Lecture 9: Savings and the Poor PDF

Lecture 9: Savings and the Poor Introduction I We have seen that the credit market does not function smoothly when it comes to the poor. I This gives rise to new products that can help the poor such as micro credit. I In this lecture we look at the savings of the poor and consid...

Lecture 9: Savings and the Poor Introduction I We have seen that the credit market does not function smoothly when it comes to the poor. I This gives rise to new products that can help the poor such as micro credit. I In this lecture we look at the savings of the poor and consider the challenges that may arise when we think about the poor saving enough to overcome the poverty trap. Motive for Savings: Intertemporal consumption smoothing I Consider an individual living two periods. In the first period, the individual can work and earn labor income. In the second period, the individual cannot work (retire). I Let’s say that I = (1, 0) is the income profile over two periods. I Suppose that the individual has a utility that shows a diminishing marginal return (concave) I As an example, suppose that the utility is u(c) = ln(c) I Let β be the discount factor that determines how much the individual discounts utility from one period later. - If the individual is fully patient so that waiting one period for consumption does not affect her utility, then β = 1. - If the individual is not fully patient, consuming 1 unit today gives greater utility than consuming 1 unit tomorrow. In this case, β < 1 Motive for Savings: Intertemporal consumption smoothing I For simplicity, we assume that the interest rate is 0, thus R = 1. I The marginal return from consumption decreases as the level of consumption increases u 0 (c) = c1 , which decreases by c I Let C = (c1 , c2 ) is the consumption choice of the individual. I Consuming x unit in period 1 gives utility of ln(x) in period 1. I Consuming x unit in period 2 gives utility of ln(x) in period 2. I Importantly, consuming x unit in period 2 gives utility of β · ln(x) from the perspective of period 1. Motive for Savings: Intertemporal consumption smoothing I Let’s evaluate the marginal utility from consumption from the perspective of period 1. I Let C = (c1 , c2 ) be the consumption profile I Marginal utility of consuming c1 in period 1: u 0 (c1 ) = c11 I Marginal utility of consuming c2 in period 2 from the perspective of period 1 (accounting for the discount factor): βu 0 (c2 ) = β c12 Motive for Savings: Intertemporal consumption smoothing I Suppose that the individual does not save and consume everything today. I Is C = (1, 0) optimal (i.e., the best way to maximize utility over two periods)? I No! u 0 (c1 ) = 1 < βu 0 (c2 ) = ∞ I She can increase the sum of utility over two periods by reducing today’s consumption and increasing tomorrow’s consumption Motive for Savings: Intertemporal consumption smoothing I The optimal consumption allocation should equalizes the marginal utility from consumption across two periods, accounting for the discount factor I If C ∗ = (c1∗ , c2∗ ) is optimal, the following equation must satisfy u 0 (c1∗ ) = βu 0 (c2∗ ) 1 I With u(c) = ln(c), the above condition implies c1∗ = β c1∗ ⇔ 2 c2∗ = βc1∗. I First, consumption level is almost the same across two periods up to β adjustment. Thus, smoothing consumption is desirable to maximize total utility. I Second, if the individual is less patient (lower β), the individual would like to consume less in the second period.Thus, patient individuals would like to save more for the future consumption. Motive for Savings: Intertemporal consumption smoothing I To sum, if you have a diminishing marginal return from consumption (concave utility), you want to smooth consumption over time. I Intertemporal consumption smoothing is an important motive for savings. Motive for Savings: Self-Insurance I Another motive for the saving is to provide self-insurance in the presence of ”uninsurable” future risk (future shock on income that cannot be hedged by a proper insurance product). I Suppose that the individual has a concave utility. If the individual does not have all relevant insurance product available in the market to hedge against future risks, her savings also depend on the extent of uncertainty in the future income. I In this case, she might want to save even more than the case without uncertainty. I This is called as ”precautionary savings” Motive for Savings: Self-Insurance I As we discussed in the previous lecture on insurance, the poor tend to face greater risks in their daily life. I Thus, if the poor cannot have proper insurance product in the market, they have to save more for the precautionary motive. I Again, the extent to which they care about future consumption depends on how patient they are. Behavioral Biases I The most recent evolution in the human brain (last 150,000 years) has been in the development of the prefrontal cortex, which is believed to control deliberative processes. Indeed, individuals with damage to the prefrontal cortex are often impaired in making decisions that require deliberation. I The development of the prefrontal cortex did not replace the earlier brain, but rather amounted to an expansion of it. The more primitive brain systems remain. These are systems shared with other mammals and regulate emotional and survival instincts; they tend to be very reactionary. I Hence, these two regions of the brain can be viewed as two types of individuals, a time-consistent one, and a more naive time-inconsistent type. Behavioral Biases I Difference between now and a year from now is not perceived to be similar, but difference between 11 years from now and 12 years from now is perceived to be similar. I It has been observed in several experimental settings that people exhibit time inconsistency in their preferences. I van Leeuwen (1998): Food experiment – Choose for Next Week: Fruit (74%) or Chocolate (26%). Choose for Today: Fruit (30%) or Chocolate (70%). I Read, Loewenstein & Kalyanaraman (1999): Video experiment – Choose for Next Week: Low-brow (37%) or High-brow (63%). Choose for Today: Low-brow (66%) or High-brow (34%). I There is also non-experimental evidence that these behavioral biases affect day to day decisions, for example people who say they would like to go the gym but don’t go. Commitment – Tying Odysseus to the mast I People may be willing to pay to be able to tie their hands in some way. I Evidence from gyms (Della Vigna and Malmendier 2004) – Average cost of gym membership is $75 per month, and the average number of visits per month is 4. I This works out to average cost per visit being $19. However, the cost of pay-per-visit is $10. I People may be paying more as a way to commit themselves to go to the gym when they realize that in absence of commitment, they would never end up going. I Are there reasons why these biases may disproportionately affect the decisions of the poor? Evidence from Phillipines I Authors partnered with the Green Bank of Caraga, a rural bank in Mindanao in the Philippines. I First, independently of the Green Bank, they administered a household survey of 1777 existing or former clients of the bank. I They asked hypothetical time discounting questions in order to identify individuals with time inconsistent preferences. I Thereafter they randomly chose half the clients and offered them a new account called a SEED(Save, Earn, Enjoy Deposits) account. Evidence from Phillipines I This account was a pure commitment savings product that restricted access to deposits, but did not compensate the client for this restriction. I The interest rate paid on the SEED account was identical to the interest paid on a normal savings account (4% per year). I The other half of the surveyed individuals were assigned to either a control group that received no further contact or a marketing group that received a special visit to encourage savings using existing savings products only (i.e., these individuals were encouraged to save more but were not offered the new product). SEED Design – Withdrawal I Individuals restricted their rights to withdraw funds until they reached a goal. Clients could restrict withdrawals until a specified month when large expenditures were expected, e.g., school, Christmas purchases, a particular celebration, or business needs. I Alternatively, clients could set a goal amount and only have access to the funds once that goal was reached (e.g., if a known quantity of money is needed for a new roof). The clients had complete flexibility to choose which of these restrictions they would like on their account. I Once the decision was made, it could not be changed, and they could not withdraw from the account until they met their chosen goal amount or date. Of the 202 clients who opened these accounts 140 opted for a date based goal and 62 opted for an amount based goal. SEED Design – Deposits I Clients could choose a locked box in exchange for a small fee. This locked box is similar to a piggy bank: it has a small opening to deposit money and a lock to prevent the client from opening it. Out of the 202 clients who these opened accounts, 167 opted for this box. I The box permits small daily deposits even if daily trips to the bank are too costly. These small daily deposits keep cash out of one? pocket and (eventually) in a savings account. I The barrier, however, is largely psychological; the box is easy to break and hence is a weak physical commitment at best. I Alternately they were offered the option to automate transfers from a primary checking or savings account into the SEED account. This feature was not popular. SEED Design – How To Identify Who Has Commitment Problem? source: Ashraf, et. al (2006) SEED Design – How to Test whether Commitment Is an Issue for the Low Savings Rate? I Test whether individuals who exhibit time inconsistent preferences are more likely to open such accounts, since theoretically these individuals may have a preference for commitment. See Takeup I Do such individuals save more as a result of opening the account? See Forced Saving 1 , Forced Saving 2 I Are other savings crowded out? See Crowding Out SEED Design – Results I Those who are time inconsistent (impatient now, but patient for future trade-offs) are in fact more likely to take up the SEED product. Little else predicts take-up of the product. I Females who exhibit hyperbolic preferences (with respect to money) are 15.8 percentage points more likely to take up the SEED product. I This effect is small (4.6 percentage points) and insignificant for men. I This result on hyperbolic preferences is robust to controlling for income, assets, education, household composition, and other potentially influential characteristics.  TABLE V Takeup DETERMINANTS OF SEED TAKE-UP PROBIT (1) (2) (3) (4) All All Female Male Time inconsistent 0.125* 0.005 0.158* 0.046 Downloaded from http://qje.oxfordjournals.org/ at Singapore Management University on (0.067) (0.080) (0.085) (0.098) Impatient, now versus 1 month !0.030 !0.039 !0.036 !0.041 (0.050) (0.050) (0.062) (0.075) Patient, now versus 1 month 0.076 0.070 0.035 0.119 (0.072) (0.072) (0.089) (0.110) Impatient, 6 months versus 7 months 0.097 0.108* 0.124 0.078 (0.065) (0.065) (0.087) (0.091) Patient, 6 months versus 7 months 0.015 0.022 0.057 !0.021 (0.064) (0.064) (0.081) (0.093) Female 0.099 0.070 (0.137) (0.138) Female X time inconsistent 0.191** (0.090) Married X female !0.113 !0.117 (0.091) (0.090) Married 0.049 0.050 !0.080 0.054 (0.077) (0.076) (0.051) (0.068) Some college 0.083** 0.081** 0.081 0.079 (0.038) (0.038) (0.050) (0.055) Number of household members 0.000 !0.000 0.003 !0.006 (0.008) (0.008) (0.010) (0.011) Unemployed 0.040 0.033 0.039 0.059 (0.109) (0.108) (0.115) (0.290) Age !0.002 !0.002 !0.001 !0.003 (0.001) (0.001) (0.002) (0.002) Lending client from bank !0.014 !0.014 !0.059 0.036 (0.036) (0.036) (0.046) (0.053) Lending client with default !0.032 !0.036 !0.019 !0.057 (0.072) (0.071) (0.088) (0.103) Total household income 0.049 0.050 0.136*** !0.026 (0.031) (0.031) (0.045) (0.043) Total household monthly income—squared !0.008* !0.008* !0.024*** 0.001 (0.004) (0.004) (0.008) (0.004) Female X Income share " 0 & #$ 25% 0.015 !0.000 source: Ashraf, et. al (2006) (0.182) (0.175) Female X Income share " 25 & #$ 50% 0.048 0.037 SEED Design – Results I Researchers measure change in total balances held in the financial institution (which includes the SEED and the preexisting ”normal” savings account) six and twelve months after the randomized intervention began. I They evaluate whether the overall positive savings response to the commitment product was merely a short term response to a new product, or rather representative of a lasting change in savings. I Researchers estimate the intent-to-treat (ITT) effect (the average effect of simply being offered the commitment product) on changes in savings balances after six and twelve months of the intervention. SEED Design – Results I Focusing on column (1), the coefficient on assignment to the commitment treatment group P235 is positive and significant at the 10 % significance level I After twelve months the coefficient estimate is P411, positive and significant at the 10% level. I The marketing effect, denoted by the coefficient on the second treatment group is insignificant in both intervention periods. SEED Design – Results I However, they find that the difference between commitment treatment effect and marketing commitment effect is not statistically significant.(not shown in the table regarding t-test associated with the difference) I Moreover, if they only compare marketing group and commitment group to estimate the treatment effect of the commitment device net of the marketing effect, they did not find significant treatment effect (Column (2) and Column (4)). I Therefore, OLS estimates do not provide a clear evidence saying that providing commitment device can significantly increase the savings. I This is a bit disappointing, but the researchers argue that this finding is driven by outlier. SEED Design – Results I To minimize the impact of outlier on the estimates, they instead look at the binary outcome variables in Probit model. I The first is equal to one if savings increases, and the second is equal to one if savings increases by more than 20 percent. I Researchers then regress these indicator variables on treatment assignment dummies to estimate the impact on the probability of increasing savings, and the probability of increasing savings by at least 20 percent. I This enables a substantial increase in savings by a wealthy individual to be muted in two ways: first, an outlier in the distribution of percentage savings increase would be no greater influence econometrically than a client with a savings increase slightly higher than the given cutoff level; second, the absolute magnitude of the savings increase is normalized by her initial savings level. SEED Design – Results I The coefficients on commitment-treatment in columns (5) and (7) can be interpreted as the impact of treatment relative to the control clients, and those in columns (6) and (8) as the impact of treatment relative to marketing group clients I All results demonstrate positive and significant impacts I For instance, column (5) tells us that a client offered the SEED commitment product will be 10.2 percentage points more likely to increase his savings after twelve months of intervention, and 10.1 percentage points more likely to increase savings by at least 20 percent. SEED Design – Results I Furthermore, the estimated coefficients on assignment into the marketing group are insignificant in every specification, compared with the control group. I This is consistent with the statistically insignificant marketing effects estimated in the OLS specifications, and suggests that the impact of the commitment product came from the product itself, and not from the door-to-door marketing. Forced Saving TABLE VI IMPACT ON CHANGE IN SAVINGS HELD AT BANK OLS, PROBIT INTENT TO TREAT EFFECT OLS Probit Length 6 months 12 months 12 months Binary outcome Binary outcome Binary outcome Binary outcome Change in Change in Change in Change in Dependent ! 1 if change ! 1 if change ! 1 if change ! 1 if change total total total total variable: in balance " in balance " in balance " in balance " balance balance balance balance 0% 0% 20% 20% Commitment & Commitment & Commitment & Commitment & Sample All (1) marketing only All (3) marketing only All (5) marketing only All (7) marketing only (2) (4) (6) (8) Commitment 234.678* 49.828 411.466* 287.575 0.102*** 0.056** 0.101*** 0.064*** treatment (101.748) (156.027) (244.021) (228.523) (3.82) (0.026) (0.022) (0.021) Marketing 184.851 123.891 0.048 0.041 treatment (146.982) (153.440) (1.56) (0.027) Constant 40.626 225.476* 65.183 189.074** (61.676) (133.405) (124.215) (90.072) Observations 1777 1308 1777 1308 1777 1308 1777 1308 R2 0.00 0.00 0.00 0.00 Robust standard errors are in parentheses. * significant at 10 percent; ** significant at 5 percent; *** significant at 1 percent. The dependent variable in the first two columns is the source: Ashraf, et. al (2006) change in total savings held at the Green Bank after six months. Column (1) regresses change in total savings balances on indicators for assignment in the commitment- and marketing-treatment groups. The omitted group indicator in this regression corresponds to the control group. Column (2) shows the regression restricting the sample to commitment- and marketing-treatment groups. Columns (3) and (4) repeat this regression, using change in savings balances after twelve months as a dependent variable. The dependent variable in columns (5)–(8) is a binary variable equal to 1 if balances increased by x percent. One hundred and fifty-four clients had a preintervention savings balance equal to zero. Twenty-four of them had positive savings after twelve months. These individuals were coded as “one,” and those that remain at zero were coded as zero for the outcome variables for columns (5) through (8). Exchange rate is 50 pesos for U.S. $1. SEED Design – Results I To test whether the SEED account balances represent new savings, or whether they represent shifting of assets between accounts held at the institution, researchers define a new outcome variable: change in balance in all non-SEED savings accounts. I Column (1) reports the regression of non-SEED change in balance on treatment indicators for the full one-year post intervention period. The estimated coefficient on both treatment indicators is positive but insignificant. I Thus, the improvement in savings is a result of new savings, not crowd-out of other financial savings. Crowding Out 668 QUARTERLY JOURNAL OF ECONOMICS TABLE IX TESTS FOR NEW SAVINGS OLS FULL SAMPLE OF CLIENTS 12 months Change in Non-SEED Change in total Dependent variable balance (1) balances (2) Commitment-treatment 220.776 411.466* (227.501) (244.021) Marketing-treatment 120.705 123.891 (153.437) (153.440) Constant 63.690 65.183 (124.234) (124.215) Observations 1777 1777 R2 0.00 0.00 Robust standard errors are in source: parentheses. Ashraf,* significant at 10 percent; ** significant at 5 percent; et. al (2006) *** significant at 1 percent. The dependent variable in the regressions in column (1) is the change in savings in all non-SEED savings accounts held at the institution. Exchange rate is 50 pesos for U.S. $1. SEED Design – Conclusion I Savings requires a delay of immediate rewards for greater future rewards. I Individuals who have a hyperbolic preference (impatient today, patient tomorrow) or self-control problem find it difficult to save. I Individuals with such preference, theoretically, should have a preference for commitment. I It is difficult to recover individual’s preference I This paper uses a hypothetical survey questions to identify individual’s time preference and test whether providing commitment device can increase savings of individuals with commitment/self-control issue SEED Design – Conclusion I The welfare implications of this project are ambiguous. I Merely demonstrating a positive increase in savings does not necessarily imply a welfare enhancing intervention. I The loss of liquidity of funds may cause harm to the individuals. I Further research should shed insight on this important question. Recommended Reading I Chapter 8, Poor Economics. I Ashraf, et. al ‘‘Tying Odysseus to the Mast: Evidence From a Commitment Savings Product in the Philippines’’ QJE 2006