3.4 Consumption Volatility and Financial Openness Does the ...contents.kocw.net/KOCW/document/2015/hanyang/namdeokwoo/11.… · If more consumption smoothing is achieved, the computed

3. GAINS FROM CONSUMPTION SMOOTHING

3.4 Consumption Volatility and Financial Openness

• Does the evidence show that countries avoid consumption volatility by embracing financial globaliza-tion?

— A simple test might be to compute the volatility of a country’s consumption divided by the volatility ofits output (where volatility is measured by the standard deviation of the growth rate).

∗ If more consumption smoothing is achieved, the computed ratio ought to fall.

· In fact, in our simple model of a small, open economy that can borrow or lend without limit, andthat prefers a perfectly smooth path of consumption, this ratio should fall to zero when the gainsfrom financial globalization are realized.

∗ In practice, this will not be true if all countries are affected by common global shocks.

· For example, if every country suffers a common negative shock, every country will want to borrow,but that is simply infeasible.

∗ However, not all shocks are global, so countries ought to be able to achieve some reductionin consumption volatility through external finance.

— With this in mind, the following figure presents some discouraging evidence.

Page: 28

3. GAINS FROM CONSUMPTION SMOOTHING

• In the figure, countries are grouped into ten groups(deciles) from those that participate least in finan-cial globalization (are least financially liberalized)to those that participate most (are most financiallyopen). The average consumption-to-output volatil-ity ratio, expressed as a percentage, in each groupis shown.

— In closed countries, we would expect thevolatility of consumption would be similar tothe volatility of output (GDP), so the ratio ofthe two would be close to 100%. But if an opencountry were able to smooth consumption inline with our simple model, this ratio ought tobe lower than 100%. The figure shows:

∗ Only the most financially open countrieshave volatility ratios less than 100%.

∗ The high ratios in groups 1 to 8 show,perversely, that consumption is even morevolatile than output in these countries.

— Why are these findings so far from the predic-tions of our simple models?

1. In poorer countries, some of the relativelyhigh consumption volatility must be unre-lated to financial openness —it is there evenin closed countries.

2. In the real world, households are not identi-cal and global capital markets do not reachevery person. In fact, some people and

firms do not or cannot participate in evendomestic financial markets, perhaps due tobackward financial systems.

3. Financial markets in emerging markets anddeveloping countries may not be fully de-veloped, or may be limited in their access.

— The evidence may not imply a failure of fi-nancial globalization, but it does not providea ringing endorsement either.

∗ Consumption-smoothing gains may proveelusive for emerging markets until they ad-vance further by improving poor governanceand weak institutions, developing their fi-nancial systems, and pursuing further finan-cial liberalization.

Page: 29

4. GAINS FROM EFFICIENT INVESTMENT

4 Gains from Effi cient Investment

• In this section, we incorporate investment into the previous model, showing how financial openness benefitsthe economy through not only consumption smoothing but also increased investment opportunities.

— That is, openness delivers gains not only on the consumption side but also on the investment sideby improving a country’s ability to augment its capital stock and take advantage of new productionopportunities.

4.1 The Basic Model

• From the previous model, we can now assume that output is produced using labor and "capital", which isaccumulated through investment, so GNE = C + I (we still assume that G = 0). Therefore, the LRBCbecomes:

0︸︷︷︸Initial wealth is zero

= Present value of TB ⇒ Present value of Q︸︷︷︸Present value of GDP

= Present value of C + Present value of I︸︷︷︸Present value of GNE

— Using this modified LRBC, we now study investment and consumption decisions in two cases:

1. A closed economy in which external borrowing and lending are not possible:

∗ TB = 0 in all periods, so the LRBC is automatically satisfied.

2. A open economy in which external borrowing and lending are possible:

∗ TB need not equal zero and can be used to finance differences between GNE = C + I andGDP = Q. As before, we need to verify the LRBC holds.

Page: 30


4.2 Effi cient Investment: A Numerical Example and Generalization

• We will consider the previous numerical example with economic shocks.

— Instead of a negative shock to output, we now suppose there is a shock to the economy in year 0 thattakes the form of a "new" investment opportunity.

∗ For example, it could be that in year 0 engineers discover that by building a new factory with a newmachine, the country can produce a certain good much more cheaply than current technology allows.Or perhaps there is a resource discovery, but a mine must first be built to extract the minerals fromthe earth.

• More specifically, we assume the investment opportunity requires 16 units of output in year 0, andwill increase output by 5 units in all subsequent periods (but not in year 0).

1. Before the investment opportunity shock, output and consumption were each 100 in all periods (QN =CN = 100), had a present value of 2, 100. Investment was zero in every period, and our modified LRBCwas clearly satisfied.

— If the country decides to not act on the investment opportunity, this situation continuesunchanged whether the economy is closed or open.

2. Now suppose the open economy undertakes the investment. Then, we consider what happens when theinvestment opportunity arises.

— To calculate the implied changes in Q, C, I, TB, NFIA, CA and W over time, we take steps similarto the ones outlined previously. Steps to computing numerical values for Q, C, I, TB, NFIA, CAand W over time are as follows:

(a) Identify how much output is needed to finance the investment project:

I0 = 16 and I = 0 in all subsequent years ⇒ PV (I) = 16

Page: 31


(b) Calculate the present value of output assuming the investment project is undertaken: by the assump-tion about a stream of output,

PV (Q) = Q0 + Qr∗

= 100 + 1050.05

= 2, 200

(c) Calculate consumption in each period, based on the assumption that the country wants to maintainthe same level of consumption each period.

— Along with Steps (1) and (2), the LRBC PV (Q) = PV (C) + PV (I) gives the present value ofconsumption PV (C) = 2, 200− 16 = 2, 184, which is now divided equally each period (i.e., to keepconsumption smooth), so C = C0 = C1 = C2 = · · · :

PV (C) = C + Cr∗

= 2, 184 ⇒ C = 104 in all years

(d) Calculate the trade balance in each period. Now that we have determined output, investment, andconsumption in each period, we have:

TB0 = Q0 − (C0 + I0) = 100− (104 + 16) = −20

and

TB = Q− (C + I) = 105− (104 + 0) = 1 in all subsequent years

Page: 32


(e) Calculate the country’s external wealth W and the implied interest payments paid or received on thisdebt to calculate NFIA in each period.

— External wealth in each period is equal to −20 (does not explode), representing the initial amountborrowed:

W0 = W1 = W2 = · · · = −20 in all years

— In each period starting from year 1, the country services this debt, paying 1 unit in interest (=20× 0.05), represented by −1 in net factor income from abroad:

NFIA0 = r∗W−1 = 0 (W−1 = 0 by the assumption)

and

NFIA = r∗W = 0.05× (−20) = −1 in all subsequent years

(f) Calculate the country’s current account in each period. Recall that the current account is the sum ofthe trade balance, net factor income, and net unilateral transfers (assumed to be zero in our model):

CA0 = TB0 +NFIA0 = −20 + 0 = −20

and

CA = TB +NFIA = 1− 1 = 0 in all subsequent years

Page: 33


• This table lays out the details.

— This outcome is preferable to any outcome the closed economy can achieve.

∗ To attain an output level of 105 from year 1 on, the closed economy cuts consumption to 84 in year0 to free up 16 units for investment. In subsequent years, consumption is 105 equal to the output levelof 105. However, this is not a smooth consumption path.

∗ The open economy could choose this path (cutting consumption to 84 in year 0), because it satisfiesthe LRBC. However, it won’t choose this path because by making the investment, it cannot onlysmooth its consumption but also do so at a higher level. The open economy is better off making theinvestment and smoothing consumption, two goals that the closed economy cannot simultaneouslyachieve.

Page: 34


• Generalizing

— The lesson of our numerical example applies toany situation in which a country confronts newinvestment opportunities. Suppose

1. A country starts with zero external wealth,constant output Q, consumption C equalto output, and investment I equal to zero.

2. A new investment opportunity appears re-quiring ∆K units of output in year 0, andthis investment will generate an additional∆Q units of output in all subsequent years(but not in year 0).

— Households care about consumption. In anopen economy, they can smooth their con-sumption, given future output. That is, thecountry’s objective is to maximize the level ofconsumption, equivalently, the present value ofconsumption, subject to the LRBC.

∗ ∆PV (Q) = ∆Q(1+r∗)

+ ∆Q

(1+r∗)2 + · · · = ∆Qr∗

and ∆PV (I) = ∆K. Thus, the changein the present value of consumption is

∆PV (C) = ∆PV (Q)−∆PV (I) = ∆Qr∗−∆K .

This means the investment will in-crease the present value of consump-tion and thus will be taken if and only

if ∆Qr∗≥ ∆K .

— Rearranging the above inequality, investmentis undertaken when

∆Q︸︷︷︸Output increase

insubsequent periods

≥ r∗∆K︸︷︷︸Interest payment duein subsequent periods tofinance initial investment

∗ Investment occurs up to the point at whichthe annual benefit from the marginal unitof capital (∆Q) equals or exceeds the an-nual cost of borrowing the funds to pay forthat capital (r∗∆K).

— Putting it another way, investment is under-taken when

∆Q

∆K︸︷︷︸MPK

≥ r∗︸︷︷︸World real interest rate

∗ This is a standard formula for the opti-mal/effi cient level of investment - firms willtake on investment projects as long as themarginal product of capital or MPK(i.e., the added output associated with aone-unit increase in the capital stock) is atleast as great as the cost, the real interestrate

Page: 35


4.3 Summary: Make Hay while the Sun Shines

• In an open economy, firms borrow and repay to undertake investment that maximizes the present value ofoutput. Households also borrow and lend to smooth consumption, but these borrowing and lending decisionsare separate from those of firms.

1. When investing, an open economy sets its MPK equal to the world real interest rate.

— If conditions are unusually good (high productivity), it makes sense to invest more capital and producemore output. Conversely, when conditions turn bad (low productivity), it makes sense to lower capitalinputs and produce less output. This strategy maximizes the present value of output minus investment.

2. Then households address the separate problem of how to smooth the path of consumption when outputchanges.

• A closed economy must be self-suffi cient.

— Any resources invested are resources not consumed. More investment implies less consumption. Thiscreates a trade-off:

∗ When investment opportunities are good, the country wants to invest to generate higher output in thefuture. Anticipating that higher output, the country wants to consume more today, but it cannot — itmust consume less to invest more.

• The key lesson is that financial openness helps countries to "make hay while the sun shines,” that is, takeadvantage of good investment opportunities (those with MPK ≥ r∗) without sacrificing current consump-tion.

Page: 36


4.4 Application: Delinking Saving from Investment

• The story of the Norwegian oil boom provides agood illustration of our theory about investmentand financial openness.

— A massive oil reserve was discovered in theNorth Sea in the 1960s, but it was very costlyto extract this oil at the time.

The world price of oil was low until the oil priceshocks in the 1970s. These shocks turned outto be largely permanent, leading to a perma-nent increase in the price of oil that then madethe extraction of North Sea oil profitable.

• This figure shows national savings (S), investment(I), and the current account (CA) for Norway — acountry that needed a large initial capital project toextract North Sea oil.

— During the mid- to late-1970s, Norway’s in-vestment increased sharply while national sav-ings decreased, and the current account de-creased by a large amount (going from a bal-ance to a deficit):

S − I = CA ⇒ S (↓)− I (↑) = CA (↓)

— Norway’s experience highlights that despitesmall changes in saving, it was able to financethe large capital projects through the currentaccount deficit (external financing). In otherwords, there was a delink between saving

and investment.

∗ In a closed economy, since CA = 0, S = I,implying the correlation between S and I isone.

• Is this delinking of saving and investment ageneral result?

— Feldstein and Horioka estimated what fractionβ of each additional dollar saved tended to beinvested in the same country: ∆I = β∆S .

∗ In a closed economy, the "saving reten-tion" measure β would equal 1, but theyargued that increasing financial opennesswould tend to push β below 1.

∗ They found that more financially opencountries have a much lower estimates of β,which indicates that these countries seemto have a greater ability to delink savingand investment.

Page: 37


4.5 Can Poor Countries Gain from Financial Globalization?

• Our analysis shows that if a country has investment projects for which the marginal product of capital exceedsthe world real interest rate (i.e., MPK ≥ r∗), then the country should borrow from the rest of the world tofinance investment in these projects.

— With this model prediction in mind, we now examine an enduring question in international macroeco-nomics: Why doesn’t more capital flow to poor countries?

4.5.1 Production Function Approach

• To look at what determines a country’s marginal product of capital (MPK), economists use a versionof a production function that maps available capital per worker, k = K/L, and the prevailing level ofproductivity, A, to the level of output per worker, q = Q/L, where Q is GDP.

— A simple and widely used production function takes the form:

q︸︷︷︸Output per worker

= A︸︷︷︸Productivity level

× kθ︸︷︷︸Capital per worker

where 0 < θ < 1 measures the contribution of capital to production, or is the elasticity of capital withrespect to output (i.e., a 1% increase in capital per worker generates a θ% increase in output per worker).

∗ In the data, θ is estimated to be 1/3. When we use θ = 1/3 and set the productivity level A equal toa reference level of 1, the production function is

q = k1/3 when A = 1 and θ = 1/3

Page: 38


— Then, MPK is the slope of the production function (i.e., the incremental change in output per worker∆q divided by the incremental change in capital per worker ∆k):

MPK = ∆q∆k

= θAkθ−1︸︷︷︸Slope of the production function

= θ × qk

∗ In this special case, MPK is proportional to output per worker divided by capital per worker (theaverage product of capital).

— Let’s see how MPK changes as k changes, given θ = 1/3 and A = 1:

from MPK = θq

kto MPK ′ = θ

2q

8k=

1

4

(θq

k

)=

1

4MPK

That is, if the country increases k by a factor of 8, then q increases by a factor of 2 because q = k1/3,and thus MPK changes by a factor of 1/4 (i.e., it falls to 1/4 its previous level).

∗ The more k, the smaller MPK (the diminishing marginal product of capital).

∗ Consider two countries, poor (P ; e.g., Mexico) and rich (R; e.g., the U.S.), assuming that theirproductions are the same (i.e., A and θ are the same for poor and rich countries).

· If the poor country’s level of output per worker is 1/2 (1/4) times the rich country’s level of outputper worker (qP is 1/2 (1/4) times qR), then the poor country’s level of capital per work is 1/8 (1/64)times the rich country’s level of capital worker (kP is 1/8 (1/64) times kR) and thus MPKP is 4(16) times MPKR.

qP =1

2qR︸︷︷︸ ⇒ kp =

1

8kR ⇒ MPKp = 4MPKR︸︷︷︸

Page: 39


4.5.2 A Benchmark Model: Countries Have Identical Productivity Levels

• Let’s consider the U.S. and Mexico in the late1980s, assuming they have an "identical" pro-ductivity level, A = 1.

— The U.S. represents rich countries while Mex-ico represents poor countries.

— Output per worker in Mexico was only 43% ofoutput per worker in the U.S. in the late 1980s.

• This figure shows the results of a comparison of theU.S. and Mexico.

— In the figure where (1) units are chosen suchthat all U.S. variables take value of 1; (2) Rrepresents the U.S. and B Mexico, we see that:

MPKUS = θ × qUSkUS

= θ ×(

11

)= θ × (1)

MPKMX = θ × qMX

kMX= θ ×

(0.430.08

)= θ × (5.4)

That is, MPKMX is 5.4 times MPKUS.

— The simple model says that the poorer thecountry, the higher its MPK because ofthe twin assumptions of diminishing marginalproduct of capital and a common productiv-ity level. Thus, investment ought to be moreprofitable in Mexico (poor countries), and soshould continue until Mexico is at Point R.Such a trajectory for Mexico is describe as con-vergence.

∗ In other words, poor countries have higherMPK than rich countries, so capital shouldflow from the rich countries into the poorcountries. As this happens, the poor coun-tries experience an increase in capital perworker, and therefore an increase in outputper work. As the output per worker risesin the poor countries, they converge to thelevels in rich countries.

Page: 40


4.5.3 An Augmented Model: Countries Have Different Productivity Levels

• The Lucas Paradox: Why Doesn’t Capital Flowfrom Rich to Poor Countries?

— The implication from the previous model isclear: capital should flow from rich to poorcountries, and poor countries should imple-ment government policies to attract capital in-flow. However, as Nobel laureate Robert Lucaspointed out that this is not the case (in fact,there is a great deal of investment in wealthiercountries), indicating the assumptions used inthe model are "drastically wrong:"

∗ The key incorrect assumption used wasthat the production functions are identicalacross countries.

• Now, we suppose that the productivity level A isdifferent in the U.S. and Mexico:

qUS = AUS × kθUS and qMX = AMX × kθMX ,

where AUS > AMX . The results for this case areshown in the figure.

— Mexico is now at C (not B) where MPKMX

is now 1.3 times MPKUS:

∗ MPKMX

MPKUS= θ×(qMK/kMK)

θ×(qUS/kUS)= (qMK/qUS)

(kMK/kUS)= 0.43

0.33

Now investment occurs only until MPKMX

falls to the U.S. level of 1 at D. Capital perworker kMX and output per worker qMX do notconverge to the levels seen in the U.S..

— In this augmented model, we found that Mex-ico did not have a high level of MPK relativeto the U.S. Thus, we would not expect largeflows of capital into Mexico but would expectit to remain relatively poor even with access toglobal financial markets.

Page: 41


• The measured MPK differentials in the augmented model do not seem to indicate a major failure of globalcapital markets to allocate capital effi ciently. But the augmented model has very different implications forconvergence.

— Mexico would borrow only enough to move from Point C to Point D, where its MPKMX equal r∗. Thisinvestment would raise its output per worker a little, but it would still be far that of the U.S..

Page: 42

5. GAINS FROM DIVERSIFICATION OF RISK

5 Gains from Diversification of Risk

• When discussing consumption smoothing previously, we saw how a small open economy that owns its ownoutput uses "borrowing and lending" to smooth the impact of output shocks on consumption.

— However, we also discussed the finding that in practice, countries seem to be unable to eliminate the effectsof output volatility. The problems seem to be especially diffi cult in emerging markets and developingcountries.

∗ We saw that reliance on borrowing and lending may create problems because there may be limits toborrowing, risk premiums, and sudden stops in the availability of credit.

• Are there other ways to cope with shocks to output? Yes. Diversification, another facet of financialglobalization, can help smooth shocks by promoting "risk sharing."

— With diversification, countries own not only the income stream from their own capital stock, but alsoincome streams from capital stocks located in other countries.

∗ In this section, we explore how, by using financial openness to trade such rights (e.g., in the form ofcapital equity claims like stocks and shares), countries may be able to reduce the volatility of theirincomes (and hence their consumption levels) without any net borrowing or lending whatsoever.

Page: 43


5.1 Diversification: A Numerical Example and Generalization

• To keep things simple and to permit us to focus on diversification, we first assume

1. There is no borrowing/lending.

2. The world consists of two countries, A and B, which are identical except that their outputs move inopposite directions. There are two possible states of the world each of which is equally likely to occur :in terms of output levels,

(a) State 1 is bad for country A and good for country B: GDP 1A = 90 and GDP 1

B = 110

(b) State 2 is good for country A and bad for country B. GDP 2A = 110 and GDP 2

B = 90

3. No investment or government spending, so output (i.e., GDP ) is equal to consumption: GDP = C

4. Output is allocated 60% to labor income and 40% to capital income in each country (As we know fromthe previous chapter, output is distributed to factors in the form of income).

— The key question here is who owns this income? Domestic residents or foreigners?

Page: 44


5.1.1 Home Portfolios

• We first consider the home portfolio case, assuming that the countries are closed, and each countryowns 100% of its own capital stock. Thus, in each country, income measured by GNI equals outputmeasured by GDP :

GNI = GDP

— The results of home portfolios are shown in the table below:

∗ (a) In State 1, A’s output is 90, of which 54 (= 90× 0.6) units are payments to labor and 36 (= 90× 0.4)units are payments to capital; (b) In State 2, A’s output rises to 110, and factor payments rise to66 (= 110× 0.6) for labor and 44 (= 110× 0.4) units for capital.

· The opposite is true in B: B’s output is higher in State 1 than it is in State 2.

∗ Thus, the variation of GNI about its mean of 100 is plus or minus 10 in each country.

· Because households prefer smooth consumption, this variation is undesirable (C = GNE byassumption, and thus C = GDP = GNI).

Page: 45


5.1.2 World Portfolios

• Two countries can achieve "partial" income smoothing if they diversify their portfolios of capitalassets.

— For example, each country owns one-half of its own capital stock but sells the other half to the othercountry in exchange for half of the other country’s capital stock. As a result, each country owns 50% ofthe domestic capital stock and 50% of the other country’s capital stock — that is, each country owns50% share of the world portfolio of capital.

∗ Indeed, this is what standard portfolio theory says that investors should try to do.

∗ Note that the countries are not able to own each other’s labor, so labor income is the sameas in the home portfolio case.

— The results of this portfolio diversification are shown in the table below:

∗ (1) Capital income for each country is smoothed at 40 units: in either State 1 or 2, [(90 + 110)× 0.4] /2 =40. (2) In each country, 54 (= 90× 0.6) and 66 (= 110× 0.6) = 66 units are payments for labor incomein low and high output states, respectively.

∗ Thus, the variation of GNI about its mean of 100 is now plus or minus 6 in each country.

· That is, the countries are able to reduce the variation in income (i.e., GNI) through owninga portion of the other country’s capital stock.

Page: 46


• How does the balance of payments work when countries hold the world portfolio? Let’s considerCountry A in State 1 which is bad for A and good for B (i.e., GDP 1

A = 90; GDP 1B = 110). Note that the

flows described below are reversed in State 2 which is good for A and bad for B.

1. In State 1, A’s income exceeds A’s output by 4:

GNI1A > GDP 1

A ⇒(GNI1

A −GDP 1A

)= (94− 90) = 4

2. This extra income is net foreign income from abroad of 4:

GNI1A = GDP 1

A +NFIA1A ⇒ NFIA1

A = 4

— This NFIA1A = 4 is the difference between the following two things:

(a) the income earned on A’s external assets (i.e., 50% of B’s payments to capital)

50%×(

40%×GDP 1B

)= 50%× (40%× 110) = 22

(b) the income paid on A’s external liabilities (i.e., 50% of A’s payments to capital)

50%×(

40%×GDP 1A

)= 50%× (40%× 90) = 18

3. What does Country A do with NFIA1A of 4? Country A runs a trade deficit (i.e., TB1

A = −4), whichmeans that A can consume 94 even when output is 90 (i.e., GNE1

A = 94 and GDP 1A = 90):—

TB1A = −4 ⇒ GDP 1

A (= 90) = GNE1A (= 94) + TB1

A (= −4)

— Thus, adding the trade balance to net foreign income from abroad gives the zero current account(i.e., CA1

A = 0), which means that there is still no need for any net borrowing or lending, asassumed:

CA1A = TB1

A +NFIA1A = −4 + 4 = 0

• The example shows that each country is able to reduce fluctuations inGNI (total income) through eliminatingvariance in capital income (i.e., diversifying portfolio of capital assets) without borrowing or lending.

Page: 47


5.1.3 Generalizing

• Let us try to generalize the concept of capital income smoothing through diversification.

— To generalize this, however, we need to recognize that not all output shocks are "asymmetric" as wehave assumed previously.

1. It has been assumed that the outputs of countries A and B are opposite, that is, output shocks arenegatively correlated (asymmetric) across countries.

∗ In this case, two countries can reduce the volatility of capital income through diversification.

∗ As seen before, holding the 50− 50 mix of the world capital income portfolio made capital incomefor each country smoothed at 40 units: reducing zero volatility of capital income.

2. If both countries suffer a negative (positive) shock at the same time (i.e., output shocks are positivelycorrelated (symmetric) across the countries), they will be unable to use diversification to reducevolatility in capital income.

∗ For example, in State 1, GDP 1A = 90 and GDP 1

B = 90; in State 2, GDP 2A = 110 and GDP 2

B = 110.

— In the real world, asymmetric and symmetric shocks are mixed.

Page: 48


• In this figure, each chart plots the volatility ofcapital income against the share of the port-folio devoted to foreign capital for each of threecases, assuming that the two countries are identicalin size and experience shocks of similar amplitude.

1. Asymmetric output shocks: perfect nega-tive correlation (-1):

— Shocks are perfectly asymmetric (correla-tion = −1), which means that capital in-come in the two countries is perfectly neg-atively correlated.

— Capital income risk can be eliminated byholding the 50 − 50 world portfolio (zerovolatility of capital income), and there arelarge gains from diversification.

2. Symmetric output shocks: perfect positivecorrelation (+1):

— Shocks are perfectly symmetric (correlation= +1), which means that capital incomein the two countries is perfectly positivelycorrelated.

— Capital income risk cannot be reduced, andthere are no gains from diversification.

3. A combination of asymmetric and symmet-ric shocks:

— When both types of shocks are present, thecorrelation is neither perfectly negative norpositive.

— Capital income risk can be partially elimi-nated by holding the world portfolio, andthere are still some gains from diversifica-tion.

Page: 49


5.1.4 Limits to Diversification: Capital versus Labor Income

• As we saw before, a country cannot achieve elimi-nation of "total" income risk by holding the worldportfolio, because labor income risk is not beingshared.

— It is not possible to trade ownership of laboracross countries (that is called “slavery” andis illegal in nearly all countries).

— However, it is worth noting that often shocksaffect capital and labor in the same way.Therefore, countries experiencing low cap-ital income likely experience low labor in-come at the same time.

∗ In fact, it is the case in our example: a good(bad) state raises (lowers) both capital andlabor income.

This means that, as a risk-sharing device,trading claims to capital income can sub-stitute for trading claims to labor income.

• This table considers a scenario in which countriesown none of their own capital stocks but rather100% of the foreign capital stock.

— We observe that countries can still reducevolatility in total income through diversifica-tion.

∗ Effectively, ownership of foreign capitalstock means that income earned on capi-tal is high when labor income is low, andvice versa.

∗ The variation of GNI about its mean of100 is now ±2 in each country, comparedwith the fluctuations of ±10 (home portfo-lio) and ±6 (the world portfolio).

Page: 50


5.2 Application: The Home Bias Puzzle

• In practice, we do not observe countries own-ing foreign-biased portfolios or even the worldportfolio.

— Countries tend to own portfolios that sufferfrom a strong home bias, a tendency of in-vestors to devote a disproportionate fractionof their wealth to assets from their own homecountry rather than assets from foreign as-sets, when a more globally diversified portfoliomight protect them better from risk.

• The figure illustrates this, showing the return (meanof monthly return) and risk (standard deviation ofmonthly return) for a hypothetical portfolio madeup from a mix of a pure home U.S. portfolio (theS&P 500) and a pure foreign portfolio (the MorganStanley EAFE) using data from the period 1970 to1996.

— U.S. investors with a 0% weight on the over-seas portfolio (point A) could have raised thatweight as high as 39% (point C) and still raisedthe return and lowered risk.

∗ Even moving to the right of point C (to-ward point D) would make sense, thoughhow far would depend on how the investorviewed the risk-return trade-off.

— However, U.S. investors actually investedroughly 8% of their wealth abroad (point B).

∗ This was considered the home bias puzzle.

Page: 51


• Broadly speaking, economists have had one of tworeactions to the emergence of the "home bias puz-zle" in the 1990s.

1. One is to propose many different theories toexplain it away:

(a) Perhaps it is costly to acquire foreign as-sets or get information about them.

(b) Perhaps there are asymmetries betweenhome and foreign countries’ consumptionpatterns (due to nontraded goods or tradefrictions or even tastes) that make domes-tic assets a better hedge against domesticconsumption risk.

(c) Perhaps home investors worry about reg-ulatory barriers and the problems of corpo-rate governance in foreign market.

However, none has been judged a completesuccess.

2. The other is to look at the evidence of homebias in the period from the 1970s to 1990s asa legacy of the pronounced deglobalizaion offinancial markets in the postwar period thatmight slowly disappear.

— Recent evidence suggests that this might behappening to some degree.

∗ There has been a dramatic increase in

overseas equity investments in a sampleof advanced countries from 1970 to 2003,with a very strong upward trend after1985 in IMF data.

∗ For example, in the U.S., the foreignshare of the U.S. portfolio has risen from5.6% in 1990 to 12.7% in 2003. Over thesame period, the U.K. portfolio saw itsforeign share rise from 33.1% to 52.4%.

∗ Furthermore, these figures might un-derstate the true diversification throughpurchasing foreign assets because largemultinational firms have capital incomestreams from many different countries.So when an American investor purchasesstock in one of these companies basedin the U.S., she is effectively buying therights to a capital income stream fromabroad.

— Recent trends in diversification within port-folio and within companies may mean thateven if the home bias puzzle cannot be ex-plained away, perhaps it will gradually goaway.

Page: 52


5.3 Summary: Don’t Put All Your Eggs in One Basket

• Our analysis highlights how diversification of income risk through the international trading of assets allows acountry to reduce volatility in income without borrowing or lending. This can be of great potential use tocountries that face high risk premiums, borrowing limits, and limited access to world capital markets.

1. In theory, if countries were able to pool their income streams and take shares from common pool of income,then all country-specific shocks would be averaged out, leaving countries exposed only to common globalshocks, the sole remaining undiversifiable shocks that cannot be avoided.

— The key lesson for countries, like households, is: don’t put all your eggs in one basket — throughownership of capital income streams from several different sources, the likelihood of suffering dramaticshocks to income is reduced.

2. In practice, however, risk sharing through asset trade is limited.

— For one thing, the number of assets is limited.

∗ The market for claims to capital income is incomplete because not all capital assets are traded (e.g.,many firms are privately held and not listed on stock market), and trade in labor assets is legallyprohibited.

— Moreover, even with the traded assets available, investors have shown very little inclination to investtheir wealth outside their own country, although that may be slowly changing in an environment ofongoing financial globalization.

Page: 53

3.4 Consumption Volatility and Financial Openness Does the ...contents.kocw.net/KOCW/document/2015/hanyang/namdeokwoo/11.… · If more consumption smoothing is achieved, the computed

Documents