SXjra pp: -- Techset Composition Ltd, Salisbury, U.K. //
J. of Modern African Studies, , (), pp. – © Cambridge University Press doi:./SX 1 2 3 The changing religious 4 5 composition of Nigeria: causes 6 7 and implications of 8 9 demographic divergence 10 11 12 MARCIN STONAWSKI 13 University of Oslo, Department of Sociology and Human Geography, 14 P.O. Box Blindern, Oslo, Norway and Cracow University 15 16 of Economics, Department of Demography, Rakowicka , - 17 Cracow, Poland 18 Email: [email protected] 19 ̌ 20 MICHAELA POTANCOKOVÁ 21 22 Wittgenstein Centre for Demography and Global Human Capital 23 (IIASA, VID/OAW, WU), Vienna Institute of Demography, 24 Welthandelsplatz , Vienna, Austria 25 Email: [email protected] 26 27 MATTHEW CANTELE 28 29 International Institute for Applied Systems Analysis (IIASA), Ecosystems 30 Services and Management, Schlossplatz ,A- Laxenburg, Austria 31 Email: [email protected] 32 33 and 34 Q1 EGARD KIRBEKK 35 V S 36 Columbia University, Columbia Aging Centre, W. th Street, 37 NY , New York, USA 38 39 Email: [email protected] 40 41 42 MARCIN STONAWSKI ET AL.
ABSTRACT 43 44 At nearly million inhabitants, Nigeria is Africa’s most populous country by 45 twofold and fertility levels remain higher than most other sub-Saharan African 46 nations. Throughout the last several decades, the fertility gap between fi 47 Christians and Muslims has widened with signi cant political implications for a nascent democracy. Where the DHS survey of revealed a non-significant 48 difference of · children, this figure had increased to · children by .As 49 the total fertility rate (TFR) of Christians decreased significantly from · to · 50 children per woman between and , the TFR of Muslims increased 51 from · to · children per woman. The timing of this divergence coincides 52 with the formal institutionalization of Sharia law in . We examine the 53 role of religion on education, contraception and family behaviour. Finally, we 54 touch upon the implications for population growth and the religious compos- 55 ition of Nigeria in the coming decades. 56 57 INTRODUCTION 58 59 At nearly million inhabitants, Nigeria is the most populous country 60 in Africa by twofold and is on course to become one of the largest coun- 61 tries in the world. It is also one of the most diverse, with a large number 62 of groups from different linguistic and ethnic backgrounds (Katzner 63 ; Lieberman and McClendon ). Nigeria is also the only state 64 in the world with equal proportions of Christians (·%in), 65 located mainly in the South, and Muslims (·%) (PEW ) who 66 are living mainly in the Northern regions of the country (Figure ). 67 The population balance between the two majority religions, however, 68 is likely to shift in the future as a result of distinctly different trajectories 69 of demographic change. While the Muslim North is currently in an early 70 phase of demographic transition with very high fertility rates, fertility 71 rates of Christians and Muslims in the South are declining as the demo- 72 graphic transition is further underway. These diverging trajectories in 73 demographic behaviour may lead to the end of the current equilibrium 74 between the two religions with significant implications for the future of 75 the country. Within this study we examine possible explanations for di- 76 vergent population trajectories along religious lines as well as the path 77 78 that these trajectories may potentially take into the future. 79 As with several other Western African countries, ethno-regional and 80 religious divides largely overlap in Nigeria and are manifested in a pro- 81 nounced north-south gap (Mancini ). Here, inequalities have 82 emerged along geographic and ethnic divisions as well as along socio- 83 economic lines between nomadic herdsmen and farmers, exacerbated 84 by climate change (e.g. Fasona & Omojola ; Sayne ). While THE CHANGING RELIGIOUS COMPOSITION OF NIGERIA
85 86 87 88 89 90 91 92 93 94 95 96 97
98 Fig. 1 - B/W online, B/W in print 99 100 Figure . Religious composition by state in Nigeria in . Shading on the 101 map indicates presence of Sharia law in states; pie charts represent religious 102 composition. 103 ethnic divides were of great importance in the past, a religious divide 104 now appears to be becoming more pronounced as religious tensions 105 compound traditional ethnic and geopolitical rivalries, magnifying 106 underlying insecurities and grievances. 107 The past decade has seen increasing reports of sectarian violence in 108 Nigeria’s Middle Belt where ethnic and religious communities meet. 109 110 And in an ominous sign of things to come, these attacks are increasingly fl – 111 framed in terms of religious and cultural con ict. In Sharia 112 laws were introduced in states in northern Nigeria (Figure ) 113 leading to religious unrest and an escalation of hostilities that resulted 114 in thousands of deaths in Kaduna state alone (Barker & Ricardo : 115 ; Ostien & Dekker ). This drastic shift is reflected in the Pew 116 Research Center’s Social Hostilities Index which documented a surge 117 from · to · on a -point scale between and – one of 118 the highest scores in the world. As part of efforts to establish an 119 Islamic caliphate, attacks on Christians and moderate Muslims by the 120 militant group Boko Haram, which translates literally to ‘Western 121 Forbidden’, may be orchestrated with the goal of stoking religious ten- 122 sions between the predominately Christian Igbo in the South and 123 Hausa-Fulani Muslims in the North, with the Yoruba in the West repre- 124 senting a mix of Christianity and Islam. 125 Within the context of an ethnically and religiously diverse country 126 such as Nigeria, a newfound emphasis on religious identity coupled MARCIN STONAWSKI ET AL. fi fi 127 with religion-speci c fertility differentials could carry signi cant political fi 128 and economic implications. Indeed, Mazrui ( ) nds that religious fl 129 divisions in sub-Saharan Africa are most likely to lead to con ict when 130 they reinforce underlying ethnic differences, precisely the situation we 131 now observe in Nigeria. It is particularly important to note that the emer- 132 gence of militant Islamic sects and possible spread of Sharia law to add- 133 itional states is potentially a response to the growing north-south gap and 134 perceived political dominance of the South (Isa ). Indeed, while its 135 implementation has ignited hostilities particularly in the border states 136 making up the Middle Belt, Sharia law was being applied to Muslims 137 in most Northern states long before its formal institutionalization 138 (Ostien & Dekker ). 139 In our paper, first, we discuss factors contributing to the widening gap 140 in the fertility differential between Muslims and Christians. Besides the 141 factors such as contraceptive use and demand for family planning and 142 differences in marital behaviour that fall within the proximate determi- 143 nants framework (Bongaarts ), which is a framework highly rele- 144 vant for understanding of fertility change in sub-Saharan African 145 context, we discuss differences in educational improvements among 146 women and changing fertility ideals. We show that religion per se is 147 not the main driver of stalling fertility among Muslims in Sharia states 148 as there is significant heterogeneity in the reproductive behaviour of 149 Muslims in Nigeria. We argue that fertility stalls and high population 150 growth are specific for Muslims in Sharia states. Therefore, it is not 151 152 useful to speak of Nigerian Muslims as a homogeneous group. 153 The second part of the paper demonstrates the consequences of the 154 increased fertility gap on the religious landscape of Nigeria. While the 155 relative similarity in size between Muslims and Christians populations 156 emerged only during the past years, Muslims have never been a 157 clear majority in a united Nigeria (Mi ). This situation may soon 158 change as we identify an ongoing demographic bifurcation that could 159 significantly alter the religious composition of Africa’s most populous 160 nation with substantial implications for future population outcomes. 161 162 163 164 THE INCREASING GAP IN RELIGION- SPECIFIC FERTILITY IN 165 NIGERIA 166 Using Demographic Health Survey (DHS) surveys, in this study we iden- 167 tify religion-based fertility differentials and project potential demo- 168 graphic futures based on a number of scenarios. We find that the total THE CHANGING RELIGIOUS COMPOSITION OF NIGERIA fi 169 fertility rate (TFR) of Christians decreased signi cantly from · [CI: – – 170 · · ]to · children [ · · ] per woman between and – 171 while the TFR of Muslims increased from · [ · · ]to · chil- – 172 dren [ · · ](Figure ). However, fertility is not rising among all 173 Nigerian Muslims. Taking a closer look we find that fertility has 174 increased from · to · children per woman between and 175 among Muslims residing in currently Sharia law states (Table I). 176 However, among those residing elsewhere fertility declined from · 177 to · children per woman. Although this is one child more than the 178 average for a Christian woman, we show that fertility change unfolds 179 in a similar way among Christians and Muslims from non-Sharia states. 180 Interestingly enough, the timing of the divergence in fertility trends 181 coincides with the formal introduction of Sharia law over the course 182 of several years following the return to civil rule and subsequent 183 reinstatement of constitutional law (Kendhammer ; for an overview 184 see Oba ; Ostien and Dekker ). This leads to questions regard- 185 ing how Sharia law or culture specific to those regions where Sharia law 186 has been institutionalized fosters high fertility. As discussed in the next 187 section, we hypothesize that the adoption of Sharia and societal norms 188 related to its practice sustains high fertility culture and contributes to in- 189 creasing fertility differentials. The widening gap in fertility by religion 190 can thus be understood as a result of persistent high fertility and prona- 191 talist culture among Muslims in Sharia law states contrasted with a pro- 192 gressing fertility transition among Christians and moderate Muslims 193 194 residing in the Southern states. 195 196 197 ISLAM, SHARIA AND FAMILY BEHAVIOUR 198 199 The institutionalization of Sharia law in northern states beginning in 200 and consequent subjugation of the English legal system repre- 201 sented a return to Islamic primacy which has not been observed since 202 the pre-colonial period (Ostien & Dekker ). However, the spread 203 of Sharia law and extended jurisdiction to cover criminal as well as 204 civil matters was the culmination of gradual and some would argue inev- 205 itable decline of the system of common law that was imposed on Nigeria 206 during colonialism (Oba ). Indeed, Sharia law was already being 207 carried out among Nigerian Muslims in the realm of civil/personal 208 law prior to official institutionalization (Nmehielle ). The full im- 209 plementation of Sharia law can thus be viewed as symptomatic of pre- 210 existing societal preferences as well as emblematic of a nascent MARCIN STONAWSKI ET AL.
211 212 213 214 215 216 217 218 219 220 221 222 223 224 Fig. 2 - B/W online, B/W in print 225 226 Figure . Total Fertility Rate among Christians and Muslims in Nigeria in 227 and . Thick black lines indicate % confidence intervals. 228 229 movement away from Western conceptions of law, education and the 230 secular state. 231 While the institutionalization of Sharia law in the northern states oc- 232 curred relatively recently, Islam has long been an integral aspect of 233 the religious and political landscape of northern Nigeria, culminating 234 in the Sokoto Caliphate of the early th century and near total replace- 235 236 ment of customary law with Islamic law (Ostien & Dekker ). During 237 238 239 T ABLE I 240 Total fertility rates by religion by states in Nigeria in , 241 and . 242 243 Sharia law states Non-Sharia law states 244 Year Christians Muslims Christians Muslims Nigeria 245 * ···· 246 ()()()()() 247 ····· 248 ( )()()()() ····· 249 ()()()()() 250 251 *Because of the small number of observations it was not possible to calculate TFR for Christians 252 in Sharia states in . THE CHANGING RELIGIOUS COMPOSITION OF NIGERIA
253 this period southern Nigeria was awarded as a protectorate to the British, 254 resulting in the introduction of English law in the South while a system 255 of indirect rule (through local proxies) led to a two-tier legal system in 256 the North which included native/customary law (with elements of 257 Islamic law) and English law. In contrast with the North, Sharia law 258 never supplanted customary personal law in the southern states of 259 Nigeria and southern Muslims have traditionally adhered to tribal 260 customs in most legal matters (Ostien & Dekker ). Thus, although 261 customary law in South Muslim communities often reflects the influence 262 of Islamic law, the norms and procedures regulating marital law have 263 remained uncodified and thus largely contingent on local customs. In 264 the North, however, the legal ease and societal norms associated with 265 marriage and divorce have been identified as key contributors to high 266 fertility by motivating women within polygynous marriages to maximize 267 their number of children in order to prevent divorce and ensure their 268 share of inheritance (Izugbara & Ezeh ). 269 Although the Quran does not prohibit birth control in marriage and it 270 is permitted for a wide range of justifiable reasons (Omran ), op- 271 position to modern contraception and family planning programmes 272 emerged in some contexts where these are seen as an import from the 273 West and a deviation from ‘the right path’ (Roudi-Fahimi ). For 274 these reasons it is not uncommon that family planning programmes 275 would be politicized in Muslim societies. In this context, strict adherence 276 to Sharia law could imply a lack of sexual health education and/or sus- 277 278 picion toward modern forms of contraception, as well as the relatively 279 low educational levels among young girls. Mazrui ( ) outlines 280 several ways in which Islamic fundamentalism can affect fertility such 281 as distrust of the West, association of contraception with prostitution, 282 Quran-based fatalistic attitudes regarding procreation, attitudes toward 283 polygyny and large families, subversion of traditional birth spacing, 284 and encouragement of a large age gap between husbands and wives. 285 Within this analysis we will focus primarily on the interplay of Sharia, 286 education, and family planning and its implications for differential 287 Christian-Muslim fertility levels. 288 Sharia law and the implicit societal norms associated with its practice 289 can influence fertility levels through a number of indirect and direct 290 causal avenues including lowered age of marriage as well as outright pro- 291 hibitions on formal education. While there is variation among differ- 292 ent schools of thought, puberty marks the age of marriage for young 293 girls under classical Sharia (Rehman ). In addition to the increased 294 prevalence of pregnancy and implications for educational attendance, MARCIN STONAWSKI ET AL.
295 child brides are frequently forbidden from attending school or lack the 296 time to continue their education (Warner ). A low marriage age 297 thus results in higher fertility by increasing the probability of pregnancy 298 while impeding access to education that could result in lower fertility 299 preferences (Westoff et al. ). Disparate age differences are also con- 300 ducive to patriarchy, resulting in increased male control over contracep- 301 tive practices and reproductive behaviour which evidence suggests leads 302 to higher realised fertility (Ezeh et al. ; Isiugo-Abanihe ; 303 Bankole & Singh ). 304 Societal norms, patriarchy, and a resistance towards Western 305 influence – which are sustained by Sharia law – seem to play a role in 306 high fertility culture among Muslims in northern Nigeria. High fertility 307 of Muslims in the Sharia states is coupled with persistently high fertility 308 ideals (Figure ). In only about % of Muslim women in Sharia 309 states were in favour of relatively smaller families (with two to four chil- 310 dren) while large family ideals remain the norm with % of women 311 considering at least seven children ideal, and an additional % stated 312 that they would accept any number of children God grants them. 313 Fertility ideals did not change much among the Muslim women in 314 Sharia states between and . Although more women now 315 give a numeric response to the question, a fairly stable share of those 316 who wish to have many children (more than or ‘any number’) sug- 317 gests that unlimited fertility and large family ideas are well embedded 318 in the states that adopted Sharia law. 319 320 In contrast, fertility ideals have shifted towards smaller families among 321 Christians and to a lesser extent also among Muslims in non-Sharia 322 states. Muslim women residing in non-Sharia states not only have a 323 lower TFR than those residing in Sharia states, but also differ in their 324 perception of the ideal number of children, with more citing four chil- 325 dren as ideal and fewer preferring large families with at least seven chil- 326 dren. The ideal family size distribution and trend towards preferences 327 for smaller families among Muslim women residing in non-Sharia 328 states mimics the pattern of Christian women, although their overall fer- 329 tility ideals are slightly higher than those of Christian women. The appar- 330 ent difference between fertility ideals of Muslim women in Sharia and 331 non-Sharia states calls for more thorough investigation of situation 332 factors of fertility and deeper understanding of the role of religious 333 leaders and locally embedded cultural practices that give rise to these 334 diverging preferences. 335 The high fertility gap between Muslims and Christians can also be 336 linked to differences in mean age at marriage and in the practice of 378 377 376 375 374 373 372 371 370 369 368 367 366 365 364 363 362 361 360 359 358 357 356 355 354 353 352 351 350 349 348 347 346 345 344 343 342 341 340 339 338 337
Fig. 3 - B/W online, B/W in print ttsmrya about at marry states during from increased has hitasadMsi oe nnnSai tts(onfrom (down states non-Sharia in to women Muslim and Christians nplgnu arae a atdi h hrasae codn to according states Sharia the in halted has the marriages polygynous in t ifrne.Nmru tde ieicesn eaeeducational female increasing cite studies fertil- Numerous of explanations differences. in ity cited conventionally is attainment Educational ( DHS of waves all about at low, very Our polygyny. r nplgnu no ete at settled union polygynous in are v n aia eaiu seuain sw hwi h etsection, next the signi in women show Christian we among As levels education. education is behaviour marital and ive com- husband are the con- economically to co-wives order husband in where the possible strain as context children many Nigerian as have to the pelled within fertility higher ntelreyMsi ot a aersle nlweuainlevels, beliefs. education pronatalist low of in retention resulted and have age the may marriage North early to in Muslim fundamentalism response largely fertility growing lower the in and in declining potentially attitudes and anti-Western and to development, use Concurrently this contributed South. contraceptive remain have Christian higher states largely may Sharia levels, fertility the education desired in women Rising Muslim low. among levels education Figure n fteudryn ehnssbhn h hne nreproduct- in changes the behind mechanisms underlying the of One among % H HNIGRLGOSCMOIINO NIGERIA OF COMPOSITION RELIGIOUS CHANGING THE da ubro hlrnb eiinadrsdne()among (%) residence and religion by children of number Ideal . and ’ neiac Iubr Ezeh & (Izugbara inheritance s – fi oe aged women dnscon ndings ’ – blt otk e ieadesr hi oto of portion their ensure and wife new a take to ability s eas e eln nplgn among polygyny in decline a see also We . H.Hr h hr fwmnaged women of share the Here DHS. er o ulmwmni h hrasae across states Sharia the in women Muslim for years erod during olds year – to – fi mta enaeat age mean that rm er,wietema g at age mean the while years, EDUCATION .Msi oe eiigi non-Sharia in residing women Muslim ). er faeaogNigeria among age of years – ,in .Plgn scnuieto conducive is Polygyny %. and – fi atyicesd while increased, cantly fi ). s araeremains marriage rst hl h decline the while ) . fi ’ s marriage rst Christians s – who % MARCIN STONAWSKI ET AL.
379 attainment as one of the key factors of fertility decline. Can improve- 380 ments in educational levels explain variation in fertility levels and the 381 opening gap between Muslims and Christians in Nigeria? As illustrated 382 in Figure , we observe a widening gap in TFR by religion and also diver- 383 ging trends over time, with Christian fertility declining and Muslim fer- 384 tility rising. If trends in education were to explain the fertility gap 385 between Muslims and Christians, we should observe (a) higher educa- 386 tional levels of Christians compared with Muslims, and (b) educational 387 improvement among Christians as opposed to stalling education or even 388 worsening educational characteristics of Muslim women. 389 Nigeria has witnessed educational expansion since the s when 390 primary schooling was made mandatory. Greater enrolment at 391 primary levels translated into increased completion of secondary and 392 higher education but this seems to be limited to the majority 393 Christian and better developed South. As expected, we find huge differ- 394 ences in educational attainment between Muslim and Christian women 395 in reproductive age (Figure ). In a mere % of Christian women 396 aged – had no formal education and a majority attained at least 397 lower secondary education (%). In contrast, two-thirds of Muslim 398 women had no formal schooling in and , a modest improve- 399 ment compared with % with no formal education in . 400 401 402 403 404 405 406 407 408 409 410 411 412 413 414 415 416 417 Fig. 4 - B/W online, B/W in print 418 419 Figure . Educational characteristics of women aged – by religion in 420 and . THE CHANGING RELIGIOUS COMPOSITION OF NIGERIA
421 Furthermore, the gap in completed mean years of schooling (MYS) 422 between Christian and Muslim women increased from · to · years fi 423 from to despite signi cant gains by non-Sharia state 424 Muslims (Table II). Consistent with the hypothesis of a negative 425 impact of Sharia culture on female educational improvement, the data 426 show that this growing rift is driven entirely by stalled educational 427 gains among Sharia state Muslims, % of which lack any formal 428 education. 429 The gap in educational attainment is partly sustained by the persistent 430 perception of Western education as being anti-Islamic (Csapo ). 431 Lincove () finds a significant effect of religion on school attend- 432 ance and Kazeem et al. () show that children from Christian house- 433 holds are five times more likely to attend school compared with those 434 from Muslim households. In particular, female education remains at 435 very low levels. Studies also suggest that many Muslim parents prefer 436 their daughters to attend traditional Islamic schools, because it guaran- 437 tees preserving religious values and traditional female roles 438 (Ogunjuyigbe & Fadeyi ). Indeed, Adiri et al. () find that a 439 Quranic education is more common than other types of school 440 among Muslim women (·%). 441 The expected inverse relationship between education and fertility 442 (Bongaarts ; Skirbekk ) holds and we find that that fertility 443 444 445 T ABLE II 446 Mean years of schooling of women – by religion and region in 447 , and . 448 449 Mean years of schooling 450 451 Christians ··· 452 ()()() 453 non-Sharia state ·· · 454 ()() () Sharia state ·· · 455 ()() () 456 Muslims ··· 457 ()()() 458 non-Sharia state · · · ()()() 459 Sharia state ··· 460 ()()() 461 Nigeria · · · ()()() 462 MARCIN STONAWSKI ET AL. 463 is lower among the better educated in both religions (Figure ). Among 464 Christians, those with primary or less education had at least six children 465 in , while those with at least secondary only have four children. 466 Between and , fertility fell by roughly half a child for all edu- 467 cational groups among Christians. Overall, fertility levels were more 468 than one child higher among Muslims compared with Christians 469 within all educational groups. The educational gradient in fertility corre- 470 sponds to the positive effect of in particular secondary and higher edu- 471 cation on the use of modern contraceptives (Ejelmi et al. ). 472 However, TFR rose substantially among Muslims in the period – 473 .Wefind the most pronounced increase among those with no edu- 474 cation, a smaller increase for those with at least some primary education, 475 and stalled fertility among women with at least completed lower second- 476 ary education. This has taken place in spite of modest improvement in 477 the educational attainment of Muslim women (Figure ). As can be 478 seen from Table III, the structural effect of increased education 479 depressed the fertility of Muslims by −· children per woman. 480 However, fertility increased by · children per woman once education 481 is controlled for, using a two-component decomposition of fertility 482 483 484 485 486 487 488 489 490 491 492 493 494 495 496 497 498 499 500 501 Fig. 5 - B/W online, B/W in print 502 503 Figure . Total fertility rate by education among Christians and Muslims in 504 Nigeria in and . THE CHANGING RELIGIOUS COMPOSITION OF NIGERIA
505 T ABLE III 506 Decomposition of TFR change into educational change and fertility 507 change. 508 509 TFR 510 Religion TFR change Educational change Fertility change 511 Christians ·· −· −· −· 512 Muslims ··· −·· 513 514 515 change among Christians and Muslims (see Appendix A; Retherford & 516 Thapa ). Decreasing age at entering first marriage can lead to a rise 517 in fertility, however, as we showed earlier the mean age at first marriage 518 519 stays low at about years but did not decline either. Therefore, other 520 factors, such as possible changes in contraceptive use and possibly the 521 institutionalization of Sharia law, stand behind the increased Muslim fer- 522 tility in Nigeria. 523 524
525 CONTRACEPTION 526 Most extant studies find that contraceptive use is more common among 527 Christian women than Muslim women in Nigeria (Adebowale et al. 528 ). Our analysis reveals that among currently married or co-habiting 529 530 Christian and Muslim women in Nigeria, there is an increase in contra- 531 ceptive practice (either using modern or traditional methods) in non- 532 Sharia states. Despite this modest increase, only around %of 533 women of reproductive age use modern contraception (see Table IV) 534 and contraceptive use in Nigeria is low even by regional standards 535 (Bongaarts & Casterline ). The shares of women using contracep- 536 tion are higher among women aged – compared with younger 537 women. 538 Differences in contraceptive use are significant also between religious 539 groups. Among Christians, the proportion of currently married women 540 not using any method to prevent pregnancy has declined from %to 541 % between and . The fraction of non-users has also 542 declined among Muslims in non-Sharia states from %to%. In con- 543 trast, contraceptive use among Muslim women in states where Sharia law 544 has been introduced has remained negligible and a vast majority of 545 married women do not use any method of birth control. The proportion 546 of non-users slightly declined towards among the younger women 588 587 586 585 584 583 582 581 580 579 578 577 576 575 574 573 572 571 570 569 568 567 566 565 564 563 562 561 560 559 558 557 556 555 554 553 552 551 550 549 548 547
T ABLE IV STONAWSKI MARCIN Percentage of sexually active married or cohabiting women using no contraceptive method, by religion and age.
All married women Christians Muslims in Sharia Muslims in non-Sharia
Year – – – – – – – – – – – – ()()()()( ()()()()()()()
AL ET ()()()()()()()()()()()()
(,)()(,)()()(,)()()()()()(,) . THE CHANGING RELIGIOUS COMPOSITION OF NIGERIA – – 589 aged years; however, this trend reversed between , fi 590 shortly after Sharia law was introduced. Ejelmi et al. ( ) nd that re- 591 ligion remains an important predictor of modern contraceptive use in 592 north-eastern and north-western Nigeria even after controlling for 593 number of individual and community level factors. 594 Further examination of the DHS surveys reveals that women in Sharia 595 states perceive unmet need for family planning. The proportion of non- 596 users who intend to use or are unsure about using some method of birth 597 control in the future has increased among all groups during –, 598 including Muslim women in Sharia states. The proportion of non-users 599 who do not intend to use any method of birth control has dropped to 600 % for Christians and % among Muslims in non-Sharia states in 601 . However, the decline has halted among Muslim women in 602 Sharia states in recent years and about % of non-users do not wish 603 to use any contraception in future, see Table V. According to 604 DHS, % of Christian women currently not using contraception and 605 % of Muslim women in non-Sharia states reported they would like 606 to use contraception. The situation is very different among Muslim 607 women in Sharia states, where only % of women below age 608 expressed a wish to use birth control, a little more than ·% among 609 – year olds. 610 The main reason given for not intending to use any birth control in 611 the future (by the non-users who do not intend to use any birth 612 control) has shifted from wanting as many children as possible, which was 613 614 the most prominent reason given by the Muslim women from Sharia 615 states, towards opposition to contraception in the period from to 616 , see Table VI. This shift is visible for both Christians and 617 618 619 T ABLE V 620 621 Trends in the intentions not to use contraception among the married women not using any method to prevent pregnancy, by 622 623 religion. 624 Year Christian Muslim non-Sharia Muslim Sharia All married 625 626 % % % % ()()()() 627 % % % % 628 ()()()() 629 % % % % ()()(,)(,) 630 MARCIN STONAWSKI ET AL.
631 T ABLE VI 632 Attitudes towards contraception among non-users, women aged 633 – years. 634 635 Christians Muslims (Sharia) Total* 636 637 N= N= N= N= N= N=
638 Wants as many children as ...... 639 possible 640 Opposed to using ...... contraception 641 respondent opposed ...... 642 husband opposed ...... 643 others opposed ...... Religious prohibition ...... 644 Attitudes and values-related ...... 645 Knowledge-related reasons ...... 646 Method-related reasons ...... Barriers-related ...... 647 Other reasons** ...... 648 Does not know ...... 649 650 651 652 Muslims in Sharia states. The growth in opposition to contraception, 653 from %to% among Christians and %to% among Muslims 654 may be related to, among other factors, more conservative religious 655 656 views that stigmatise modern contraception. A smaller proportion of 657 the respondents explicitly cited religious prohibition as the main 658 reason for not using birth control in the future. Religious prohibition 659 was cited more frequently by Muslims ( %in and %in ) 660 than by Christians where the share actually declined (from %in 661 to %in). 662 Compared with , more Muslim non-users from Sharia states cited 663 husband’s opposition to birth control – up from about %to%in 664 , while Christians saw a stagnation of around %. Due to this 665 shift towards presumably more direct pressure not to take contracep- 666 tion, fewer cited disapproval of others as influential in their decision. 667 This suggests that both the internalization of attitudes leading to disap- 668 proval of birth control and social pressure against use may have inten- 669 sified in Sharia law states. Moreover, this shift is consistent with a 670 significant body of literature documenting the pivotal role male fertility 671 preferences play on fertility decisions in Nigerian society, particularly 672 among the Hausa ethnic group predominant in northern Nigeria THE CHANGING RELIGIOUS COMPOSITION OF NIGERIA 673 (Isiugo-Abanihe ; Bankole & Singh ; Duze & Mohammed 674 ; Izugbara et al. ). 675 Another difference between all Christian and Muslim non-users in 676 Sharia states is illustrated in the reason for intention not to use contra- 677 ception in the future. Here we observe that among Christians the ration- 678 ale for not intending to use is more frequently related to barriers such as 679 lack of knowledge or a lack of access to contraception (e.g. due to high 680 price, no access, or health concerns). The aforementioned reasons have 681 been increasingly cited by Christian non-users, by about % compared 682 with about % of Muslims non-users in (this percentage has not 683 fluctuated). We assume that respondents citing barriers-related reasons 684 are not principally opposed to birth control. If so, a larger proportion of 685 Christian and a smaller proportion of Muslim women would intend to 686 use some methods of preventing pregnancy if they were better 687 informed, had financial means to buy contraceptives and had (better) 688 access to facilities offering methods of family planning. Problematic 689 access to family planning facilities and services may indeed be a 690 reason why Muslim women in Sharia states who perceive unmet need 691 have little choice, in particular if reproductive choices are being made 692 in highly unpredictable context (see Johnson-Hanks et al. ; 693 Johnson-Hanks ). DHS data on the family planning services 694 network in Nigeria do not allow for a more detailed analysis. However, 695 we can conclude that the current opposition to modern contraceptive 696 use is an important reason not to use birth control for Muslim women 697 698 in Sharia states and it is as much embedded in as it sustains high fertility 699 culture in the region. 700 701 702 POPULATION PROJECTIONS 703 Will Nigeria become a majority Muslim country? 704 705 We argue in the previous sections that the institutionalization of Sharia 706 law has led to persistent high fertility as a result of the protective role it 707 plays in sustaining a high fertility culture. With a small share of the 708 Muslim population residing in non-Sharia states, the fertility of Sharia- 709 state Muslims drives the overall high fertility of Nigerian Muslims. The 710 lack of convergence in Muslim-Christian fertility differentials will not 711 only change the religious landscape of Nigeria, but can also affect 712 overall population growth. If the greater fertility of Muslims is sustained, 713 this will lead to faster population growth over time, particularly as the 714 proportion of Muslims grows. However, if fertility rates for both MARCIN STONAWSKI ET AL.
715 Muslims and Christians were to converge and decline, this would imply a 716 lower population growth in the coming decades. We study the possible 717 consequences of different fertility trajectories on religious composition 718 in Nigeria using a demographic projection model that takes into 719 account differences in fertility, intergenerational transmissions of reli- 720 gion and age-distribution of religious beliefs (Goujon et al. ; 721 Q2 Hackett et al. a; Stonawski et al. ). 722 The projection of population by age, sex and religion in Nigeria is 723 based on the demographic method of multi-state population projec- 724 tions that is widely accepted by technical demographers (Samir et al. 725 ; Hackett et al. a, b). In order to illustrate the long-term 726 consequences of fertility differentials by religion on the religious land- 727 scape of the country, we introduce four scenarios in the period of 728 years, from to . The scenarios only differ in terms of fertility 729 assumptions (see Figure ): 730 fi – – 731 ( ) in the rst scenario Fert(conv) we assume slow convergence in 732 fertility between Muslims and Christians from initial levels observed 733 in that would lead to the same level and age-pattern for these 734 groups by (within years). At each step of this scenario, 735 736 737 738 739 740 741 742 743 744 745 746 747 748 749 750 751 752 753 Fig. 6 - B/W online, B/W in print 754 755 Figure . Total fertility rate and fertility gap between Muslims and Christians 756 in Nigeria in – by scenario. THE CHANGING RELIGIOUS COMPOSITION OF NIGERIA
757 TFR for the total population follows the UN medium assumptions 758 (United Nations ); – – 759 ( ) in the second scenario Fert(c ) the assumption is that fertil- 760 ity by religion is constant during the whole projection period. This 761 we estimate using fertility age-patterns based on the DHS and 762 adjust them to meet baseline fertility level from the UN (United 763 Nations ); 764 () similar to Fert(c) – Fert(c) – fertility is constant but on 765 the level estimated from the DHS that are adjusted to be in 766 line with the UN level in –; 767 () the last scenario – Fert(o) – is designed to examine how religious 768 composition would change if there was no difference in fertility 769 between religious groups, and TFR of total population followed 770 the UN medium scenario assumptions (United Nations ). 771 772 fi 773 Because of a lack of empirical evidence on religion-speci c mortality, we 774 assume no mortality difference between religious groups. The course of – 775 change in mortality in follows the UN medium assump- 776 tions. Baseline inflow and outflow of migrants in Nigeria are calculated 777 using estimates of international migration bilateral flows (Abel ). 778 Religion of migrants was estimated using information from the Pew 779 Research Center’s Faith on the Move project (Connor ). In all scen- 780 arios we assume religious switching between Christians and Muslims. 781 Evidence from the Global Attitudes Project of Pew Research Center sug- 782 gests that around –% of males and –% females (both Muslims 783 and Christians) convert during their life. We model this explicitly in 784 the projections by introducing transition rates. 785 The results from our projection model suggest that constant fertility 786 levels from the baseline could lead to a dramatic change in population 787 size in Nigeria. During the coming years the population size is pro- 788 jected to increase from million in to million if relative dif- 789 ference in fertility by religion is kept constant as observed in (Fert 790 (c)), and to million, if fertility differentials from is used 791 (Fert(c)). Fertility following the UN medium trajectory (United 792 Nations ) leads to million people by (Fert(conv) and 793 Fert(o)). As a result of strong population growth, Nigeria may have 794 the second largest Christian and fourth largest Muslim population by 795 (Hackett et al. a). 796 797 In addition to strong overall population growth, it is likely that ’ fi 798 Nigeria s population will undergo signi cant changes in religious MARCIN STONAWSKI ET AL. 799 composition (Figures and ). A shift away from the current shares 800 seems inevitable, placing the country on a clear path towards becoming 801 a majority Muslim country in the coming decades. Even if there were no 802 differences in fertility between the two religious groups (Fert(o) scen- 803 ario), the proportion of Christians would decrease from ·%in 804 to ·%in, whereas Muslims would gain · percentage 805 points and for the first time in the history of a united Nigeria reach a ma- 806 jority of ·%. In this scenario the differences in population structure, 807 migration, and religious switching between Muslims and Christians have 808 a minor impact on religious composition (Fert(o)). 809 Several other scenarios suggest a stronger increase in the size of the 810 Muslim population resulting in a growing majority share of the total 811 population. If the relative difference in fertility is projected from the 812 level, when the differential was small, the proportion of 813 Christians drops to % and Muslims rise to %. In a situation of 814 more pronounced and slowly diminishing fertility gap (Fert(conv)), 815 the share of Muslims increases to almost % of Nigeria’s population 816 and Christians drops to %in. Finally, the scenario with the 817 most pronounced fertility gap at the level of the DHS and constant 818 fertility – Fert (c) TFR: Muslims – ·, Christians – · – shows that 819 820 821 822 823 824 825 826 827 828 829 830 831 832 833 834 835 836 837 838 Fig. 7 - B/W online, B/W in print 839 840 Figure . Proportion of Christians in Nigeria in – by scenario. THE CHANGING RELIGIOUS COMPOSITION OF NIGERIA
841 842 843 844 845 846 847 848 849 850 851 852 853 854 855 856 857 Fig. 8 - B/W online, B/W in print 858 859 Figure . Proportion of Muslims in Nigeria in – by scenario. 860 861 862 the Muslim population can reach as much as % whereas Christians 863 can decline to as little as %. In , and DHS data we 864 see stable fertility differential between Christians and Muslims and the 865 data so far do not support the convergence scenario. This does not 866 mean that convergence is not possible in the future but for the 867 coming years the share of the Muslim population will likely follow the 868 Fert(c). 869 We assume in our scenarios, conservatively, that the fertility differen- 870 tials between Christians and Muslims could decrease or remain constant 871 at the baseline level. However, it is also possible that in the future the fer- 872 tility of Christians will decrease following the demographic transition tra- 873 jectory and Muslim fertility stagnates at the current level (e.g. due to a 874 lack of modernisation and sustained high fertility culture). Such a diver- 875 gence would lead to a far greater share of Muslims in the religious land- 876 scape of Nigeria – reaching as high as –%by. 877 878 879 880 CONCLUSIONS AND DISCUSSION 881 Nigeria is most likely going to become a majority Muslim country in no 882 more than a couple of decades unless the differential between the MARCIN STONAWSKI ET AL.
883 Muslim and Christian fertility quickly converges in the near future. 884 However, this is not likely to happen given the high fertility ideals, low 885 prevalence of family planning, and slow educational improvements in 886 Sharia states. Quite to the contrary, legal implementation of Sharia 887 may sustain the high fertility culture by preventing modernisation as a 888 result of opposition towards the Western influence and increased 889 desires to preserve cultural differences in respect to the rest of 890 Nigeria’s population. 891 In this regard, the institutionalisation of Sharia law should be viewed 892 as a symptom of societal preferences that have long been present 893 throughout northern Nigeria. It is likely however that the ongoing nor- 894 malisation of Sharia law will lead to greater societal pressure to conform 895 to underlying tenets, thus reinforcing many practices contributing to 896 higher fertility while forming a bulwark against Western influences con- 897 ducive to lower fertility (on normalisation, see e.g. May & Finch ). 898 In Nigeria we are therefore witnessing the confluence of distinctly pro- 899 natalist traditions in sub-Saharan Africa (Bongaarts & Casterline ) 900 with a conservative religious ideology, a synthesis which will likely con- 901 tinue to result in high ideal family size and sustained higher fertility. 902 Fertility intentions and preferences can be powerful in terms of 903 shaping behaviour and they can also be important as a marker of how 904 potentially unrestricted pronatalism penetrates religious and cultural 905 norms following Sharia. Therefore, high fertility can be seen as an inte- 906 gral part of the local traditions and institutionalisation of Sharia law may 907 908 be seen as a way to preserve local culture. High fertility ideals are a key 909 aspect of high fertility culture and as we showed earlier more than % 910 of women in Sharia states prefer large families of and more children ‘ ’ 911 or any number of children God gives them . Only % think that four or – 912 fewer children is ideal this contrasts sharply with declining fertility 913 ideals throughout the rest of Nigeria. Marriage customs, polygyny and 914 in particular high divorce rates that result from liberal divorce regula- 915 tions under Sharia are among other possible factors responsible for sus- 916 tained high fertility among the majority of Muslims in Nigeria. We could 917 not empirically investigate these potential influences due to data limita- 918 tions, however, existing research supports this interpretation (e.g. 919 Izugbara & Ezeh ). 920 Nigeria is an outstanding case due to the uniquely balanced popula- 921 tion share of the two main religions and the clear divergence in demo- 922 graphic trends. The high relevance of religion within the Nigerian 923 context further underscores the usefulness of considering religion in 924 demographic studies. According to the Afrobarometer, both THE CHANGING RELIGIOUS COMPOSITION OF NIGERIA
925 Muslims and Christians claim that religion is very important in their life 926 ( % of Muslims and % of Christians). Nigerian society is divided 927 across religious, ethnic and regional lines with a long history of rivalry 928 between the largely Muslim North and mostly Christian South. 929 Moreover, religion is the most frequently cited reason of violent 930 conflict in Nigeria besides land disputes, according to Afrobarometer 931 ( and ). Religious conflict has intensified since the time of 932 the survey; Boko Haram began committing atrocities in and there- 933 fore it is likely that religion has become an even more important driver 934 of political instability. 935 Our projections show that should this fertility gap remain constant, or 936 slowly decrease (in the convergence scenario), Nigerian Muslims will 937 have a clear % majority by , making it one of the largest 938 Muslim states in the world. However, it is very likely that the gap in fer- 939 tility can grow even further leading to a higher share of Muslims in 940 Nigerian society. This discrepancy can ultimately bring significant 941 change to democratic outcomes, redefine the political dynamics 942 between traditionally comparable religious groups, and potentially call 943 into question power-sharing agreements such as the rotating presidency. 944 Indeed, within the context of an ethnically and religiously diverse 945 country such as Nigeria, a newfound emphasis on religious identity 946 coupled with religion-specific fertility differentials could very well be pre- 947 cursors to future sectarian conflict and carry significant political and 948 economic implications for the region and the world. 949 950 951 NOTES 952 . Source: Own calculations based on DHS . 953 . Nigerian States: Bauchi, Borno, Gombe, Jigawa, Kaduna, Kano, Katsina, Kebbi, Niger, Sokoto, 954 Yobe and Zamfara ’ 955 . In the census, which is the earliest reliable data source we found, . % of Nigeria s total population was Muslim, .% Animist, and a mere .% Christian. In the – census, the pro- 956 portion of Muslims rose to .% and in the census to .% while the share of Christians sur- 957 passed Animists, most likely due to conversion to other religions (.% of Christians in ). The share of Muslims grew most quickly in the Northern regions, amounting to about % of the popu- 958 lation in – and censuses (Mi : ). 959 . We base our analysis on Demographic and Health |Surveys , , and . DHS 960 was excluded because of the data quality issues that include underreporting of recent births in the order of about % and a sample bias towards more educated women (DHS report ). Thus 961 the results of the survey diverge from the trends obtained using the three other waves. 962 . TFR estimations were done using tfr STATA module (Schoumaker ). 963 . Christians are comprising % of the population in the northern regions while about %of Muslims reside in the southern states (DHS ). 964 . Source: own estimates computed for -year period prior survey based on DHS and . 965 . We consider states from north-central, north-east and north-west as northern, and south-east, south-central and south-west as southern. This division differs from the Sharia, non-Sharia distinction 966 we introduce later in the text because not all northern states have adopted Sharia. MARCIN STONAWSKI ET AL.
. Source: DHS , and . Own estimates computed for -year period prior survey. 967 Number of observations is in brackets. Sharia was introduced as source of civic and criminal law 968 from . Thus, DHS indicates results in pre-Sharia law states. 969 . We refer explicitly to formal education to differentiate from Quranic education that is deeply rooted in northern Nigeria. 970 . There is no significant change for low fertility ideals and the increase in the preference for – 971 and + children can be attributed to the decline in non-numeric responses (any number, God’s will). This indicates that preference for large families remains stable between the two surveys. 972 . Source: DHS and ; Note that DHS cannot be used due to differently phrased 973 response options. In the category ‘God’s will’ was explicitly stated but in and it was 974 removed and non-numeric response is recorded only if numeric is not given after probing. As a result the fraction of non-numeric responses dropped from more than % to about %. 975 . Gross enrolment rate (GER) at primary increase from to %in– and GER at 976 secondary level rose from to % (UNESCO – UIS database). GER are not available by regions but differences in educational attainment point towards nearly universal primary school enrollment 977 in the majority Christian South. 978 . Source: DHS and . Note: MYS stands for mean years of schooling. The indicator was 979 computed using duration of schooling in single years of the respondents. Secondary and higher edu- cation means that the person completed at least years of schooling (lower secondary completed). 980 . Source: DHS , and . 981 . Source: DHS and DHS . . Source: own calculations based on data from DHS and computed for -year period 982 prior survey. 983 . Source: DHS , and . 984 . Source: DHS , and . . Unfortunately, the question was not asked in DHS . 985 . Source: DHS and . 986 . Part of DHS questionnaire inspecting the presence and kind of family planning facilities in the ’ 987 proximity of women s place of residence was discontinued since DHS III ( ) and therefore we could not address this issue. 988 . In our article we present results of population projections which present scenarios of possible 989 future changes of a population when certain assumptions are made about the future trajectories of demographic processes. In contrast to projections, demographers also build forecasts which aim to 990 present most probable future development of population using the most realistic assumptions of 991 demographic processes (for more see Preston et al. ). 992 . The method is an extension of cohort-component based population projections (Preston et al. ). Rather than projecting only age and sex, this method extrapolates additional dimensions such 993 as health status, marital status, education, political views – or in this case, religion (e.g. Philipov ; 994 Rogers ; Stonawski et al. ). . In our projection we follow the UN World Population Prospects, rather than revi- 995 sion because the revision is closer to TFR observed in DHS (. for – compared 996 with . in DHS). In contrast, the latest revision assumed fertility stalls in Nigeria at . per – – 997 woman in and also pre-estimate for is at . children per woman, well above the DHS level. 998 . Note: Fertility gap = TFR(Muslims)–TFR(Christians). 999 1000 1001 1002 REFERENCES
1003 Abel, G.J. . ‘Estimating global migration flow tables using place of birth data’, Demographic 1004 Research , : –. ‘ 1005 Adebowale, S.A., F.A. Fagbamigbe & E.A. Bamgboye. . Contraceptive use: implication for com- pleted fertility, parity progression and maternal nutritional status in Nigeria’, African Journal of 1006 Reproductive Health , : –. 1007 Adiri, F., H.I. Ibrahim, V. Ajayi, H.U. Sulayman, A.M. Yafeh & C.L. Ejembi. . ‘Fertility behaviour of men and women in three communities in Kaduna State, Nigeria’, African Journal of Reproductive 1008 Health , s: –. THE CHANGING RELIGIOUS COMPOSITION OF NIGERIA
Bankole, A. and S. Singh. . ‘Couples’ fertility and contraceptive decision-making in developing 1009 countries: hearing the man’s voice’, International Family Planning Perspectives , : –. 1010 Barker, G. and C. Ricardo. . Young Men and the Construction of Masculinity in sub-Saharan Africa: fl 1011 implications for HIV/AIDS, con ict, and violence. Washington, DC: World Bank. Bongaarts, J. . ‘A framework for analyzing the proximate determinants of fertility’, Population 1012 and Development Review , : –. 1013 Bongaarts, J. . ‘Completing the fertility transition in the developing world: the role of education- al differences and fertility preferences’, Population Studies , : –. 1014 Bongaarts, J. & J. Casterline. . ‘Fertility transition: is sub‐Saharan Africa different?’, Population 1015 and Development Review ,s: –. ‘ fi ’ 1016 Connor, P. . Faith on the Move: The Religious Af liation of International Migrants . Washington, DC: Pew Forum on Religion and Public Life. http://features.pewforum.org/ 1017 religious-migration/Faithonthemove.pdf. 1018 Csapo, M. . ‘Religious, social and economic factors hindering the education of girls in northern Nigeria’, Comparative Education , : –. 1019 Duze, M.C. and I.Z. Mohammed. . ‘Male knowledge, attitude, and family planning practices in 1020 northern Nigeria’, African Journal of Reproductive Health , : –. fl 1021 Ejelmi, C.L., T. Dahiru & A.A. Aliyu. . Contextual Factors In uencing Modern Contraceptive Use in Nigeria (DHS Working Paper No. ). 1022 Ezeh, A.C., U.C. Isiugo-Abanihe, S.E. Tolnay, A. Woodcock, K. Stenner, R. Ingham, C. Reichert et al. 1023 . ‘The influence of spouses over each other’s contraceptive attitudes in Ghana’, Studies in Family Planning , : –. 1024 Fasona, M.J. and A.S. Omojola. . ‘Climate change, human security and communal clashes in 1025 Nigeria’, Proceedings of the International Workshop on Human Security and Climate Change, – 1026 Asker, Oslo, June . Hackett, C., P. Connor, M. Stonawski, V. Skirbekk, M. Potancoková̌ & G. Abel. a. The Future of 1027 World Religions: Population Growth Projections, –. Washington, DC: PEW Research Center. 1028 Hackett, C., M. Stonawski, M. Potancoková,̌ B.J. Grim & V. Skirbekk. b. ‘The future size of reli- fi fi ’ – 1029 giously af liated and unaf liated populations , Demographic Research , : . Isa, M.K. . ‘Militant Islamic groups in northern Nigeria’, in W. Okumu & A. Ikelegbe, eds. 1030 Militias, Rebels and Islamic Militants: human security and state crises in Africa. Pretoria: Institute for 1031 Security Studies, –. Isiugo-Abanihe, U.C. . ‘Reproductive motivation and family-size preferences among Nigerian 1032 men’, Studies in Family Planning , : –. 1033 Izugbara, C.O. and A.C. Ezeh. . ‘Women and high fertility in Islamic Northern Nigeria’, Studies – 1034 in Family Planning , : . Izugbara, C., L. Ibisomi, A.C. Ezeh & M. Mandara. . ‘Gendered interests and poor spousal 1035 contraceptive communication in Islamic northern Nigeria’, Journal of Family Planning and 1036 Reproductive Health Care , : –. Johnson-Hanks, J. . ‘Natural intentions: fertility decline in the African demographic and health 1037 surveys’, American Journal of Sociology , : –. 1038 Johnson-Hanks, J., J. Caldwell, S. Dennis, J. Guyer, H. Miyazaki, C. Notermans, S. Randall et al. . ‘ ’ 1039 When the future decides: uncertainty and intentional action in contemporary Cameroon , Current Anthropology , : –. 1040 Katzner, K. . The Languages of the World. London: Routledge. 1041 Kazeem, A., L. Jensen & C.S. Stokes. . ‘School attendance in Nigeria: understanding the impact and intersection of gender, urban‐rural residence, and socioeconomic status’, Comparative 1042 Education Review , : –. 1043 Kendhammer, B. . ‘The Sharia controversy in northern Nigeria and the politics of Islamic law in ’ – 1044 new and uncertain democracies , Comparative Politics , : . Lieberman, E.S. & G.H. McClendon. . ‘The ethnicity–policy preference link in sub-Saharan 1045 Africa’, Comparative Political Studies , : –. 1046 Lincove, J.A. . ‘Determinants of schooling for boys and girls in Nigeria under a policy of free ’ – 1047 primary education , Economics of Education Review , : . Mancini, L. . ‘Comparative trends in ethno-regional inequalities in Ghana and Nigeria: evidence 1048 from demographic and health surveys’, Working Paper . Oxford: CRISE. 1049 May, C. and T. Finch. . ‘Implementing, embedding, and integrating practices: an outline of nor- malization process theory’, Sociology , : –. 1050 MARCIN STONAWSKI ET AL.
Mazrui, A.A. . ‘Islamic doctrine and the politics of induced fertility change: an African perspec- 1051 tive’, in J.L. Finkle & C.A. McIntosh, eds. The New Politics of Population: conflict and consensus in family 1052 planning. New York, NY: Population Council, –. 1053 Mi, I. . Population Censuses of Nigeria from Colonial Times: an evaluation of their coverage and accuracy [online]. Retrieved from http://www.popline.org/node/. 1054 Nmehielle, V.O.O. . ‘Sharia law in the northern states of Nigeria: to implement or not to imple- 1055 ment, the constitutionality is the question’, Human Rights Quarterly , : –. Oba, A.A. . ‘Islamic law as customary law: the changing perspective in Nigeria’, International and 1056 Comparative Law Quarterly , : –. 1057 Ogunjuyigbe, P.O. & A.O. Fadeyi. . ‘Problems of gender differentials in literacy and enrolment ’ – 1058 among the Yorubas of South- west Nigeria , Journal of Social Sciences , : . Omran, A.-R. . Family Planning in the Legacy of Islam. London: Routledge. 1059 Ostien, P. & A. Dekker. . ‘Sharia and national law in Nigeria’, in J.M. Otto, ed. Sharia Incorporated: 1060 a comparative overview of the legal systems of twelve Muslim countries in past and present. Leiden: Leiden University Press, –. 1061 PEW. . The Global Religious Landscape. A Report on the Size and Distribution of the World’s Major 1062 Religious Groups as of . [online]. Washington, DC: PEW Forum on Religion and Public Life. 1063 Retrieved from http://www.pewforum.org/The-Future-of-the-Global-Muslim-Population.aspx. Philipov, D. . ‘Multistate population projections’, in A. Rogers, ed. Advances in Multiregional 1064 Demography. Laxenburg: International Institute for Applied Systems Analysis. 1065 Preston, S.H., P. Heuveline & M. Guillot. . Demography: measuring and modeling population processes. Oxford: Blackwell. 1066 Rehman, J. . ‘The sharia, Islamic family laws and international human rights law: examining the 1067 theory and practice of polygamy and talaq’, International Journal of Law, Policy and the Family , : – 1068 . Retherford, R.D. & S. Thapa. . ‘Recent trends and components of change in fertility in Nepal’, 1069 Journal of Biosocial Science , : –. 1070 Rogers, A. . Multiregional Demography: principles methods and extensions. Chichester: Wiley. 1071 Roudi-Fahimi, F. . Islam and family planning. Washington, DC: Population Reference Bureau. Samir, K.C., B. Barakat, A. Goujon, V. Skirbekk, W.C. Sanderson & W. Lutz. . ‘Projection of 1072 populations by level of educational attainment, age, and sex for countries for –’, 1073 Demographic Research : –. Sayne, A. . Climate Change Adaptation and Conflict in Nigeria. USIP Special Report No. . 1074 Washington, DC: US Institute of Peace. 1075 Schoumaker, B. . ‘A Stata module for computing fertility rates and TFRs from birth histories: ’ – 1076 tfr . , Demographic Research : . Skirbekk, V. . ‘Fertility trends by social status’, Demographic Research , : –. 1077 Stonawski, M., V. Skirbekk, E. Kaufmann & A. Goujon. . ‘The end of secularisation through dem- 1078 ography? Projections of Spanish religiosity’, Journal of Contemporary Religion , : –. United Nations . World Population Prospects. New York, NY: United Nations Population Division. 1079 Warner, E. . ‘Behind the wedding veil: child marriage as a form of trafficking in girls’, American 1080 University Journal of Gender, Social Policy & the Law , : –. ‘ 1081 Westoff, C.F., T. Pullum, S.E. Adamchak, K. Hill, P. Stupp, J.T. Bertrand, M.T. Brown et al. . Age at marriage age at first birth and fertility in Africa.’, Journal of Biosocial Science , : –. 1082 1083 1084 1085 1086 1087 1088 1089 1090 1091 1092 THE CHANGING RELIGIOUS COMPOSITION OF NIGERIA
APPENDIX 1093 1094 A. Decomposition method for disentangling fertility change from educational 1095 change. 1096 Two components: () educational change, () fertility change 1097 1098 ΔTFR ¼ TFRðÞ t þ h TFRðÞ t 1099 X X 1100 wa;e ðÞt þ h wa;e ðÞt ¼ 5 ASFR ; ðÞ t þ h 5 ASFR ; ðÞ t 1101 a e ðÞþ a e ðÞ a;e wa t h a;e wa t 1102 X 1103 ASFR ; ðÞþt þ h ASFR ; ðÞt w ; ðÞt þ h w ; ðÞt ¼ 5 a e a e a e a e 1104 ðÞþ ðÞ a;e 2 wa t h wa t 1105 20 1 3 ðÞþ ðÞ 1106 wa;e t h þ wa;e t X6B ðÞþ ðÞC 7 1107 6B wa t h wa t C 7 þ 5 4@ A ASFRa;e ðÞ t þ h ASFRa;e ðÞt 5 1108 a;e 2 1109 1110 – fi 1111 where: ASFRa,e (t + h) age speci c fertility rate at age a and educational 1112 group e at time t+h, wa,e (t + h)- number of females at age a and educa- 1113 Q3 tion e at time t+h, 1114 1115 1116 1117 1118 1119 1120 1121 1122 1123 1124 1125 1126 1127 1128 1129 1130 1131 1132 1133 1134