The Saving–Investment Nexus
Is there a market mechanism that allows people to actually make the tradeoffs
discussed in the previous section? This is a critical question – one that
lies at the heart of macroeconomic debate and one whose answers separate
the different schools of thought. The question can be posed in a way that
highlights the macroeconomic concerns: is there a market mechanism that
brings saving and investment in line with one another without at the same
time having perverse effects (for example widespread resource idleness) on
the macroeconomy? The alternative answers have clear implications for the
viability of market economies and for the proper role of the policy maker.
9.4.1 A detour through monetarism
Some macroeconomists would answer the critical question in the affirmative,
taking the market’s allocation of resources to the production of consumption
goods and the production of investment goods, the latter financed by saving,
to be on a par with the market’s allocation of resources to the production of
fruits and the production of vegetables. In other words, within the overall
output aggregate, the allocation issue – whether among narrowly defined
goods (peaches and potatoes) or among broad-based sub-aggregates (consumption
and investment) – is largely the province of microeconomics.
Macroeconomics, in this view, should focus on the overall output aggregate
itself as it relates to other macroeconomic variables, such as the general
price level and the money supply. These macroeconomic variables, symbolized
as Q, P and M, come together in the familiar equation of exchange:
MV = PQ (9.1)
This equation, of course, was ground zero for the monetarist counter-revolution
against the Keynesianism of the 1950s (see Chapter 4). The velocity of
money, V, is defined by the equation itself, and before the early 1980s its
empirically demonstrated near-constancy in different countries and in different
time periods established a strong relationship between the money supply
and some index of output prices. What is commonly known as the quantity
theory of money is more descriptively called the quantity-of-money theory of
the price level.
The monetarists argued that the long-run consequence of a change in the
money supply is an equiproportional change in the general level of prices – a
consequence tempered only by ongoing secular changes in real output and in
the velocity of money. Allowances were made for short-run variations in real
output. That is, overall output Q may rise and then fall while P is adjusting to
an increased M. However, the monetarists paid little attention to the relative
movements of the major sub-aggregates (consumption and investment) during
the adjustment process and no attention at all to the sub-aggregates
(stages of production) that make up aggregate investment. Whether dealing
with long-run secular growth or with short-run money-induced movements in
484 Modern macroeconomics
real output, the focus was on the summary output variable Q. Whatever
change is occurring within the output aggregate – as might be tracked by the
Austrians in terms of the Hayekian triangle – was taken to be irrelevant to the
greater issues of macroeconomics.
The saving–investment perversity of Keynesianism
It was Keynesian economics, of course, that the monetarist counter-revolution
was intended to counter. But on the issue of a saving–investment nexus,
the counter could be more accurately described as a cover-up. In his General
Theory Keynes (1936, p. 21) had explicitly faulted his predecessors and
contemporaries for ‘fallaciously supposing that there is a nexus which unites
decisions to abstain from present consumption with decisions to provide for
future consumption’. According to Keynes, there is no simple and effective
way of coordinating these two decisions. Rather, the mechanisms that do
eventually bring saving into line with investment are indirect and perverse.
The saving–investment perversity, in fact, is central to the Keynesian vision
of the macroeconomy (see Leijonhufvud, 1968).
The equation of exchange can be rewritten in a way that uncovers the
issues on which the Keynesian revolution was based. Aggregate output Q
consists of the output of consumption goods plus the output of investment
goods. That is, Q = QC + QI, the QI reckoned as the ‘final’ output of investment
goods – so as to avoid double counting. The equation of exchange, then,
can be rewritten as:
MV = P(QC + QI) (9.2)
emphasizing that the problem as seen by Keynes (the volatility of QI and its
impact on all other macroeconomic magnitudes) is a problem that is simply
not addressed by the monetarists. Rather, replacing the Keynesian QC + QI
with the monetarist Q served only to cover up the primary locus of perversity.
The question of just how the output of investment goods gets squared with
preferred trade-off between current consumption and future consumption is
not answered by the monetarists – nor is it even asked.
In the Keynesian vision, which will be dealt with at some length in section
9.11, movements in the investment aggregate impinge in the first instance on
incomes, which in turn impinge on consumption spending. That is, QC and QI
move in the same direction, the movements in QI being unpredictable and the
corresponding same-direction movements in QC being amplified by the familiar
Keynesian multiplier (see Chapter 2). Similarly, autonomous changes in
current consumption, if any, would tend to affect profit expectations and hence
cause investment spending to change in the same direction. Here, the principle
of derived demand is in play. With the two major sub-aggregates moving up
and down together (though at different rates), the Keynesian theory precludes
by construction any possibility of there being a trade-off of the sort emphasized
by the Austrians. Further, considerations of durable capital and the so-called
investment accelerator imply the absence of a generally binding supply-side
constraint. There is simply no scope in the Keynesian vision for investment to
rise at the expense of current consumption. Similarly, market participants willing
to forgo current consumption (that is, to save) in order to be able to enjoy
greater future consumption would find their efforts foiled by the market mechanisms
that link saving and investment. Rather than stimulating investment,
increased saving would impinge on overall spending and hence on overall
income. This perverse negative income effect, which Keynes identified as the
paradox of thrift, is discussed at length in section 9.9.
Austrian disaggregation
The Austrian perspective on Keynesianism and monetarism in the context of
the equation of exchange is revealing. Keynesianism adopts a level of aggregation
that suggests a potential problem – one of dividing resources
appropriately between consumption and investment – but without allowing
for a non-perverse market solution to that problem. Monetarism, as well as
most strands of new classicism, increases the level of aggregation, obscuring
this central issue and hence relegating the problem as well as its solution to
the realm of microeconomics. Pre-dating both monetarism and Keynesianism,
the Austrians were inclined to work at a lower level of aggregation than either
of these schools, one in which both the problem and a potentially viable
market solution could be identified.
Again, the equation of exchange can serve as the common denominator of
the different schools of thought. For the Austrians, the investment aggregate
in the Keynesian rendition must be disaggregated so as to bring the stages of
production into play. QC is consumable output, or goods of the first order – to
use Menger’s terminology. Investments distributed across the nine preceding
stages are identified as Q2 through Q10. The equation of exchange thus becomes:
MV = P(QC + Q2 + Q3 + Q4 + Q5 + Q6 + Q7 + Q8 + Q9 + Q10) (9.3)
Just as QI is reckoned as ‘final’ output in conventional macroeconomic theorizing,
the second- and higher-order goods (Q2 through Q10) in equation (9.3)
are similarly reckoned so as to maintain the integrity of the equation of
exchange. Double counting is thus avoided, and the sum of the output
magnitudes (in equations 9.2 and 9.3) is equal to total output and, equivalently,
to total income. But with the Austrian disaggregation, the focus of the
analysis is on the relative movements among the Qs as well as on their sum.
In the Keynesian construction, it might well seem implausible that an
increase in saving and a corresponding decrease in spending on QC could
cause QI to increase. If business firms are having problems selling out of their
current inventories, they are unlikely to be inspired to commit additional
resources to an expanded capacity and hence to the further overstocking of
these inventories. The doctrine of derived demand suggests that the demand
for productive capacity will mirror the demand for output. In the Austrian
construction, the doctrine of derived demand is tempered by considerations
of time discount. The multiple stages of production allow for enough degrees
of freedom for the consequences of a fall in consumer spending to be described
in terms of a change in the pattern of investment spending rather than
exclusively in terms of an opposing movement in an all-inclusive investment
aggregate. The story of how the market can plausibly work can be squared
with the doctrine of derived demand, but as told by the Austrians, the story is
not dominated by it. The analysis draws on microeconomics as well as
macroeconomics and, as indicated earlier, the main character is the entrepreneur.
Derived demand and time discount
An increase in saving sends two market signals to the business community.
Both must come into play if a change in intertemporal preferences is to get
translated successfully into corresponding changes in the economy’s multistage
production process. Changes in output prices together with changes in
the interest rate have consequences that affect the various stages of production
differentially. A non-perverse reallocation of resources in the face of
increased saving hinges critically on two principles: the principle of derived
demand and the principle of time discount. It is worth noting here that
perceived perversities in the saving–investment nexus of market economies
stem from an implicit denial of the second-mentioned principle. If derived
demand is taken to be the only principle in play, then it follows almost
trivially that the market cannot adapt to an increase in saving.
Increased saving means reduced current demand for consumer goods. (Of
course, for a growing economy in which both saving and consumption are
increasing, we would have to think in terms of changes in the relative rates of
increase. More rapidly increasing saving means a less rapidly increasing
demand for consumer goods.) A decrease in the demand for goods of the first
order – again, Menger’s terminology – has straightforward implications for
the demand for goods of the second order. The demand for coffee beans
moves with the demand for coffee. Menger’s Law prevails. More generally,
the demand for inputs that are in close temporal proximity to the consumable
output moves with the demand for that output. The demand for goods of the
second order is a derived demand. Under strict ceteris paribus conditions,
which would entail no change in the rate of interest, derived demand would
be the whole story.
The more favourable credit conditions brought about by the increase in
saving is the basis for the rest of the story. A lower interest rate allows
businesses to carry inventories more cheaply. But how important is this
change in supply conditions? In gauging the relative changes in the demands
for goods of the first order and goods of the second order (coffee and coffee
beans), the time-discount effect is weak. Inventories of coffee beans are held
for only a short period of time, and consequently, the time-discount effect –
in this case, the reduced costs of carrying inventory – is trivial compared to
the derived-demand effect. The demand for coffee beans falls almost as much
as the demand for coffee. The strength of the time-discount effect is greater –
and increasingly greater – for the higher orders of goods. Consider, say, a
tenth-order good in the form of durable capital equipment. Testing facilities
and laboratory fixtures devoted to product development are good examples.
More favourable credit conditions could easily tip the scales toward creating
or expanding such a facility. In early stages of production, the time-discount
effect can more than offset the derived-demand effect.
Considerations of time discount draw resources into early stages of production.
Further, in gauging the profitability of early-stage activities, the
derived-demand effect itself can be augmenting rather than offsetting. Here,
the entrepreneurial element comes into play in a special way. What counts as
the relevant derived demand is not based on the current demand for goods of
the first order but rather on the anticipated demand at some future point in
time – a demand that may well be strengthened precisely because of the
accumulation of savings. The increased saving need not be taken as an
indication that the demand for consumables is permanently reduced. Rather,
savers are saving up for something. And entrepreneurs who best anticipate
just what they will be inclined to buy with their increased buying power stand
to profit from the intertemporal shift in spending.
The interplay between derived demand and time discount accounts for the
change in the pattern of resource allocation brought about by an increase in
saving. A judgement might be made that this account saddles the entrepreneurs
with a greater burden than they can bear. Yet those same entrepreneurial
skills were already in play in maintaining the intertemporal capital structure
before the increase in saving. That is, even in the absence of a change in
intertemporal preferences, market conditions throughout the economy are
continuously changing in every other respect – changes in tastes, in technology,
in resource availabilities. Entrepreneurs must continuously adapt to those
changes, while maintaining the temporal progression from early-stage to latestage
activities. An increase in saving simply requires that they make use of
those same skills – but under marginally changed credit conditions. A more
disaggregated into nine stages of production. The arrows indicate the direction
and relative magnitude of the change in the output quantities brought
about by an increase in saving. The reduction in the output of first-order
goods (QC) is echoed in the reduction in the output of second-through-fifthorder
goods (Q2, Q3, Q4 and Q5), the magnitude of the reduction attenuated
by the time-discount effect for the increasingly higher orders of goods. Starting
(in this illustration) with sixth-order goods, the time-discount more than
offsets the derived-demand effect. There are increases in the output levels of
the sixth and earlier stages of production (Q6, Q7, Q8, Q9 and Q10), the timediscount
effect becoming more dominant with increasingly higher-order goods.
The increased saving frees up resources, which are then allocated to the
different stages of production in a pattern governed by the more favourable
credit conditions. Grouping Q2 through Q10 together in Figure 9.4, we see that
overall investment rises as the current demand for consumable output (QC)
falls. Contrary to Keynes’s paradox of thrift, consumption and investment can
move in opposite directions. Attention to the intertemporal pattern of investment
allows us to resolve the paradox and to show how changes in investment
can be consistent with changes in saving behaviour. The wholesale neglect of
the pattern of investment underlay an early judgement by Hayek (1931) that
‘Mr. Keynes’s aggregates conceal the most fundamental mechanisms of
change’. It is significant that those fundamental mechanisms are set into
motion by the supply and demand for loanable funds – because it was
loanable funds theory, a staple in the pre-Keynesians’ toolkit, that Keynes
specifically jettisoned.
Is there a market mechanism that allows people to actually make the tradeoffs
discussed in the previous section? This is a critical question – one that
lies at the heart of macroeconomic debate and one whose answers separate
the different schools of thought. The question can be posed in a way that
highlights the macroeconomic concerns: is there a market mechanism that
brings saving and investment in line with one another without at the same
time having perverse effects (for example widespread resource idleness) on
the macroeconomy? The alternative answers have clear implications for the
viability of market economies and for the proper role of the policy maker.
9.4.1 A detour through monetarism
Some macroeconomists would answer the critical question in the affirmative,
taking the market’s allocation of resources to the production of consumption
goods and the production of investment goods, the latter financed by saving,
to be on a par with the market’s allocation of resources to the production of
fruits and the production of vegetables. In other words, within the overall
output aggregate, the allocation issue – whether among narrowly defined
goods (peaches and potatoes) or among broad-based sub-aggregates (consumption
and investment) – is largely the province of microeconomics.
Macroeconomics, in this view, should focus on the overall output aggregate
itself as it relates to other macroeconomic variables, such as the general
price level and the money supply. These macroeconomic variables, symbolized
as Q, P and M, come together in the familiar equation of exchange:
MV = PQ (9.1)
This equation, of course, was ground zero for the monetarist counter-revolution
against the Keynesianism of the 1950s (see Chapter 4). The velocity of
money, V, is defined by the equation itself, and before the early 1980s its
empirically demonstrated near-constancy in different countries and in different
time periods established a strong relationship between the money supply
and some index of output prices. What is commonly known as the quantity
theory of money is more descriptively called the quantity-of-money theory of
the price level.
The monetarists argued that the long-run consequence of a change in the
money supply is an equiproportional change in the general level of prices – a
consequence tempered only by ongoing secular changes in real output and in
the velocity of money. Allowances were made for short-run variations in real
output. That is, overall output Q may rise and then fall while P is adjusting to
an increased M. However, the monetarists paid little attention to the relative
movements of the major sub-aggregates (consumption and investment) during
the adjustment process and no attention at all to the sub-aggregates
(stages of production) that make up aggregate investment. Whether dealing
with long-run secular growth or with short-run money-induced movements in
484 Modern macroeconomics
real output, the focus was on the summary output variable Q. Whatever
change is occurring within the output aggregate – as might be tracked by the
Austrians in terms of the Hayekian triangle – was taken to be irrelevant to the
greater issues of macroeconomics.
The saving–investment perversity of Keynesianism
It was Keynesian economics, of course, that the monetarist counter-revolution
was intended to counter. But on the issue of a saving–investment nexus,
the counter could be more accurately described as a cover-up. In his General
Theory Keynes (1936, p. 21) had explicitly faulted his predecessors and
contemporaries for ‘fallaciously supposing that there is a nexus which unites
decisions to abstain from present consumption with decisions to provide for
future consumption’. According to Keynes, there is no simple and effective
way of coordinating these two decisions. Rather, the mechanisms that do
eventually bring saving into line with investment are indirect and perverse.
The saving–investment perversity, in fact, is central to the Keynesian vision
of the macroeconomy (see Leijonhufvud, 1968).
The equation of exchange can be rewritten in a way that uncovers the
issues on which the Keynesian revolution was based. Aggregate output Q
consists of the output of consumption goods plus the output of investment
goods. That is, Q = QC + QI, the QI reckoned as the ‘final’ output of investment
goods – so as to avoid double counting. The equation of exchange, then,
can be rewritten as:
MV = P(QC + QI) (9.2)
emphasizing that the problem as seen by Keynes (the volatility of QI and its
impact on all other macroeconomic magnitudes) is a problem that is simply
not addressed by the monetarists. Rather, replacing the Keynesian QC + QI
with the monetarist Q served only to cover up the primary locus of perversity.
The question of just how the output of investment goods gets squared with
preferred trade-off between current consumption and future consumption is
not answered by the monetarists – nor is it even asked.
In the Keynesian vision, which will be dealt with at some length in section
9.11, movements in the investment aggregate impinge in the first instance on
incomes, which in turn impinge on consumption spending. That is, QC and QI
move in the same direction, the movements in QI being unpredictable and the
corresponding same-direction movements in QC being amplified by the familiar
Keynesian multiplier (see Chapter 2). Similarly, autonomous changes in
current consumption, if any, would tend to affect profit expectations and hence
cause investment spending to change in the same direction. Here, the principle
of derived demand is in play. With the two major sub-aggregates moving up
and down together (though at different rates), the Keynesian theory precludes
by construction any possibility of there being a trade-off of the sort emphasized
by the Austrians. Further, considerations of durable capital and the so-called
investment accelerator imply the absence of a generally binding supply-side
constraint. There is simply no scope in the Keynesian vision for investment to
rise at the expense of current consumption. Similarly, market participants willing
to forgo current consumption (that is, to save) in order to be able to enjoy
greater future consumption would find their efforts foiled by the market mechanisms
that link saving and investment. Rather than stimulating investment,
increased saving would impinge on overall spending and hence on overall
income. This perverse negative income effect, which Keynes identified as the
paradox of thrift, is discussed at length in section 9.9.
Austrian disaggregation
The Austrian perspective on Keynesianism and monetarism in the context of
the equation of exchange is revealing. Keynesianism adopts a level of aggregation
that suggests a potential problem – one of dividing resources
appropriately between consumption and investment – but without allowing
for a non-perverse market solution to that problem. Monetarism, as well as
most strands of new classicism, increases the level of aggregation, obscuring
this central issue and hence relegating the problem as well as its solution to
the realm of microeconomics. Pre-dating both monetarism and Keynesianism,
the Austrians were inclined to work at a lower level of aggregation than either
of these schools, one in which both the problem and a potentially viable
market solution could be identified.
Again, the equation of exchange can serve as the common denominator of
the different schools of thought. For the Austrians, the investment aggregate
in the Keynesian rendition must be disaggregated so as to bring the stages of
production into play. QC is consumable output, or goods of the first order – to
use Menger’s terminology. Investments distributed across the nine preceding
stages are identified as Q2 through Q10. The equation of exchange thus becomes:
MV = P(QC + Q2 + Q3 + Q4 + Q5 + Q6 + Q7 + Q8 + Q9 + Q10) (9.3)
Just as QI is reckoned as ‘final’ output in conventional macroeconomic theorizing,
the second- and higher-order goods (Q2 through Q10) in equation (9.3)
are similarly reckoned so as to maintain the integrity of the equation of
exchange. Double counting is thus avoided, and the sum of the output
magnitudes (in equations 9.2 and 9.3) is equal to total output and, equivalently,
to total income. But with the Austrian disaggregation, the focus of the
analysis is on the relative movements among the Qs as well as on their sum.
In the Keynesian construction, it might well seem implausible that an
increase in saving and a corresponding decrease in spending on QC could
cause QI to increase. If business firms are having problems selling out of their
current inventories, they are unlikely to be inspired to commit additional
resources to an expanded capacity and hence to the further overstocking of
these inventories. The doctrine of derived demand suggests that the demand
for productive capacity will mirror the demand for output. In the Austrian
construction, the doctrine of derived demand is tempered by considerations
of time discount. The multiple stages of production allow for enough degrees
of freedom for the consequences of a fall in consumer spending to be described
in terms of a change in the pattern of investment spending rather than
exclusively in terms of an opposing movement in an all-inclusive investment
aggregate. The story of how the market can plausibly work can be squared
with the doctrine of derived demand, but as told by the Austrians, the story is
not dominated by it. The analysis draws on microeconomics as well as
macroeconomics and, as indicated earlier, the main character is the entrepreneur.
Derived demand and time discount
An increase in saving sends two market signals to the business community.
Both must come into play if a change in intertemporal preferences is to get
translated successfully into corresponding changes in the economy’s multistage
production process. Changes in output prices together with changes in
the interest rate have consequences that affect the various stages of production
differentially. A non-perverse reallocation of resources in the face of
increased saving hinges critically on two principles: the principle of derived
demand and the principle of time discount. It is worth noting here that
perceived perversities in the saving–investment nexus of market economies
stem from an implicit denial of the second-mentioned principle. If derived
demand is taken to be the only principle in play, then it follows almost
trivially that the market cannot adapt to an increase in saving.
Increased saving means reduced current demand for consumer goods. (Of
course, for a growing economy in which both saving and consumption are
increasing, we would have to think in terms of changes in the relative rates of
increase. More rapidly increasing saving means a less rapidly increasing
demand for consumer goods.) A decrease in the demand for goods of the first
order – again, Menger’s terminology – has straightforward implications for
the demand for goods of the second order. The demand for coffee beans
moves with the demand for coffee. Menger’s Law prevails. More generally,
the demand for inputs that are in close temporal proximity to the consumable
output moves with the demand for that output. The demand for goods of the
second order is a derived demand. Under strict ceteris paribus conditions,
which would entail no change in the rate of interest, derived demand would
be the whole story.
The more favourable credit conditions brought about by the increase in
saving is the basis for the rest of the story. A lower interest rate allows
businesses to carry inventories more cheaply. But how important is this
change in supply conditions? In gauging the relative changes in the demands
for goods of the first order and goods of the second order (coffee and coffee
beans), the time-discount effect is weak. Inventories of coffee beans are held
for only a short period of time, and consequently, the time-discount effect –
in this case, the reduced costs of carrying inventory – is trivial compared to
the derived-demand effect. The demand for coffee beans falls almost as much
as the demand for coffee. The strength of the time-discount effect is greater –
and increasingly greater – for the higher orders of goods. Consider, say, a
tenth-order good in the form of durable capital equipment. Testing facilities
and laboratory fixtures devoted to product development are good examples.
More favourable credit conditions could easily tip the scales toward creating
or expanding such a facility. In early stages of production, the time-discount
effect can more than offset the derived-demand effect.
Considerations of time discount draw resources into early stages of production.
Further, in gauging the profitability of early-stage activities, the
derived-demand effect itself can be augmenting rather than offsetting. Here,
the entrepreneurial element comes into play in a special way. What counts as
the relevant derived demand is not based on the current demand for goods of
the first order but rather on the anticipated demand at some future point in
time – a demand that may well be strengthened precisely because of the
accumulation of savings. The increased saving need not be taken as an
indication that the demand for consumables is permanently reduced. Rather,
savers are saving up for something. And entrepreneurs who best anticipate
just what they will be inclined to buy with their increased buying power stand
to profit from the intertemporal shift in spending.
The interplay between derived demand and time discount accounts for the
change in the pattern of resource allocation brought about by an increase in
saving. A judgement might be made that this account saddles the entrepreneurs
with a greater burden than they can bear. Yet those same entrepreneurial
skills were already in play in maintaining the intertemporal capital structure
before the increase in saving. That is, even in the absence of a change in
intertemporal preferences, market conditions throughout the economy are
continuously changing in every other respect – changes in tastes, in technology,
in resource availabilities. Entrepreneurs must continuously adapt to those
changes, while maintaining the temporal progression from early-stage to latestage
activities. An increase in saving simply requires that they make use of
those same skills – but under marginally changed credit conditions. A more
disaggregated into nine stages of production. The arrows indicate the direction
and relative magnitude of the change in the output quantities brought
about by an increase in saving. The reduction in the output of first-order
goods (QC) is echoed in the reduction in the output of second-through-fifthorder
goods (Q2, Q3, Q4 and Q5), the magnitude of the reduction attenuated
by the time-discount effect for the increasingly higher orders of goods. Starting
(in this illustration) with sixth-order goods, the time-discount more than
offsets the derived-demand effect. There are increases in the output levels of
the sixth and earlier stages of production (Q6, Q7, Q8, Q9 and Q10), the timediscount
effect becoming more dominant with increasingly higher-order goods.
The increased saving frees up resources, which are then allocated to the
different stages of production in a pattern governed by the more favourable
credit conditions. Grouping Q2 through Q10 together in Figure 9.4, we see that
overall investment rises as the current demand for consumable output (QC)
falls. Contrary to Keynes’s paradox of thrift, consumption and investment can
move in opposite directions. Attention to the intertemporal pattern of investment
allows us to resolve the paradox and to show how changes in investment
can be consistent with changes in saving behaviour. The wholesale neglect of
the pattern of investment underlay an early judgement by Hayek (1931) that
‘Mr. Keynes’s aggregates conceal the most fundamental mechanisms of
change’. It is significant that those fundamental mechanisms are set into
motion by the supply and demand for loanable funds – because it was
loanable funds theory, a staple in the pre-Keynesians’ toolkit, that Keynes
specifically jettisoned.
No comments:
Post a Comment