Consumption, Saving, and Investment

1
Consumption, Saving, and Investment

• Norman Loayza
• World Bank

Macroeconomic Management for Financial Stability
and Poverty Reduction
2
 Overview

• Introduction
• Why do we care? Saving, investment and growth
• Measurement problems
• Consumption and saving
• Some facts
• Basic theories
• Recent extensions and policy issues

3

• Investment
• Some facts
• Basic theories
• Recent extensions Irreversibility and
  uncertainty
• Summary and conclusions

4

• Saving and Investment Why do we care?
• Fact 1 Saving and growth rates are strongly
  correlated
• Three alternative views
• (A) Saving ? investment ? growth
• (virtuous circles and poverty traps)
• (B) Growth ? saving
• (C) Some third factor (firm investment?) drives
  both
• If (A) is true, growth-oriented policies should
  target saving
• Even if (A) is false, avoiding large
  saving-investment imbalances contributes to macro
  stability and, thus, to growth (and East Asia?)

5

• Fact 2 Investment and growth are strongly
  correlated
•   Again three views
• (A) Investment is the only source of growth
• capital fundamentalism
•   The simplest (most popular?) version
• growth (Investment rate)/ICOR
• (B) Investment is one among several sources of
  growth
• The others Human capital investment
• Technical progress (embodied?)
• Productive efficiency
• (C) Investment is a consequence of growth
• If (A) or (B) are true, growth policies should
  target investment among other things.

6

• Fact 3 Saving and investment ratios are
  strongly correlated
• Many explanations
• (A) There is low international capital
  mobility(Feldstein-Horiokka) Legal/regulatory
  barriers, credit rationing...
• (B) Governments target (successfully) the current
  account deficit (S-I)
• (C)There is low domestic capital mobility
• Financial system imperfections force firms to
  finance Investment with retained earnings
• D)Third factors (demographics, taxation...) move
  saving and investment in the same direction

7

• Severe measurement problems with macro data
• (1) Statistical
• Consumption residual from GDP expenditure
  identity
• Saving residual of the residual
• (2) Conceptual
• Human capital accumulation, durable purchases
  classified as consumption rather than investment
  
• Standard saving measures ignore capital gains ?
  saving very different from change in net worth

8
Consumption and Saving
9
(No Transcript)
10
(No Transcript)
11
(No Transcript)
12
(No Transcript)
13
Basic Consumption/Saving Theory

• The beginnings The Keynesian consumption
  function- only current disposable income matters
• Ct ? Yt
• Contemporary mainstream Two major approaches
• The Permanent-Income Hypothesis
• The Life-cycle model

14

• The Permanent-Income Hypothesis (Friedman)
• Identical consumers looking into the indefinite
  future. Consumption equals permanent income 
  the annuity value of present and future
  discounted income.
• Three major implications...
• (1) Consumption is unaffected by predictable
  changes in income. In simplified version (Hall
  1978)
• (2) Transitory income changes are (dis-)saved
  permanent changes are consumed
• (3) Anticipated income growth reduces saving

15

• and some empirical problems
• (1) PIH fails excess sensitivity of
  consumption to income
• (2) is clearly violated in the data

16

• The Life-Cycle model (LCH) (Modigliani-Brumberg)
  Cohorts of finite-lived consumers with
  hump-shaped income over their lifetime, saving
  for retirement. Consumption depends on lifetime
  disposable income.
• Three major implications
• (1) Saving is hump-shaped, too Low when young,
  peak at midlife, negative when old.
• (2) Income growth across cohorts (but not
  otherwise), as well as population growth, raise
  aggregate saving.
• (3) Demographics affect aggregate saving Higher
  fractions of old or very young reduce total
  saving.

17
Life-cycle Profiles of Earnings and Consumption
Consumption and Earnings
earnings
work-related consumption
consumption
borrowing and dissaving
death
retirement
Age
18

• ... with empirical problems, too
• (1) LCH fails Consumption tracks income too
  closely there is not enough hump-saving in
  the cross- section data.
• (2) LCH is not sufficient to account for the
  observed growth- saving correlation.
• - Besides, old people do not dis-save at least
  not much.
• - Empirical evidence on the predicted
  relationship between age demographics and saving
  is not conclusive.
• There is growing evidence that bequests are a
  major saving motive ? the distinction between LCH
  and PIH is weakened.

19
Consumption/Saving Recent Developments

• Precautionary Saving
• Prudent consumers set aside resources to
  shield future consumption against possible future
  unfavorable changes in income, interest rates,
  etc
• Three implications
• Prudent consumers will lower their current
  consumption when future consumption is more
  uncertain ? higher uncertain raises saving
• With high future income uncertainty, consumption
  will track current income very closely as in a
  keynesian framework ? can account in part for the
  excess sensitivity
• Precautionary saving may explain the failure of
  the elderly to dissave they face uncertain
  horizons and uncertain health costs

20

• Evidence
• Limited empirical work Precautionary saving
  models are analytically hard and yield no
  closed-form solutions.
• Recent evidence suggests that precautionary
  wealth may account for a major fraction of total
  consumer wealth (Carroll and Samwick 1997).

21

• Borrowing constraints
• PIH and simple version of LCH unrealistically
  assume that consumers can freely borrow at going
  rates. In reality, many (most) consumers cannot
  borrow, particularly in LDCs
• Interest rates do not clear credit markets.
• Assets such as human capital cannot be used as
  collateral.
• The simplest approach Constrained consumers
  (denoted c) gear consumption to current income
  
• Ct ? Cut (1-?) Yct
• Empirically, ? is always significantly less than
  1 ? many consumers are constrained.
• But this approach forbids them from ever saving!

22

• More elaborate approach
• - Combined with precautionary saving, borrowing
  constraints generate buffer-stock saving (Deaton
  1992) Accumulate in good times to buffer
  consumption in bad times when they cannot borrow.
• - Buffer stocks are accumulated and run down
  over short intervals
• ? they are quantitatively small.
• Borrowing constraints on housing finance raise
  required levels of pre-purchase saving.
  Empirically, they have a significant positive
  effect on aggregate saving (Jappelli and Pagano
  1994).

23

• Subsistence consumption (not really a new
  approach)
• Below a certain income threshold, no saving is
  possible, all income is spent.
• This view can explain two facts
• (1) The positive relation between income levels
  and saving rates
• (2) The positive saving-growth correlation As
  growth rises, more consumers are pushed beyond
  the subsistence threshold and can save

24

• Consumption habits
• Standard models of consumption/saving assume
  inter-temporal separability of preferences. With
  habits, todays happiness depends on
• (1) Todays consumption and,
• (2) A stock of past consumption (internal
  habits) or others consumption (external habits).
• Habits act as a drag on consumption levels.
  They are consistent with three facts
• Excess sensitivity. Consumption changes are
  predictable (by past consumption changes)
• Positive impact of growth on saving consumption
  lags behind income growth due to the habit drag
• Low consumption early in the life cycle, due to
  the anticipated cost of having to feed the habit
  forever

25
Policy Issues Related to Saving

• Is there a distortion causing socially too low
  saving?
• Self-perpetuating traps of under-saving and
  poverty
• Saving offers insurance in imperfect world
  financial markets (sunspots, herd behavior)
• Moral hazard in retirement saving
• Too low public saving (if there is no Ricardian
  equivalence)
• ... but countries can also save too much
• Forced saving (the Soviet bloc)
• Excessive mandatory saving
• Too much risk/too few diversification
  possibilities

26

• If national saving is too low, how to raise it?
• Public saving
• Tax incentives
• Institutional/regulatory reform
• Financial system
• Pension system

27

• Does higher public saving raise national saving?
• Ricardian equivalence implication of PIH What
  matters for private consumption is the present
  value of current and future public recurrent
  expenditures ( NPV of taxes).
• ? A change in the time profile of taxes without
  changing their present value does not change
  private consumption.
• ? Tax and deficit financing are fully equivalent
• ? An increase in public saving will be fully
  offset by a matching decline in private saving

28

• This is strictly true only under stringent
  conditions
• (i) Consumers look far into the future
• (ii) They face no borrowing constraints
• (iii) Taxes are not distortionary
• (iv) Public and public consumption are
  independent
• ? It cannot be literally true. But is it a good
  approximation?
• The bulk of evidence is against full Ricardian
  equivalence.
• Offset coefficients are around .23-.65
• They are somewhat lower for expenditure cuts
  than for tax increases.
• Provisional conclusion Public saving is an
  effective tool to raise national saving.

29
Private/Public Saving Offset Coefficients
Study Sample Offset
Corbo and Schmidt-Hebbel(1991) 13 Developing 1980-1987 .47-.50 .23-.26
Masson, Bayoumi and Samei(1995) 21 Industrial 1971-1993 .40-.53
Edwards (1995) 11 Industrial 25 Developing 1970-1992 .36-.65
Dayal-Ghulati and Thimann (1997) 14 Developing 1970-1995 .23-.42
Bailliu and Reisen(1997) 7 Industrial 4 Developing 1982-1993 .46-.54
Loayza, Schmidt-Hebbel and Servén(2000) 18 Industrial 46 Developing 1970-1994 Short run 0.28 Long-run 0.60
30

• Do tax incentives raise national saving?
• - General tax incentives
• Raise the after-tax rate of return
• ? ambiguous impact on private saving
• - Targeted incentives on specific assets (IRAs,
  401k, etc.)
• Encourage portfolio reshuffling to collect the
  tax break
• Unclear if they have any effect on total
  private saving
• ? Evidence for US, OECD is inconclusive.
• Two points to note
• Even if there is a positive impact on private
  saving, there is also a fiscal cost
• ? still unclear impact on national saving 
• Distributive aspect Asset reshuffling
  easier for wealthier consumers
• ? possibly regressive redistribution

31
Digression Saving and Interest Rates   

• Three effects of real interest rate changes
• (1) Substitution effect Tomorrows
  consumption becomes relatively cheaper ? positive
  impact on saving, bigger the higher
  inter-temporal consumption substitutability
• (2) Income effect Less needs to be saved to
  enjoy a given level of consumption tomorrow ?
  negative impact on saving
• (3) Human wealth effect The present value of
  tomorrows labor income declines ? positive
  impact on saving
• In the LCH, further complications, as interest
  rate changes trigger off distributive effects
  across cohorts.

32

• Then, what is the net effect of interest rates
  on saving?
• Total net effect is theoretically ambiguous, and
  likely small.
• Empirically, little if any effects on saving.
• A recent approach Consumption-level-dependent
  substitution elasticity (Ogaki, Ostry, and
  Reinhart 1996)
• At low consumption levels, little substitution
• At high consumption levels, stronger
• ? Interest rate effect stronger for richer
  consumers
• Some empirical support for this view.

33

• Do institutional reforms raise national saving?
• (1) Financial Reform
• Typically, three aspects
• Higher real interest rates
• Easing of borrowing constraints assets become
  more easily collateralizable and, hence, more
  liquid
• Enhanced portfolio choices

34

• ? Clear impact on portfolio allocation
• - Financial savings rise
• - Flight capital returns (? measured saving
  may rise)
• ? Likely positive impact on the efficiency of
  intermediation
• ? But likely adverse direct impact on volume of
  saving
• Easier access to consumer credit (Mexico 1990s)
• Better access to risk diversification ? reduced
  precautionary saving
• ? Empirically,
• - Negative effects cross-country regressions
  (Loayza, Schmidt-Hebbel, and Serven 2000)
• - No systematic effect 8 country-case studies.
  Negative in Korea and Mexico, Positive in Ghana
  and Turkey, and Negligible in the rest (Bandiera
  et. al 2000)
• - No effect when expenditures on durable goods
  are considered as saving (negative otherwise)
  India (Loayza and Shankar 2000)

35

• (2) Pension reform Transition to fully-funded
  pension system
• Effects studied mostly with simulation models
  but also with cross-country regressions and case
  studies
• The direct effect
• Transition financed by issuing public debt No
  direct effect (its only a conversion from
  implicit to explicit public liability)
• Transition financed by reducing the non-pension
  public deficit (lowering net benefits to current
  retirees, imposing higher taxes on current
  generations, or lowering government
  expenditures) Saving levels of current
  generations decline while those of future
  generations rise, although their saving rates
  will not necessarily change.
• Higher mandated saving (with possible over-saving
  by borrowing-constrained consumers) Singapores
  Central Provident Fund - minimum 25.

36

• The side effects (mostly in the long run)
• Higher growth (from reduced distortions)
• Capital market development
• Higher rates of return on pension assets
• What does the evidence say?
• - Chile Private saving/GNP rose from 1 (1980)
  to 17 (1992). But was it due to pension reform?
• - General Countries that increase the funding of
  their mandatory retirement programs tend to
  achieve higher private saving rates
• Chile 3.8 p.p. of the increase in saving rate
  attributed to pension reform (Schmidt-Hebbel
  1999)
• Cross-country pay-as-you go systems had negative
  effects on saving
• Overall, the effect on saving is mainly in the
  long run and through mandated saving

37
(No Transcript)
38
Investment
39
(No Transcript)
40
(No Transcript)
41
Basic Investment Theory

• Two conventional views
• (1) The neoclassical model (Jorgenson)
• F(K) ucK
• where the user cost of capital depends on
  interest rates and the relative price of capital
  ucK (rd)pI ignoring taxes
• (2) Tobins q model Investment depends on the
  ratio of the value of the firm to the replacement
  cost of its productive capital (q)

42

• In both approaches, (convex) adjustment costs
  need to be assumed to transform a static problem
  into a dynamic one
• Under appropriate assumptions, the two approaches
  are essentially equivalent (Hayashi 1982)
• Empirically, these models are not successful.
• Main reason The estimated effects of the user
  cost of capital are usually very small and often
  insignificant.
• Empirical models explain only a small part of
  the observed variation in investment.

43
Recent Developments

• Financial Factors and Investment
• Modigliani-Miller (1958) Internal and external
  funds are perfect substitutes ? firms investment
  decisions are independent from their financing
  decisions
• In reality, informational asymmetries, cost
  monitoring, and contract enforcement problems
  lead to imperfect substitutability.
• Two main results from this literature (e.g.
  Fazzari et al 1988)
• (i) Unless all loans are fully collateralized,
  external finance is more costly than internal
  finance
• (ii) The premium on external finance is
  decreasing in net worth.

44

• Implications
• The financial accelerator Adverse shocks to net
  worth raise the premium and reduce investment (?
  shocks are amplified)
• Interest rate changes affect investment not only
  through (i) the cost of capital, but also through
  (ii) net worth (impacting on the cost spread),
  and through (iii) the availability of internal
  funds.
• Taxes have marginal effects (on marginal
  profitability and cost of capital) and also
  average (income) effects on investment, through
  internal funds availability.
• Financial reform can reduce the cost spread and,
  thus, raise investment.
• Empirically, there is strong evidence that the
  availability of internal funds affects investment
  positively. The impact seems to be larger for
  small firms.

45

• Irreversibility and Uncertainty The Options
  Approach
• Three basic facts (Dixit-Pindyck 1994)
• (i) Investment is irreversible
• ? adjustment costs are asymmetric Projects
  cannot be undone - without high cost
• (ii) Future returns are uncertain
• ? waiting allows new information about future
  profits
• (iii) Investors can choose the timing of
  investment
• ? investment can be postponed
• (i)(ii)(iii)? Option value of waiting
  Investing kills the option to wait

46

• New Investment rule Invest if naive NPV gt
  value of option to wait (and not just NPV gt 0)
• Properties of the option value
• It only depends on bad news (good news is
  irrelevant)
• It can be very large even with small uncertainty
• ? Uncertainty can be a powerful investment
  deterrent.

47

• The bad news principle
• Projects costs pk
• Yields R0 today and an uncertain R every future
  period let ER R0
• Discount rate is r

48

• Reversible investment
• ? invest now iff current profit ? user cost
• R0 ? r pk
• Irreversible investment, and waiting yields
  information
• ? invest now iff current profit ? user cost
  user cost EPV of bad news.
• If uncertainty is fully resolved next period
  invest now iff
• R0 ? rpk Pr R lt rpk (1/r) Erpk-R
  Rltrpk
• current profit ? user cost of capital Prob.
  irrev. mistake EPV mistake

49
The bad news principle with multiple projects
a productivity improvement leading to inaction
Return per period
Return per period
user cost of capital
user cost of capital
Real wage
Real wage
50
Irreversibility and Uncertainty Implications

• At the firm level
• (1) Higher uncertainty raises the profitability
  threshold for positive investment (enlarges the
  range of inaction)
• But
• (2) Higher uncertainty may also raise expected
  profitability
• (1) (2) ? the short term impact on investment
  is ambiguous
• And
• (3) Even if (1) outweighs (2), long-run impact
  is ambiguous due to hangover effect Higher
  uncertainty implies more excess K
• At the aggregate level
• (4) Investment displays considerable inertia -
  the response to aggregate shocks/policy changes
  is gradual
• (5) The response depends on initial conditions -
  i.e., where individual firms are relative to
  their investment thresholds.

51
Irreversible Investment Applications

• Interest rat uncertainty (Ingersoll-Ross)
• Tax uncertainty (Hasset-Metcalf)
• Real exchange rate uncertainty (Dixit-Krugman)
• Debt overhang (uncertain debt service/anticipated
  tax)
• Main message Stability of incentives is as
  important as their level.
• Reform credibility
• Why the sluggish investment response to policy
  reform?
• Possibility of reform reversal ? value of
  waiting
• Imperfect credibility of reform leads to
  investment pause
• Problem Endogenous reform collapse may occur
  due to lack of investment (? coordination
  failure)

52
Irreversibility Empirical Evidence

• Few structural models too difficult to do
• Many reduced-form studies of uncertainty-investmen
  t link (Almost always significantly negative)
• Economic instability (volatility of RER, TOT,
  inflation...)
• Political/social instability (political
  violence, civil unrest...)
• But is irreversibility the reason? Other
  possible explanations
• Risk aversion/incomplete markets
• uncertainty has a negative impact on investment
  (Zeira 1989)
• Disappointment aversion
• Bad outcomes are most important
  (Aizenman-Marion 1995)
• negative effect on investment

53
Private Investment and Economic Instability
54
(No Transcript)
55
(No Transcript)
56

• Socio-Political/Institutional Factors
• Property Rights
• Poorly defined or poorly enforced property
  rights
• (1) Raise transaction costs
• (2) Make investment returns more uncertain
• Corruption
• Informal taxes on investment
• Like formal taxes, they raise project cost
• But are hard to predict ? raise uncertainty as
  well
• Empirically Reduced-form models including
  indicators of
• Property rights/contract enforcement
  (Knack-Keefer 1995)
• Corruption (Mauro 1996)
• Overall institutional quality (civil liberties
  indicators)
• ? Almost always find significant effects

57
Private Investment and Socio-Political Variables
58

• Coordination Failures
• Strategic interactions among investors can lead
  to sub-optimal (typically under-) investment.
• Two examples
• 1. Increasing returns (Kiyotaki 1990)
  Profitability of individual firms depends on
  overall activity  ? on overall investment.
  Individual firms ignore their contribution to
  overall investment. ? Multiple equilibria with
  self-fulfilling expectations
• If each firm anticipates low aggregate
  investment, it does not invest.
• If it anticipates high aggregate investment, it
  will invest.
• ? The economy may get stuck in a pessimistic
  equilibrium

59

• 2. Irreversible investment with private
  information (Gale 1994)
• Investment by others is informative (about
  profitability, technology)
• It pays to wait for them to invest
• ? Equilibrium with too much waiting and
  under-investment
• Implication With coordination failures, the
  social value of investment differs from
  (exceeds?) its private value
• ? Role for investment-oriented policies

60
Conclusions

• Public saving is an effective tool to raise
  national saving
• Financial reform will improve the efficiency of
  intermediation and foster investment and growth.
  It will likely reduce saving in the short to
  medium run
• Pension reform likewise has beneficial effects on
  efficiency and growth. Its direct impact on
  saving appears to be positive.
• The level of income is one the most important
  determinants of the saving rate. Any reform that
  increases income will have a long-run positive
  effect on saving.

61

• Investment is highly sensitive to economic
  instability. The stability of the incentive
  framework is at least as important as the level
  of the incentives themselves
• Coordination failures may lead to
  under-investment and justify policy intervention
  to raise investment
• Elimination of macro-instability (large deficits,
  high and volatile inflation...) is a key
  requirement for private investment to take off
• Socio-political instability, corruption and weak
  institutions are major investment deterrents.
  Institutional reform may play a major role in
  fostering investment