____________________________________________________________________________________________________

1. Learning Outcomes

After studying this module, you shall be able to

1. Know about the flaws of the permanent income hypothesis and life cycle hypothesis.
2. Understand the role of liquidity constraints behind the consumption behavior.
3. Understand the idea of buffer-stock savings.
4. Examine the relationship between the accumulated wealth by the consumers and their
level of consumption.
5. Know about the empirical evidences on liquidity constraints and buffer-stock savings
models of consumption.

2. Introduction

According to Permanent Income Hypothesis and the Life-Cycle Hypothesis, the

consumer is assumed to determine its level of consumption on basis of its income which
it earns throughout its life time on a stable path. But the alternative theories of
consumption say that the optimal inter-temporal consumption behaviour of consumers is
actually restricted due to their ability to borrow to finance consumption. Chief
modifications in this direction have been pioneered by the works of Deaton (1991) and
Carroll et al. (1992) and Carroll (1996). This model takes in to account the savings for
precautionary motive or say, as a buffer stock. It is generally presumed that the
precautionary or the buffer stock savings arise mainly due to two assumptions held by
Deaton as well as Carroll. Deaton (1991) assumes that the agent is always employed and
hence has a positive income. But Carroll assumes that the agent may face unemployment
spell at some point in future and therefore will receive a zero labour income in that time
period. Deaton says that the savings act like a buffer stock, protecting consumption

ECONOMICS Paper No. 6 : Advanced Macroeconomics

Module No. 15: Empirical Issues: Evidence on Liquidity
Constraints and Buffer Stock View of Savings
____________________________________________________________________________________________________

against the bad draws of income when the consumers are

not permitted to borrow. The precautionary demand for savings interacts with borrowing
constraints to provide a motive for holding assets. Hence, this module is concerned with
the optimal inter-temporal consumption behavio ur of consumers who are restricted in
their ability to borrow to finance consumption.

3. Liquidity Constraints

The permanent-income hypothesis assumes that the lending and the borrowing rates are
the same for the consumers but we know that the consumers borrow at a higher rate than
their savings earn. Besides, some individuals are unable to borrow more at any rate of
interest. These facts pose as the liquidity constraints on the consumers which influence
their consumption decisions. If they face high interest rates on borrowings, they may
choose not to borrow to smooth their consumption when their current resources are low
and if they do not borrow at all, they have no choice but to have low consumption when
their current resources are low. The presence of liquidity constraints causes individuals to
save as insurance against the effects of future falls in income. The liquidity constraints
also imply that the current income is more important determinant of consumption than
predicted by the life-cycle and permanent income hypothesis. These theories actually fail
to show that in reality, when the permanent income is higher than the current income, the
consumers may be unable to borrow to consume at the higher level in the expectation of
higher income in future. The liquidity constraints, on the other hand, can raise saving in
two ways. Firstly, if the liquidity constraints are binding, the individuals will consume
less and secondly even if the liquidity constraints are not currently binding but they may
bind in future, this will again reduce their current consumption. The second case may
happen if there is a chance of low income in the next time period, it may force the
consumer to consume less in the current period if there are liquidity constraints unless the
individual has enough savings to compensate that fall in future income. Thus, the
presence of liquidity constraints forces individuals to save for precautionary motives. In
other words, the savings are used as buffer stocks which are accumulated when times are
ECONOMICS Paper No. 6 : Advanced Macroeconomics
Module No. 15: Empirical Issues: Evidence on Liquidity
Constraints and Buffer Stock View of Savings
____________________________________________________________________________________________________

good in order to maintain consumption when times are

bad. In simple words, the liquidity constraint implies that the consumption in any time
period must be less than or equal to one.

Thus, under the borrowing constraint, the consumers can not borrow and has to make his
consumption decisions on basis of their current income and budget constraint. This can
be observed from figure 1.
Figure 1: The Borrowing Constraint

Source: Mankiw (2009).

Under the borrowing constraint, we can see in the figure that apart from the budget
constraint, the consumer faces the additional constraint that first-period consumption
cannot exceed first-period income. The shaded area represents the combinations of first-
period and second-period consumption the consumer can choose. Thus, the liquidity
constraint affects the consumption decision but these decisions will be different in case of
binding and non-binding constraints. This can be observed from figure 2.

Figure 2: The Binding and Non-binding Liquidity/Borrowing Constraint

ECONOMICS Paper No. 6 : Advanced Macroeconomics

Module No. 15: Empirical Issues: Evidence on Liquidity
Constraints and Buffer Stock View of Savings
____________________________________________________________________________________________________

Source: Mankiw (2009).

Figure 2 shows that there are two possibilities that the liquidity constraint can be binding
or not binding. In panel (a), the consumer wishes to consume less in period one than he
earns, so even if there is some liquidity constraint, the consumer may continue with
his/her existing level of consumption. The borrowing constraint is not binding and,
therefore, does not affect consumption. In panel (b), the consumer would like to choose
point D, where the consumer likes to consume more in period one than he earns, but the

ECONOMICS Paper No. 6 : Advanced Macroeconomics

Module No. 15: Empirical Issues: Evidence on Liquidity
Constraints and Buffer Stock View of Savings
____________________________________________________________________________________________________

borrowing constraint prevents this outcome. The best the

consumer can do is to consume all of his first-period income, as shown by point E. Thus,
we can say that the existence of liquidity constraints leads us to conclude that there are
two consumption functions. For some consumers, the borrowing constraint is not binding,
and consumption in both periods depends on the present value of lifetime income i.e.
Y1 + [Y2/(1 +r)]
ty constraint is
binding, and the consumption function is C1 =Y1 and C2 =Y2. Hence, for those consumers
who would like to borrow but are not able to do so, the consumption decisions depend
only on current income or in other words, when the borrowing constraint is binding, first-
period consumption
equals first-period income and same is true for the second period.
Deaton, describes the consumption behaviour of the consumer with liquidity
constraint in three period model the end of period 1, period 2 and period 3, simply
-
temporal utility maximization.
i.e. (1)
Where, , is the time rate of discount, and is the instantaneous utility function
assumed to be increasing, strictly concave, and differentiable. The evolution of assets is
given by:
(2)
-the consumption
). The real interest rate is treated as fixed
and known, and all the uncertainty is focussed on labour income yt; labour is inelastically
supplied and and the marginal utility function , is a positive
monotone decreasing function. Considering the case of impatient consumer (i.e. ),
the cash in hand of the consumer at a point of time is
(3)

ECONOMICS Paper No. 6 : Advanced Macroeconomics

Module No. 15: Empirical Issues: Evidence on Liquidity
Constraints and Buffer Stock View of Savings
____________________________________________________________________________________________________

i.e. is the maximum amount that a consumer can spend

t is the maximum that can be spent on consumption in period t.
Consumption in periods t and t + 1 must satisfy

If the consumer is constrained, consumption can be no higher than x t, and the marginal
utility no lower than . The constraint will bind if marginal utility at x t is higher than
the discounted expected marginal utility next period; otherwise the two marginal utilities
are equated in the usual way. From equation 2 and 3, x t+1 can be evolved as:

In this constrained function the marginal utility today is equated to the maximum value of
marginal utility in the constrained situation and the discounted expected value of
tomorrow's marginal utility. The marginal utility of money (price of consumption)

or
Imagining a series of time periods from 0 to n, the marginal utility of money in initial
period will be equal to
is set by the borrowing constraints or to equate marginal utilities, and also back in
time. Without borrowing constraints, it is the convexity of that controls the degree
of precautionary saving. With borrowing constraints, the same role is played by p(x), so
the inherited convexity means that the same arguments for prudence and precautionary
savings go through when borrowing is prohibited. Indeed, p(x) is more convex than ;
the inability to borrow in adversity reinforces the precautionary motive. The general
properties of the solution are clear. Starting from some initial level of assets, the
household receives a draw of income. If the total value of assets and income is below the
critical level x*, everything is spent, and the household goes into the next period with no
assets. If the total is greater than x*, something will be held over, and the new, positive
level of assets will be carried forward to be added to the next period's income. Note that
there is no presumption that saving will be exactly zero; consumption is a function of x,
not of y, and f(x) can be greater than, less than, or equal to y. Assets are not desired for
ECONOMICS Paper No. 6 : Advanced Macroeconomics
Module No. 15: Empirical Issues: Evidence on Liquidity
Constraints and Buffer Stock View of Savings
____________________________________________________________________________________________________

their own sake, but to buffer fluctuations in income.

When income is low, there will be dissaving, and when it is high, there will be saving.
3.1 Empirical Evidence of Liquidity Constraints: There are many important works

savings etc. One such work is by Jappelli and Pagano (1994). They investigated if the
cross-country differences in liquidity constraints are the main explanatory factors behind
the cross country differences in aggregate savings. They tried to capture the differences in
conditions regarding taking loans, down-payments required to purchase, restrictions on
kruptcy and foreclosure laws etc., then they tried to find if these
differences are associated with the differences in saving rates. They tried to examine the
relationship between loan-to-value ratio (which is defined as one minus the required
down-payment) for home purchases and the saving rate. It was found that there exist a
strong negative relation between the two. The regression equation results in this case
pointed out that an increase of 10 per cent in the required down-payments is associated
with a rise in the saving rate of 2 per cent of NNP. With these results they tried to prove
that similar results exist if we replace the loan-to-value ratio with availability of
consumer credit. Thus, the liquidity constraints have an important influence on aggregate
savings.
The works of Hubbard and Judd (1986) tried to find the impact of liquidity
constraints on consumption as well as the fiscal policy. They emphasised that policy
simulation models that ignore "liquidity constraints" result in flawed tax policy analysis.
They found that the existence of liquidity constraints is important in the determination of
consumption behaviour in a life-cycle model. Forced lifetime saving due to liquidity
constraints is substantial, and if the inter-temporal elasticity of substitution in
consumption is small, the incorporation of borrowing constraints enables the life-cycle
model to generate more realistic predictions about the size of the aggregate capital stock.
They argued that reduced capital income taxation financed by increased labour income
taxation raises individual welfare depend on a substantial interest sensitivity of saving in
the life-cycle framework and on the ability of consumers with low current earnings to

ECONOMICS Paper No. 6 : Advanced Macroeconomics

Module No. 15: Empirical Issues: Evidence on Liquidity
Constraints and Buffer Stock View of Savings
____________________________________________________________________________________________________

borrow to finance higher labour income taxes. With

borrowing restrictions, the gains from higher saving rates and output must be weighed
against the efficiency losses from the reduced consumption of constrained individuals.
Thus, elimination of capital income taxation compensated by higher labour income
taxation can reduce the welfare of a representative individual. There are many other
empirical studies like Flavin (1981) and many subsequent authors which have found
evidence that changes in consumption are positively related to predictable changes in
income. The microeconomic evidence, primarily from the Panel Study of Income
Dynamics (PSID), is more mixed, but Hall and Mishkin (1982) and Zeldes (1985) found
a relationship between changes in food consumption and previously predictable income
changes. Similarly, Bradford DeLong and Lawrence Summers note that from 1899 to
1916, essentially all consumption was done by liquidity-constrained consumers.

4. Buffer Stock Savings

Buffer-stock saving behaviour can emerge from the standard dynamic optimization
framework when consumers facing important income uncertainty are both impatient, in
the sense that if income were certain, they would like to borrow against future income to
finance current consumption, and prudent, in Miles Kimball's sense that they have a
precautionary saving motive. The buffer-stock behaviour arises because impatience
makes consumers want to spend down their assets, while prudence makes them reluctant
to draw down assets too far. If wealth is below the target, fear (prudence) will dominate
impatience and the consumer will try to save, while if wealth is above the target,
impatience will be stronger than fear and consumers will plan to dissave. In this context
Carroll et al. (1992) says that unemployment expectations are important in this model
because when consumers become more pessimistic about their employment situation,
their uncertainty about future income increases, so their target buffer-stock increases, and
they increase their savings to build up wealth towards the new target. While Deaton
(1991) takes in to account the liquidity constraint, the model by Carroll et al. says that if
there is uncertainty about employment situation, the consumer will try to maintain the
ECONOMICS Paper No. 6 : Advanced Macroeconomics
Module No. 15: Empirical Issues: Evidence on Liquidity
Constraints and Buffer Stock View of Savings
____________________________________________________________________________________________________

buffer stock savings even if there is no liquidity

constraint. While Deaton did not treat unemployment in his model, the simulation

probably a crucial factor in determining the amount and characteristics of buffer-stock

saving. This model finds a high degree of uncertainty as the households may occasionally
experience very bad outcomes in which their income drops essentially to zero. Even with
unchanging expectations about the average future level of income, changes in the
expected probability of bad events (i.e. spells of unemployment) have a major impact on
current consumption and saving. Interestingly, while standard life-cycle and permanent-
income models imply that the interest elasticity of saving should be strongly positive, in
the buffer-stock model the interest elasticity of saving is approximately zero.
4.1 The Basic Model: Considering a standard inter-temporal consumption model:

Such that

Here, YLt is total family non-capital income (labour income for short); V t is a
t

which can also be defined as the value of labour income if no transitory shocks occur i.e.
Vt=1; Nt is period t's multiplicative shock to permanent labour income. G = 1 + g, where
G is the growth factor and g is the growth rate; W is net wealth; R = 1 + r, where r is the
interest rate and R is the interest factor; is the discount factor, where is the

discount rate; and C is consumption. The standard Constant Relative Risk Aversion
(CRRA) utility function is of the form where, is the coefficient
of relative risk aversion.
In order to generate buffer-stock saving behaviour, it is necessary that consumers
be impatient in the sense that if they faced no income uncertainty, they would want to

ECONOMICS Paper No. 6 : Advanced Macroeconomics

Module No. 15: Empirical Issues: Evidence on Liquidity
Constraints and Buffer Stock View of Savings
____________________________________________________________________________________________________

borrow against their future income in order to consume

more today. Since the consumer is assumed to be impatient, the condition required to
exhibit impatience is

If income is perfectly certain, the growth of consumption would be approximately equal

to

If the consumer has no wealth at all, the present discounted value (PDV) of consumption
will be equal to the PDV of income. if the growth of consumption is less than the income
at same PDV, the level of consumption must be higher than the level of income. Thus, if
income were certain and , the consumers would wish to spend more than
income, i.e., they would like to borrow and consume more than the income. Thus, the
consumers are impatient in their initial life but as they grow older and g falls with age,
they will be no longer be impatient. But in reality, we know that people always save for
the rainy day. This fact raises the question why does buffer stock behaviour arises?
Carroll et al. says that if shocks to consumption are lognormally distributed, using the
Euler equation, the consumption will grow according to

Here, the variance in consumption i.e is negatively related to wealth

i.e. the consumers with less wealth have less ability to buffer consumption against
income shocks. Thus, they have higher variance and faster consumption growth. The
growth rate of consumption is high when wealth is small because the level of
consumption is being depressed by precautionary saving. Therefore, over time, as
precautionary saving adds to wealth, consumption will become less depressed. This
relationship can also be understood with help of the following figure.
Figure 3: The Buffer-Stock Relationship between Growth Rates and the Gross
Wealth Ratio

ECONOMICS Paper No. 6 : Advanced Macroeconomics

Module No. 15: Empirical Issues: Evidence on Liquidity
Constraints and Buffer Stock View of Savings
____________________________________________________________________________________________________

Source: Carroll et al. (1992)

Figure 3 shows the relationship between the gross wealth ratio, expected consumption
growth, and income growth. Consumers are impatient as . The line with
arrows traces out the relation between expected consumption growth ( in the
next period and gross wealth in the current period i.e. xt. As xt approaches its minimum
possible level , expected consumption growth approaches infinity (here, is meant to
signify the negative of the minimum possible present discounted value of future labour
income). This is because as gross wealth xt approaches , Ct must approach zero. The
expected level of Ct+1 is positive even if the consumer enters period t + 1with no assets,
therefore as xt approaches , ( approaches infinity. On the other hand, as xt
approaches infinity, uncertainty about future labour income becomes essentially
irrelevant to consumption. Thus as xt approaches infinity, the consumption growth rate
approaches the growth rate under certainty, . As everything in this model is
assumed to be continuous and monotonic, the expected consumption growth rate will
cross the income growth rate curve at one point. The gross wealth ratio at this point will
be called x*, or the target gross wealth ratio. It is a target ratio in the sense that, if actual
gross wealth is below x*, the consumer will spend an amount small enough so that gross

ECONOMICS Paper No. 6 : Advanced Macroeconomics

Module No. 15: Empirical Issues: Evidence on Liquidity
Constraints and Buffer Stock View of Savings
____________________________________________________________________________________________________

wealth will be expected to increase; however, if actual

gross wealth is greater than x*, the consumer will spend enough so that expected gross
wealth next period will decline (as shown by the arrows on the expected growth rate
curve). The model's results differ under alternative parameter values. For instance,
increasing g, the growth rate of future income, will decrease the target wealth stock by
shifting the intersection with the curve to the left. Thus, higher future income
results in higher current consumption, hence lower saving and lower wealth. Increasing
the discount rate will shift down; therefore the curve will also
shift down, decreasing target wealth. Increasing the interest rate will shift up,
pushing up and increasing target wealth. Increasing the degree of uncertainty
in income will increase the variance of consumption growth for any level of wealth,
directly shifting the curve up and increasing target wealth. Increasing the
coefficient of relative risk aversion will have two effects. It will shift the
curve up as [ ] increases, tending to increase wealth (as a direct effect of
increased risk aversion). It also will shift the curve down, thus tending to
reduce wealth (this is the effect of a lower inter-temporal elasticity of substitution).

4.2 The Buffer-Stock Model and Some Empirical Evidence: There is sufficient
empirical evidence to suggest that the Keynesian and the standard Permanent Income
Hypothesis (PIH) and Life Cycle Hypothesis (LCH) do not hold true in life and the
consumers behave according to the Buffer-stock model. Carroll and Summers (1991)
across countries, and within the same country over time, the growth rate of consumption
tends to be very close to the growth rate of income. The mechanism for this is adjustment
of the capital stock. If consumption is too high and growing more slowly than income,
the capital stock will be declining. As the capital stock declines, the interest rate
increases, and as a result, the growth rate of consumption tends toward the growth rate of
income. Conversely, if consumption is low and growing faster than income, the capital
stock will be increasing, driving interest rates down and reducing the consumption
growth rate. Hence, the steady state is eventually achieved with consumption growth
ECONOMICS Paper No. 6 : Advanced Macroeconomics
Module No. 15: Empirical Issues: Evidence on Liquidity
Constraints and Buffer Stock View of Savings
____________________________________________________________________________________________________

equal to income growth. Campbell (1987) also showed

that the current savings are equal to the present discounted value of expected declines in
income. Campbell and Mankiw (1989) further provided evidence to prove that for both
total consumption and consumption of nondurables and services, spending responds
significantly to predictable changes in income. In this study the unemployment
expectations variable has a negative coefficient, implying that consumption growth is
slower in periods when consumers are pessimistic about future unemployment
conditions. Campbell and Deaton (1989) and many others have found that consumption
appears to exhibit "excess smoothness" with respect to changes in permanent income.
They emphasised that there is some sort of "excess
smoothness" of consumption with respect to changes in unemployment expectations. In
this perspective, Ricardo Caballero (1992) interprets it as a model of near-rationality, but
it could just as easily be interpreted as one in which even nondurable consumption
spending has fixed adjustment costs. He finds that such a model can generate excess
smoothness of consumption with respect to shocks to permanent income. Such a model
would also generate excess smoothness with respect to changes in uncertainty, thus
providing at least a potential for a joint explanation of both kinds of excess smoothness.
Thus, we can find many paper which provide macro evidence in support of buffer-stock
saving model.

5. Summary

According to the Permanent Income Hypothesis and the Life-Cycle Hypothesis, the
consumer is assumed to determine its level of consumption on basis of its income which
it earns throughout its life time on a stable path. But the alternative theories of
consumption say that the optimal inter-temporal consumption behaviour of consumers is
actually restricted due to their ability to borrow to finance consumption. These models
such as the model based on liquidity constraint and the buffer-stock savings take in to
account the savings for precautionary motive or say, as a buffer stock. The precautionary
or the buffer stock savings arise mainly due to uncertainty about future income or the
ECONOMICS Paper No. 6 : Advanced Macroeconomics
Module No. 15: Empirical Issues: Evidence on Liquidity
Constraints and Buffer Stock View of Savings
____________________________________________________________________________________________________

capacity to repay the borrowings in future time period.

The precautionary demand for savings interacts with borrowing constraints to provide a
motive for holding assets. Hence, for these consumers, the current income and its growth
is a stronger factor determining the current level of consumption and its growth than
assumed by the PIH and LCH.

ECONOMICS Paper No. 6 : Advanced Macroeconomics

Module No. 15: Empirical Issues: Evidence on Liquidity
Constraints and Buffer Stock View of Savings