National Chengchi University Faculty Attendance at International Conferences--Report

19/80/2007 (dd/mm/yyyy) Name Chu –Chia Lin Administrative Unit

and Job Title

Department of Economics Professor

Location of

Conference Macau Duration of

Conference 09~12/07/2007

Name of Conference

(Chinese) 第十三屆亞洲不動產學會與美國不動產與都市經濟學會聯合年會

(English) The 12^th AsRES Annual Conference & The 2007 AREUEA International Conference

Title of Presented Manuscript

Chinese)住宅環境與兒童教育表現關係的新證據：台灣的個案分析 (English) New Evidence on the Link between Housing Environment and

Children’s Educational Attainments：The case of Taiwan

The report should include:

1.Type of participation in the conference

2.Reflections deriving from conference participation 3.Suggestions

4.Name and content of the materials brought back 5.Other

2007 年出席國際學術會議心得報告書

一．會議名稱：the 12^th AsRES Annual Conference and The 207 AREUEA International Conference

二．會議地點：Macau

三．參與人數與論文數目：約三百人，約八十場，約二百三十篇論文四．本人論文發表場次 7/12, 1400-1530, Session I-5

題目：New Evidence on the Link between Housing Environment and Children’s Educational Attainments: The Case of Taiwan 五．重要結論與研究成果：

(1) 國際學術文獻中有許多探討居住環境對小孩讀書效果影響的文章，但是這些文章中都一直缺乏嚴謹的實証研究。我們的文章利用台灣 2000 年的住宅普查資料，我們可以實際的來檢視居住環境對於小孩讀書成效的正面影響效果。因為台灣的住宅普查資料中，有完整的地址資料，所以我們可以用來控制無法觀察到的家庭異質變異的問題，然後我們可以進一步的來比較居住在鄰近的同年級小孩的讀書成效。結果我們發現，16 與 17 歲的青少年及 19 與 20 歲的年輕成年人的學術表現與其家庭的住宅面積、居住時間與是否自有等變數有正且顯著的關係；而與住宅年齡有負的關係。

在現在的國際學術文獻當中，本研究可能使用住宅資料來討論這個問題最完整的文章，所以本文的學術貢獻應該是很可觀的，因此本文應該有很大的機會在國際學術期刊上發表。

(2) 在本人發表的場次上，也有許多學者提出問題，顯示他們對於本文的議題都很有興趣，對於本文的改進建議也有不少，對本文的修改也有很顯大的助益。

六．相關聯結：Faculty of Business Administration, University of Macau http://www.umac.mo/fba

七．附件：(1) 本人論文全文一份 (2) 大會手冊一份

New Evidence on the Link between Housing Environment and Children’s Educational Attainments⁺

Hsien-Ming Lien^*, Wen-Chieh Wu^**, Chu-Chia Lin^***

Abstract

There is an extensive literature that posits the hypothesis that a better housing environment enhances a child’s educational attainments. However, there is little causal evidence demonstrating the presence of this effect. Using the census files covering the entire population of Taiwan, we examine the effect of housing environment on children’s educational attainments. Because the Taiwan census data contain unique address information for every household, we are able to control for unobserved family heterogeneity by comparing a child with his or her peers of the same age cohort in the same neighborhood. After controlling for neighborhood using tens of thousands of area dummies, the chance of high school enrollment for teens (ages 16 and 17) and college enrollment for young adults (ages 19 and 20) is found to be positively correlated with increases in floor space, increases in residential stability, and ownership status, but negatively correlated with increases in building age. In addition, we found that the effect of a child’s private space on the chance of school enrollment is nonlinear and different across age and gender. The results are robust even when we account for the potential endogeneity between sibship size and educational outcome using the instrumental variable method.

Keywords: quantity–quality trade-off, housing, educational attainment JEL classification: R0, I2

# We thank David H. Autor, Charles Leung, and Jin-Tan Liu for their comments and suggestions. We are grateful to Directorate General of Budget, Accounting, and Statistics for providing the census data. Supports from National Science Council (NSC-93-2415-H-004-013) and National Health Research Institute (NHRI-EX93-9204PP) for Hsienming Lien are greatly appreciated. The usual disclaimer applies.

* Department of Public Finance, National Cheng-Chi University, 64 Sec 2, Zhi-Nan Rd, Taipei, Taiwan; email:

[email protected].

** Corresponding author, Department of Public Finance, National Cheng-Chi University, 64 Sec 2, Zhi-Nan Rd, Taipei, Taiwan; email:[email protected].

*** Department of Economics, National Cheng-Chi University, 64 Sec 2, Zhi-Nan Rd, Taipei, Taiwan; email:

[email protected].

1. Introduction

One long-standing area of interest in the social sciences is to understand the connection between the family environment and a child’s outcome, particularly educational attainments. It is generally believed that a larger family size may negatively affect a child’s outcome through resource dilution [e.g., Blake (1981, 1989)]. The best-known economic theory that links family circumstances and a child’s educational outcome is perhaps the quantity–quality trade-off model [Becker and Lewis (1973) and Becker and Tomes (1976)]. This theory claims that as parents become richer because of the interaction between quantity and quality in the budget constraint, they demand higher quality children, but not necessarily more children. Thus, a reduction in family size leads to an improvement in a child’s schooling.

The majority of early studies confirm this trade-off relationship, with a negative relationship between family size and a child’s educational attainments being widely observed in regression results.¹ While this negative correlation is often interpreted as evidence supporting the quantity–

quality trade-off theory, the conclusions are facing serious criticism. Most problematic is that the apparent negative relationship is not necessarily indicative of a causal effect. That children raised in larger families have less schooling than those in smaller families is not necessarily because of the sibship size per se, but may reflect the omission of other unobserved characteristics, such as parental preferences, household resources, neighborhood conditions, and quality of schooling. In light of this potential bias, several studies have sought to uncover the causal effect of family size on a child’s educational outcome using the instrumental variable method (IV) [e.g., Angrist, Lavy, and Schlosser (2005), Caceres (2004), and Conley and Glauber (2005)], or controlling for family fixed effects [e.g., Guo and VanWey (1999) and Black, Devereux, and Salvanes (2005)]. Notably, these studies generally found the coefficient of sibship size becomes insignificant after controlling for unobserved family characteristics.

1 For a review in the economic literature on the link between family size and children’s outcomes, see Schultz

Why are results of OLS estimation so different from those of IV or the family fixed effects model?

One likely explanation, as pointed out by Phillips (1999), is that sibship size does not produce a negative impact on a child’s educational outcome, but the type of family resources it dilutes does.² Furthermore, Goux and Maurin (2005) investigated the effect of household crowdedness on a child’s school performance, one key resource likely to affect a child’s education. Using exogenous variations of family size and household crowdedness as instruments, they found the importance of sibship size becomes negligible under IV estimation, but the private space each child has is negatively associated with a child’s educational attainments. In other words, children in large families perform less well not because of their family size, but because of the smaller private space each child has available to them.

In the same spirit as Goux and Maurin (2005), this paper seeks to explore the underlying relationship between the housing environment and a child’s educational attainments. Unlike Goux and Maurin (2005), which controls for unobserved family heterogeneity using instruments, we overcome this difficulty by comparing a child with his or her peers of the same age in the same very small neighborhood: “lin,” the smallest government jurisdiction in Taiwan that usually covers less than 0.1 square kilometer. Families residing in the same lin often share similar housing preferences and family incomes. In addition, youths raised in the same lin generally have experienced the same neighborhood effect. Furthermore, under the current regulation, children in the same lin typically attend the same school for compulsory education. Thus, by comparing youths with peers of the same age in the same lin, we control to some extent for unobserved family heterogeneity such as parental preferences, earning potential, neighborhood conditions, and, most importantly, quality of compulsory schooling.

Our data are derived from the census files that cover the entire Taiwanese population, more than 22 million, in the year 2000. The census data not only record detailed family and housing information, but also include unique address information for every household. The large sample size, together with detailed address information, allows us to examine the chances of high school enrollment for teens (ages 16 and 17) and college enrollment for young adults (ages 19 and 20) while controlling for family

2 Black, Devereux, and Salvanes (2005) offer a different explanation: family size itself might have little impact on the quality of every child, but more likely impacts the marginal child through the effect of birth order. In their results, children of higher birth orders are likely to have worse educational attainments.

heterogeneity. After including tens of thousands of area dummies, our results confirm the importance of housing environment in determining a child’s educational attainments. Specifically, our estimates show that youths’ educational attainment is positively associated with an increase in housing floor space, an increase in residential stability, and ownership status, but negatively related to an increase of building age. The results continue to hold even accounting for the endogeneity between sibship size and a child’s education using twin births or sex-composition as instruments.

An important difference between our study and Goux and Maurin (2005) is that we include a wide range of housing variables. Aside from each child’s private space, this study also considers a house’s floor space, building age, residential stability, and ownership status as various determinants of housing environment. Therefore, the analysis is able to provide a more complete picture about the impact of housing on a child’s education. Another key difference is that we obtain a different effect of household crowdedness. While Goux and Maurin (2005) found that a reduction in a child’s private space resulted in a negative effect on his or her schooling, our estimates indicate that the effect may be nonlinear:

conditional on a household’s size, reducing each child’s private space does not always lead to an decrease in the chance of school enrollment. Moreover, this crowdedness effect is likely to differ according to the child’s gender and age.

Our paper also relates to another line of literature exploring the effect of housing variables on children’s outcomes, including tenure status [e.g., Green and White (1997), Boehm and Schlottmann (1999), Aaronson (2000), and Haurin, Parcel, and Haurin (2002)] and residential mobility [e.g., Lee, Oropesa, and Kanan (1994), Green and White (1997), Aaronson (2000), and Haurin, Parcel, and Haurin (2002)].³ Although some studies have demonstrated the importance of housing environment, few of them controlled for the endogeneity problem caused by various housing variables.⁴ To our knowledge, this paper is the first study that investigates the effect on a child’s educational attainments of a wide range of housing variables.

3 For a complete review on the tenure status literature, see Haurin, Dietz, and Weinberg (2003).

4 A number of studies have attempted to control for the endogeneity of housing variables. For instance, Green and White (1997) adopted the bivariate probit model to solve the selection bias problem between tenure decision

The paper is organized as follows. In the next section, we outline the estimation problem and discuss the existing identification strategies as well as our strategies. Section 3 describes the data source, sample selection, and measures of educational attainments, along with the basic statistics of our sample. Section 4 shows results of the basic specification, the effect of area dummies, as well as comparisons with IV estimates. Section 5 offers concluding remarks.

2. Conceptual Framework

where is the child’s educational attainments, is a vector of observed characteristics of the child and his or her family (e.g., age, sex, birth order, and father’s and mother’s education and working status), is a variable of child i’s sibship size, is the family-specific unobserved determinant (e.g., parental preferences or quality of schooling), and

edui X_i

Ni v_i

ε

i represents the idiosyncratic shock that is assumed to be independent across other factors.

The central parameter of interest is , which is viewed as a measure of the trade-off between quantity and quality of children. Early studies primarily found this coefficient to be negative in OLS estimation and therefore inferred that substantial quality improvements can be gained by controlling for family size. However, the regression results are likely to be confounded by the existing observed factors (e.g., parental education) as well as the unobserved determinants (e.g., quality of schooling). The omitted variable formula suggests that the OLS coefficient from the regression is:

cov( , ) cov( , )

Therefore, even if children raised in larger families have less schooling than those in smaller families, the strength of the relationship could be driven by the correlation between sibship size and other observed and unobserved factors, not necessarily the quantity–quality trade-off.

B. Existing Identification Strategy

In light of the potential bias, the existing literature has adopted several methods to uncover the underlying relationship between a child’s education and sibship size. Early studies attempted to account for this potential bias by including more controls, such as parental IQ, and better measures of household income. However, adding more controls cannot rule out the possibility of an association between family size, educational attainment, and something immeasurable, such as housing environment, neighborhood conditions, or quality of schooling. As a result, recent studies have taken different approaches to account for unobserved family heterogeneity. For instance, Guo and VanWey (1999) and Black, Devereux, and Salvanes (2005) include the household’s dummies, i.e., family fixed effects, to control for the unobserved family-level heterogeneity. Angrist, Lavy, and Schlosser (2005), Caceres (2004), and Conley and Glauber (2005) employ exogenous variations in family size, such as multiple births or preferences of a mixed sibling-sex composition, as instruments to investigate the causal effect of family size on a child’s education. Notably, studies using IV estimations or fixed family effects found weaker correlations between family size and a child’s education, many of which turn out to be negligible.

The inconsistency of results between OLS and other estimation methods cast doubts over the link between family size and a child’s outcome. One likely explanation, as pointed out by Phillips (1999), is that sibship size per se does not affect the child’s educational attainments, but the type of resources it dilutes does. Goux and Maurin (2005) extended this line of thought by exploring the impact of a child’s private space, one important kind of resource likely to be affected by additional children, on the child’s schooling. Specifically, they considered the following equation:

i i i i i ,

edu =X

α β

+ N +

γ

h + +

ν ε

(3)

where is the average number of rooms per person in the household, used as a proxy for a child’s private space. Notice that equation (3) also includes the sibship size variable to account for the effect caused by family size. Because sibship size and the child’s private space are likely to be endogenous, they employ two instruments, gender of the first two children and of the last two children, respectively, to control for unobserved family heterogeneity. Consistent with previous studies, they found that the coefficient of sibship size becomes insignificant under IV estimation. Interestingly, the coefficient of the average number of rooms per person in IV estimates is significantly negative, suggesting that children in large families perform less well, not because of their family size, but because of the smaller private space available to each child.

C. Our Identification Strategy

In contrast to Goux and Maurin (2005), our study seeks to identify the effect of a variety of housing variables on a child’s educational outcome, such that:

i i i i i .

edu =X

α β

+ N +H

γ ν

+ +

ε

(4)

The biggest difference between (3) and (4) is that the housing environment is now a vector of multiple variables ( ) instead of a single variable ( ). There are substantial difficulties in using existing identification strategies for this specification. Because these housing variables do not change within a household, including household dummies essentially eliminates the effect of housing environment. Another possible strategy is to find instruments for housing variables, as Goux and Maurin (2005) did for household crowdedness. Nevertheless, controlling for the unobserved heterogeneity in this setting requires us to find many more instruments.

Hi h_i

We take a different approach to identify the causal link. Apart from including a detailed set of important variables of a child’s family background used in previous studies (e.g., a child’s birth order, parental age, work status, and education), we account for unobserved family heterogeneity by adding dummies of the child’s residential neighborhood. Our unique data are derived from the census data that collects information on the entire Taiwanese population, with detailed address information.

Therefore, we are able to compare a child with his or her peers of the same age in the same very small neighborhood, the lin. Families residing in the same lin tend to share similar housing preferences and parental incomes, as well as earning potentials. Moreover, youths raised in the same lin generally encounter the same neighborhood effect. Furthermore, youths in the same lin typically attend the same elementary or junior high schools, allowing us to control for the quality of compulsory schooling prior to high school or college. In fact, given Taiwan’s current school regulation, it is almost certain that youths in the same lin go to the same school.^5,6 Thus, by controlling for neighborhood fixed effects, we account for the neighborhood effect, quality of schooling, and parental incomes and preferences.

Nevertheless, it is still possible that our approach may not fully capture unobserved family heterogeneity. We will discuss this point in the results section.

To be more specific, we estimate the following equation:

在文檔中台灣地區住宅存量與住宅價格之動態調整：失衡模型之應用 (頁 91-162)