Research Background Paper

Transcription

1 Research Background Paper Measuring Savings Mark Schreiner Research Background Paper 01-4 Sepember 2001 Cener for Social Developmen

2 Measuring Savings Mark Schreiner CYSAPD 01-4 Sepember 2001 Prepared for he Research Design Projec Children and Youh Savings Accoun Policy Demonsraion Cener for Social Developmen Washingon Universiy in S. Louis One Brookings Drive, Campus Box 1196 S. Louis, MO (314) The Universiy of Kansas School of Social Welfare 1545 Lilac Lane, Room 207 Lawrence, KS (785) Absrac Developmen depends on saving. Bu wha exacly is saving, and how is i measured? This paper defines saving and describes several measures of financial savings. The measures accoun for he passage of ime and for he hree sages of saving: puing in (deposiing), keeping in (mainaining a balance), and aking ou (wihdrawing). Togeher, he differen measures capure how people move financial resources hrough ime. Acknowledgemens I am hankful for conribuions o relaed work from Sondra Beverly, Margare Clancy, Suzie Fragale, Lissa Johnson, and Michael Sherraden. I am also graeful for financial suppor from he Division of Asse Building and Communiy Developmen of he Ford Foundaion.

3 MEASURING SAVINGS 1. INTRODUCTION Producion requires naural resources, ools, and human capial. These facors of producion come from saving, he choice o move resources hrough ime raher han o use hem up now. Wihou saving, people are huners and gaherers who live hand-o-mouh. Wih saving, people can build seadily on he pas o improve he fuure. In shor, saving drives developmen and progress. Alhough saving is required for long-erm improvemen in well-being, measures of savings are rudimenary. For example, he mos imporan form of savings is human capial (Schulz, 1979), bu measures of he quaniy of human capial such as age, educaion, or job experience are a bes oblique proxies. Measures of qualiy are also imperfec and usually boil down o wages, a proxy available only for people who work for pay. Measuring financial savings is more sraighforward. Dollars are quanified, have uniform qualiy, and change forms a known imes. Even if measuring financial savings is simple in relaive erms, however, i is sill complex in absolue erms. How o measure saving? And wha is saving in he firs place? This paper proposes a definiion saving is he movemen of resources hrough ime and a series of measures ha accoun for he passage of ime and for he hree sages of saving: puing in (deposiing), keeping in (mainaining a balance), and aking ou (wihdrawing). The specific conex is measuring financial savings in he Children and Youh Savings Accoun Policy Demonsraion (CYSAPD), bu he conceps are general. Beer measuremen may help policymakers o design beer savings incenives. This maers because savings incenives cos billions each year because savings is cenral o developmen (Sherraden, 1991). The paper proceeds as follows. Secion 2 defines saving and oher basic conceps. Secion 3 proposes a series of measures of financial savings in CYSAPD. Secion 4 discusses issues in measuring changes in savings caused by he CYSAPD. Secion 5 wraps up. 2. Basic conceps This secion defines saving, discusses he hree sages of saving, and explains why measures of saving mus accoun explicily for ime. 2.1 Income, asses, saving, and asse accumulaion Resources received in a given ime period are income; resources conrolled a a poin in ime are asses. Boh income and asses refer o resources; hey differ only in he frame of reference. If resources received as income are no immediaely consumed, hen hey become asses. 1 Cener for Social Developmen Washingon Universiy in S. Louis

4 Moving resources hrough ime is saving. The definiion includes boh conscious and unconscious failure o consume. For example, people wih cash migh make bank deposis raher han buy hamburgers, fail o wihdraw exising bank balances o buy hamburgers, or use so lile peroleum ha reserves are lef unexploied. Mos people would recognize only he firs example as saving, bu all hree examples move resources hrough ime. The use of resources (consumpion) is dissaving. If, in a long ime frame, saving exceeds dissaving, hen he resul is asse accumulaion. Everyone saves, and everyone dissaves. For example, a person may use a paycheck o pay bills over ime. The person firs saves (even if he check is no immediaely deposied or cashed) and hen dissaves. From a high-frequency view (pay-day unil he nex day, for example), almos all income is saved, and almos all asses are soon dissaved. Asse accumulaion occurs if saving consisenly exceeds dissaving. Slivers of differences in high-frequency behavior can produce large differences in asse accumulaion. For example, suppose ha wo people each earn $100 per day bu ha one saves $2 more per day. Wih a 3 percen annual reurn, he difference in asse accumulaion in 20 years is abou $20,000. Furhermore, accumulaion gaps end o grow because asses bege asses (Schreiner e al., 2001). Tha is, greaer asses be hey physical, social, financial, or human lead o greaer producion, greaer income, and hus greaer resources. Once asses pu someone ahead, she ends o say ahead. 2.2 Sages of financial saving Moving dollars hrough ime is financial saving. Financial saving has hree sages. The firs is puing in. This changes non-financial resources ino dollars or when puing in means deposiing changes cash ino accoun balances. Alhough many people equae deposiing wih saving, saving is far broader. The second sage of financial saving is mainaining balances, or keeping in. Alhough no always recognized as saving, failure o consume asses does move resources hrough ime. The hird sage is aking ou. Resources aken ou may be consumed (dissaved) or kep in anoher form (saved). For accoun balances, aking ou means making wihdrawals. Each sage is a disinc aspec of financial saving. Savings migh be high in one sage bu low in anoher, so measuremen should look a all hree sages. For example, savers wih large deposis may have high savings in erms of puing in, bu, if hey make quick wihdrawals, hey may have low savings in erms of keeping in. Likewise, savers wih low deposis migh noneheless mainain balances for a long ime. Finally, savers wih high savings in erms of puing in and/or keeping in migh if wihdrawals are consumed raher han convered o oher asses have low savings in erms of aking ou. Cener for Social Developmen Washingon Universiy in S. Louis 2

5 Measuremen should cover all hree sages because a narrow focus on one or wo sages would miss some faces of behavior. For example, when workers swich jobs, abou half of he number of disribuions from 401(k) plans are no rolled ino oher reiremen plans (Samwick and Skinner, 1997; Poerba, Veni, and Wise, 1995). Wha does his mean for saving? For aking ou, abou half he non-rollover is kep as asses in some form. For puing in, people who do no make rollovers usually had smaller deposis. For keeping in, mos people who do no make rollovers are young and have small balances. Measures of savings ha omi sages miss imporan pars of he sory. 2.3 Saving and ime Saving moves resources hrough ime, so measures of financial savings mus refer explicily o ime. (For example, deposis per monh refers o ime, bu balance does no.) Changes in resources in a period of ime are flows, and resources a a poin in ime are socks. Socks and flows describe wo sages of financial saving, puing in as flows of deposis and aking ou as flows of wihdrawals (or socks of balances o be wihdrawn laer). Socks and flows, however, describe keeping in inadequaely. Measuring he resources held hrough ime requires a flowified sock. Wih unis of dollar-monhs, such a measure is he Average Balance. For example, suppose a saver in he CYSAPD deposis $10 on he firs of he monh for a year and wihdraws i all for consumpion a year s end. Wha is CYSAPD savings? Deposis pu in are $120; wihdrawals aken ou are zero. The average balance kep in is 65 dollar-monhs. Tha is, he saver moved resources hrough ime equivalen o $65 per monh. 3. Measures of financial savings This secion describes measuring savings wih explici reference o ime and o all hree sages of financial savings. The formulae are in erms of monhly deposis and wihdrawals. They are framed in erms of CYSAPD accouns, bu hey would apply jus as well o Individual Developmen Accouns (IDAs), Individual Reiremen Accouns (IRAs), or 401(k) plans. 3.1 Savings measures Sources of deposis CYSAPD deposis have four sources. The firs are paricipans, he second are ineres paymens, and he hird are friends and family. In CYSAPD, deposis by family and friends are acknowledged and encouraged. While rarely discussed, friends and family also conribue o IDAs, IRAs, and 401(k) plans. 3 Cener for Social Developmen Washingon Universiy in S. Louis

6 The fourh source of CYSAPD deposis are hird paries beyond family or friends. Jus as employers can conribue o 401(k) plans, hird paries can conribue o CYSAPD accouns. Alhough paricipans do no own maches unil hey make a mached wihdrawal, hey do own hird-pary deposis from he momen of deposi. Alhough hird-pary deposis may be rewards for paricipan asks (such as compleion of a year of school), from he poin of view of he paricipan, heir presence and size are arbirary. For example, a CYSAPD program migh reward aendance a financial educaion wih a hird-pary deposi of $100 or $1,000. Thus, he size of hird-pary deposis in a CYSAPD accoun depends on program donors. In conras, paricipans choose wheher resources from hird-pary deposis are kep in he accoun and wheher hey are convered o oher asses upon wihdrawal. Thus, hird-pary deposis are no couned for measures of puing in bu are couned for measures of keeping in and aking ou Gross deposis Gross Deposis in a CYSAPD accoun by saver i in monh are denoed as g i. (From now on, he subscrip i is suppressed.) Gross Deposis include boh Own Deposis o (from he paricipan, from ineres, and from family and friends) and Third-Pary Deposis q : g o + q. The measure of saving as puing in excludes Third-Pary Deposis. The sum of Own Deposis hrough monh is Cumulaive Own Deposis O : O o j. j 1 Cumulaive Own Deposis O should no be compared across people a differen sages in paricipaion. To accoun for ime as a paricipan, use Own Deposis per Monh O : O O. The principal measure of saving should no be he firs sep of puing in. Firs, CYSAPD (like oher saving incenives) maches deposis only up o he annual cap c. Excess deposis are sill savings, bu hey are no CYSAPD savings. (To keep formulae simple, his paper ignores excess deposis.) Second, deposis may be wihdrawn o finance consumpion or oher forms of asses. Third, people migh rea CYSAPD like a checking accoun, making frequen deposis and wihdrawals wihou plans for long-erm accumulaion. This churning leads o high puing in bu low keeping in and low aking ou. The bes measures look a deposis and wihdrawals ogeher Wihdrawals Gross Wihdrawals w are Mached Wihdrawals m (approved for incenives) plus Unmached Wihdrawals u (unapproved for incenives): w m + u. Cener for Social Developmen Washingon Universiy in S. Louis 4

7 CYSAPD has wo ypes of wihdrawals because as in IDAs, IRAs, and 401(k) plans only some uses of savings qualify for incenives. For purposes of measuremen, a furher assumpion is ha unmached wihdrawals are consumed bu mached wihdrawals are convered ino oher forms of asses. Cumulaive Unmached Wihdrawals U measure resources aken ou of CYSAPD and assumed consumed: U u j. j 1 Cumulaive Mached Wihdrawals M measures resources aken ou of CYSAPD and assumed convered ino oher of forms of asses: M m j. j 1 As a measure of saving, mached wihdrawals aken ou is useful bu incomplee. Firs, a a given poin in ime, some mached wihdrawals have ye o happen. Second, resources are fungible, so i is pure assumpion ha people save all mached wihdrawals and consume all unmached wihdrawals Paricipan accumulaion Paricipans may accumulae hree forms of resources hrough CYSAPD. The firs are balances ha may be aken ou in fuure mached wihdrawals. The second are resources already aken ou in mached wihdrawals and assumed convered o asses in anoher form. The hird are maches. Accoun balances and mached wihdrawals are kep in by paricipans and so are included he measure of Paricipan Accumulaion P. Maches are excluded because hey do no come from he paricipan. Assuming ha he mach cap never binds, Paricipan Accumulaion P is Cumulaive Own Deposis O minus Cumulaive Unmached Wihdrawals U : P O U, j 1 o j u. Accumulaion depends on he lengh of paricipaion. To conrol for his, divide Paricipan Accumulaion by monhs o ge Paricipan Accumulaion per Monh P : O U P. Paricipan Accumulaion per Monh shows how fas resources accumulae. (Schreiner e al. (2001) and Sherraden e al. (2000) call his measure average monhly ne deposis.) I is a beer measure of savings han Own Deposis per Monh because i accouns for unmached wihdrawals. I is beer han Cumulaive Mached Wihdrawals because i couns boh mached j 5 Cener for Social Developmen Washingon Universiy in S. Louis

8 wihdrawals and balances ha may ye be mached. And i is beer han Paricipan Accumulaion because i conrols for he lengh of paricipaion. Paricipan Accumulaion per Monh is he bes summary measure of paricipan saving in he sages of puing in deposis and aking ou wihdrawals Toal accumulaion For wo reasons, paricipan accumulaion does no include he mach iself. Firs, paricipans do no own he mach unil afer a mached wihdrawal. Second, he mach rae is deermined no by he paricipan bu by he program. If paricipan accumulaion included he mach, hen arbirary program choices would affec measures of paricipan behavior. Alhough Paricipan Accumulaion before mached wihdrawals should no include he mach, Toal Accumulaion A afer mached wihdrawals should. Paricipan Accumulaion P, plus any maches a mach rae : A P + M ( 1 + ρ). Toal Accumulaion is useful as a measure of asses buil hrough CYSAPD afer all sages of financial saving are compleed. I includes deposis from hird paries Dollar-monhs saved Suppose wo people join CYSAPD on January 1. The firs deposis $10 on he firs day of each monh for a year, and he second makes a single deposi of $120 on December 1. Neiher makes any wihdrawals. Who saved more? Alhough inuiion suggess ha he slow-and-seady person saved more, he savings measures described so far are idenical for each saver. Each has Cumulaive Own Deposis of $120, Own Deposis per Monh of $10, Cumulaive Unmached Wihdrawals of 0, Cumulaive Mached Wihdrawals of $0, Paricipan Accumulaion of $120, Paricipan Accumulaion per Monh of $10, and Toal Accumulaion of $120. The measures are he same because hey look a only he puing in and aking ou sages and ignore he keeping in sage. Measuring he movemen of resources hrough ime requires a flowified sock such as he sum of Paricipan Accumulaion P in all monhs, Dollar-Monhs Saved, D : D j 1 P, j j j 1 k 1 o k u. k Cener for Social Developmen Washingon Universiy in S. Louis 6

9 In he example, he firs person saved $780 dollar-monhs: 10 dollar-monhs in he firs monh, 20 dollar-monhs in he second monh, and so on. The second person saved 120 dollar-monhs, all in December. Consisen wih inuiion, Dollar-Monhs Saved suggess ha he firs person saved more. Dollar-Monhs Saved disinguishes beween he wo savers because i looks a boh size and iming of deposis and wihdrawals and hus accouns for keeping in. Oher measures look a only size via puing in and aking ou. Dollar-Monhs Saved is especially useful for savers who make unmached wihdrawals. Dropous are he exreme; all heir deposis become unmached wihdrawals. Savings measures ha ignore keeping in show zero savings for resources once removed as unmached wihdrawals. Even drop-ous, however, moved some resources hrough ime, and Dollar-Monhs Saved reflecs his Dollar-monhs per monh Comparisons of Dollar-Monhs Saved beween savers work bes when boh have paricipaed for he same lengh of ime. One approach is o divide by by monhs of paricipaion o ge Dollar- Monhs per Monh, D : D D. This is jus he average balance. The erm Dollar-Monhs per Monh is preferred because i shows beer ha he measure is a flowified sock Dollar-monhs saved raio Dollar-Monhs per Monh sill depends on he lengh of paricipaion. For example, saving $10 a monh for one year gives an average balance of $65, bu saving $10 a monh for wo years gives an average balance of $125. Anoher way o conrol for lengh of paricipaion is o compare acual Dollar-Monhs Saved wih Dollar-Monhs Saved if deposis were c/12 each monh. (These equal-sized monhly deposis would jus add up o he annual mach cap c.) This is he Dollarr Monhs Saved Raio, D : r D D, j c j 1 k 112 D 24. c ( + 1) In Figure 1, he Dollar-Monhs Saved Raio is he area under line A (D Α24) o he area under line B (cαα(+1)). In he figure, he raio is Wih no excess deposis, he maximum rae is (a single deposi in he firs monh of (Αc)/12). The minimum rae of obains if here are never any deposis. The rae is if he paern of size and iming of cash flows 7 Cener for Social Developmen Washingon Universiy in S. Louis

10 produces Dollar-Monhs Saved equal o wha Dollar-Monhs Saved would be wih deposis of c/12 each monh. (A rae of is possible even if deposis are no c/12 each monh. For example, someone wih a mach cap of $120 migh deposi 0 in monh 1 and $30 on he firs day of monh 2.) In he example wih wo savers, suppose he annual mach rae c is $120. The slow-and-seady saver saved $10 a monh, exacly he amoun o reach he mach cap a year-end. Accordingly, he Dollar-Monhs Saved Raio is ($780Α24) / ($120Α12Α13) The raio for he saver who made only one deposi of $120 in December is much lower, ($120Α24) / ($120Α12Α13) Summary Why boher wih so many savings measures? Firs, he effecs of asse use (such as a down paymen on a house) depend on Toal Accumulaion. In urn, Toal Accumulaion depends on he mach rae and on Paricipan Accumulaion and hus on lengh of paricipaion and Paricipan Accumulaion per Monh. This means measuring puing in and aking ou. Second, he social/psychological/behavioral effecs of asse ownership wha Sherraden (1990) calls asse effecs depend on moving resources hrough ime. When people hink abou heir asses and how hey will use hem when hey savor heir savings hey may be happier and make beer choices (Schreiner e al., 2001). A measure of asse ownership is monhs of paricipaion, and a measure of he amoun of asses kep in is Dollar-Monhs Saved. 3.2 Accumulaion comparisons This secion describes measures ha compare Paricipan Accumulaion wih wo benchmarks, income and he mach cap Savings rae Deposis as a share of income is he savings rae. This radiional erm is a misnomer; mos savings in a period come no from newly acquired resources bu exising asses lef unconsumed. Also, mos paricipans have non-cysapd savings, so he measure here is really he CYSAPD savings rae. Paricipan Accumulaion per Monh P (annualized) as a share of annual income y is he savings rae, r : r P 12. y The difficuly is measuring income. As defined earlier, resources received in a period are income. Income is hus more han jus financial resources; for example, he main form of income for mos people is ime o live. This paper, however, follows convenion and couns as income only financial inflows. Cener for Social Developmen Washingon Universiy in S. Louis 8

11 Sill, some issues remain. Should income include in-cash public assisance? Wha abou in-kind public assisance? Alhough public assisance is indeed par of income, i is very difficul o measure. Thus, his research, like almos all oher work, does no include i in he measure of income. Should measures use gross income, adjused gross income, or afer-ax income? As new disposable resources, he bes measure is afer-ax income. Should income be measured for he monh prior o enrollmen, for he 12 monhs prior o enrollmen, or for he prior ax year? Because income flucuaes from monh o monh and because here are no official 12-monh records of income, measuremen should use ax reurns for he year prior o enrollmen Mach use Do paricipans save all he way up o he mach cap? Paricipan Accumulaion P divided by he pro-raed mach cap is Mach Use, X : P X. ( c ) / 12 Mach Use shows he pace of saving relaive o poenial maching. Someone on pace o use exacly all mach eligibiliy has a raio of Someone behind (ahead) of his pace is below (above) Deposi consisency Savings incenives like CYSAPD aim o promoe asse accumulaion and savings habis. Alhough here is lile hard evidence, many people believe ha consisen savers boh become beer savers and accumulae more. This secion presens wo measures of deposi consisency, one ha focuses on he presence of monhly deposis and one ha focuses on he size and iming of deposis Deposi frequency The share of monhs wih a deposi is Deposi Frequency, f. (Ineres is no couned; oherwise, all monhs would have a deposi.) If he funcion I(o ) is uniy if Own Deposis o is posiive and zero if o is zero, Deposi Frequency is: f j 1 I( oj ). Higher frequencies indicae greaer consisency. If a saver makes a deposi in all monhs (maximum frequency), hen Deposi Frequency is The minimum (wih no deposis) is Cener for Social Developmen Washingon Universiy in S. Louis

12 The srengh of Deposi Frequency is is simpliciy. This is also is weakness; he measure is he same wheher someone deposis $10 each monh or $1, $19, $15, and hen $4. This weakness may be unimporan if, for learning o save, wha maers is no he size of he deposis bu is mere presence Deposi enropy A measure of he disribuion of deposis size and iming is Deposi Enropy, e. (As above, ineres is excluded.) Based on he classic enropy measure (Golan, Judge, and Miller, 1996), Deposi Enropy is closer o zero as deposis are more concenraed (less consisen) and is closer o uniy as deposis are more uniform (more consisen). The formula is: 1 e 1 + oj lno j, ln j 1 oj where oj. o The weaknesses of enropy is is newness and is difficul-o-inerpre unis. For example, 0.8 is a more-uniform deposi paern han 0.6, bu he precise meaning of he 0.2 difference is no clear. The srengh of enropy is ha i summarizes he enire disribuion of deposis. Unlike oher summary measures of he uniformiy of disribuions (such as variance or coefficien of variaion), enropy is bounded beween zero and uniy and depends only on he shape of he disribuion, no he heigh. For example, suppose a saver deposis $10 and $20. The deposi shares are o 1 10/ and o 20/ Deposi Enropy is hen: 2 e ln 2 k 1 k ( 0.33 ln ln 0.66) ( 0.637) For comparison, he variance is [(10!15) 2 +(20!15) 2 ] / Wih a mean of (10+20) / 2 15, he coefficien of variaion is 50 / Wha if deposis were $1 and $2? Deposi Enropy is unchanged, as only heigh changes, no shape. The variance and coefficien of variaion, however, are now 0.5 and Unlike hese wo summary measures of disribuion, enropy is invarian o unis. 4. Measuring Impacs in CYSAPD Measuring impac in CYSAPD is sraighforward, he difference in average oucomes beween reamen and conrol groups. Bu wha oucomes o measure? In CYSAPD, he cenral oucomes are Paricipan Accumulaion and Toal Accumulaion in he name of he child or youh. Dollar- Monhs Saved because of is links wih asse effecs is also an imporan oucome. Cener for Social Developmen Washingon Universiy in S. Louis 10

13 4.1 New savings and asse shifs Deposis in CYSAPD come from wo sources (Schreiner e al., 2001). The firs is an increase in difference beween income and consumpion, or new saving. New saving comes from increased income (hrough increased ime or effor in work) and/or from decreased consumpion. The second source of deposis is a reshuffling of already-saved resources, or asse shifs. There are wo ypes of asse shifs. In he firs, already-saved resources (for example, exising balances in a checking accoun) are moved o CYSAPD. In he second, resources ha would have been saved anyway are pu in CYSAPD. Savings programs be hey CYSAPD, IDAs, IRAs, or 401(k) plans aim boh o increase new saving and o increase asse accumulaion. These wo goals are someimes conflicing, someimes complemenary. (CYSAPD also has non-financial goals, bu reaching hem presumably requires financial success.) Savings incenives for example, maches reward new saving and boos asse accumulaion. These same incenives, however, also arac asse shifs. Afer all, i may be easier o conver resources saved in he pas o CYSAPD balances han o increase curren income and/or decrease curren consumpion. 4.2 Asse accumulaion in CYSAPD Impac measuremen for asse accumulaion in CYSAPD focuses on wo oucomes, Paricipan Accumulaion and Toal Accumulaion. As argued below, boh ake he poin of view of he child or youh. Paricipan Accumulaion ells how much resources were pu in he CYSAPD accoun by paricipans, family, and friends and eiher kep in or aken ou as mached wihdrawals. Toal Accumulaion adds hird-pary deposis and maches and so measures all asse accumulaion ha occurs hrough he CYSAPD accoun. I is useful o measure accumulaion boh wih and wihou hird-pary deposis and maches because hey do no depend compleely on paricipan behavior. 4.3 New Savings Wha abou new savings? Paricipan Accumulaion is no likely all new savings, and Toal Accumulaion cerainly is no. Measuring new savings requires measuring differences in ne worh beween reamens and conrols. (Wih an experimenal design, his requires no beforeand-afer measures, jus afer (Schreiner, 2000)). Measuring ne worh requires measuring oal asses and liabiliies. This is a all order, and i may be even more difficul han usual in CYSAPD because a large share of deposis may come from beyond he household, for example, from uncles or grandparens. Thus, measuring new savings due o CYSAPD would require measuring ne worh for all family and friends who migh conribue. 11 Cener for Social Developmen Washingon Universiy in S. Louis

14 This is probably impossible. Wha o do? One opion is o assume ha family or friends ouside he immediae household do no make CYSAPD conribuions. Then measuring ne worh requires only measuring oal asses and liabiliies in he paricipan s household. This is he approach adoped in experimenal componen of he American Dream Demonsraion (ADD) of IDAs. A second opion is o look only a impacs on asses kep in he name of he paricipan. Tha is, CYSAPD may be less ineresed in new savings by all poenial conribuors han in new savings for he child or youh. Then he quesion is wheher family and friends shif o CYSAPD gifs ha hey would have made o he paricipan anyway. For example, grandparens who would buy educaional savings bonds for heir grandchildren anyway migh insead pu he resources in CYSAPD. Likewise, parens migh spend less on books or compuers o free up resources o pu in CYSAPD. This approach aemps o finesse he quagmire of measuring ne worh for all possible conribuors, subsiuing measuring only asses in he name of he CYSAPD paricipan. Graned, he approach does no measure impacs on oal ne savings caused by CYSAPD; i only measures impacs on ne savings in he name of he paricipan. This assumes ha CYSAPD does no care abou asse shifs if hey lead o greaer asses in he name of he child or youh. Evaluaors mus judge wheher his is an accepable compromise. 4.5 Asse effecs Asse effecs depend no on resource use bu on resource ownership. If all CYSAPD did were help children and youh o consume more, hen i would resemble radiional ransfer programs. Bu CYSAPD aspires o much more; in any case, in he space of he demonsraion, very few paricipans will reach an age when hey can use CYSAPD resources. Thus, CYSAPD will no increase resource use by paricipans during he demonsraion; in fac, because CYSAPD encourages saving, resource use will probably decrease. Thus, any nonfinancial effecs will depend on how ownership of financial resources as opposed o use changes behavior. Measuring non-financial impacs will require measuring how differen levels of oucomes are associaed wih differen levels of ownership. How o measure ownership? Two key dimensions are presence and amoun. The mere fac of accumulaion, regardless of amoun, may affec non-financial oucomes. A simple measure of he exen of he presence of ownership in CYSAPD is monhs of paricipaion,. (ADD suggess ha few paricipans ever carry zero balances, so paricipaion is a very good proxy for ownership.) The amoun of accumulaion may also affec non-financial oucomes. A measure is Dollar- Monhs Saved, D. If one person has $100 in CYSAPD for one year and a second person has $50 for four years, Dollar-Monhs Saved reflecs ha he second person owned more resources hrough ime. Cener for Social Developmen Washingon Universiy in S. Louis 12

15 Because greaer resource ownership evenually leads o greaer resource use, few (if any) sudies unangle heir disinc forces. Because CYSAPD paricipans will no use heir accumulaions unil afer he demonsraion ends, CYSAPD presens a unique chance o measure he effecs of ownership unained by he effecs of use ha ownership evenually confers. 5. Conclusion This paper defines saving as moving resources hrough ime. Saving has hree sages: puing in, keeping in, and aking ou. Each sage maers because some paricipans save more in one sage han in anoher. For example, boh paricipans and hird paries pu in deposis in CYSAPD, and measuremen should disinguish among hese wo sources (boh disinc from maches). Also, some paricipans (drop-ous in paricular) keep resources in CYSAPD ha are evenually consumed. Finally, CYSAPD aims o increase accumulaion hrough convering resources aken ou o asses in oher forms. The basic measure of resources pu in by paricipans is Own Deposis per Monh. The measure Paricipan Accumulaion per Monh recognizes ha some deposis are consumed and ha differen people have differen lenghs of paricipaion. To measure resources kep in CYSAPD hrough ime, use Dollar-Monhs Saved or he Dollar-Monhs Saved Raio. The Savings Rae compares Paricipan Accumulaion wih income, and Mach Use compares Paricipan Accumulaion wih he mach cap. Savings consisency is indicaed by Deposi Frequency and Deposi Enropy. Paricipans will no conver heir CYSAPD resources o oher forms of asses during he life of he demonsraion. Thus, any non-financial effecs mus derive no from asse use bu from asse ownership. A measure of he exen of he presence of ownership is he lengh of paricipaion, and a measure of he amoun of resources owned is Dollar-Monhs Saved. 13 Cener for Social Developmen Washingon Universiy in S. Louis

16 REFERENCES Golan, Amos; Judge, George; and Douglas Miller. (1996) Maximum Enropy Economerics: Robus Esimaion wih Limied Daa, New York: Wiley, ISBN Poerba, James M.; Veni, Sephen F.; and David A Wise. (1995) Lump-sum Disribuions from Reiremen Saving Plans: Receip and Uilizaion, Naional Bureau of Economic Research Paper 5298, papers.nber.org/papers/w5298.pdf. Samwick, Andrew A.; and Jonahan Skinner. (1997) Abandoning he Nes Egg? 401(k) Plans and Inadequae Pension Saving, in Sylveser J. Schieber and John Shoven (eds) The Economics of U.S. Reiremen Policy: Curren Saus and Fuure Direcions, Cambridge, MA: MIT Press. Schreiner, Mark. (2000) A Framework for Financial Benefi-Cos Analysis of Individual Developmen Accouns a he Experimenal Sie of he American Dream Demonsraion, Research Design, Cener for Social Developmen, Washingon Universiy in S. Louis, gwbweb.wusl.edu/users/csd/. Schreiner, Mark; Sherraden, Michael; Clancy, Margare; Johnson, Lissa; Curley, Jami; Grinsein-Weiss, Michal; Zhan, Min; and Sondra Beverly. (2001) Savings and Asse Accumulaion in Individual Developmen Accouns, Cener for Social Developmen, Washingon Universiy in S. Louis, gwbweb.wusl.edu/users/csd/. Schulz, Theodore W. (1979) Nobel Lecure: The Economics of Being Poor, Journal of Poliical Economy, Vol. 88, No. 4, pp Sherraden, Michael. (1991) Asses and he Poor: A New American Welfare Policy, Armonk, NY: M.E. Sharpe, ISBN Sherraden, Michael; Johnson, Lissa; Clancy, Margare; Beverly, Sondra; Schreiner, Mark; Zhan, Min; and Jami Curley. (2000) Saving Paerns in IDA Programs, Cener for Social Developmen, Washingon Universiy in S. Louis, gwbweb.wusl.edu/csd/users/. Cener for Social Developmen Washingon Universiy in S. Louis 14

17 Figure 1: The Dollar-Monhs Saved Raio Line B (deposi c/ 12 each monh) Balance Line A (deposi (0.6*c)/ 12 each monh) 0 Monhs 15 Cener for Social Developmen Washingon Universiy in S. Louis