Good Schools Make Good Neighbors: Introduction Human Capital - - PowerPoint PPT Presentation

Good Schools Make Good Neighbors: Human Capital Spillovers in Early 20th Century Agriculture John Parman Introduction Human Capital and Agriculture Iowa’s Farms and Schools Data Results

Private Returns Spillovers Spillovers and Social Networks

Conclusions

Good Schools Make Good Neighbors: Human Capital Spillovers in Early 20th Century Agriculture

John Parman February 23, 2010

Good Schools Make Good Neighbors: Human Capital Spillovers in Early 20th Century Agriculture John Parman Introduction Human Capital and Agriculture Iowa’s Farms and Schools Data Results

Private Returns Spillovers Spillovers and Social Networks

Conclusions

Brief Overview

Public school expansion in the American Midwest

• ccurred at a time when human capital was becoming

increasingly important in agriculture. The public school system in Iowa developed with a focus on improving agricultural productivity. Estimates of the private returns to education for farmers suggest that Iowa achieved this goal; an additional year

• f high school increased a farmer’s earnings by 5%.

A farmer’s education also benefited his neighbors; an increase in the mean education level of his neighbors by

• ne year increased a farmer’s annual earnings by roughly

3%. The size of these spillovers depended on social group membership and a farmer’s own level of education.

Good Schools Make Good Neighbors: Human Capital Spillovers in Early 20th Century Agriculture John Parman Introduction Human Capital and Agriculture Iowa’s Farms and Schools Data Results

Private Returns Spillovers Spillovers and Social Networks

Conclusions

Outline of Presentation

Review of modern evidence of the role of education in agriculture Brief history of Iowa agriculture and education Construction of a dataset with education and income for farmers and their adjacent neighbors Estimates of the private returns to education in agriculture Estimates of human capital spillovers across farms Estimates of human capital spillovers within and across social networks Concluding remarks

Good Schools Make Good Neighbors: Human Capital Spillovers in Early 20th Century Agriculture John Parman Introduction Human Capital and Agriculture Iowa’s Farms and Schools Data Results

Private Returns Spillovers Spillovers and Social Networks

Conclusions

Previous studies of human capital and agriculture

6 2 3 4 5 Frequency 1 ‐3 ‐2 ‐1 1 2 3 4 5 6 Estimated percent increase in output with an additional year of schooling

Good Schools Make Good Neighbors: Human Capital Spillovers in Early 20th Century Agriculture John Parman Introduction Human Capital and Agriculture Iowa’s Farms and Schools Data Results

Private Returns Spillovers Spillovers and Social Networks

Conclusions

Previous studies of human capital and agriculture

The empirical evidence of the returns to education in agriculture is very mixed. In very stable environments, the returns to formal schooling are small. In these environments, experience appears more effective than schooling for improving productivity. Schooling is more important in dynamic settings: educated farmers are more likely to seek out information, experiment with and adopt new technologies and adapt to changing market conditions.

Good Schools Make Good Neighbors: Human Capital Spillovers in Early 20th Century Agriculture John Parman Introduction Human Capital and Agriculture Iowa’s Farms and Schools Data Results

Private Returns Spillovers Spillovers and Social Networks

Conclusions

Previous studies of human capital and agriculture

There is a small body of evidence suggesting that diffusion of information through social networks is important:

Bandiera & Rasul (2006): adoption of new crops depends on decisions of family and friends Conley & Udrey (2001): farmers learn from experimentation of members of their social network Foster & Rosenzweig (1995): farmers learn how to successfully adopt new seed varieties from neighbors’ experimentation Ryan & Gross (1943): hybrid seed corn diffusion in Iowa based largely on neighbors talking to each other

As with the formal schooling, these social networks will be more important when the agricultural sector is more dynamic.

Good Schools Make Good Neighbors: Human Capital Spillovers in Early 20th Century Agriculture John Parman Introduction Human Capital and Agriculture Iowa’s Farms and Schools Data Results

Private Returns Spillovers Spillovers and Social Networks

Conclusions

Innovation in Iowa Agriculture

Traditional view, Cochrane (1993): mechanization accounted for nearly all technological advance on American farms, latter half of the 19th century was not a period of innovation New view, Olmstead & Rhode (1993, 2002, 2008): significant portion of growth in labor productivity was from biological advances and settlement patterns Tasks facing the Iowa farmer: constructing proper drainage systems, applying advances in soil science, introducing new crops, breeding experiments with corn, pest and disease control, etc.

Good Schools Make Good Neighbors: Human Capital Spillovers in Early 20th Century Agriculture John Parman Introduction Human Capital and Agriculture Iowa’s Farms and Schools Data Results

Private Returns Spillovers Spillovers and Social Networks

Conclusions

A Basic Timeline of Iowa Education

1858: Iowa Agricultural College founded, center for research, educated rural teachers, directly educated farmers through short courses) 1862: Morrill Act passed, land grant institutions to “promote the liberal and practical education of the industrial classes” 1887: Hatch Act passed, federal land grants for agricultural experiment stations 1888: Iowa’s Agricultural Experiment Station founded 1914: Smith-Lever Act passed, established cooperative extension services School consolidation and the introduction of high schools is occurring between 1880 and 1910

Good Schools Make Good Neighbors: Human Capital Spillovers in Early 20th Century Agriculture John Parman Introduction Human Capital and Agriculture Iowa’s Farms and Schools Data Results

Private Returns Spillovers Spillovers and Social Networks

Conclusions

Rural Schools at the Turn of the Century

Good Schools Make Good Neighbors: Human Capital Spillovers in Early 20th Century Agriculture John Parman Introduction Human Capital and Agriculture Iowa’s Farms and Schools Data Results

Private Returns Spillovers Spillovers and Social Networks

Conclusions

Experimental Plots

Excerpt from a 1903 study guide for school garden work: Experimental Work

• 1. Plant potatoes at different depths, from just

under the surface to four inches. Which method gives best results?

• 2. Use seed pieces of varying sizes, from those with

no eye up to the whole potato.

• 3. Any difference in value of cutting from stem to

stem end and bud end?

• 4. Which is better, one large piece or several small
• nes?
• 5. What is the order of sprouting of the eyes?
• 6. How [to best] prevent potato scab?

Good Schools Make Good Neighbors: Human Capital Spillovers in Early 20th Century Agriculture John Parman Introduction Human Capital and Agriculture Iowa’s Farms and Schools Data Results

Private Returns Spillovers Spillovers and Social Networks

Conclusions

Constructing a Dataset of Linked Neighbors

1915 Iowa state census: income and education information (unique for pre-1940 period), occupation, ancestry, church affiliation, farm value Plat maps: farm boundaries, ability to measure areas and distances County superintendents of schools reports: public school locations, school characteristics Basic approach: digitize plat maps and link geographical data to census records and school district data

Good Schools Make Good Neighbors: Human Capital Spillovers in Early 20th Century Agriculture John Parman Introduction Human Capital and Agriculture Iowa’s Farms and Schools Data Results

Private Returns Spillovers Spillovers and Social Networks

Conclusions

1915 Iowa State Census

Good Schools Make Good Neighbors: Human Capital Spillovers in Early 20th Century Agriculture John Parman Introduction Human Capital and Agriculture Iowa’s Farms and Schools Data Results

Private Returns Spillovers Spillovers and Social Networks

Conclusions

Township Plat Maps

Good Schools Make Good Neighbors: Human Capital Spillovers in Early 20th Century Agriculture John Parman Introduction Human Capital and Agriculture Iowa’s Farms and Schools Data Results

Private Returns Spillovers Spillovers and Social Networks

Conclusions

Township Plat Maps

Good Schools Make Good Neighbors: Human Capital Spillovers in Early 20th Century Agriculture John Parman Introduction Human Capital and Agriculture Iowa’s Farms and Schools Data Results

Private Returns Spillovers Spillovers and Social Networks

Conclusions

Digitizing Plat Maps

Good Schools Make Good Neighbors: Human Capital Spillovers in Early 20th Century Agriculture John Parman Introduction Human Capital and Agriculture Iowa’s Farms and Schools Data Results

Private Returns Spillovers Spillovers and Social Networks

Conclusions

Digitizing Plat Maps

Good Schools Make Good Neighbors: Human Capital Spillovers in Early 20th Century Agriculture John Parman Introduction Human Capital and Agriculture Iowa’s Farms and Schools Data Results

Private Returns Spillovers Spillovers and Social Networks

Conclusions

Digitizing Plat Maps

Good Schools Make Good Neighbors: Human Capital Spillovers in Early 20th Century Agriculture John Parman Introduction Human Capital and Agriculture Iowa’s Farms and Schools Data Results

Private Returns Spillovers Spillovers and Social Networks

Conclusions

Linking Census Records to Maps

1 Each name from the maps is searched for in the census

(search criteria are the name and township)

2 If a unique match is found, an image of the census

record is transcribed and the farmer’s characteristics are linked to the farm polygon on the map

3 GIS software is then used to calculate acreage of each

farm and distance of farm to schools and towns

4 GIS scripts are used to identify adjacent neighbors and

calculate neighbors’ average earnings and educational attainment

Good Schools Make Good Neighbors: Human Capital Spillovers in Early 20th Century Agriculture John Parman Introduction Human Capital and Agriculture Iowa’s Farms and Schools Data Results

Private Returns Spillovers Spillovers and Social Networks

Conclusions

Sample Counties

Variable Chickasaw Poweshiek Ringgold State Average Number of farms 1,905 2,142 1,854 1,996

• Ave. farm size (acres)

152 160 168 161

• Ave. monthly wage paid

farm help, summer months 30.80 33.00 28.43 32.70

• Ave. monthly wage paid

farm help, winter months 20.69 25.72 28.77 24.61 Corn, acres 63,194 110,557 69,328 98,463 Corn, bushels per acre 3 38 23 28 Winter wheat, acres 179 860 13,245 5,929 Winter wheat, bushels per acre 16 23 9 19 Spring wheat, acres 1,607 780 6 1,495 Spring wheat, bushels per acre 12 14 7 14 Cattle, cows and heifers kept for milk 17,367 9,877 7,987 11,053 Statistics are compiled from the 1915 Annual Iowa Yearbook of Agriculture . Sample county agricultural statistics, 1915

Good Schools Make Good Neighbors: Human Capital Spillovers in Early 20th Century Agriculture John Parman Introduction Human Capital and Agriculture Iowa’s Farms and Schools Data Results

Private Returns Spillovers Spillovers and Social Networks

Conclusions

Farmer Characteristics

Variable Mean Standard Deviation Age 46.70 11.68 Annual earnings 1199.62 1175.47 Farm value 17439.32 13079.76 Incumbrance on farm 3357.70 5399.68 Farm acreage 153.03 106.92 Earnings per acre 9.80 10.77 Farm value per acre 126.08 90.98 Distance to nearest town (miles) 2.25 1.52 Distance to nearest high school (miles) 6.51 3.28 Foreign born (yes=1) 0.14 0.35 Number of neighbors 7.84 3.26 Number of observations

Notes: All dollar values are in 1915 dollars. Total schooling is defined as the sum of years of common school, grammar school, high school and college.

Farm owner characteristics, 1915 2410

Good Schools Make Good Neighbors: Human Capital Spillovers in Early 20th Century Agriculture John Parman Introduction Human Capital and Agriculture Iowa’s Farms and Schools Data Results

Private Returns Spillovers Spillovers and Social Networks

Conclusions

Farmers’ Educational Attainment

Variable Mean Standard Deviation Total schooling 8.43 2.65 Common school 7.91 2.65 Grammar school 0.23 1.26 High school 0.19 0.72 College 0.10 0.54 Total schooling in Iowa 5.77 4.46 Total schooling in US 6.30 4.31 Total schooling outside US 2.13 3.51 Graded schooling 0.52 1.80 Graded schooling in Iowa 0.41 1.53 Graded schooling in US 0.44 1.63 Graded schooling outside US 0.08 0.72

Notes: Total schooling includes years of common school, grammar school, high school and college. Graded schooling includes years of grammar school, high school and college only.

Educational attainment of farm owners, 1915

Good Schools Make Good Neighbors: Human Capital Spillovers in Early 20th Century Agriculture John Parman Introduction Human Capital and Agriculture Iowa’s Farms and Schools Data Results

Private Returns Spillovers Spillovers and Social Networks

Conclusions

Geographical Distribution of Total Schooling

Distribution of total schooling for Poweshiek county, 1915

Farmer's Years of Total Schooling

0-5 5-8 8-10 10-13 13+

School Locations

Towns with high schools Towns with graded schools

Good Schools Make Good Neighbors: Human Capital Spillovers in Early 20th Century Agriculture John Parman Introduction Human Capital and Agriculture Iowa’s Farms and Schools Data Results

Private Returns Spillovers Spillovers and Social Networks

Conclusions

Geographical Distribution of Earnings

Distribution of annual earnings for Poweshiek county, 1915

Farmer's Annual Earnings (1915 $)

0 - 200 201 - 578 579 - 852 853 - 1260 1261 - 1750 1751 - 2340 2341 - 3500 3501 - 5500 5501 - 9000 9001 - 21000

School Locations

Towns with high schools Towns with graded schools

Good Schools Make Good Neighbors: Human Capital Spillovers in Early 20th Century Agriculture John Parman Introduction Human Capital and Agriculture Iowa’s Farms and Schools Data Results

Private Returns Spillovers Spillovers and Social Networks

Conclusions

Farmers and Their Neighbors

Variable Correlation Total years of schooling 0.1656 Years of graded schooling 0.1494 Annual earnings 0.2274 Farm value 0.2309 Age 0.0409

Notes: Neighbor characteristics are measured as the mean value of the characteristic over all observed neighbors.

Correlations between a farmer and his neighbors

Good Schools Make Good Neighbors: Human Capital Spillovers in Early 20th Century Agriculture John Parman Introduction Human Capital and Agriculture Iowa’s Farms and Schools Data Results

Private Returns Spillovers Spillovers and Social Networks

Conclusions

Estimating the Returns to Education

Standard Mincer regression to estimate private returns to education

Education measures: total schooling, schooling by type, schooling by location received Personal characteristics: nativity, years in the US, age, township dummies, religion dummies, land quality (proxied by land value per acre)

Spillovers are estimated by including a measure of neighbors’ education

Several measures of neighbors’ education are used (mean years of schooling, mean years of high school, maximum years of schooling, etc.) Distinguish between neighbors who are within and

• utside of a farmer’s social network (defined by religion,

ancestry and birth cohort)

Good Schools Make Good Neighbors: Human Capital Spillovers in Early 20th Century Agriculture John Parman Introduction Human Capital and Agriculture Iowa’s Farms and Schools Data Results

Private Returns Spillovers Spillovers and Social Networks

Conclusions

Private Returns to Education

Measure of schooling used: All land owners Farmers only Total schooling 0.017*** 0.013** (0.005) (0.006) Graded schooling 0.022*** 0.010 (0.008) (0.010) Common school 0.010* 0.011* (0.006) (0.006) Grammar school 0.005

• 0.004

(0.014) (0.015) High school 0.046** 0.052** (0.020) (0.022) College 0.064** 0.009 (0.025) (0.029) Numbers of observations 2410 2219

Standard errors in parentheses. All regressions control for age, religion, land value per acre and township.* significant at 10%; ** significant at 5%; *** significant at 1%

Private returns to education, log annual earnings as dependent variable

Good Schools Make Good Neighbors: Human Capital Spillovers in Early 20th Century Agriculture John Parman Introduction Human Capital and Agriculture Iowa’s Farms and Schools Data Results

Private Returns Spillovers Spillovers and Social Networks

Conclusions

Human Capital Spillovers

Mean Maximum Mean Maximum Mean Maximum Neighbors' education

0.023*** 0.026*** 0.010 0.011**

0.038 0.028***

(0.008) (0.006) (0.011) (0.005)

(0.029) (0.011) Own education: Common school

0.009 0.008 0.009 0.009

0.009 0.008

(0.006) (0.006) (0.006) (0.006)

(0.006) (0.006) Grammar school

• 0.007
• 0.008
• 0.008
• 0.008
• 0.007
• 0.008

(0.016) (0.015) (0.016) (0.016)

(0.016) (0.016) High school

0.055** 0.052** 0.057*** 0.057***

0.057*** 0.056**

(0.022) (0.022) (0.022) (0.022)

(0.022) (0.022) College

0.021 0.021 0.023 0.023

0.022 0.023

(0.031) (0.031) (0.031) (0.031)

(0.031) (0.031) Number of obs. 2158 2158 2158 2158 2158 2158

Standard errors in parentheses. All regressions control for age, religion, land value per acre and township.* significant at 10%; ** significant at 5%; *** significant at 1%

Private returns to education and spillovers, log annual earnings as dependent variable Neighbors' education measure: Total schooling All graded schooling High school

Good Schools Make Good Neighbors: Human Capital Spillovers in Early 20th Century Agriculture John Parman Introduction Human Capital and Agriculture Iowa’s Farms and Schools Data Results

Private Returns Spillovers Spillovers and Social Networks

Conclusions

Human Capital Spillovers with Interactions

Neighbors' education measure: Mean years of high school Max years of high school Own schooling: Common school 0.008 0.007 (0.007) (0.007) Grammar school

• 0.006
• 0.006

(0.017) (0.017) High school 0.063*** 0.069*** (0.023) (0.024) College 0.037 0.045 (0.029) (0.029) Neighbors' schooling: Neighbors' high school 0.041* 0.029*** (0.021) (0.009) Own HS x

• 0.018
• 0.011*

Neighbors' HS (0.014) (0.006) Number of obs. 2158 2158 Private returns to education and spillovers with own-neighbor education interactions, log annual earnings as dependent variable

Standard errors in parentheses. All regressions control for age, religion, land value per acre and township.* significant at 10%; ** significant at 5%; *** significant at 1%

Good Schools Make Good Neighbors: Human Capital Spillovers in Early 20th Century Agriculture John Parman Introduction Human Capital and Agriculture Iowa’s Farms and Schools Data Results

Private Returns Spillovers Spillovers and Social Networks

Conclusions

Human Capital Spillovers Within and Across Social Networks

Parents' place of birth for Chickasaw county, 1915

Austria Bohemia Canada Connecticut Denmark England Germany Illinois Iowa Ireland Maine New York New Hampshire Norway Ohio Pennsylvania Scotland Sweden Wisconsin Other Not Reported/Not Found

Good Schools Make Good Neighbors: Human Capital Spillovers in Early 20th Century Agriculture John Parman Introduction Human Capital and Agriculture Iowa’s Farms and Schools Data Results

Private Returns Spillovers Spillovers and Social Networks

Conclusions

Human Capital Spillovers Within and Across Social Networks

Church affiliations for Chickasaw county, 1915

Baptist Other Catholic Congregational Evangelical Lutheran Methodist Presbyterian Not Reported/Not Found

Good Schools Make Good Neighbors: Human Capital Spillovers in Early 20th Century Agriculture John Parman Introduction Human Capital and Agriculture Iowa’s Farms and Schools Data Results

Private Returns Spillovers Spillovers and Social Networks

Conclusions

Human Capital Spillovers Within and Across Social Networks

Social group defined by: Church affiliation Parents' birthplaces Birth cohort Percentage of neighbors that are in social group 38.80% 21.8% 17.4% Correlation of own education with that of similar neighbors 0.258 0.299 0.131 Correlation of own education with that of dissimilar neighbors 0.125 0.155 0.057 Correlation of own years HS/college with that of similar neighbors 0.242 0.103 0.048 Correlation of own years HS/college with that of dissimilar neighbors 0.053 0.078 0.084 Correlation of own log earnings with that of similar neighbors 0.194 0.312 0.270 Correlation of own log earnings with that of dissimilar neighbors 0.254 0.248 0.207 Correlations between neighbors based on group membership

Good Schools Make Good Neighbors: Human Capital Spillovers in Early 20th Century Agriculture John Parman Introduction Human Capital and Agriculture Iowa’s Farms and Schools Data Results

Private Returns Spillovers Spillovers and Social Networks

Conclusions

Human Capital Spillovers Within and Across Social Networks

Social group defined by: Neighbors' education measure: Total schooling Graded schooling High school and college Neighbors' education: Mean education within group 0.010** 0.027* 0.045 (0.004) (0.015) (0.035) Mean education outside group 0.012** 0.005 0.018 (0.006) (0.010) (0.018) Own education: Years of common school 0.008 0.009 0.008 (0.006) (0.006) (0.006) Years of grammar school

• 0.007
• 0.008
• 0.007

(0.014) (0.015) (0.015) Years of high school 0.053** 0.054** 0.053** (0.026) (0.026) (0.026) Years of college 0.023 0.024 0.021 (0.033) (0.031) (0.032) Observations 2148 2148 2148

Robust standard errors clustered by township in parentheses, all regressions control for age, nativity, years in the United States, religion, land value per acre and township. * significant at 10%; ** significant at 5%; *** significant at 1%

Private returns to education and spillovers by social group membership, log annual earnings as dependent variable Parents' birthplaces

Good Schools Make Good Neighbors: Human Capital Spillovers in Early 20th Century Agriculture John Parman Introduction Human Capital and Agriculture Iowa’s Farms and Schools Data Results

Private Returns Spillovers Spillovers and Social Networks

Conclusions

Human Capital Spillovers Within and Across Social Networks

Social group defined by: Neighbors' education measure: Total schooling Graded schooling High school and college Neighbors' education: Mean education within group 0.013*** 0.019** 0.037** (0.003) (0.009) (0.017) Mean education outside group 0.017** 0.004 0.016 (0.007) (0.009) (0.018) Own education: Years of common school 0.008 0.008 0.008 (0.006) (0.006) (0.006) Years of grammar school

• 0.006
• 0.009
• 0.007

(0.014) (0.014) (0.014) Years of high school 0.051* 0.054** 0.052** (0.027) (0.025) (0.026) Years of college 0.025 0.025 0.023 (0.033) (0.031) (0.032) Observations 2158 2158 2158 Private returns to education and spillovers by social group membership, log annual earnings as dependent variable Birth cohort

Robust standard errors clustered by township in parentheses, all regressions control for age, nativity, years in the United States, religion, land value per acre and township. * significant at 10%; ** significant at 5%; *** significant at 1%

Good Schools Make Good Neighbors: Human Capital Spillovers in Early 20th Century Agriculture John Parman Introduction Human Capital and Agriculture Iowa’s Farms and Schools Data Results

Private Returns Spillovers Spillovers and Social Networks

Conclusions

Human Capital Spillovers Within and Across Social Networks

Social group defined by: Neighbors' education measure: Total schooling Graded schooling High school and college Neighbors' education: Mean education within group 0.005

• 0.003

0.018 (0.004) (0.014) (0.034) Mean education outside group 0.016** 0.041*** 0.070*** (0.006) (0.010) (0.018) Own education: Years of common school 0.014 0.015* 0.014 (0.009) (0.009) (0.009) Years of grammar school 0.004 0.001 0.006 (0.021) (0.023) (0.021) Years of high school 0.049* 0.054** 0.050* (0.026) (0.026) (0.027) Years of college 0.014 0.015 0.011 (0.032) (0.032) (0.033) Observations 1284 1284 1284 Church affiliation

Robust standard errors clustered by township in parentheses, all regressions control for age, nativity, years in the United States, religion, land value per acre and township. * significant at 10%; ** significant at 5%; *** significant at 1%

Private returns to education and spillovers by social group membership, log annual earnings as dependent variable

Good Schools Make Good Neighbors: Human Capital Spillovers in Early 20th Century Agriculture John Parman Introduction Human Capital and Agriculture Iowa’s Farms and Schools Data Results

Private Returns Spillovers Spillovers and Social Networks

Conclusions

Summary of Findings

The organization of Iowa’s schools and research institutions suggests that formal education was designed in part to improve agricultural productivity. The data confirm that these efforts succeeded at the individual level. Returns to education were high for farmers - 5% increase in annual earnings from an additional year of high school. Spillovers were also significant - 2-7% increase in annual earnings from an additional year of high school for adjacent neighbors. The results suggest that there were large public benefits to public education in agricultural communities.