Internal educational and economic migration and the presence of universities in Peru

Variable	Description	Sample Mean	Sample Std Dev	Estimated Mean	Estimated Std Error
Dependent variable
migrant	Dichotomous variable that takes the value of 1 if the household has migrated to a different department, and 0 if it is not a migrant household	0.024	0.152	0.023	0.001
Main explanatory variables (Xji from the model shown previously)
ccdd_univ_pm	Number of university campuses per million inhabitants in the department where the household is located	16.076	8.037	14.159	0.075
ccdd_hquniv_pm	Number of high-quality university campuses per million inhabitants in the department where the household is located	3.454	4.934	4.328	0.055
Household characteristics (c1 from the model shown previously)
head_age	Age of the head of the household	49.463	13.679	49.894	0.148
head_educ	Years of education of the head of the household	8.858	4.835	9.089	0.050
head_employment	Dichotomous variable that takes the value of 1 if the head of the household is employed, and 0 otherwise	0.868	0.338	0.849	0.004
dep_ratio	Dependence ratio between the number of unemployed and employed household members	0.356	0.233	0.345	0.002
ln_income	Yearly log-income (PEN)	10.351	0.788	10.452	0.008
ln_rent_exp	Yearly rent log-expenditure (PEN). For households that do not pay rent or own their homes, we impute the estimated income that would be earned if the house were rented out	7.684	1.155	7.908	0.0127
Department characteristics (c2 from the model shown previously)
ccdd_rural	Percentage of people residing in rural areas in the destination department	0.257	0.192	0.198	0.003
ccdd_water	Percentage of people with access to public network water supply in the destination department	0.848	0.105	0.869	0.001
ccdd_employment	Employment rate of the destination department	0.947	0.027	0.936	0.000
ccdd_crime	Percentage of population older than 14 that have been the victim of a crime in the destination department	0.234	0.076	0.252	0.001

[i] Source: own study

Table 2.

Model estimation results with the number of universities per million inhabitants, including Lima

	With Lima
	All households				Households comprising members with higher education
Variables	(1)	(2)	(3)	(4)	(5)	(6)	(7)	(8)
*Odds Ratios – variables of interest*
ccdd_univ_pm	1.021*** (0.003)		1.018*** (.003)	1.126 (.155)	1.017*** (.007)		1.017*** (.007)	.737 (.285)
ccdd_employment		6049.902*** (7929.715)	2766.937*** (3603.884)	10918.55 ***(28106.05)		4.343 (15.010)	1.943 (6.710)	.028 (.193)
ccdd_univ_pm ## ccdd_employment				.899 (.130)				1.401 (.568)
*Average Marginal Effects*
*ccdd_univ_pm*	.00036* (.00005)**		.00031* (.00005)**		*.00035 (.00015)**		*.00034 (.00015)**
*ccdd_employment*		.15353* (.02278)**	.13967* (.02270)**			.03037 (.07113)	.01373 (.07124)
head_age	−.00119*** (.00005)	−.00120*** (.00005)	−.00119*** (.00000)		−.00202*** (.00016)	−.00202*** (.00016)	−.00202*** (.00016)
head_educ	.00079*** (.00012)	.00080*** (.00012)	.00078*** (.00012)		−.00029 (.00042)	−.00027 (.00042)	−.00029 (.00042)
head_employment	−.00917*** (.00122)	−.00952*** (.00122)	−.00942*** (.00122)		−.00122 (.00447)	−.00135 (.00448)	−.00123 (.00448)
dep_ratio	−.01184*** (.00165)	−.01191*** (.00165)	−.01195*** (.00165)		−.01725** (.00552)	−.01713** (.00552)	−.01726** (.00555)
ln_income	−.0014* (0.0001)	−.00156** (.00061)	−.00164** (.00061)		−.00460* (.00196)	−.00458* (.00198)	−.00463* (.00196)
ln_rent_expenditure	.00156** (.00052)	.00182** (.00052)	.00187*** (.00052)		.07461*** (.01883)	.00746*** (.00182)	.00749*** (.00182)
ccdd_rural	.00261 (.00236)	−.00949** (.00278)	−.00809** (.00279)		.00034 (.00757)	−.00175 (.00888)	−.00030 (.00895)
ccdd_water	.00055 (.00399)	.00111 (.00416)	−.00017 (.00385)		.01053 (.01089)	.01227 (.01157)	.01042 (.01078)
ccdd_crime	−.02467*** (.00631)	−.01074 (.00617)	−.01518* (.00615)		.01689 (.01907)	.02306 (.01934)	.01782 (.01943)
Time FE	Yes	Yes	Yes		Yes	Yes	Yes
Sample	270,381	270,381	270,381		26,285	26,285	26,285
Wald Test – Statistic	85.84	84.50	84.34		28.59	28.35	27.14
Wald Test – Pvalue	0.00	0.00	0.00		0.00	0.00	0.00
Sensitivity	0.5619	0.5522	0.5684		0.6981	0.6949	0.6981
Specificity	0.8184	0.8233	0.8131		0.8402	0.8454	0.8402
Correctly classified	0.8130	0.8177	0.8079		0.8368	0.8418	0.8368
AUC	0.6901	0.6877	0.6907		0.7691	0.7701	0.7691

Notes: Delta-method or robust standard errors in parenthesis.

*** p<0.01,

** p<0.05,

* p<0.1

Source: own study

Table 3.

Model estimation results with the number of universities per million inhabitants, excluding Lima

	Without Lima
	All households				Households comprising members with higher education
Variables	(1)	(2)	(3)	(4)	(5)	(6)	(7)	(8)
*Odds Ratios – variables of interest*
ccdd_univ_pm	1.019*** (.003)		1.019*** (.003)	1.064 (.160)	1.019*** (.008)		1.019*** (0.008)	.865 (.354)
ccdd_employment		16500.63*** (28099.76)	12858.91*** (21430.55)	26545.38*** (87638.4)		93.032 (372.544)	61.370 (240.440)	3.996 (32.337)
ccdd_univ_pm ccdd_employment*				.956 (.151)				1.187 (.510)
*Average Marginal Effects*
*ccdd_univ_pm*	.00034* (.00005)**		.00034* (.00005)**		.00038 (.00014)**		*.00037 (.00014)**
*ccdd_employment*		.17368* (.03066)**	.16910* (.02999)**			.08940 (.07838)	.08110 (.07664)
head_age	−.00112*** (.00004)	−.00112*** (.00004)	−.00112*** (.00004)		−.00186*** (.00015)	−.00186* (.00015)**	−.00185* (.00015)**
head_educ	.00104*** (.00011)	.00106*** (.00011)	.00101*** (.00011)		.00007 (.00035)	.00009 (.00035)	.00007 (.00035)
head_employment	−.00843*** (.00118)	−.00880*** (.00121)	−.00867*** (.00122)		−.00080 (.00420)	−.00102 (.00420)	−.00112 (.00420)
dep_ratio	−.00815*** (.00154)	−.00826*** (.00154)	−.00830*** (.00154)		−.00245 (.00537)	−.00237 (.00539)	−.00256 (.00537)
ln_income	−.00077 (.00059)	−.00073 (.00060)	−.00088 (.00060)		−.00060 (.00208)	−.00053 (.00209)	−.00061 (.00207)
ln_rent_expenditure	.00110* (.00049)	.00118* (.00049)	.00117* (.00048)		.00380* (.00153)	.00388* (.00153)	.00339* (.00154)
ccdd_rural	.00122 (.00244)	−.00825** (.00273)	−.00593* (.00274)		.00480 (.00758)	−.00154 (.00814)	.00119 (.00815)
ccdd_water	−.00006 (.00390)	−.00067 (.00420)	−.00276 (.00384)		.00983 (.01030)	.01143 (.01121)	.00873 (.01021)
ccdd_crime	−.02105** (.00680)	−.01276 (.00658)	−.01931** (.00667)		.00470 (.01981)	.01374 (.01941)	.00591 (.01978)
Time FE	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes
Sample	239,639	239,639	239,639	239,639	22,699	22,699	22,699	22,699
Wald Test – Statistic	111.04	107.64	107.20		23.50	23.35	22.40
Wald Test – Pvalue	0.00	0.00	0.00		0.00	0.00	0.00
Sensitivity	0.6981	0.5687	0.5870		0.7106	0.7050	0.7124
Specificity	0.8402	0.8163	0.8045		0.8326	0.8384	0.8323
Correctly classified	0.8368	0.8110	0.7999		0.8297	0.8352	0.8295
AUC	0.6939	0.6925	0.6958		0.7716	0.7717	0.7724

Notes: Delta-method or robust standard errors in parenthesis.

*** p<0.01,

** p<0.05,

* p<0.1

Source: own study

Table 4.

Model estimation results with high-quality universities per million inhabitants

	With Lima				Without Lima
	All households		Households comprising members with higher education		All households		Households comprising members with higher education
Variables	(1)	(2)	(3)	(4)	(5)	(6)	(7)	(8)
*Odds Ratios – variables of interest*
ccdd_hquniv_pm	1.00883 (.00551)	1.76362* (.42924)	1.02472 (.01415)	3.15456 (1.93880)	1.00679 (.00560)	1.51493 (.39645)	1.02565 (0.01474)	2.17189 (1.39837)
ccdd_employment	8489.087*** (11343.70000)	334249.50*** (661539.80)	10.67778 (37.84495)	30282.80* (155374.50})	21476.780*** (36577.700)	153989.80*** (320713.90)	276.35070 (1097.77899)	13048.570 (66789.360)
ccdd_hquniv_pm ## ccdd_employment		.55533* (.14235)		.30605 (.19799)		.65026 (.17918)		0.45388 (0.30737)
*Average Marginal Effects*
*ccdd_hquniv_pm*	.00016 (.00010)		.00050 (.00029)		0.00012 (0.00010)		0.00045 (0.00028)
*ccdd_employment*	.15950* (.02325)**		.04895 (.07272)		0.17839* (0.02999)**		0.11080 (0.07751)
head_age	−.00119*** (.00049)		−.00202*** (.00016)		−0.00112*** (0.00004)		−0.00186* (0.00015)**
head_educ	.00080*** (.00012)		−.00026 (.00040)		0.00106*** (0.00011)		0.00090 (0.00035)
head_employment	−.00951*** (.00122)		−.00129 (.00448)		−0.00879*** (0.00122)		−0.00113 (0.00420)
dep_ratio	−.01195*** (.00165)		−.01719** (.00552)		−0.00381*** (0.00154)		−0.00260 (0.00539)
ln_income	−.00159** (.00061)		−.00466* (.00198)		−0.00075 (0.00060)		−0.00060 (0.00208)
ln_rent_expenditure	.00178** (.00052)		.00731*** (.00184)		0.00119* (0.00048)		0.00371* (0.00155)
ccdd_rural	−.00876** (.00273)		.00125 (.00877)		−0.00795** (0.00274)		0.00087 (0.00815)
ccdd_water	.00394 (.00457)		.02112 (.01335)		0.00172 (0.00384)		0.02108 (0.01346)
ccdd_crime	−.01342* (.00652)		.01529 (.02077)		−0.01462** (0.00667)		0.00612 (0.02069)
Time FE	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes
Sample	270.381	270.381	26.285	26.285	239.639	239.639	22.699	22.699
Wald Test – Statistic	79.98		27.00		101.92		22.15
Wald Test – Pvalue	0.00		0.00		0.00		0.00
Sensitivity	0.5506		0.6949		0.5656		0.7069
Specificity	0.8239		0.8449		0.8161		0.8387
Correctly classified	0.8182		0.8414		0.8108		0.8356
AUC	0.6872		0.7699		0.6908		0.7728

Notes: Delta-method or robust standard errors in parenthesis.

*** p<0.01,

** p<0.05,

* p<0.1

Source: own study

Table A.

Variance Inflation Factor

Variable	(1) VIF (all population)	(2) VIF (filtered & no Lima)	(3) VIF (all population)	(4) VIF (filtered & no Lima)
*ccdd_univ_pm*	1.20	1.22
*ccdd_hquniv_pm*			1.48	1.44
head_age	1.64	1.32	1.64	1.31
head_educ	1.76	1.46	1.76	1.46
head_employment	1.23	1.10	1.23	1.10
dep_ratio	1.25	1.07	1.26	1.07
ln_income	2.01	1.75	2.01	1.75
ln_rent_exp	2.46	1.93	2.48	1.94
ccdd_rural	2.53	1.94	2.44	1.90
ccdd_water	1.78	1.81	2.32	2.36
ccdd_employment	2.48	2.00	2.48	2.00
ccdd_crime	1.47	1.38	1.51	1.38
year: 2016	1.86	1.81	1.87	1.81
year: 2017	1.87	1.79	1.87	1.80
year: 2018	1.96	1.86	1.98	1.88
year: 2019	1.94	1.83	1.93	1.82
year: 2020	2.89	2.54	2.88	2.51
year: 2021	2.47	1.90	2.47	2.10
year: 2022	2.07	1.86	2.03	1.84
Mean VIF	2.00	1.74	2.05	1.78

[i] Source: own study

Table B.1

Odds-ratio results for models with universities per million inhabitants

	With Lima				Without Lima
	All households		Households comprising members with higher education		All households		Households comprising members with higher education
Variables	(1)	(2)	(3)	(4)	(5)	(6)	(7)	(8)
*ccdd_univ_pm*	1.01759* (0.00279)**	1.12569 (0.15495)	1.01653 (0.00736)**	0.73716 (0.28488)	1.01899* (0.00289)**	1.06403 (0.16012)	1.01886 (0.00759)**	0.86478 (0.35445)
*ccdd_employment*	2766.937* (3603.88400)**	10918.55* (28106.05)**	1.94268 (6.70959)	0.02760 (0.19322)	12858.91* (21430.55)**	26545.38* (87638.40)**	61.36969 (240.43990)	3.99559 (32.33698)
*ccdd_univ_pm ## ccdd_employment*		0.89949 (0.13001)		1.40072 (0.56800)		0.95566 (0.15068)		1.18749 (0.51037)
head_age	0.93459*** (0.00215)	0.93458*** (0.00214)	0.90706*** (0.00507)	0.90713*** (0.00505)	0.93932*** (0.00199)	0.93932*** (0.00192)	0.91039*** (0.00554)	0.91048*** (0.00554)
head_educ	1.04553*** (0.00736)	1.04545*** (0.00736)	0.98623 (0.02015)	0.98650 (0.02016)	1.06049*** (0.00653)	1.06046*** (0.00653)	1.00333 (0.01801)	1.00355 (0.01804)
head_employment	0.58603*** (0.03943)	0.58600*** (0.03943)	0.94201 (0.20368)	0.94114 (0.20351)	0.61578*** (0.04103)	0.61580*** (0.04106)	0.94458 (0.20147)	0.94394 (0.20139)
dep_ratio	0.50759*** (0.04655)	0.50789*** (0.04659)	0.43396*** (0.11279)	0.43090*** (0.11240)	0.62843*** (0.05365)	0.62851*** (0.05366)	0.87831 (0.23941)	0.87559 (0.23955)
ln_income	0.91117*** (0.03145)	0.91097*** (0.03144)	0.79935** (0.07627)	0.79948** (0.07632)	0.95237 (0.03128)	0.95247 (0.03173)	0.96708 (0.10191)	0.96579 (0.10155)
ln_rent_expenditure	1.11169*** (0.03265)	1.11234*** (0.03277)	1.43688*** (0.12525)	1.43359*** (0.12262)	1.06782** (0.02916)	1.06800** (0.02916)	1.21519** (0.09287)	1.21408** (0.09312)
ccdd_rural	0.63201*** (0.09926)	0.64357*** (0.10060)	0.98557 (0.42627)	0.92070 (0.39555)	0.71750** (0.11007)	0.72039** (0.11095)	1.06213 (0.43885)	1.04383 (0.43583)
ccdd_water	0.99950 (0.21648)	1.00484 (0.22203)	1.65523 (0.85832)	1.57893 (0.82309)	0.85703 (0.18483)	0.86149 (0.18704)	1.55745 (0.80951)	1.52543 (0.79660)
ccdd_crime	0.42264** (0.14889)	0.43665** (0.15623)	2.36848 (2.20126)	2.16734 (2.06353)	0.33938*** (0.12756)	0.34267*** (0.12972)	1.34993 (1.35003)	1.30759 (1.33685)
Time FE	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes
Sample	270,381	270,381	26,285	26,285	239,639	239,639	22,699	22,699

[i] Source: own study

Table B.2

Odds ratio results for models with high quality universities per million inhabitants

	With Lima				Without Lima
	All households		Households comprising members with higher education		All households		Households comprising members with higher education
Variables	(1)	(2)	(3)	(4)	(5)	(6)	(7)	(8)
*ccdd_hquniv_pm*	1.00883 (.00551)	*1.76362 (.42924)**	1.02472 (0.0141494)	*3.15456 (1.93880)**	1.00679 (0.00560)	1.51493 (0.39645)	*1.02565 (0.01474)**	2.17189 (1.39837)
*ccdd_employment*	8489.087* (11343.700)**	334249.5* (661539.8)**	10.67778 (37.84495)	30282.8 (155374.5)**	21476.78* (36577.7)**	153989.8* (320713.9)**	276.35070 (1097.77900)	*13048.57 (66789.4)**
*ccdd_hquniv_pm ## ccdd_employment*		*.55533 (.14235)**		*0.30605 (0.19799)**		0.65026 (0.17918)		0.45381 (0.30737)
head_age	.93448*** (.00214)	.93447*** (.00214)	0.90709*** (0.00507)	0.90693*** (0.00508)	0.93909*** (0.00192)	0.93906*** (0.00192)	0.91011*** (0.00555)	0.91001*** (0.00556)
head_educ	1.04662*** (.00737)	1.04634*** (.00737)	0.98757 (0.02021)	0.98673 (0.02247)	1.06130*** (0.00553)	1.06100*** (0.00553)	1.00502 (0.01804)	1.00430 (0.01806)
head_employment	.58301*** (.03915)	.58398*** (.03924)	0.93960 (0.20311)	0.94607 (0.20493)	0.61184*** (0.04066)	0.61273*** (0.04073)	0.94451 (0.20131)	0.94982 (0.20163)
dep_ratio	.50763*** (.04656)	.50728*** (.04652)	0.43532*** (0.11345)	0.43428*** (0.11303)	0.62848*** (0.05364)	0.62849*** (0.05361)	0.87647 (0.23997)	0.87721 (0.23988)
ln_income	.91382** (.03161)	.91359 (.03160)	0.79810** (0.07630)	0.79699** (0.07602)	0.95879 (0.03198)	0.95835 (0.03196)	0.96982 (0.10266)	0.96767 (0.10273)
ln_rent_expenditure	1.10481** (.03243)	1.06715* (.02955)	1.42479*** (0.12209)	1.43050*** (0.12171)	1.06519** (0.02927)	1.06722** (0.02955)	1.20685** (0.09287)	1.21224** (0.09307)
ccdd_rural	.60845** (.09474)	.59282** (.09174)	1.06231 (0.45069)	0.98286 (0.46769)	0.64118*** (0.09856)	0.64196*** (0.09818)	1.00441 (0.41512)	1.00690 (0.41240)
ccdd_water	1.25070 (.32367)	1.20696 (.31357)	2.77830* (1.77982)	2.66212* (1.70838)	1.10064 (0.28926)	1.07988 (0.28417)	2.91389 (1.98477)	2.82990 (1.92127)
ccdd_crime	.46701* (.17386)	.50528 (.18790)	2.09499 (2.08826)	2.56419 (2.56214)	0.44147** (0.17005)	0.46980* (0.18132)	1.36419 (1.42652)	1.56543 (1.65093)
Time FE	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes
Sample	270,381	270,381	26,285	26,285	239,639	239,639	22,699	22,699

[i] Source: own study

References

Authors

Metrics

Articles in this issue

DOI: https://doi.org/10.2478/mgrsd-2025-0049 | Journal eISSN: 2084-6118 | Journal ISSN: 0867-6046 (formerly 2084-6118)

Journal RSS Feed

Language: English

Page range: 135 - 148

Submitted on: Mar 21, 2025

Accepted on: Oct 9, 2025

Published on: Apr 20, 2026

Published by: Faculty of Geography and Regional Studies, University of Warsaw

In partnership with: Paradigm Publishing Services

Keywords:

Related subjects:

© 2026 Carolina Guevara, Marco Ríos-Luna Ruiz, José Luis Herrera-Hinojosa, published by Faculty of Geography and Regional Studies, University of Warsaw
This work is licensed under the Creative Commons Attribution 4.0 License.

Volume 30 (2026): Issue 3 (July 2026)

Internal educational and economic migration and the presence of universities in Peru

Figures & Tables

Figure 1.

Table 1.

Table 2.

Table 3.

Table 4.

Table A.

Table B.1

Table B.2

Paradigm

My account