DE ECONOMIST 145, NO. 3, 1997

HOW DUTCH TEENAGERS SPEND THEIR MONEY** BY MARCEL WARNAAR AND BERNARD VAN PRAAG* Key words: school children, consumption by adolescents, Almost Ideal Demand system, zero demand problem

1 INTRODUCTION

It is argued that the way adults manage their money is strongly influenced by the way they managed their money when they were children. So there is a good reason to study in how far children grasp economic concepts and how they deal with them. Contrary to what is known about economic behaviour of primary school pupils ~up to 12 years!, in for instance Furnham and Thomas ~1984!, Jahoda ~1979!, and Abramovitch et al. ~1991!, surprisingly little can be found about the expenditures of adolescents ~from 13 to 20 years!, although they can spend a lot of money. One of the striking differences between adult and adolescent behaviour is that adolescents are hardly constrained by fixed costs like rent and energy bills, as adults are. They can choose freely among various categories without any fixed obligations and, as their incomes are sizeable, thay have a lot of purchasing power. For an economist there is also a methodological point of interest. Is it possible to explain adolescents’ expenditures by a similar structure as household expenditures ~by adults! and what are the main explanatory variables? In this article we will describe how Dutch secondary school pupils, 1 ranging in age from 13 to 20, manage their money; what expenditure categories they spend their money on and how much they spend in each category. The 1.3 million secondary school pupils in The Netherlands have in total about Dfl. 4.8 billion 2 to spend per year, which corresponds to about 1% of Dutch GNP. Esti* The first author is research associate at the National Institute for Family Finance Information ~NIBUD!. At the time of writing the second author was Professor of Economics, Erasmus University Rotterdam and member of the Scientific Council for Government Policy ~WRR!, The Hague. From September 1, 1992, Van Praag has been managing director of the SEO, Foundation for Economic Research, University of Amsterdam. ** The authors wish to thank Raoul de Zwart and René Diekstra who co-designed the questionnaire, and NIBUD and Keesings Publishing Company for financial support. 1 We will use the words pupils, adolescents and teenagers indiscriminately. We will also refer to the average adolescent as ‘he’. 2 The exchange rate of one Dfl. varied in the range from $0.40 to $0.60. One ECU corresponds to Dfl. 2.50. De Economist 145, 367–397, 1997. © 1997 Kluwer Academic Publishers. Printed in the Netherlands.

368

M.F. WARNAAR AND B.M.S. VAN PRAAG

mates in the USA range from $4 billion to $30 billion per month ~Hall, 1987!. This amounts to 0.1% of USA GNP. Greenberger and Steinberg ~1986! report $143 per month for working sophomores and $272 per working seniors. We shall use three comparable data sets, gathered in 1984, 1990, and 1992. Comparison will also shed light on developments between 1984 and 1992. The height of expenditures on commodity groups is the primary point of investigation; not whether or why an adolescent buys specific brands like NIKE shoes instead of ADIDAS shoes. In section 2 we describe what sources of income adolescents have and what their opinions are about their own way of managing money. In section 3 we posit a complete demand system for adolescents. Such a model is frequently used for adults, but it has never been specified or estiùmated for adolescents. Disturbing factors on utility-maximizing behaviour that hold for adults, like taxes and intertemporal decisions, are not or in any case less important for adolescents. The econometric implications of zero-expenditures are also dealt with. Although the data sets are cross-sectional, we are able to distinguish price variation for some spending categories, as some categories are co-financed by teenagers and their parents. In section 4 the estimated model is presented, while conclusions are drawn in section 5. Although our results only describe Dutch school-going teenagers, we expect that results for other OECD counttries will not be particularly different. However, we do not know of any comparable data or results elsewhere. This fact makes our results of more general relevance. 1.1 The Data Set In 1984, 1990, and 1992 a National School Survey 3 was held in The Netherlands. The sample was a cluster sample, where the clusters were school classes. In 1984 about 400 schools of all levels and denominations were willing to participate with one or more classes. Primarily in order to reduce costs, we finally selected 250 schools for participation. In 1990 a,d 1992 a similar sample was created. In this way representative samples of Dutch secondary school pupils ranging in age from 13 to 20 were determined. Questions were asked about various matters ranging from their incomes and expenditures to the way they spend their leisure time, or their expectations for the future. The answers on these questions had to be filled in during school time on a 20-page questionnaire. Some of the issues had already been examined separately on a smaller scale, but this was the 3 The first National School Survey, held in 1984, was an initiative of the Econometric Institute of the Erasmus University Rotterdam, the National Institute for Budget Education ~NIBUD!, and Keesings Publishers. The questionnaire was developed and the survey organized by Raoul de Zwart and Bernard van Praag. Apart from the institutes and persons mentioned earlier, René Diekstra, Professor of Adolescent Psychology at Leiden University, took part in the design of the second and third surveys ~1990 and 1992!, adding a couple of questions on adolescents’ health and well-being. The Dutch Social and Cultural Planning Bureau also participated in the 1992 survey.

DUTCH TEENAGERS

369

first time these subjects were investigated simultaneously. The size of the samples is impressive, i.e., about 15,000 each. A thorough screening was applied to filter out ‘jokers,’ which present a non-negligible but surmountable problem in anonymous surveys among pupils. A weighting procedure made the sample representative for the Dutch secondary school population ~de Zwart and Sevinga, 1986 and Warnaar and de Zwart, 1991 and 1994!. The screened samples consisted of about 12,000 pupils each. Apart from a nation-wide published report by de Zwart et al. ~1984, 1990, 1993!, each participating school also received its own scores in terms of averages per class. Because leaving school is permitted after the age of 16 in The Netherlands, not all adolescents of 17 are attending school anymore. The results in this article refer to school-attending teenagers only. It goes without saying that full-time working teenagers may have different spending behaviour. Although a survey inquiry like this has its disadvantages, e.g. ‘open’ questions have to be avoided as they are too costly and time-consuming, it appears to be a rather cheap and ~after cleaning and reweighting! reliable way to obtain a lot of information. The allegations that survey research is difficult for the less-educated are partly justified. During the process of screening we did find that pupils following lower education levels filled in the questionnaire less accurately than pupils at higher level education. However, by reweighting we corrected for this selective non-response. 2 TEENAGERS’ INCOME AND WAYS OF SPENDING

Let us give a description of some primary facts. 2.1 The Income Teenagers may derive their income from various sources. In Table 2.1 the percentages of incidence and average monthly amounts of money from the various sources are presented. The averages ~in Dutch guilders! are calculated with exclusion of the zero-observations to facilitate the comparison over time. That is why the income categories do not add up to total income. The first source of income is an allowance from the adolescents’ parents ~‘pocket-money’!, either in exchange for some household duties or unconditionally ~Miller and Yung, 1990!. A second type of parental allowance is ‘tied’ money intended for special purposes, like clothing or travelling allowances. This is partly done to educate children how to deal with larger amounts of money, partly to avoid unpleasant frictions between parents and children, e.g., in the case of money for clothing. The incidence of travel allowances fell in those 8 years from 20% in 1984 to 11% in 1992. The use of clothing allowances is less stable, ranging from 23% in 1984 to 36% in 1990, while the amount of the allowance steadily increased.

370

M.F. WARNAAR AND B.M.S. VAN PRAAG TABEL 2.1 – THE INCOMES OF ADOLESCENTS

1984 %

1990 amount

1992

%

amount

%

amount

pocket money clothing allowance travelling allowance extra money from parents weekly job holiday job study grant

89% 23% 20% 50% 44% 47% –

48 85 57 13 128 91 –

86% 36% 22% 60% 45% 55% 15%

64 116 66 32 229 61 413

77% 30% 11% 40% 49% 49% 17%

58 126 55 60 233 80 337

total income

98%

182

97%

317

98%

329

A third source of income is wage income. Almost half of the teenagers have a weekly job. Furthermore, a considerable number of adolescents work during holidays. Compared to the United States, in The Netherlands fewer adolescents work; in 1985 59% of the American sophomores and 76% of the seniors had paid work, according to Greenberger and Steinberg ~1986!. Furthermore, Dutch adolescents work fewer hours than their American counterparts. American adolescents worked on average 16 hours a week in 1985, while Dutch adolescents worked on average 6.7 hours a week in 1984. However, the Dutch are catching up: in 1990 the average working time increased to 7.5 hours a week and in 1992 it further increased to 9.8 hours. The amount of ‘homework’ in preparing for school lessons surely plays a part with respect to this difference. Whereas Dutch adolescents spend on average two hours a day on homework, American adolescents only spend three hours a week on it ~Lewin-Epstein, 1981!. We also see the time spent on ‘homework’ decreasing in The Netherlands. In 1986, a fourth source of income came into being in The Netherlands: a so-called ‘basic grant,’ which is given by the Dutch government to every fulltime student between 18 and 30 years of age. Pupils of secondary schools at 18 years of age and over also receive this basic grant. Pupils living with their parents receive about Dfl. 275 per month under this programme and students and students living on their own receive about Dfl. 600. Depending on the height of their parents income this grant may be further increased by a loan. Before 1986, study grants were given to the student’s parents, but in the new system the grant was provided directly to the pupil. This extra amount of money strongly influenced the management of money, as we shall see. 4

4

Since 1991, the study grant has been cut by about Dfl. 40 in exchange for free public transport.

DUTCH TEENAGERS

371

It is reported that 30 to 40 years ago working teenagers had to hand over most of their income out of jobs to their parents ~e.g. Diederich, 1951, Greenberger and Steinberg, 1986!. Since then, adolescents became more and more financially independent from their parents, having part-time jobs to generate more purchasing power for their own spending. Teenagers are quite satisfied with their spending power. In 1990 and 1992, the teenagers were asked what their opinion was about their level of income. For both years the results were comparable: a fraction of 10% considered their income ‘very good,’ 35% considered it ‘good’ and 24% considered it ‘fairly good.’ Only 12% considered their income moderate and 6% thought they had insufficient income. 2.2 Ways of Spending Money It might be that adolescents, when they start spending larger amounts of money, use their money unwisely in the beginning. It is sometimes said that they spend all their money the same day they get it, that they buy expensive things they cannot afford, and so on. Casual reports of adolescents behaving like this can be found, but research on this topic on a large scale base has, to our knowledge, hardly been done. Our results show that the majority of Dutch teenagers evaluated their way of spending money as considerate, at least according to themselves. It stands to reason that there is a difference between what parents call ‘wise’ and what the teenagers themselves think. In 1984 24% of the adolescents considered the way they handled their money to be ‘good,’ 41% considered it as ‘normal,’ 27% thought ‘they could do better’ and only 4% thought their way of spending was ‘bad.’ In 1990 and 1992 the results were almost the same. Differences in age, gender, and educational level were not significant. In all three surveys questions were asked about the adolescents’ expenditure patterns. They were asked to give the approximate average spendings per month on about 20 categories of goods. The expenditure categories in the three surveys are not exactly the same, as Table 2.2 shows. The table shows that in 1992 teenagers spend their money on a wider array of expenditure categories than in 1984. Apparently, the freedom of the teenager how to spend his own money increased from 1984 to 1992. Striking is the increase of the percentage of teenagers spending their money on gambling. This increase is concomitant with the sharp rise in gambling possibilities in The Netherlands in 1986. Not only do teenagers spend their money on a wider range of commodities in 1992 than in 1984, they also spend more money on those categories and far more than accumulated inflation would predict ~accumulated inflation since 1984 was about 7% in 1990 and 13% in 1992!. Spending on most categories more than doubled or even tripled. Especially expenditures on outdoor activities as food, drinks and going out increased considerably. It seems that these spendings are for

372

M.F. WARNAAR AND B.M.S. VAN PRAAG

more regular in 1992 than in 1984, when teenagers only occasionally spent money in these categories. On the other hand, the monthly amount of money saved did not increase from 1990 to 1992; although it rose sharply from 1984 to 1990.

TABLE 2.2 – EXPENDITURE CATEGORIES

1984

Snacks, crisps, etc. Non-alcoholic drinks Alcoholic beverages Sweets Smoking Travelling expenses* Cinema, theatre, disco, etc.** Educational material Gambling Drugs Records, music cassettes*** Gifts Magazines ~1984 also books! Sports equipment Clothing and shoes Contributions/subscriptions Computer material School books Insurance Holiday ~on a monthly basis! Cosmetics Other regular costs Saving

1990

1992

%

Dfl./ month

%

Dfl./ month

%

Dfl./ month

47% 40% 30% 51% 18% 18% 61% 17% 1% 3% 57% 65% 36% 19% 23% 11% – – – – – 50% 51%

13 14 35 12 22 20 20 7 8 45 12 9 7 18 59 16 – – – – – 15 34

58% 54% 38% 60% 19% 25% 69% 33% 11% 4% 60% 76% 37% – 42% 18% 8% 6% 7% 70% – 60% 49%

20 24 54 17 37 32 24 9 17 61 21 14 8 – 75 22 24 39 35 13 – 28 76

64% 48% 37% 63% 24% 20% 58% 27% 13% 6% 50% 61% 33% – 45% 12% 8% 7% 7% 80% 38% 45% 75%

35 37 95 25 63 41 46 18 29 72 35 20 10 – 103 26 43 67 47 30 20 43 76

The first column denotes the percentage of adolescents who spend part of their disposable income on that particular item. The average monthly expenditure of these adolescents can be found in the second column. Because of this, the total of expenditures does not equal total income. * in 1984 explicitly mentioned: public and own transport. ** in 1990 and 1992 explicitly mentioned: without consumptions. *** in 1984 including photographic material, in 1990 and 1992 including CDs. – not asked.

DUTCH TEENAGERS

373

3 CONSUMER THEORY

In principle there is no reason to assume that the economic behaviour of adolescents structurally differs from adult economic behaviour. Adolescents’ preferences may differ from adults’ preferences, but the structure to make choices from those preferences may be the same. Let us assume that teenager j, j 5 1, ... , J has preferences that can be described by the utility function: U j 5 U~c 1j, c 2j, ... , c nj!

~3.1!

where c ij denotes the quantity of good i consumed by person j, and n denotes the number of goods. We assume that c ij . 0 ;ij and that the utility function is quasiconcave and twice differentiable. Unlike American pupils, Dutch adolescents do not have to pay considerable amounts of money for college education, education being almost free. So, we do not have to insert future consumption as a factor into the utility function. Adolescents do save, but that is mainly to buy durable consumption goods ~e.g. hi-fi equipment!. We allow for this possibility of financial behaviour ~and the expenditures on durables! by inserting the monthly amount of saving into the utility function as if it were a consumption good. 3.1 The AID System We chose to use the Almost Ideal Demand ~AID! system of Deaton and Muellbauer ~1980! to describe the expenditures of Dutch adolescents. Assuming that all respondents face the same prices, we set prices equal to a constant. Then, the Working-Leser budget share equations result: w ij 5 a i 1 b i ln y j

i 5 1, ... , n

~3.2!

where w i denotes commodity i’s budget share for teenager j, (a k 5 1, (b k 5 0 and y j stands for income of teenager j. In daily practice parents make arrangements with their children that part of the spending on e.g. clothing is paid by the teenager himself out of allowances, while another part is paid by the parents themselves. In that case we speak of ‘co-spending’ analogous to co-insurance. For 1990 and 1992 the rate of co-spending was asked for three categories, viz. clothing, school costs and travelling expenses. 5 Formally, we can compare co-spending to price variation. If there is 5 In 1992 possible co-spending was also asked in the category magazines, which brings the number of commodities where co-spending is possible on four. In this chapter the theoretical explanation of the estimation technique is laid out for 1990, but can easily be adjusted for 1992.

374

M.F. WARNAAR AND B.M.S. VAN PRAAG

1/3 co-spending by the parents for clothing, this may be interpreted as a price reduction 6 for the buying teenager on clothing by 1/3. So we see that respondents are in practice faced with price variation. For these three categories we call relevant prices pseudo-prices. We assume the order of commodities to be such that commodities 1, 2, and 3 are clothing, school costs and travelling expenses. We inserted these pseudo-prices in the AID system. Equation ~3.2! now changes into: 3

w ij 5 a i 1 b i ln y j 1

(

m51

g im p mj 1 e ij

~3.3!

where an error term e ij is introduced. This term is assumed to be N~0, S!distributed. This system has to satisfy the usual restrictions, indicating that the sum of the budget shares should equal one. m

n

( a 51

i51

;m

( b 50

i

j51

i

g im 5 g mi

(g

kl

5 ( g lk 5 0

i, m 5 1, ... , 3

3

(

m51

g im 5 0 ~3.4!

As we do not know the pseudo-prices for all categories, we assume constant prices for the categories where co-spending is not observed. Hence we do not need to impose the additivity restrictions on the parameters g il of the unknown pseudo-prices. As we assume that preferences and consequently elasticities vary over individuals, we parameterize the preference parameters a i, i.e. we assume: K

a ij 5 w i0 1

(w

k51

ik

~3.5!

x kj

where ~the vector! x j stands for the personal characteristics of j. This technique is known as demographic translation ~Pollak and Wales, 1981!. Other techniques are also possible, but demographic translation preserves linearity and keeps estimation simple. Equation ~3.3! now turns into: K

w ij 5 w i0 1

(

k51

3

w ik x kj 1 b i ln y j 1

(

m51

g im p mj 1 e ij .

~3.6!

6 Another point of view for this phenomenon may be that parents may be willing to pay for a certain amount of clothing beyond which the adolescent has to pay for clothing himself. In that case the adolescent faces a non-linear price schedule. However, it is questionable of such strict agreements between parents and children exist. In quite a lot of cases we see that spending on clothing by adolescents is less than the clothing allowance received from the parents. So, money that is meant for spending on clothes is also spent on other commodities.

DUTCH TEENAGERS

375

Combining ~3.5! with ~3.4! yields the additional restrictions on w: n

(

i51

n

w i0 5 1

(w

i51

ik

50

k 5 1, ... , K .

~3.7!

These restrictions have an impact on the variance-covariance matrix S. As is usual in the analysis of complete demand systems the variance-covariance matrix has rank ~n 2 1!, because according to ~3.6! the following holds: n

(e

i51

ij

50 .

~3.8!

One way to circumvent this problem is to reduce the N-equations system by one equation ~Barten, 1969!. The choice of the equation to be eliminated does not alter the results. 3.2 Zero-expenditures Still, more adaptations have been made. Due to the fact that we split up expenditures into 15 categories, which is an unusually fine differentiation, and due to the fact that many things are entirely paid for by the parents, we have a lot of zero-expenditures. Especially for young teenagers, some expenditure categories are totally covered by their parents. We can interpret this as a reduction of the individual’s choice set. Instead of ~3.1!, their utility function will look like: U j 5 U~q 1j, ... , q k 2 1, j, q k, j, q k 1 1, j, ... , q l 2 1, j, q l, j, q l 1 1, j, ... , q n, j! ~3.9! when parents pay commodity k and commodity l. We assume that the preferences for goods are the same for all adolescents, whether in one’s budget set or not. Some goods are surely not paid for by the parents, but still a zero-expenditure is reported. In this case the good is not wanted by the adolescent himself. Then we have a corner solution, i.e. the restriction q i . 0 turns into q i 5 0. Houthakker ~1954! was the first to recognize this problem. He treated it as a special way of rationing. Wales and Woodland ~1983! proposed two econometric models dealing with these non-negativity constraints. With their first approach, the AmemiyaTobin approach, a specification for the utility-maximizing shares is given. Stochastic elements are incorporated by adding truncated multivariate disturbances to these deterministic shares. These ‘real’ shares are mapped onto the boundaries of the unit simplex, if they are outside this unit simplex, i.e. if they are below zero or above one. However, Wales and Woodland preferred another approach, the Kuhn-Tucker approach. One derives conditions on the marginal utilities of each good via the Kuhn-Tucker conditions for constrained maximization. It is as-

376

M.F. WARNAAR AND B.M.S. VAN PRAAG

sumed that preferences are randomly distributed over the population. This assumption is imposed by separating the marginal utilities into a deterministic and a random component, which describes the differences in taste. The resulting density function is supported by all possible consumption patterns, with the probability of a zero-consumption expressed as the integral of the multivariate normal density over the non-bought goods. This means the computation of many multi-dimensional integrals which is only possible if the number of goods is small ~three or less!. Ransom ~1983! showed the relationship between the Kuhn-Tucker approach of Wales and Woodland and the simultaneous equation system with limited dependent variables discussed by Amemiya ~1974!. Amemiya proposed a two-step estimation strategy that is more practical than maximum likelihood, especially when the number of goods is large. This twostep estimator was generalized by Lee ~1978!. Following Ransom’s suggestion we will use Lee’s model to estimate the share equations. We have to deal with a model in which the dependent variables are assumed to be censored by a subset of unobservable latent variables z. Each dependent variable, in this case the budget share w ij, is either a positive amount or zero. Those shares that are zero are censored by an unobservable latent variable I ij* which indicates the decision to buy a specific good or not. 7 If I ij* is positive one buys the good, if I ij* is negative, one does not. We cannot observe the latent variable I ij*, but we can observe the binary indicator I ij, which indicates if someone buys the good or not. I ij 5 1 if j buys good i 5 0 if j does not buy good i

~3.10!

I ij* can be explained by means of a probit model with personal characteristics z ij as independent variables and y ij as a normally distributed disturbance term. L

I ij* 5

(d

i51

z 1 v ij

il lj

~3.11!

Because the disturbance term of the budget share equation ~3.6! is censored, its expectation is not zero. The estimation technique to overcome this problem is described in Appendix A. The choices of explanatory variables and the results of the estimations are presented in the next section.

7 Zero-expenditures may arise in this situation for two reasons: because the expenditure is totally covered by the adolescents’ parents or because the adolescent does not want to buy the commodity. Zero-expenditures might also arise because of infrequency of purchase. As we asked the teenagers to estimate their monthly spendings on the various categories, infrequency of purchase does not cause zero-expenditures in our case. For a general analysis of zero-expenditures, see Van Soest, Kooreman, and Kapteyn ~1993!.

DUTCH TEENAGERS

377

4 IMPLEMENTATION

In this section we will describe how we implemented the estimation technique on the three data sets. We screened the data sets carefully and removed adolescents with missing or inconsistent replies on at least one of the variables needed. Furthermore we checked if the total of reported expenditures was about as high as the reported income. As a result we discarded many observations. We estimated the model for 9,861 adolescents in 1984, for 10,558 adolescents in 1990 and for 13,457 adolescents in 1992. As we want to make a comparison of the rsults of all three data sets, we had te redefine some expenditure categories. We ended up with 15 expenditure categories, as shown in Table 4.1. One of these equations should not be estimated for identification reasons, as explained in section 3.2. We chose the category ‘other.’ 4.1 Explanatory Variables Both for the probit equations and for the estimation of the AID system, we used the same explanatory variables. We will give a short description of them. One of the main factors influencing the decision to buy a good is of course the teenager’s income. To satisfy the adding-up criterion of the AID system we do not use the reported income of the teenager, but rather the total of the observed expenditures and savings as a measure for income. 8 The higher the teenager’s income, the higher the probability that an expenditure category is covered by the teenager himself. We specified a quadratic function in the logarithm of the monthly total expenditures ~INC and INC2!. Other used variables are demographic variables, such as the logarithm of age ~AGE!, a dummy variable, which is one for boys and zero for girls ~GENDER!, and dummy variables for type of schooling ~MAVO ~lower general secondary education!, HAVO ~higher general secondary education!, VWO ~pre-university education!, MBO ~intermediate vocational training!!, the lowest education level ~LBO! ~lower vocational training! being the reference level. To get insight into the attitude of the parents towards letting their children pay for particular goods, the family background of the adolescent was also taken into account. The way the father of the adolescent earns his income ~wage earner ~WAGE!, or self-employed ~SELF!! was included into the group of explanatory 8 The assumption of exogeneity of income may be questionable, as a large part of the income is earned from work. The decisions about the amount of total income, the number of working hours, the choice of spending categories and the amount of spendings on these categories are made simultaneously – together with the choice of time spending – conditional on background characteristics of both the teenager and his parents. These other decision areas fall beyond the scope of this paper. For the decisions about the amount of income and working hours in the 1984 data set we refer to de Zwart and Van Praag ~1989!.

378

M.F. WARNAAR AND B.M.S. VAN PRAAG TABLE 4.1 – EXPENDITURE CATEGORIES USED IN THE AID-EQUATIONS 9

new category 1984

1990

1992

Snacks

Snacks, crisps, etc.

Snacks, crisps, etc.

Snacks, crisps, etc.

Non-alcohol

Non-alcoholic beverages

Non-alcoholic beverages

Non-alcoholic beverages

Alcohol

Alcoholic drinks

Alcoholic drinks

Alcoholic drinks

Sweets

Sweets

Sweets

Sweets

Tobacco

Smoking

Smoking

Smoking

Travel

Travelling expenses ~public transport and own transport!

Travelling expenses

Travelling expenses ~public transport and own transport!

Going out

Cinema, theater, disco, etc.

Cinema, theater, disco, etc. ~Consumptions excluded!

Cinema, theater, disco, etc. ~Consumptions excluded!

School

School supplies

School supplies 1 school books

School supplies 1 school books

Vice

Gambling 1 Drugs

Gambling 1 Drugs

Gambling 1 Drugs

Audio

Records, Music casRecords, CDs, Music settes 1 Photo material cassettes

Records, CDs, Music cassettes

Gifts

Gifts

Gifts

Gifts

Magazines

Books and magazines

Magazines

Magazines

Clothing

Clothing and shoes

Clothing and shoes

Clothing and shoes

Saving

Saving

Saving

Saving

Other

Sports equipment 1 Contributions 1 Other costs

Computer needs 1 Insurance 1 Contributions 1 Other costs

Computer needs 1 Insurance 1 Contributions 1 Other costs

The first column contains the ‘new’ expenditure categories, the second, third and fourth colum state the expressions given in the questionnaires of 1984, 1990, and 1992, respectively. A ‘1’ means that two separate expenditure categories have been combined. 9 As can be deduced from Table 4.1, not all ‘new’ expenditure categories are completely identical. The category ‘other’ consists of all other expenditures, viz., the ones that appeared in only one of the questionnaires, the category ‘contributions’ ~i.e. membership fees!, which are hardly paid by pupils themselves and the category labeled ‘others.’ We left out the 1990 and 1992 expenditure category ‘summer holidays’ as this was probably closely related to monthly saving.

DUTCH TEENAGERS

379

variables. A dummy for a two-earner family ~TWOEARN! is included. Further, a dummy variable was inserted for religious behaviour ~RELIGIOUS!. Regional differences may also cause differences in spending patterns. With the central Dutch provinces as a reference, dummies were specified for the regions NORTH, EAST, WEST and SOUTH. Also the degree of urbanization ~URBAN! was inserted, ranging from 1 for cities with over 100,000 inhabitants to 5 for villages with less than 5000 inhabitants. Special characteristics of the adolescent himself can also lead to differences in spending behaviour. Perhaps ‘going steady’ ~STEADY! leads to different spending behaviour. This variable was not available for the 1984 data set. The influence of having a part-time job on the behaviour of teenagers is widely discussed. Greenberger and Steinberg ~1986! conclude that in the USA working adolescents are likely to have more freedom in spending decisions than their non-working peers. Manning ~1990! draws the conclusion that in the USA woeking adolescents have more disagreements with their parents about spending behaviour than non-working adolescents. Furthermore, she concludes that the more an adolescent earns, the more his parents try to control his spending pattern. Note that working and the amount of income are correlated. A significant influence of the dummy variable JOB indicates that having a job has an impact on the spending pattern ceteris paribus, i.e., compared to an adolescent who gets the same income from other sources. Greenberger and Steinberg ~1986! and Manning ~1990! did not take that difference into account, although they speculate that income may have an influence. The negative influences on psychological behaviour might not be caused by working as such, but by having a large amount of money at one’s disposal. Without some parental control, adolescents will be tempted to spend their ~high! income in a way which may be undesirable in the view of parents and/or society. The last variable we take into account is a dummy variable A18, which equals one for adolescents of 18 years and older. In this way we can test whether the introduction of the basic grant in 1986 changed spending behaviour, or whether the age of 18 is a ‘magic’ age anyway. The effect of free public transport for students in 1992 can also be investigated by means of this variable. 4.2 Probit Equations Decisions about buying or not buying a particular good are modeled by probit equations. Not buying a good may come about because of two reasons. First, parents can buy the good for their child and secondly, even if the children are allowed to buy a particular item themselves, they may not want to do so. We do not have exact information what the reason is for not buying. So, for some goods a mixture of the two reasons may prevail. We will not go into detail about all effects; we will only highlight the main results. Most commodities can be classified either as typical boy or typical girl

snacks 84 2 1 2 2 2

90

92

84

90

1

2 1

1 2

2 2

2 2

2

1

1 2

2 1 2 1 2

1 1 1 2 2 1 1 1 1 1 1 2

1 1 1

alcohol 92

1 1 1 1 1

1 1 1 1

1

1 1 1

1 2

1 1 1 2

1 2

travel

school

90

92

84

90

92

84

90

92

84

90

92

84

90

92

1 1

1 1

1 1

2 2 1 1

2 2

2 2

1 2 2 2 2 2

1 2 2 2 2 2 2

1 2 2 2 2 2

1

1 2 1 1 1 1

1

2 2 1

2 2

2

1

1 1

2

2

2

1 2 2 1 1 1

1 1

1 1

2 2 1 1 2

2

1 1 1 1 2 2 1 1

1 2

smoking

84

2 2 2 1 2

sweets

2 1 2

1 1 2 1 1 1 1 1 1 2 1 2

1 1 1 1 1 2 1 1 1

2 1

1

1 2 1 1

1

1 1

1 2 1

1 1 2

1 1 1 1

1

2 2 2 1 1

1 2

1 2

1 2

1 2

2 1

2 1

1 2

1 1 2

2 1 2

1

M.F. WARNAAR AND B.M.S. VAN PRAAG

age gender mavo havio vwo mbo wage self twoearn religious north south east west urban steady job a18 inc inc2

non-amc

380

TABLE 4.2 – RESULTS OF THE PROBIT EQUATIONS

TABLE 4.2 ~CONTINUED!

vice 84

90

92

2 1

1

1

2 2 2

2 2 2

audio

84

90

92

84

90

1 2 1 1 1

1 2 1 1 1

1 2 1 1 1

2 1 1 1 1 1

2 1 1 1 1

2

2

magazines

92

84

90

1 1 1 1 1

2 1 1 1 1

2 2 1 1 1 1

1

2

2

2

2 1 1 1

2 1

gifts

1 2 2 1

2 2 1

1 2

2 1 1 1 1 1

84

90

2 2 1 1 1 1

2 2 1

2

92

2 1 1 1 1 1 2

saving

84

90

92

84

90

92

1 2 1 1 1

1 2

2 2

1 1

1 1

1 2

2 2 1 1 1 1 1

2

2

2

2

2 2

1

1

1 1 2

1 1

2 2 2

1

2 1 2

2

2

2

2

2

2

1 1 2

1 1 2

2 1

2

1

1 1

2 1 2

2 1 2

2

2 2

1 2

1 2

1

2 1

1 2

92

clothing

1 2

1 2

1 1 2

1 1 1

2 1 2

2 2 2 1 1 2

1 2

2 2 1 1 2

1 2

1 1 2

1 1 1 1

1 1 1 1 1 1

DUTCH TEENAGERS

age gender mavo havio vwo mbo wage self twoearn religious north south east west urban steady job a18 inc inc2

going out

1 1 1 2 381

The influences are significant at the 1% level. Blanks are not significant. The exact values may be obtained from the authors. The shaded effects could not be estimated because of missing data.

382

M.F. WARNAAR AND B.M.S. VAN PRAAG

expenditures. For almost all commodities an increase in the teenager’s income leads to a higher probability of buying these commodities. Age also plays an important part in whether or not to purchase particular goods of not. Differences in type of schooling do exist, but family characteristics hardly have an impact on the spending decision. Adolescents from the southern, formerly Catholic, provinces have a slightly different expenditure pattern. They have to pay less for clothing and travelling, but they go out and drink alcohol at an earlier age. Going steady is correlated with most spending categories; going out, travelling and the use of alcohol, tobacco, drugs and gambling machines increase when going steady, while ‘home activities’ like reading magazines, listening to music and using school supplies is less favourite. The introduction of the basic grant made adolescents financially more independent at the age of 18. Compared to 1984, more had to pay for school supplies, magazines and clothing themselves. In 1990, they also had to pay travelling expenses themselves, but free public transport in 1992 lowered the cost for transport for teenagers over 18. Having a job hardly causes differences in the expenditure patterns ceteris paribus. We do see that adolescents with a job have a larger propensity to save, probably to buy durable consumption goods after a couple of months. 4.3 The AID System Estimated We used the same explanatory variables in the AID system to explain the budget shares as in the probit equations. Differences in the budget share of a particular good may be caused by age, gender, type of schooling, family characteristics, region, and urbanization. It is also possible that having a job may shift the amount of the budget shares. We will also investigate if the basic grant has changed relative spending in the expenditure categories. Co-spending is analyzed by means of pseudo-price elasticities. Income elasticities can be calculated from income parameters and average budget shares. In order to calculate pseudo-price and income elasticities for both boys and girls, we inserted interaction terms of pseudoprices and gender, and income and gender, respectively. The estimation results for the AID equations can be found in Appendix B. The first important fact to note is the significance of the correction terms in nearly every equation. It shows that censoring is an important factor, which should clearly not be neglected. It indicates that there is a correlation between the error terms of the probit equations and the AID equations. Demographic effects We will not describe all significant effects. It is clear that gender 10 has its influence, just as age. The shares of the first goods children buy ~sweets, gifys, etc.! 10 One should keep in mind that gender influences the budget shares in two ways, viz., also through the interaction with income.

DUTCH TEENAGERS

383

decrease as they get older, because more expenditures have to be covered. Family characteristics hardly have any influence. Adolescents in the southern, formerly Catholic, part of The Netherlands have a slightly different expenditure pattern. As less of their income is spent on clothing and travelling expenses, a larger share of their income is spent on alcohol and going out. Religious teenagers spend less on alcohol, tobacco and ‘vice,’ but more on school supplies and they save more of their income. The introduction of the basic grant made students of 18 year and older pay relatively more for school costs and in 1990 also for travelling. In 1992 however, the amount spent on travelling is much lower for the 18-year old students, because of their free public transport ticket. We do not find the hypothesized negative influences of having a job. The only difference between workers and nonworkers is the higher spending on clothing for non-workers and the higher saving of workers. Pseudo-prices We use the fraction of expenditure adolescents have to pay themselves as a proxy for perceived prices. As parents are co-spending partners for varying amounts we have a good variation of pseudo-prices. We will investigate the influences of these pseudo-prices by means of the calculated price elasticities. These price elasticities could only be calculated for the 1990 and the 1992 data set. The equation for the uncompensated own-price elasticity of expenditure i ~p i! equals:

pi 5 wi 1

g ii wi

21

~4.1!

and the cross-price elasticity of expenditure i from expenditure j ~p ij! equals:

p ij 5 w j 1

g ij wi

.

~4.2!

The price elasticities ~at the means w i! can be found in Table 4.3. We see that most cross-price elasticities are not significant. However, the own-price elasticities cannot be neglected. Most of them have the expected negative sign. It means that adolescents who have to pay for travel costs, clothing and magazines themselves spend less on these categories than adolescents whose parents pay for these costs. Clothing allowances may not only end disagreements between parents and teenagers, they may also save the parents some money! Quite remarkable is the positive own-price elasticity of school costs. However, it stands to reason that these school costs are unavoidable and that one can hardly save on these costs. We do not have a good explanation for the increase in the own-price elasticity from 1990 to 1992. The 1992 own-price elasticity estimate for travel expenses

384

M.F. WARNAAR AND B.M.S. VAN PRAAG TABLE 4.3 – PSEUDO-PRICE ELASTICITIES

Gen- Clothing der 1990 1992

School 1990

Travel 1992

1990

1992

Magazines 1992

20.14* 20.02

0.05 20.08*

20.01 0.12*

0.11* 20.10

0.02 0.01

20.02 20.13

20.10 20.07

Non-alcohol B G

20.05 20.09

20.20** 20.24**

0.01 0.08

20.10 20.19*

20.18** 20.07 20.10 20.02

20.02 20.09*

Alcohol

B G

20.16** 20.27** 20.21** 0.06 20.11 20.20** 0.07 20.17

20.09 20.24*

0.07 0.10

20.12** 20.20**

Sweets

B G

20.01 0.16**

0.19 0.12

20.07 0.00

20.12 0.02

0.04 20.12**

Smoking

B G

20.12 0.03

0.09 20.13*

0.20 20.08

0.14 20.05

20.24* 0.11

0.03 20.04

Travel

B G

20.18 20.13

20.06 20.19

Going out

B G

20.21** 20.15* 20.15 20.21** 20.21** 0.02

School

B G

20.28* 20.13

20.19 20.08

Vice

B G

20.37* 20.13

20.28** 20.18 20.40 0.33

Audio

B G

20.06 20.13

20.13* 20.18*

0.06 0.14

Gifts

B G

0.02 0.07

20.09 20.12*

Magazines

B G

20.02 0.01

20.03 20.00

Clothing

B G

20.43** 20.44** 20.56** 20.71** 20.24** 20.24** 20.07 20.09** 0.17** 20.60** 20.61** 20.26** 20.24** 20.15**

Saving

B G

20.16** 20.18** 20.10** 20.15**

Snacks

B G

0.06 20.00 0.14** 0.19* 20.22 0.13 0.62** 0.24

0.28 0.31*

0.03 0.09

20.40** 21.90** 0.45** 20.58** 21.83** 20.03

20.22** 20.10 20.32** 20.06

20.08 20.08

20.12** 20.01

1.09** 20.40** 20.53** 20.06 0.67** 20.08 20.06 20.04 0.20 0.53

0.00 0.01

0.02 0.26

20.26** 0.10

20.17* 20.02

20.16 20.08 20.22** 20.06

20.22 0.06

20.08 0.05

20.08 20.00

20.15 0.06

20.14* 0.09**

20.01 0.03

20.16 20.01

20.17 0.06

20.16 20.14

20.95** 20.76**

0.07 0.14*

0.05 0.14*

20.11* 20.06

20.04 20.08

0.01 0.02

20.12* 20.16**

Own-price elasticities are printed in bold. The pseudo-price elasticity for magazines was only available in 1992. * means significant at 5% level; ** means significant at 1% level. The standard errors are assessed by the well-known d-method.

DUTCH TEENAGERS

385

changed a lot from 1990 to 1992. This is probably caused by the free public transport programme for students 18 years and over, which was started by the government in 1991. The influence of income The income elasticities may be calculated from the AID system as follows: ei 5

­q i ? y ­y ? q i

511

bi wi

S

5 a i 1 b i ln y 1 .

n

(

m51

g im p mj 1 b i

D S y

qi

5 wi 1 bi

D

1 wi ~4.3!

To calculate them, we use the average budget shares. These average shares are tabulated in Table 4.4 for boys and girls separately just as the income elasticities. We can label some expenditure categories as ‘luxury goods.’ Alcohol, tobacco, travelling expenses; ‘vice’ and clothing have an income elasticity above one in all three surveys for both girls and boys. However, most goods are necessities: their income elasticity is between zero and one. Very low income elasticities are found for sweets, school material, and magazines. Inferior goods are not found within these expenditure categories. We do not see striking differences for income elasticities between the sexes. The income elasticities of most expenditure categories are quite stable over time. Exceptions are the increasing income elasticities for school supplies and audio material and the decreasing income elasticities for sweets and saving. So, on the one hand we see that teenagers with higher incomes are urged to spend their higher income on ‘necessary’ categories, like travelling expenses and clothing, which had formerly been paid by their parents. However, on the other hand, there is a tendency to spend extra money on alcohol, tobacco and gambling. Having a job implies extra revenues, and spending on ‘vice’ increases accordingly, but there is no additional increase due to having a job. 5 CONCLUSION

The income of Dutch teenagers attending secondary schools has risen considerably in the eight years from 1984 to 1992. Monthly, these adolescents received Dfl. 180 in 1984, Dfl. 320 in 1990 and Dfl. 350 in 1992, while accumulated inflation has been 13% over these 8 years. The rise in income has been caused by all three sources of income. Not only did adolescents receive more allowances from their parents, they also worked more and received higher wages. Above all, the introduction of a study grant for 18-year old students, directly sent to their own bank account, increased these students’ monthly earnings. Because most of

Budget share ~in %! Boys

Income elasticity Girls

Boys

1990

1992

1984

1990

1992

Snacks

8.6

6.5

7.7

5.2

4.7

Non-alcohol

5.9

6.2

5.6

6.7

6.3

Alcohol

9.9

9.5

9.9

5.9

Sweets

6.8

6.2

6.4

Smoking

2.8

2.5

3.0

Going out School

Girls

1984

1990

1992

1984

1990

1992

9.4

0.60**

0.79**

0.85**

0.77**

0.78**

0.94**

5.6

0.87**

0.75**

0.96

0.88**

0.80**

1.07*

4.3

4.6

1.36**

1.32**

1.44**

1.33**

1.30**

1.39**

8.4

7.2

8.9

0.61**

0.51**

0.50**

0.64**

0.35**

0.34**

4.3

2.6

3.9

1.20**

1.26**

1.12**

1.17**

1.24**

1.30**

3.1

2.7

2.2

2.1

2.1

2.2

1.09

1.06

1.15**

1.20**

1.18**

1.36**

10.0

7.5

6.8

10.2

8.3

6.7

1.17

0.98

1.14**

0.97

0.87**

1.11**

1.2

2.1

1.6

1.5

2.6

1.9

0.83**

0.86**

1.01

0.71**

0.74**

0.90*

Vice

1.9

1.8

2.0

0.4

0.5

0.5

1.40**

1.64**

1.70**

1.84**

2.13**

2.04**

Audio

8.7

9.4

7.1

6.4

5.8

4.6

0.72**

0.75**

0.94*

0.74**

0.86**

1.14**

Gifts

5.5

5.3

3.3

9.5

9.1

6.8

0.58**

0.59**

0.65**

0.51**

0.48**

0.66**

Magazines

2.7

1.5

1.1

4.2

2.7

2.0

0.41**

0.67**

0.32**

0.46**

0.54**

0.41**

Clothing

5.1

8.1

8.5

12.5

15.7

13.2

1.98**

1.51**

1.60**

1.77**

1.40**

1.40**

Saving

14.6

18.1

21.2

13.8

19.9

18.8

1.07**

1.25**

0.70**

0.92**

1.30**

0.75**

Other

13.2

12.7

9.0

7.9

8.1

10.9

1.05

0.82

1.59

1.28

1.06

1.32

* means: significantly different from 1 ~5% level!, ** significantly different from 1 ~1% level!. Standard errors assessed by d-method.

M.F. WARNAAR AND B.M.S. VAN PRAAG

1984

Travel

386

TABLE 4.4 – AVERAGE BUDGET SHARES AND INCOME ELASTICITIES ~SEE EQUATION 3.9!

DUTCH TEENAGERS

387

these adolescents still live with their parents, they do not have to pay fixed costs, such as rent, insurance and so on. They can spend most of their income freely on the things they want. So it is hardly surprising that most teenagers do not have financial difficulties. Adolescence is a stage of life in which one has to make the first steps towards adult life. Handling money is a part of that. Most of the adolescents report to manage their money well. About 25% of all adolescents state that their managing of money should improve and only 3% report that they spend their money badly. Adolescents have become financially more independent from their parents in these eight years. This is clear from the increase in the percentage of adolescents spending on almost all types of commodities. Because of the increased income the average spending on most commodities also increased. Expenditure on most commodities more than doubled in these eight years, especially costs related to outdoor activities. Clothing and saving remain the main expenditure categories. The introduction of the basic grant has led to a shift in the amount of contributions the parents make. One of the goals of introducing the study grant was to make 18-year old adolescents financially independent. It appears that a lot of the expenditure categories are now indeed covered by 18-year old adolescents themselves, but still a fair amount is paid by the parents. In sections 3 and 4, a model was built based on the AID system of Deaton and Muelbauer. It was assumed that the decision to buy a certain commodity was censored by a variety of variables. On the one hand it was possible that parents made the purchase for their child, on the other hand it could be a decision of the adolescent himself not to spend that money even if he was allowed to. Probit equations were estimated for the decision to spend money on a particular expenditure category or not. The results of these estimations were used to construct a correction variable that was included in the AID equations. With the AID system the budget shares were modeled. We saw, besides some demographic effects, that adolescents whose parents pay a smaller part of some expenditure categories are inclined to reduce their spendings on these categories, relatively spealing. Adolescents whose parents do not pay for clothing will not spend twice as much on it as adolescents whose parents pay half of the expenses for clothing. Having a job does not lead to more ‘vice’ spending ceteris paribus. However, having a job is mingled with a higher income. This higher income may have negative consequences. When the income of an adolescent rises, a relatively large part of this higher income is spent on both necessary commodities as clothing and travelling expenses and luxury commodities like alcohol, tobacco, gambling, and drugs. This article deals with the financial behaviour of Dutch teenagers. We do not know of any research on this topic in other OECD countries, but it is our opinion that in those countries teenagers face comparable situations and choices on their way to become financially independent citizens.

388

M.F. WARNAAR AND B.M.S. VAN PRAAG

APPENDICES APPENDIX A: ESTIMATION TECHNIQUE TO ACCOUNT FOR ZERO-EXPENDITURES

It holds that:

S( S( S( S( L

SU D SU

L

E e ij I ij 5 0 5 E e ij v ij $

(d

l51

z

il lj

D

f

l51

5 si

d il z lj i

D

L

12F

l51

and also that:

d il z lj

L

SU D SU

L

E e ij I ij 5 1 5 E e ij v ij ,

(d

l51

z

il lj i

D

f 5 2 si

l51

d il z lj

L

F

l51

D D D

~A.1!

d il z lj

~A.2!

where s i denotes the covariance between the error terms v ij and e ij. A way to tackle this problem is by introducing an error term z ij, the expectation of which is zero. The equation for the budget share according to the AID specification ~3.9! now turns into: K

w ij 5 w i0 1

si G

(w

k51 L

3

ik x kj 1 b i ln y j 1

S( U D l51

(

m51

g im p mj 1

d il z lj I ij 1 z ij

~A.3!

L d il z lj u I ij! is defined as: where G ~(l51

S( D S( D L

S( U D L

G

l51

d il z lj I ij 5

f

l51

d il z lj

L

12F

l51

~A.4!

d il z lj

if the good is bought and

S( D S( D L

S( U D

f

L

G

l51

d il z lj I ij 5 2 2 F

l51 L

l51

if the good is not bought.

d il z lj

d il z lj

~A.5!

DUTCH TEENAGERS

389

The way to proceed is now as follows: First we estimate a probit model on the binary variables I ij, which yields us estimates of the parameter vector d, say d |. This d | is plugged into ~A.3!: K

w ij 5 w i0 1

si G

3

(

w ik x kj 1 b i ln y j 1

k51 L

S( U D

(

m51

g im p mj 1

|d il z lj I ij 1 z ij .

l51

~A.6!

Then, the AID equations ~A.6! can be estimated. The only problem is that specification ~A.6! does not satisfy the adding-up restriction. If all n equations were specified according to ~A.6! the budget shares would not add up to one. In that case, the following should hold: n

S

L

( s G ( |d

i51

i

l51

z

il lj

UD

I ij 5 0 .

~A.7!

L |d il z lj u I ij! depends on I and z lj for each teenager j, such a restriction As G ~(l51 is not possible in general. Therefore the nth equation, which is not estimated, is differently specified, as suggested by Heien and Wessels 11 ~1990!. K

w nj 5 w n0 1

(

k51

n21

(

i51

3

w nk x kj 1 b n ln y j 1

S( U D

(

m51

g nm p mj 2

L

si G

l51

d| il z lj I ij 1 xz nj .

~A.8!

The resulting error term of ~A.6! xz ij equals

F S ( U D S ( U DG L

xz ij 5 z ij 2 s i G

l51

L

|d il z lj I ij 2 G

l51

d il z lj I ij

~A.9!

so that the variance of xz ij does not equal the variance of z ij. Maddala ~1987! gives an expression of the variance of the estimators wˆ , b | , gˆ and s ˆ . it is:

11 Note that it is possible that some estimated budget shares are below 0 or above 1. This could be solved by using a Dirichlet distribution instead of a normal distribution ~Woodland, 1979!. For obvious complexity we have abstained from doing this ~see e.g. Wunderink, 1988!.

390

M.F. WARNAAR AND B.M.S. VAN PRAAG

34 wˆ

var

b | gˆ

s ˆ

5 ~CJI K15! ~P8P! 21 2 ~SJI K15! ~P8P! 21P8@D 2 DH ~H8LH! 21H8D#P~P8P! 21 ~A.10!

with

C the @N 3 N#-diagonal matrix of the variances of the regressions ~A.4!

F

P the @NJ 3 N~K 1 5!#-matrix x ij, ln y j, p mj, G

S ( U DG L

l51

d| il z lj I ij

,

S the @N 3 N#-diagonal matrix with the estimated s i , D a @NJ 3 NJ#-diagonal matrix with terms

S( U D F S( U D S D ( G L

G

l51

L

d| il z lj I ij ? G

l51

d| il z lj I ij 1 2 I ij 2

1

L

2

l51

|d il z lj ,

H the @NJ 3 L#-matrix Z ,

L a @NJ 3 NJ#-diagonal matrix with terms

S( D S( D S( D S( D L

f

l51

L

d| il z lj

?

L

F

l51

f

d| il z lj

l51

d| il z lj

L

12F

l51

|d il z lj

This expression is used to calculate the standard errors of the AID equations.

DUTCH TEENAGERS

391

APPENDIX B: RESULTS OF THE AID SYSTEM ~EQUATION 20, COEFFICIENT MULTIPLIED BY 100!

Snacks 1984

Non-alcohol 1990

age –0.2 0.4 gender 13.5** 3.8** MAVO –1.7** –0.7pp HAVO –2.6** –1.8** MBO –1.8** –1.7** VWO –3.5** –2.6** wage 0.1 –0.5** self 0.2 –0.1 urban 0.3** –0.2** twoearn –0.3* –0.0 religious –0.2** 0.1* north –1.3* –0.2 east –0.2 –0.2 west –0.5 –0.2 south –0.6 –0.5 job –0.2 –0.3** a18 –0.3 –0.8 steady 0.3* ln ~inc! –1.2** –1.0** gender* –2.3** –0.3** ln ~inc! P ~school! 0.1 gender* 0.1 P ~school! P ~travel! –0.4 gender* –0.4 P ~travel! P ~cloth! –0.2 gender* –0.4 P ~cloth! P ~magaz! gender* P ~magaz! Heckman –10.5** –7.0** constant 11.5** 12.2** R2 0.56 0.47

1992 –1.3** 6.1** –0.7** –0.3** 0.5 –3.0** –0.8** 0.0 –0.9** 0.3 0.8** 0.5 0.2 2.7** 1.4** –1.8** 0.0 0.2 –0.5** –0.9**

1984

1990

2.4** –0.6 –0.1 0.5** 0.5* 0.5 –0.2 0.2 0.1 0.7** 0.4** 3.3** 1.3** 0.4 2.2** –1.1** –0.9**

–1.0** 1.5* –0.5** 0.2 0.8** 0.6** 0.1 0.8** 0.0 0.4* 0.1 0.7 –0.1 –0.2 0.8** 0.1 1.4** –0.1 –0.8** –1.3** 0.0 –0.3

Alcohol 1992 3.1** 2.6** –0.5** 0.7** 0.6* 1.0** 0.3 0.2* –0.0 –0.4** 0.0 0.9** 0.5* 0.4 1.4** –0.5** –0.8** 0.0 0.4** –0.6**

1984 7.6** –3.6** –1.9** –0.9** 0.7 –1.9** 0.5* 2.3** 1.2** –0.9** –0.9** 1.7** 0.5 0.8 5.9** 0.3 –2.0** 2.0** 1.6**

1990

1992

6.6** –4.2** –0.8** –0.2 2.4** 0.3 –0.7** 2.3** 0.9** 0.5* –0.9** 2.0** 1.4** 0.4 3.4** 0.3 –2.9** 0.9** 1.3** 1.7**

4.2** –9.6pp 0.3 2.1** 2.4** 1.5** 0.6** 0.8** 1.1** 0.0 –1.4** 2.8** 0.8** 1.0** 2.3** –0.0 –2.5** 0.5** 1.8** 2.5**

–1.2 0.9

0.3 –0.4

–1.2** 0.6

0.2 –2.4**

–0.9 1.3

–0.3 –1.2*

–0.8* –0.5

–0.2 –0.3

–1.1** 0.0

0.3 0.1

–1.8** –0.1

–1.5** –2.1** 0.7 0.5

–1.2** –1.1*

–1.5** –2.0**

–0.8** 1.2**

–0.6** 0.4

–1.0** –0.3

–9.1** –10.4** –7.9** –6.6** –13.6** –10.6** –12.0** 19.0** 6.0** 13.9** 3.4** –22.5** –18.8** –19.3** 0.38 0.56 0.46 0.42 0.59 0.54 0.61

392

M.F. WARNAAR AND B.M.S. VAN PRAAG

Tobacco 1984

Sweets 1990

age 3.5** 2.1** gender –1.3* –0.7 MAVO –1.0* –0.8** HAVO –2.4** –1.8** MBO –2.6** –1.9** VWO –3.6** –2.4** wage –0.2 –0.2 self –0.5** –0.1 urban –0.1** –0.1** twoearn 0.1 0.5** religious –0.9** –0.5** north –0.7 –0.3 east 0.2 –0.2 west 0.2 –0.3 south –0.1 –0.3 job –0.2 –0.4** a18 –0.8** –0.6** steady 1.1** ln ~inc! 0.7** 0.6** gender* –0.2 0.0 ln ~inc! P ~school! 0.3 gender* –0.9* P ~school! P ~travel! –0.2 gender* 0.5 P ~travel! P ~cloth! –0.3 gender* –0.2 P ~cloth! P ~magaz! gender* P ~magaz! Heckman –10.8** –8.1** constant –4.0** –3.2** 2 R 0.61 0.60

1992 2.8** 2.7** –0.8** –2.5** –2.4** –3.2** –0.6** –0.3** –0.1** 0.4** –0.9** –0.6* –0.2 –1.0** –0.6** –0.6** –1.4** 1.0** 1.2** –0.8**

1984

Travel 1990

1992

–5.3** –5.3** –3.1* –2.3** –7.9** –15.7* 1.2** 20.3 0.1 0.3 –0.4 –0.3 –0.7* 0.1 0.0 –1.0** –0.7** –1.3** –0.4* 0.0 –0.9** –0.6* –0.7** 0.0 –0.1* –0.1 –0.1 0.5** 0.5** –0.0 0.3** 0.1 –0.2* –0.5 –0.3 0.3 –0.2 0.0 1.5** –0.7 –0.1 1.1** –0.6 –1.0** 0.4 –0.1 –0.1 0.5** 0.8** 1.9** 1.4** 0.1 0.5** –3.0** –4.7** –5.8** 0.4* 1.7** 2.6**

–0.4 0.9*

1.2* –1.3

0.9 0.2

0.3 –1.1**

–0.1 –0.4

–0.0 –0.9

–1.0** 1.0**

0.0 –0.6

0.0 –0.2

–0.2 0.3 –9.9** –10.5** –4.9** 31.5** 0.62 0.58

–1.2** 1.4** –7.3** 38.9** 0.50

1984

1990

3.8** 1.6* –0.5** –0.8** –0.6* –1.1** 0.2 –0.0 0.1** 0.7** 0.3** –0.5 –0.1 –0.1 –1.5** 0.4** 1.8**

1.7** 0.8 0.0 0.4* 3.1** –0.4** 0.1 –0.1 0.3** 0.2 0.0 –1.3** –1.1** –0.8** –1.6** –0.4** 1.6** 0.5** 0.4** 0.4** –0.2 –0.2 0.5 1.2*

1992 –0.6** 2.1* –1.0** –1.1** 2.1** –0.7** –0.5* –0.1 –0.0 0.7** 0.2* –0.4 –0.3 0.5 –0.3 –1.7** –2.8** 0.2 0.8** –0.5* 0.2 –0.1

0.8** –1.9** 0.7* –0.2 –0.6** –0.7* –0.1 0.4 –0.1 1.1**

–7.4** –8.1** –5.4** –0.6** 43.3** –4.9** –2.6** 1.3* 0.44 0.42 0.46 0.05

DUTCH TEENAGERS

School

age gender MAVO HAVO MBO VWO wage self urban twoearn religious north east west south job a18 steady ln ~inc! gender* ln ~inc! P ~school! gender* P ~school! P ~travel! gender* P ~travel! P ~cloth! gender* P ~cloth! P ~magaz! gender* P ~magaz! Heckman constant R2

393

Vice

Going out

1984

1990

1992

1984

1990

1992

–1.3** –1.1** 0.8** 0.3** 0.8** 0.3** –0.4** –0.5** 0.1** 0.3** 0.3** 0.4 –0.2 –0.0 –0.3 –0.5** 0.6**

–0.4* –2.3** –0.1 –0.4** 0.8** –0.3* –0.2 –0.3* –0.1 0.2 0.4** 0.1 –0.1 –0.3 –0.5* –0.5** 1.8** –0.1 –0.7** 0.4**

–0.5** –1.4** –0.1 –0.2 0.4** –0.1 –0.2 –0.2** –0.1** –0.0 0.2** 0.4** 0.2 0.1 0.1 –0.1 1.7** –0.3** –0.2** 0.2**

–0.4* –0.5 –0.1 –0.2 –0.2 –0.1 –0.2 0.2 –0.0 0.1 –0.0 0.0 0.0 –0.0 0.2 –0.2 –0.4*

0.2 –1.6** –0.0 –0.6** –1.3** –0.8** –0.2 –0.3* 0.1* 0.1 –0.1 0.6* 0.1 0.0 0.2 –0.4** –0.4* 0.2 0.5** 0.6**

–0.3* –3.3** –0.5** –0.7** –0.4** –0.9** –0.4** 0.0 –0.1** –0.0 –0.2** 0.4* –0.1 0.0 0.3* –0.3** –0.7** –0.0 0.5** 0.9**

3.4** –0.8

3.2** 0.1*

0.1 –0.5

–0.3 –0.6*

–0.2 –0.7**

–0.8** 0.0

–0.4** –0.0

–0.4** 0.2**

0.3** 0.5**

1990

1992

3.2** –5.6** –0.3 1.1** 1.0** 0.2 –0.1 0.2 –0.0 0.6** –0.7** 0.6 1.3** 1.3** 1.8** –0.1 –0.6 1.2** –1.1** 1.0**

4.3** –2.0** –0.7** –0.5 –0.8** –2.0** –0.9** –0.2 0.4** 0.0 –0.2 0.3 1.0** 0.6* 0.3 –0.7** –0.8** 0.5** 0.7** 0.2

0.2 0.1

–0.0 –1.2

–2.3** 0.7

–0.0 –0.0

0.1 –0.1

–0.7 –0.2

–0.7 0.0

–0.1 –0.7*

–0.3 –0.5

–3.0** 0.8

–2.3** 0.7

20.1 0.0 –4.8** 6.1** 0.49

–3.9** 8.4** 0.36

–3.4** 4.2** 0.41

1984 4.3** –9.2** 0.5 0.9** 3.3** 2.0** –0.1 0.9** 0.4** –0.3 –0.1** –0.2 –0.0 –0.4 1.9** –0.9** –1.4** –0.3 2.0**

0.0 –0.6** –5.2** –0.1 0.37

–4.6** –2.2** 0.26

–4.5** –1.6** 0.31

–0.2 –0.7* –10.2** –0.5 0.38

–7.6** 5.4** 0.30

–7.7** –2.1** 0.37

394

M.F. WARNAAR AND B.M.S. VAN PRAAG

Audio

age gender MAVO HAVO MBO VWO wage self urban twoearn religious north east west south job a18 steady ln ~inc! gender* ln ~inc! P ~school! gender* P ~school! P ~travel! gender* P ~travel! P ~cloth! gender* P ~cloth! P ~magaz! gender* P ~magaz! Heckman constant R2

Gifts

Magazines

1984

1990

1992

1984

1990

1992

1984

1990

1992

–1.1** 5.6** 2.7** 2.6** 1.2** 3.1** 0.7** 0.3 –0.6** –0.4* –0.3** 1.0 –0.1 –0.5 0.3 0.0 –0.2

–2.3** 11.8** 1.0** 1.5** –0.3 1.2** 0.3 0.5 –0.0 –0.8** –0.5** –0.7 –1.2** –0.7 0.0 –0.4* 0.9* –0.7** –0.8** –1.5**

0.4* 8.3** 0.5** 0.3 –0.6* 0.4 0.3 –0.0 –0.1* 0.0 –0.4** –1.3** –1.1** –1.7** –0.3 –0.2 –0.8** –0.8** 0.6** –1.1**

0.0 –13.5** 1.3** 1.9** 1.3** 2.4** –0.4* 0.2 –0.3** 0.1 0.2** –1.9** 0.0 0.5 –0.2 0.6** 0.4

–0.4 –15.3** 0.6* 1.7** 0.8* 0.7** –0.4* –0.3 –0.2** –0.4* 0.4** –2.0** –1.6** –0.1 –1.0** 0.4** 0.8* 0.2 –4.7** 2.6**

0.2 –8.5** 0.4* 0.6** 0.6** 1.2** 0.6** 0.2* –0.1* –0.0 0.1 –0.1 –0.9** –0.6** –0.9** –0.3* 0.6** 0.4** –2.3** 1.1**

–0.1 –3.8** 0.4* 0.9** 0.4 0.9** –0.6** –0.4* –0.1 0.0 0.0 –0.7 –0.7* –0.3 –0.8** –0.4** 0.2

–0.3 –4.6** 0.1 –0.0 –0.3* –0.3* –0.0 –0.1 –0.1** –0.1 0.1 –0.2 –0.2 –0.1 –0.2 –0.4** 0.2 –0.5** –1.2** 0.7**

0.1 –3.0** 0.4** 0.2 0.5** 0.0 –0.1 –0.1 –0.1* 0.1 –0.1** 0.4** –0.1 0.2 –0.2 –0.1 0.4* –0.3** –1.2** 0.5**

0.7 –0.3

–0.2 –1.1

0.3 –1.6*

0.2 –0.4

0.0 –0.1

–0.1 –0.1

–1.4** –0.3

–0.4 –0.4

–0.2 –0.1

0.3 –0.9*

0.1 –0.4

–0.3* 0.1

–1.6** 0.3

–1.4** –0.1

–0.8* 0.4

–1.7** 1.1**

–0.4* 0.3

–0.3 0.2

–1.7** –0.7**

–4.7** 2.4**

–0.0 0.0 –9.1** 14.3** 0.41

–7.9** 12.5** 0.35

–8.1** 2.1** 0.43

–2.3** 0.7**

0.5** –1.0** –8.0** 29.7** 0.42

–6.6** 34.6** 0.39

–5.8** 18.8** 0.42

0.4** –0.4** –6.4** 15.2** 0.47

–3.7** 10.3** 0.43

–3.1** 8.0** 0.37

DUTCH TEENAGERS

Clothing 1984 age gender MAVO HAVO MBO VWO wage self urban twoearn religious north east west south job a18 steady ln ~inc! gender* ln ~inc! P ~school! gender* P ~school! P ~travel! gender* P ~travel! P ~cloth! gender* P ~cloth! P ~magaz! gender* P ~magaz! Heckman constant R2

1.1* 11.5** 0.3 2.2** –1.0* 3.6** 1.2** –1.0** –1.2** 1.0** –0.0 –2.3** –0.4 –1.4** –3.9** –1.8** –1.8** 9.6** –4.6**

Saving 1990

1992

1.3** 3.7** 0.4 2.3** –1.5** 3.6** 0.5 –1.5** –0.7** 0.4 0.2 –0.0 0.1 0.2 –2.9** –0.8** 0.3 –1.1** 6.3** –2.2**

–0.6 –5.0** 0.3 1.6** –1.9** 2.0** –0.2 –0.8** –0.8** 0.2 0.3** –0.4 –0.6 –1.0** –2.3** –1.8** 1.5** –0.4* 5.3** 0.2

–9.8** 5.1**

1984 –10.1** –8.9** –2.4** –2.7** –2.1** –1.9** –0.9* –1.0 0.5** –1.2** 1.4** 1.0 –1.4 1.0 –2.8** 4.1** 2.2**

Other 1990

1992

1984

1990

1992

–4.2 12.6 0.5 0.0 0.8 0.3 0.7 –0.3 –0.3 –0.3 0.3 0.7 1.4 1.0 0.3 0.0 1.8

0.1 16.4 0.7 –1.2 –0.9 –0.7 –0.6 –0.6 –0.1 –1.1 0.1 –0.2 0.2 0.9 –1.9 0.0 –4.3 –0.1 0.4 –2.7

–3.2 15.6 2.4 2.5 –3.2 4.1 0.2 –0.1 –0.3 –0.5 0.4 –3.2 –2.0 –2.1 –5.2 0.4 1.0 –1.0 3.4 –2.6

–5.9** 4.2* 0.7 –0.8 –1.1 1.6** 1.9** 0.3 0.3** –1.0** 1.2** 1.5 1.6* 0.0 1.2 3.0** 0.7 –1.9** 6.0** –1.5**

–5.5** 11.1** –0.1 –2.4** 2.2** 1.0 1.8** 0.6* 1.2** –0.8* 1.4** 0.0 1.1 1.6* –0.4 7.2** 3.2** –0.5 –4.7** –1.7**

–8.3** 2.2*

2.3* –1.5

2.2 –1.5

0.4 4.6

7.7 –3.8

–4.5** 2.3**

–3.4** 1.2

–1.6* –0.8

–1.6 0.2

10.4 0.2

8.0 4.2

11.9** –7.9**

13.8** –9.7**

–5.2** 0.9

–5.2** –0.3

3.8 8.2

4.3 9.0

–1.1** 2.1**

–2.2** 1.5** –19.0** –25.4** 0.66

395

–15.6** –15.8** 0.61

–14.0** –8.9** 0.61

2.5 –1.9

–3.4** 0.6 –17.7** 38.7** 0.44

–18.9** 0.6 0.49

–14.6** 51.5** 0.33

8.9 –3.5

4.4

5.8

–14.8

The dispendent variables are the budget shares in percentages. * 5 significant at 5% level; ** 5 significant at 1% level. The coefficients to the category ‘other’ were calculated afterwards via the additivity restrictions ~3.10!.

396

M.F. WARNAAR AND B.M.S. VAN PRAAG

REFERENCES Abramovitch, R., J.L. Freedman and P. Pliner, 1991, ‘Children and Money: Getting an Allowance, Credit versus Cash, and Knowledge of Pricing,’ Journal of Economic Psychology, 12, pp. 27–45. Amemiya, T., 1974, ‘Multivariate Regression and Simultaneous Equation Models when the Dependent Variables are Truncated Normal,’ Econometrica, 42, pp. 999–1012. Bachman, J., 1983, ‘Premature Affluence: Do High School Students Eat Too Much?,’ Economic Outlook USA, Summer, pp. 64–67. Barten, A.P., 1969, ‘Maximum Likelihood Estimation of a Complete System of Demand Functions,’ European Economic Review, 1, pp. 7–73. Chesher, A. and M. Irish, 1987, ‘Residual Analysis in the Grouped and Censored Normal Linear Model,’ Journal of Econometrics, 34, pp. 33–61. Deaton, A. and J. Muellbauer, 1980, ‘An Almost Ideal Demand System,’ American Economic Review, 70, pp. 312–326. Diedrich, J., 1951, Werkende jeugd en zakgeldbesteding ~Working Youth and Spending of Allowances!, Leiden, Stenfert Kroese. Diekstra, R.F.W., N. Gamefski, P. de Heus, R. de Zwart, B.M.S. van Praag, and M.F. Warnaar, 1991, Scholierenonderzoek 1990, Gedrag en Gezondheid van Scholieren uit het Voortgezet Onderwijs ~Behavior and Health of Pupils Attending Secondary Schools!, NIBUD, The Hague. Fumham, A. and P. Thomas, 1984, ‘Pocket Money: A Study of Economic Education,’ British Journal of Developmental Psychology, 2, pp. 205–212. Greenberger, E. and L. Steinberg, 1986, When Teenagers Work, New York, Basic Books, Inc. Hall, C., 1987, ‘Teen Power, Youth’s Middle Tier Comes of Age,’ Marketing & Media Decisions, Oct. 1987. Helen, D. and C.R. Wessels, 1990, ‘Demand Systems Estimation with Microdata: A Censored Regression Approach,’ Journal of Business and Economic Studies, 8, pp. 365–371. Houthakker, H.S., 1954, ‘La forme des Courbes d’Engel,’ Cahiers de seminaire d’économetrie, 2, pp. 59–66. Jahoda, G., 1979, ‘The Construction of Economic Reality by some Glasgowian Children,’ European Journal of Social Psychology, 9, pp. 115–127. Lee, L.F., 1986, ‘Simultaneous Equations Models with Discrete and Censored Varuables,’ in: C.F. Manski and D. McFadden ~eds.!, Structured Analysis of Discrete Data with Econometric Applications, Cambridge, Mass., MIT Press. Lewin-Epstein, N., 1981, Youth Development During High School, Washington, National Center for Education Statistics. Maddala, G.S., 1987, Limited-Dependent and Qualitative Variables in Econometrics, Cambridge, Cambridge University Press. Manning, W.D., 1990, ‘Parenting Employed Teenagers,’ Youth & Society, 22, pp. 184–200. Miller, J. and S. Yung, 1990, ‘The Role of Allowances in Adolescent Socialization,’ Youth & Society, 22, pp. 137–159. Pollak, R.A. and T.J. Wales, 1981, ‘Demographic Variables in Demand Analysis,’ Econometrica, 49, pp. 1533–1551. Ransom, M.R., 1987, ‘A Comment on Consumer Demand Systems with Binding Non-negativity Constraints,’ Journal of Econometrics, 34, pp. 355–359. Soest, A. van, A. Kapteyn, and P. Kooreman, 1993, ‘Coherency and Regularity of Demand Systems with Equality and Inequality Constraints,’ Journal of Econometrics, 57, pp. 161–188.

DUTCH TEENAGERS

397

Wales, T.J. and A.D. Woodland, 1983, ‘Estimation of Consumer Demand Systems with Binding NonNegativity Constraints,’ Journal of Econometrics, 21, pp. 263–285. Warnaar, M.F., 1991, The Expenditures of Dutch Adolescents, Erasmus University, Rotterdam. Warnaar, M.F., 1994, Selectie, Schoning en Weging van het Scholierenonderzoek 1992 ~Selection, Cleaning and Reweighting of the National School Survey 1992!, NIBUD, The Hague. Warnaar, M.F. and R. de Zwart, 1992, Selectie, Schoning en Weging van het Scholierenonderzoek 1990 ~Selection, Cleaning and Reweighting of the National School Survey 1990!, NIBUD, The Hague. Woodland, A.D., 1979, ‘Stochastic Specification and the Estimation of Share Equations,’ Journal of Econometrics, 10, pp. 361–383. Wunderink-Van Veen, S., 1988, Verdere Toepassingen van een Budget-Verdeelmodel ~Further Applications of a Budgetshare Model!, Erasmus University, Rotterdam. Zellner, A., 1962, ‘An Efficient Method of Estimating Seemingly Unrelated Regressions and Tests of Aggregation Bias,’ Journal of the American Statistical Association, 57, pp. 348–368. Zwart, R. de, A. Luten, and B.M.S. van Praag, 1985, Scholierenonderzoek 1984, Een Eerste Reportage van de Belangrijkste Resultaten ~National School Survey 1984; A First Report of the Main Results!, NIBUD, The Hague. Zwart, R. de and I. Sevinga, 1987, Schonings- en Wegingsprocedures van een Scholierenonderzoek ~Cleaning and Weighting Procedures for a School Survey!, Econometrisch Instituut, report 8914/c. Zwart, R. de, and B.M.S. van Praag, 1989, Earnings of Dutch Adolescents, Econometrisch Instituut, Erasmus University Rotterdam, report 8913/A. Zwart, R. de, B.M.S. van Praag, R.F.W. Diekstra, and M.F. Warnaar, 1990, Scholierenonderzoek 1990, Een Eerste Rapportage van de Belangrijkste Resultaten ~National School Survey 1990; A First Report of the Mean Results!, NIBUD, The Hague. Zwart, R. de and M.F. Warnaar, 1993, Scholierenonderzoek 1992, Landelijke Uitkomsten ~National School Survey 1992; Nation-wide Outcomes!, NIBUD, The Hague.

Summary HOW DUTCH TEENAGERS SPEND THEIR MONEY Adolescents have relatively large sums of money at their disposal; How and on what goods they spend their money is the subject of this paper. An AID system describing the influences of demographic characteristics, pseudo-prices and income on various expenditure categories is estimated. The system is estimated in two steps with the first step explaining zero-expenditures. Income elasticities and pseudo-price elasticities are presented. A comparison is made on the basis of three large data sets, gathered in 1984, 1990, and 1992.