crete, ensuring that by 1928, no geographical department remained below the replacement threshold.

These national trends were underpinned by significant regional variations that reveal a complex demographic mosaic. In 1921, Northern Greece (Macedonia, Thrace, and Epirus) exhibited high TFRs driven by a pattern of early and universal marriage, as evidenced by the high Coale’s index of marriage (I_m). Interestingly, these same northern regions initially recorded TMFRs below the national average, suggesting that their high fertility was due to marital timing rather than the number of children per marriage. Conversely, Southern Greece and the Aegean Islands exhibited lower marriage indices but higher TMFRs, meaning that, while fewer women married, those who did were more prolific. By 1928, however, data show that this discrepancy vanished; TMFRs surged across nearly all departments—particularly in the north—aligning high marital fecundity with traditional early marriage patterns.

Refugee women married significantly earlier than natives, with a SMAM of 22 compared with 24.2. Although the percentage of married refugees fell below that of natives after age 30 due to higher mortality and a greater proportion of widows in harsh settlement conditions, their early marriage “intensity” was demographically decisive. The national I_m for refugees (0.60) was higher than for natives (0.56), confirming a higher overall “marital potential”. By out-marrying natives during their most biologically fertile years (15–29), the refugee population produced more children per capita and effectively shielded Greece from the fertility declines observed elsewhere in Europe.

However, this fertility surge in the 1920s proved short-lived. From the 1930s onward, all fertility indicators—specifically Crude Birth Rates (CBRs), TFRs, and TMFRs—entered a decline that has continued almost monotonically to the present day, with the sole exception of a slight uptick in the TFR during the 1960s (Gavalas 2025). Therefore, the 1920s was a phenomenal decade in the history of the modern Greek state. The Asia Minor Catastrophe marked the end of a decade of successive wars—the Balkan Wars, World War I, and the Asia Minor Campaign—and, in the most tragic manner, brought to a close the vision of the “Great Idea”. This irredentist concept, which envisioned a Greek state spanning two continents and five seas, had served as the cornerstone of Greek foreign policy from the foundation of the modern Greek state until 1922. The influx of refugees gave a more homogeneous ethnic composition to the Greek territory and boosted the fertility of Greece for the last time. Thereafter, a succession of historical disruptions, including the conflicts of the 1940s (WWII and the Greek Civil War) and the massive emigration waves of the 1950s and 1960s, coupled with declining fertility, undermined the natural balance of births and deaths, and resulted in the demographic shrink of Greece in the 21st century despite its conversion from a migrant-sending to a migrant-receiving nation.

To advance our understanding of interwar demographic dynamics, future historical demographic research should prioritize two key avenues. First, there is a clear need for granular archival investigations utilizing micro-level regional records—such as local municipal registers and parish books—to examine fine-grained local variations in age at marriage, birth spacing, and refugee–native intermarriage patterns that macro-level census data cannot fully capture. Second, applying Coale’s indirect standardization framework ($I_f,I_g,I_m$) to other Mediterranean populations during the interwar period would provide a critical comparative perspective. Expanding this methodology across the broader region will determine whether the reproductive behaviors, nuptiality timing, and natural fertility trajectories observed in Greece were part of a shared Mediterranean demographic regime or represented a distinct national experience.