The Impact of the Hajj Religion and Tolerance in Islam’s Global Gathering

The Impact of the HajjThe Impact of the Hajj Religion and Tolerance in Islam’s Global GatheringReligion and Tolerance in Islam’s Global Gathering

David Clingingsmith, Case WesternDavid Clingingsmith, Case WesternAsim Ijaz Khwaja, KSGAsim Ijaz Khwaja, KSGMichael Kremer, HarvardMichael Kremer, Harvard

Malcolm X’s Account of Hajj

“Never have I witnessed such sincere hospitality and overwhelming spirit of true brotherhood as is practiced by people of all colors and races here in this ancient Holy Land … There were tens of thousands of pilgrims, from all over the world. They were of all colors, from blue-eyed blondes to black-skinned Africans. But we were all participating in the same ritual, displaying a spirit of unity and brotherhood that my experiences in America had led me to believe never could exist between the white and non-white…On this pilgrimage, what I have seen, and experienced, has forced me to rearrange much of my thought-patterns previously held, and to toss aside some of my previous conclusions … I have eaten from the same plate, drunk from the same glass, and slept on the same rug - while praying to the same God - with fellow Muslims, whose eyes were the bluest of blue, whose hair was the blondest of blond, and whose skin was the whitest of white. (The Autobiography of Malcolm X, Alex Haley)

Views of Impact of Hajj

Pilgrim accounts stress feelings of unity Sparked concern of outsiders

British colonial authorities Recent MI5 monitoring of British pilgrims after it

emerged that two 7/7 bombers had recently undertaken the Hajj

Concerns that Saudi version of Islam promoted at expense of local religious traditions (Naipaul, 1981)

Isolating the Impact of the Hajj

Two views could be consistent In-group vs. out-group?

Impact on other identities?

Identifying causal impact is difficult Selection problem – those who choose to perform Hajj differ

from those who do not Iannacone (1992): costly religious practices screen for

commitment Decision to undertake Hajj may be correlated with other changes

We address by comparing successful and unsuccessful applicants to Pakistan’s Hajj visa allocation lottery

Outline Background on Hajj

Statistical approach, visa application process, survey, sample

Results Religion Tolerance Gender Health and well-being Socio-political participation Channels – group exposure

Conclusion Micro mechanisms? Macro impact?

Background on Hajj

Formal rituals last 5 days, but most stay 40 days One of five pillars of Islam; obligatory if good health,

sufficient means Focus on religious practice (not knowledge) Physically taxing (80 km distance, infectious disease) Traditionally seen as conferring social prestige Social interaction during Hajj

Emphasis on equity/unity: standardized rituals, similar dress, head shaving, title at completion may promote common identity

Participants observe others from all over the world, but in-depth interaction unlikely

Men pray alongside women, are exposed to societies with different gender norms

Hajj Visa Allocation Logistical limits on Hajj participation

Saudis allocate Pakistan 150,000 Hajj visas

Pakistan allocates 66,500 visas to: Private travel operators, who charge US $2,000 - $5,000 Special quotas for military, gov’t employees, two-time lottery losers

Remaining 83,500 visas for subsidized government scheme US$2,000 to US$2,400 per person (3 x GDP/capita) 138,000 applicants Lottery

We focus on Sunni (90%) lottery applicants from 9 districts in Punjab province

Selection Process Application and deposit in bank branch

Formation of groups (1-20); lottery over groups

Mehrem (close male relative) required for Sunni women

Randomization by strata Sect (2) Place of departure (8) Accommodation class (2)

Group size effect

Randomization checks

Lottery Consistent with Random Assignment

Success in LotteryApplicant Characteristic Coef. SEFemale -0.001 (0.004)Application NumberA 0.001 (0.003)Travel Party NumberA 0.005 (0.006)Year of Birth 0.000 (0.000)Married 0.009 (0.008)Middle School -0.001 (0.005)High School 0.000 (0.006)Intercollege and Up 0.002 (0.008)Branch of Application 0.000 (0.000)Provided Phone Number -0.001 (0.011)Constant 1.142 (0.264)Observations 134,948 R-squared 0.02 Joint F-Test of Individual Characteristics (p-value) 0.98

Data and Survey

Surveyed 1605 applicants to January 2006 Hajj from northern districts of Punjab province

Survey took place summer to early fall 2006 Early wave: June to mid-July Late wave: September to mid-October No baseline

Sample to get roughly equal winners/losers; male/female

Lottery Compliance

Hajj Lottery and Actual Hajj Travel

Randomization Status

2006 Hajj StatusSuccessful

Unsuccessful

Went 99% 11%

Did not go 1% 89%

LATE Estimate Estimate:

Ykic = Outcomes of interest

HAJJic is an indicator for whether individual i went on Hajj in 2006.

c is a cell fixed effect. Cells: place of departure X accommodation category X group size.

Instrument for HAJJic with indicator Successi

LATE = average impact on lottery compliers

Captures impact of Hajj provided no direct lottery effect

In our case LATE roughly captures counterfactual impact of Hajj on a person who was unable to go due to losing the lottery

iccick

ick HAJJY

Identification Strategy: Advantages & Limits

Internal validity: randomization helps overcome OVB problem in estimating impact of Hajj on applicants

External validity: other populations Higher SES Shia Other regions/countries Nonapplicants; nonmarginal

External validity: other time scales We look at 5-8 months after Immediate Long-run

External Validity

Overall similar Applicants more likely to be older, married Literacy, high school, and log expenditure similar Excludes tails of distribution: slightly higher literacy but lower

college

Adult Pakistani

Population (restricted >20 year old)

Full Sample Mean Std. Dev. Mean Std. Dev.Age 40.16 16.244 54.575 13.240Female 0.499 0.500 0.490 0.500Married 0.703 0.497 0.943 0.232Illiterate 0.482 0.458 0.402 0.490Intercollege and Up 0.201 0.43 0.178 0.383City 0.400 0.490Periurban/large village 0.274 0.460Rural 0.325 0.470Ballot Success 0.533 0.499Monthly Expenditures 8.678 0.641 8.832 0.783

Survey Completion

Full sample: 63% overall completion rate 8% refused 26% could not be located/contacted 2% dead

Completion rate among applicants we could contact: 87%

Possible concern: differential completion rates 67% of successful completed compared to 60% of unsuccessful Successful show slightly lower refusal rates and more likely to be

found Reject null hypothesis that rates are equal, p = 0.01

But no difference along observables

Survey Completion, ctd. Restricted sub-sample:

Balanced on survey completion and reasons for non-completion Difference in ease of survey across regions, quality of survey team Excludes tehsils with more than 25 selected applicants where

completion rate for successful applicants >7% higher than that for unsuccessful applicants

Excludes 9/49 tehsils 81% of overall sample 66% of winners completed compared to 64% of losers. Fail to reject

null that completion rates are equal, p=0.66

Robustness: Redo results with demographic controls, etc. Redo results in restricted subsample

Observables Do Not Predict Lottery Success in Our Samples

Success in Lottery Among

Interviewed

Success in Lottery Among Interviewed,

Restricted Sub-sample

Applicant Characteristic Coef. SE Coef. SEFemale -0.014 (0.021) -0.026 (0.024)Application NumberA -0.009 (0.014) -0.015 (0.016)Travel Party NumberA 0.094 (0.066) 0.037 (0.072)Year of Birth 0.001 (0.001) 0.001 (0.001)Married -0.015 (0.062) 0.006 (0.073)Middle School -0.027 (0.036) -0.012 (0.041)High School -0.047 (0.045) -0.050 (0.051)Intercollege and Up -0.009 (0.051) -0.006 (0.060)Branch of Application 0.000 (0.000) 0.000 (0.000)Provided Phone Number 0.063 (0.060) 0.094 (0.064)Constant -1.863 (2.394) -1.995 (2.725)Observations 1,605 1,295 R-squared 0.06 0.10 Joint F-Test of Individual Characteristics (p-value) 0.68 0.81

Thematic Indices

Average effect size (AES) Also report averages over individual

variables, components

Global Islamic Practice Praying five times daily Performing Tasbih Praying in congregation Praying Tahajjud Namaz Reciting Qu’ran Religious dialogue

Regarded as Religious Do others regard you as religious?

Base Controls

Restricted

Sub-Sample

(1) Regarded as Religious 0.238*** 0.230***0.258*** (0.06) (0.055) (0.061)(2) Global Islamic Practice 0.163*** 0.166***0.171*** (0.030) (0.029) (0.033)

Religion

Localized religious practices Sufi teachers Amulets Chaliswan (40-day ceremony following a death) Maulud Mehfil Prayer cap Widow remarriage views

Base Controls

Restricted

Sub-Sample

(3) Belief in Localized Muslim Practices -0.101***

-0.094*** -0.074**

(0.032) (0.031) (0.035)(4) Participation in Localized Muslim Practices -0.097** -0.097** -0.085* (0.046) (0.045) (0.052)

Religion, ctd.

Positive views of others Countries

Is your view of Saudi people very positive to very negative?

Biggest effect on Indonesia No effect on Europeans

Religious/social/ethnic equality Religion effect biggest

Harmony Religious, social, ethnic

Can people of different groups live in harmony?

Religion effect biggest Praying in mosque of

different school Particularly large effect

Base ControlsRestricted Sub-Sample

(1) Views of Other Countries 0.150*** 0.147*** 0.151*** (0.04) (0.04) (0.04)(2) Views of Other Groups 0.131*** 0.108** 0.122 ** (0.05) (0.05) (0.06)(3) Harmony 0.128*** 0.117*** 0.126*** (0.04) (0.04) (0.05)

Tolerance

Peaceful Inclination Opposing:

Osama Bin Laden--Hajj doubles opposition (7% to 13%)

Physical punishment: 17% effect

Suicide bombing Supporting:

Importance of peace with India: 14% effect

Political Islam Religion’s role in the state Sig effects on (not) enforcing

religion (not) having religious leaders dispense justice

Views of West Oppose West social values/tech West against interests of Pak. Western/Jewish role in 9/11


(4) Peaceful Inclination 0.111*** 0.121*** 0.128*** (0.03) (0.03) (0.04)(5) Political Islam Index -0.050 -0.044 -0.043 (0.04) (0.03) (0.04)(6) Views of West 0.029 0.039 0.011 (0.04) (0.04) (0.04)

Tolerance, ctd.

Views: relative to men Intellectual Morally Spiritually

Biggest effect (+51%)

Quality of life Relative to Pakistan:

Saudi Arabia (smallest incr.) Indonesia/Malaysia (biggest incr.) West (second biggest incr.)

Crimes against women in Pakistan

Pak. relatively conservative

Base Controls

Restricted Sub-

Sample

(1) Views toward Women0.120*** 0.116*** 0.139***

(0.04) (0.04) (0.04)

(2) Women's Quality of Life0.158*** 0.138*** 0.166***

(0.05) (0.05) (0.06)

Gender

Education for women School for girls

Increase off high base Coeduction

8% incr. both girls/boys Equal schooling girls/boys

(no effect)

Women in workforce and professions Daughters and grand-

daughters Working (up 8%) Professional job (up 7%) Daughter-in-law have job

(up 12%)

Base Controls

Restricted Sub-

Sample(3) Girls' Education 0.092** 0.089** 0.097** (0.04) (0.04) (0.04)(4) Women in Workforce/Professions 0.119*** 0.112*** 0.091** (0.04) (0.04) (0.04)

Gender, ctd.

Islamic rules Unequal inheritance Women seeking

divorce Unequal witnesses

Household Management Number of children Marriage against

parents wishes

Base Controls

Restricted Sub-

Sample(5) Gender Authority -0.005 -0.010 0.005 (0.02) (0.02) (0.03)

Gender, ctd.

K6 Index 6 questions on negative

feelings in past month e.g. last month felt

hopeless, restless, worthless

Positive Feelings Joyful, peaceful, content feelings

Overall Life and Finances No evidence for income effect Also no effect on monthly consumption

Base Controls

Restricted

Sub-Sample

(1) Rescaled K6 Index-0.206***

-0.206*** -0.200***

(0.05) (0.05) (0.05)(2) Positive Feelings -0.109** -0.098** -0.079 (0.05) (0.04) (0.05)(3) Index of Satisfaction with Life and Finances -0.010 0.006 0.011 (0.04) (0.04) (0.04)

Well-Being

Gender difference K6 effect primarily for women

Main EffectMale

Interaction(4) Rescaled K6 Index -0.369*** 0.326*** (0.08) (0.09)(5) Positive Feelings -0.149** 0.079 (0.07) (0.08)

Well-Being, ctd.

Physical health Lack of injury/illness and self-

reported health Self-reported current

measure, 5-8 mos. post Hajj Rate of having been ill/injured

nearly doubles

Gender difference Mainly on women, but still -0.1 SD for

men Physical/mental

connection? Hard to rule out but think unlikely to

be only factor

Base Controls

Restricted

Sub-Sample

(6) Self-Rated Physical Health-0.213***

-0.219*** -0.239***

(0.05) (0.05) (0.06)

Main Effect

Male Interactio

n

(7) Self-Rated Physical Health-0.320*** 0.210**

(0.07) (0.10)

Well-Being, ctd.

Possible Mechanisms? Status Change? (little/no personal change during Hajj)

Hajji’s status changes after return and so responds differently Unlikely: no change in socio-economic or political engagement in

medium term

Political engagement Voting, party membership, interest in

and following of national affairs, opinions about politicians

Hajjis less likely to follow national affairs Otherwise no effects outcome-by-

outcome Only medium run, though

Socioeconomic engagement Receive/make visits Give advice to others Participate in religious, prof, ed.

organizations Employment status

Small positives on out-of-town visitors and self-employment


(1) Socioeconomic Engagement -0.002 -0.008 0.011 (0.02) (0.02) (0.02)(2) Engagement in Politics -0.011 -0.010 -0.024 (0.03) (0.03) (0.03)

Engagement and Exposure

Saudi Influence?

Increased tolerance of different sects Gender views

Formal Islamic Knowledge?

Transformation via Islamic knowledge? Mix of easy to difficult questions about the Qu’ran Outcome-by-outcome effects small and insignificant


(3) Formal Knowledge of Islam 0.004 0.000 -0.003 (0.04) (0.03) (0.04)

Exposure

Exposure and Group Observation/Interaction on Hajj? Increase in experiential knowledge (diversity,

gender, global)

Diversity awareness Divorce/prayer cap questions whose

answers depend on schools of thought Global awareness

Countries bordering Pakistan, % Nigerians Muslim, largest countries

Gender awareness Gender and marriage issues within

Islam (dowry, marriage to Christians/Jews, Hudood)

Having an opinion on women’s issues


(4) Diversity Knowledge 0.146*** 0.139*** 0.133***

(0.04) (0.04) (0.05)

(5) Gender Knowledge 0.125*** 0.116*** 0.104**

(0.04) (0.03) (0.04)

(6) Global Knowledge 0.083** 0.086** 0.072

(0.04) (0.03) (0.05)

Exposure, ctd.

Exposure and Group Observation/Interactions on Hajj, ctd.

Hajjis return with more positive views of Indonesians (0.32 SD) and Saudis (0.14 SD).

Smaller Hajj parties (more exposed?): Bigger gains in Islamic gender, global, and (same

sign but not sig) diversity awareness indices Larger increase in positive views of other countries

(signs on other tolerance indices also in same direction)

Some evidence for bigger shift away from local religious

Global Influence? Results on tolerance, gender, suggest Hajjis influenced by pilgrims

from around globe, not just Saudis

Social psychology of group interaction Empathy versus antipathy to others Depends on context: cooperative versus

competitive

Hajj as “glue” that makes for a cooperative environment? Common goal for all, overall unity emphasized More than that? Religion legitimizes/limits change

Conclusions

Micro mechanisms that facilitate strengthening of common identity Removal from usual environment Change in dress; physical appearance Common ritual Change in title Similar mechanisms in other social institutions that build common identity?

Medical education; police training; military basic training; international peace camps – similar effect?

Macro: role of pilgrimage in evolution of belief systems Pilgrimages very common How survive in competition among religions? Club goods models of religion

Practices screen out people who provide lower public goods BUT evolutionary pressures toward productive signals in competitive religious market

World religions: how to adapt to local conditions while retaining coherence--danger of division and schism Hierarchy one solution (not present in Sunni Islam) Hajj (and single text, etc.) may be another – convergence Elimination of non-Islamic influence from the Hajj What is impact on pilgrims from other areas?

END

END

Extra/Old stuff

Iannaccone (1992)

Q: Why do religions often require adherents to perform costly activities (e.g., behavior prohibitions, restrictions on dress, pilgrimages)?

A: Religions are clubs in which people receive positive externalities from committed co-religionists. Costly activities screen for commitment.

Viewing Hajj through lens of Ianaccone model Hajj involves substantial personal sacrifice

Financial (>$2000 in sample) Time (40 days average in sample) Physical and mental strain

Costly activities like Hajj could flourish in competitive marketplace for adherents if useful to the religion E.g., by promoting its survival

Islam is extraordinarily successful as a world religion. Has survived, flourished under competitive pressure.

Is Hajj just about screening, or does it have a treatment effect? If so, what?

Hajjis are exposed to Muslims from many other countries. Hajj may thereby promote a standardization of religious belief

Suppose a group of dynasties indexed by i. Let the religious beliefs and practices of dynasty i in generation t be xi,t.

Suppose that xi,t+1 = αxi,t + (1-α)xt +εi,t+1 , where xt is the average belief/practice of all Muslims and εi,t+1 is an i.i.d. random variable.

The Hajj reduces α.

It is straightforward to show var (xi,t+1 ) = var(xi,t)+2 and that in the steady state

This implies that the Hajj reduces variance in belief.

x2

2

1 2

Application and Selection Process

1. Applications made at bank branches around Pakistan. Deposits submitted

2. Applicants grouped into parties of up to 20. 3. Application data entered into computer. 4. Parties chosen randomly to win the ballot.

• Selection continues until total number of visas to be allocated is exhausted.

• Selection nominally stratified by place of departure and accommodation category, sect

5. Data we have was generated immediately after ballot, before anyone found out the results.

6. Results are announced. 7. Some adjustments may be need; for example following withdrawals.

Survey Sample

Random sample of Sunni applicants to the 2006 Hajj

Stratified by gender and randomization into win/loss status

Northern districts of Punjab province chosen for convenience

Interview status by lottery outcome

Total Successful Unsuccessful Total Successful UnsuccessfulSelected for interview 2512 1266 1246 1896 955 941

Raw completed interviews 1605 859 766 1269 652 617Completion rate 63.9% 67.9% 61.5% 66.9% 68.3% 65.6%

Not completedDead/ ill 2.0% 2.0% 2.0% 1.5% 1.5% 1.6%Lives elsewhere 9.9% 9.4% 10.4% 9.3% 9.5% 9.2%Not found 7.9% 6.1% 9.8% 7.9% 6.8% 9.0%Not home 7.6% 8.3% 7.0% 7.8% 8.4% 7.1%Refused 7.5% 6.0% 9.0% 6.3% 5.4% 7.4%

Table 5: Survey Completion Statistics

Lottery StatusPanel A: Full Sample Panel B: Subsample

Lottery Status

Survey Development Survey development based on literature,

discussions with scholars, pretesting

Wide range of questions on health, religion, social relations, political beliefs, and economic outcomes.

Effects on individual and on others’ perceptions, Individual and family.

Survey Mechanics

Survey administered at the household level.

Randomly select a party, and then a member of that party. If they live with a party member of the opposite gender, select them also. Average 1.6 applicants per household.

Survey approximately 2.5 hours duration.

Two survey teams worked out of Rawalpindi and Faisalabad cities.

Areas varied in the difficulty of surveying.

Introduction

We present some initial results that speak to the role of Hajj as an institution within Islam.

These results relate to: A literature that suggests Hajj may promote orthodoxy, unity with

Muslims worldwide. Debate on local practices in South Asia, elsewhere. Concerns regarding relations between Muslims and others and

the role of Hajj.

Engagement and exposure

Participation and interest in political and national affairs

Social interaction Formal religious knowledge Awareness of diversity within Islam Views on gender and marriage in Islam Global awareness

Views on Tolerance

How do Hajjis view people from other countries and groups?

How do Hajjis view the possibility of social cooperation and compromise across groups?

Moving toward what?

Political/Radical Islam?Unlikely (religion & govt views unchanged)

Saudi practice?Possible BUT

Religious tolerance; later gender results

Original intent vs Global norm?Hard to separate

Survey Summary Statistics

Mean Std. Dev. Mean Std. Dev.Age 54.575 13.240 55.347 13.246Female 0.490 0.500 0.494 0.500Married 0.943 0.232 0.951 0.218Illiterate 0.402 0.490 0.424 0.494Middle School Completer 0.253 0.435 0.255 0.436High School Graduate 0.167 0.373 0.167 0.373College Graduate 0.178 0.383 0.154 0.361City 0.400 0.490 0.361 0.481Periurban/ large village 0.274 0.460 0.284 0.451Rural 0.325 0.470 0.355 0.479Ballot Success 0.533 0.499 0.516 0.500

Table 7: Summary StatisticsFull Sample Subsample

Robustness

Controls Individual characteristics, district dummies, surveyor and

supervisor dummies Only coefficient to drop of views of other social groups

Balanced subsample Few changes – nothing statistically different Local beliefs and practices, global awareness smaller magnitude &

not significant. But Subsample doesn’t cover all regions--heterogeneity?

Related literature

Beliefs and Exposure How immutable are (adult) beliefs on religion, attitudes

towards others? What induces/facilitates change? Exposure & Experience

Group interactions – Social Psychology (Pettigrew and Tropp, 00; Stephan, 78)

Lab settings – Empathy (Cooperative settings) vs. Antipathy (Competitive)

Religion as the necessary “glue” (Putnam, 07)?

Institutions and belief systems Order w/o hierarchy - Hajj & Islamic cohesion?

Results Summary: Religion:

Global religious practices increase relative to locally specific religious beliefs & practices.

Tolerance More positive on other countries and social groups; belief in

compromise Belief in nonviolence; no increase in political Islam or West suspicion

Gender: Improved views of women; view rel. quality of women’s lives in Pakistan

lower Increased desire for girls to be educated and have professions No challenge to male authority in household matters, etc.

Health: Psychological strain for women & worse physical health

Channels? Engagement & Exposure - interaction in context established by Hajj? Increase knowledge of Islamic diversity, gender, and the world No change in formal religious knowledge No evidence of change in social role once return

Survey Topics

Religious knowledge and practice Tolerance Views on gender Physical and mental health Political involvement and beliefs Social interaction and roles Business and employment

Results Summary: Religion:

Global religious practices increase relative to locally specific religious beliefs & practices.

Tolerance More positive on other countries and social groups; belief in

compromise Belief in nonviolence; no increase in political Islam or West suspicion

Gender: Improved views of women; view rel. quality of women’s lives in Pakistan

lower Increased desire for girls to be educated and have professions No challenge to male authority in household matters, etc.

Health: Psychological strain for women & worse physical health

Channels? Engagement & Exposure - interaction in context established by Hajj? Increase knowledge of Islamic diversity, gender, and the world No change in formal religious knowledge No evidence of change in social role once return

Documents

The Impact of the Hajj Religion and Tolerance in Islam’s Global Gathering