Ageing and Global Capital Flows - Is it Optimal to Dissave?

A preliminary apology is in order. What follows is uber wonkish and should be consumed preferably in small quantities. In fact, I am not sure that I have gotten everything right yet. Of course, we never are but in this case it is an important point to make up front. The argument is loosely built on my upcoming master's thesis which seeks to explore the connection between ageing and capital flows and specifically how and whether the former may lead to a state of export dependency; where export dependency is defined as a high and increasing sensitivity of the rate of change of national income to the rate of change of the current account. This is something I have discussed extensively on AS, but in this case I am trying to give it a thorough theoretical spin which means that any non econ-wonks are likely to be lost in translation.

I should stress immediately that this is not a virtue but rather a vice, but I do think that exploring  conventional economic theory (and moving beyond?) is an important part of the process. In the end, the argument should be amendable to plain English and, as it were, plain common sense and intuition as well as, of course, empirical falsification. One thing which I can promise though is that there will be no mathematical models; at least there, I should have provided some comfort. If you want the math, you will have to wait for the thesis.

The best way to frame the following argument is perhaps to insert it in the chronology of the thesis where it appears, in the end, as a perspectivation on aggregate global capital flows. What precedes this is thus an, hopefully convincing, account of the fact that Germany and Japan are effectively dependent on exports to grow as a result of their demographic profiles.


Some Thoughts on Export Dependency


First it would be worthwhile taking a look at the following sketch (click to enlarge) which captures a lot of the issues that will be discussed below. Essentially, it attempts to show, as nastily and brutishly short as possible, what export dependency means in the context of what we could call conventional economic theory.



For non-econ wonks it is likely to be quite difficult to read, but in fact; it is fairly simple. The X-axis  represents the age of an economy here exemplified in the phases of the age transition derived from a study by Malmberg and Sommestad that discusses the demographic transition as a transition in age structure (and not population growth). But really, it could just as well be median age where a median age of 40-50 would the limit to the right and, let us say, about 20-25 to the left. The Y-axis is adopted directly from Higgins (1998,  e.g. table 1, p. 350) who uses empirical estimations to model the effect from age on, in this case, the current account. As I, Higgins also use age on the x-axis (age distributions) and on the y-axis he has age coefficients. This basically means that a negative value signifies that the age structure in the economy influences the current account negatively (i.e. pushing the CA towards a deficit) while the reverse is true for a positive value. Essentially, we need not, initially, bother with the numerical value of such age coefficients here, but merely note whether it is positive or negative and then secondarily we may look at the level effect (i.e. whether the effect is numerically large).

I should immediately point out that there are no direct empirical basis for the curves. The line which is labeled trajectory of export dependency is basically a cubed function designed to show a more less linear association between export dependency and ageing with the added and important feature (hence the cubed functional form) that export dependency tends to increase exponentially when you move into very old age. The blue line naturally do contain some empirical foundation in that it originates from an empirical study Higgins (1998) where it is constructed based on empirical estimations. I shall not go into Higgin's empirical framework here since it would take us into the dark world of time series econometrics but merely point out that Higgins manages to come up with what we could call the textbook representation of the effect of ageing on the current account and it is worth pointing out that he is not the only one. Also Supan et al. (2007), Bryant (2006), Henriksen (2002) and Summers et al. (1990) (among a myriad of other studies) postulate either through theoretical elaborations or empirical estimations a relationship which may be approximated by the chart above.

In theoretical terms, the basis for this "hump-shaped" relationship between ageing and the current account is derived in the context of the simple, yet crucial, intuition derived from Modigliani's life cycle hypothesis which states that consumers spend their working age years saving for retirement where they will dissave those accumulated assets. Then, at some point during working age there is a "peak" which is characterised by the time when the saving rate is highest and thus also, indirectly, where the effect on the current account should be largest. Following convention, this is modeled in economics through the idea of overlapping generations and often in the form of a neo-classical growth theory framework or simply a general equlibrium representative agent framework. I shall not open pandora's box and discuss the merit of these methods here but merely point out that I think it is very difficult to argue against the the basic intution which lies behind these models and thus, as it were, the intuition from the life cycle hypothesis.

In essence of course, the real issue is one of calibration and thus one of empirical analysis to see just how this postulated hump may materialise as well as of course realizing that ageing is not the only variable which influences the current account. It is also here that the fun begins and where things very quickly become very complicated. 

In this respect, it is worthwhile focusing the attention on the so-called dissaving phase which should be a natural result of the move towards an ever higher share of the elderly in the population. The basic mechanism here is simply that in standard economic models "old" economic agents will dissave their entire asset and thus as the old cohorts increasingly will outnumber the young cohorts the dissaving of the former will trumph the saving of the latter and lead to dissaving on an aggregate level. All sorts of ill prophecies have been proposed in the context of this dissaving hypothesis, not least that we are facing an asset meltdown scenario in 2050 because there will be far too many elderly wanting to offload their assets to a much smaller base of younger cohorts who cannot support a satisfactory price (yield) level.

Now, the problem here is that empirical studies have shown that the idea of dissaving, while intuitively strong, is difficult to verify to the extent that theoretical models suggest. This is not difficult to imagine I think. By very nature of the uncertainty of the mortality schedule people do not (cannot) dissave to 0 and beyond this there are may be bequest motives. In an open economy context this further creates the rather dubious situation in which economies well into their old age will have to run persistent external deficits because, presumably, savings will have decline far faster than domestic investment demands. I say dubious here because this is exactly where I have chosen to take my stab at trying to amend the theoretical framework.

Consequently, I am not so sure that this is a plausible end point in the context of continuing population ageing. Specifically, I would like to ask the simple question of whether it is actually optimal for any society to dissave as the theory postulates. I don't think it is and while it is certainly not unlikely that economies may actually dissave (defacto) they will still be dependent on exports to grow and thus the difference between the two curves into the latter age transitions represents an externality. This is also why I think that economies, in stead of responding with dissaving, will fight the point at which they reach this stage since when they do it is effectively game over. Imagine for example how the likes of Germany and Japan would ever be able to finance an external deficit brought about solely on the basis of the fact that savings has declined so fast as to not even be able to meet domestic investment demand which in itself will be declining. In this situation, wouldn't it be much smarter to maintain savings persistently higher than domestic investment demand which can of course only be materialized in an external surplus. I think it will and it is this point which you need to keep in mind as we move forward.


Implications for Global Capital Flows

With these considerations in mind, the key question then becomes; what happens when more and more economies grow to become increasingly like Germany and Japan? (As we know they will, at least in the context of the OECD). Naturally, not everyone can maintain excess exports over imports at the same time so something, as they say, has got to give and it is this something, as it were, which is the topic of this entry.

The focus on the implication of ageing on aggregate global capital flows is not new and is, in fact, an integral part of the analysis in Supan et al. (2007), Higgins (1998) and Bryant (2006) which were also mentioned above. However, in the following we are going to relax the condition of dissaving normally assumed in e.g. OLG models and accept that the propensity to run an external surplus will increase as an economy ages.

If we do this, it should not require too much imagination to see the issues that may rise. For starters, I want to reiterate yet again the trivial fact that not all economies can run an external surplus at one and the same time. This means, quite naturally, that what might be optimal from the point of view of a single economy (i.e. maintaining a surplus as it ages) may not be viable or optimal from the point of view of the global economy.

In order to frame the discussion it is natural to take our point of departure in the discourse on global macroeconomic imbalances.

As so many other things, the financial crisis has completely dislocated this system but it still worthwhile to ponder the nature of the global financial system in a post Asian crisis perspective and up until now. Consequently, the global macroeconomic landscape has long been characterized by what many has termed Bretton Woods II in which a large batch of especially Asian and oil exporting economies have been pegging their currencies to the US dollar who in turn have been running a large current account deficit to match the savings surplus in emerging markets such as China, South Korea, the Petroexporters, Brazil and Russia. In fact, if we cut a lateral line through this argument we could say that the world has hitherto been characterized by the Anglo-Saxon economies running external deficits to match surpluses in big emerging markets as well as Japan and Germany.[1] That however changed abruptly with the advent of the financial crisis and it is interesting to note the initial response by market participants and many scholars in their interpretation. Consequently, as it became clear that the US economy had been mortally wounded on the back of the subprime mortgage debacle the US Fed slashed nominal interest rates significantly. As a result the USD plummeted which led many commentators to hail the US economy’s fall from grace and specifically coined the notion of decoupling in which the Eurozone economy and Japan were pinned as the ones taking up the slack in steering forward global demand. Initial versions of the decoupling thesis thus centered on the shift in emerging market exports from the US to Japan and, especially, the Eurozone and thus in the process also a shift from the US dollar to the Euro as a global reserve currency. As it turned out this was nothing but a mirage masked by the fact that US policy makers essentially acted preemptively to a crisis which turned global during the summer 2007 and now most major central banks in the OECD have slowly bitten the bullet and followed Bernanke into quantitative easing to combat the risk of deflation which would be devastating in the context of the debt overhangs some economies face. Moreover, and as a general point, the global economy already decoupled from the US, and indeed OECD, economy a long time ago. Consequently, it is an irrefutable fact that the global economy is undergoing a fundamental change in which emerging economies such as India, Turkey, Brazil, China; Chile etc will ascend to account for an ever larger share of global GDP and growth. The crucial question is then; how will this process and the process of global ageing be transmitted to the global economy through capital flows?

As a starting point to answer this question I would like to draw the attention to comments made by two of the most prominent members of the global financial punditry in the form of US economist Paul Krugman (PK) and the Financial Times’ chief economics commentator Martin Wolf (MW).

Starting with the former[2] he recently pointed to the fact, in the context of Japan, that external demand was instrumental in ending the slump and providing a relative bounce between 2003 and 2007. As PK further goes to argue, this may present a rather ominous outlook since the extent to which we are all, in the OECD, currently stuck in a “Japan-style” liquidity trap the way out may constitute a rather crowded route. As PK poignantly points out at the end of his small piece;


(...) needless to say, we can’t all export ourselves out of a global slump. So, how does this end?

Krugman 2009


This is indeed a good question and MW makes a similar argument in a recent column[3] where he points towards the fact that the global imbalances themselves may prove to be an impediment to a swift global recovery.

 In short, if the world economy is to get through this crisis in reasonable shape, creditworthy surplus countries must expand domestic demand relative to potential output. How they achieve this outcome is up to them. But only in this way can the deficit countries realistically hope to avoid spending themselves into bankruptcy.

Martin Wolf (2008)

This is of course a very appealing proposition and also goes to heart of idea that, at least, one part of the solution of the current global crisis lies in the resolution of global macroeconomic imbalances. But prey tell, how are these surplus countries going to revert towards a growth path characterized by a more balanced external account and perhaps even an external deficit?

It is in this context that the argument presented in this thesis becomes important. Consequently, there is a big risk that these surplus economies (e.g. Japan and Germany) simply will not be able to heed the call of MW. The main reason for this inability is then, in part, exactly to be found in the economic profile of a rapidly ageing economy with a median age pushing 40 year mark and beyond. Japan and Germany as well as the economies next in line to reach their age bracket cannot achieve growth based on domestic demand in a way which would allow them to suck up excess global capacity through an external deficit.

This is a very important point to stress in the context of the global economy and must be stressed with great emphasis.


Some Charts to Go With This

In order to try to make sense of all this consider the supply/demand chart below which plots the supply and demand for savings in the global economy. Following convention, the X-axis represents quantity and the Y-axis represents price. In this specific case, the X-axis can be seen as the total demand (from deficit nations) for excess investment beyond the level which can be achieved through domestic savings. The Y-axis then becomes the price[4] (interest rate) which equates this demand with the level (supply) of excess savings provided by the surplus nations beyond the level which can be absorbed by domestic investment demand

(click to enlarge)

As a natural consequence of the intuition underlying this small model, equilibrium is a forced (and always binding) condition since, by definition, the sum of external deficits must equal the sum of external surpluses in the global economy. 

If we accept the idea behind the theoretical framework presented in this thesis it is very easy to see the implications of a sustained global process of ageing. As is shown in the diagram the supply of excess savings (external surpluses) will increase with ageing [S(1) to S(2)]. But this is not the only effect. Following the simple intuition of a closed system an increase in supply must be meet by a decrease in demand too since we assume that economies are moving from a position as external deficit nations to a position of external surplus nations. In this sense, the constant level of output (quantity) is largely a simplifying trick in the sense that we let the entire adjustment process occur on the return of the excess savings of surplus nations rather than the quantity of excess widgets they can produce to sell abroad.[5] This produces an effect whereby ageing reduces the price of excess savings in equilibrium.

In order to move forward from here we need to mentally relax, as it were, the idea that deficits need to equal surpluses in equilibrium. In concrete terms, we need to understand the idea of equilibrium does not capture the notion of dependency on exports/foreign asset income to grow.

This is amended in the following graph; (click to enlarge).

The key here is the notion of the critical price level [6]. This should be seen as the level needed to sustain an acceptable level of growth in ageing economies and is thus a direct proxy for export dependency. As the global economy ages and assuming that equilibrium must hold at all times, the supply of excess savings and the demand for these savings decrease both lowering the equilibrium price and quantity. However, the critical price level remains. One key implications of this is a systematic oversupply of savings, or glut if you will, produced by the process of ageing and it is very important to understand that this oversupply is very tangible. It represents the value of external surpluses which would be enough for the likes of Germany, Japan etc to maintain a growth rate consistent with expectations and essentially the maintenance of their market economies. In the jargon of the theory, it represents the point at which ageing economies are optimally smoothing consumption and saving as a function of their intertemporal preference for the latter over the former. Of course, it cannot exist as a real entity but it may still have real implications.

The first obvious effect is to make the variation of ageing economies’ output very sensitive to the variation in out of deficit nations and thus global output. In its strictest form, this is how export dependency emerges. Another notable effect would be that it drives down the return in ageing economies to such an extent that it may fuel so called carry trade flows in which traders borrow in low interest rates currencies and invest in high interest rate currencies. Another example would be how these savings may be used to fund temporary and unsustainable build up of credit expansion in economies running external deficits. This is to say that if the equilibrium depicted above essentially is binding in the long run the implied existence of this excess pool of savings may lead to sudden outward jumps of the demand curve and thus the creation of credit bubbles. The main key to take away from this small economic model is thus the idea of an externality of ageing on a global level. This externality arises as a direct function of the implied existence of an excess of savings over demand as the global economy ages. In the context of the theoretical framework above the externality should be seen as function of the crowding of economies in one end of the spectrum on intertemporal preferences for consumption and saving. Crucially, it also means that what we might find to be optimal in the context of a single economy is not optimal on a global level a point which is certain to make standard economic modeling of aggregation from the representative economy level to the global economy very difficult. In empirical terms it means that what one might find to be the optimal path in a time series perspective of one economy may turn out to have radically different implications in the cross section when more or all global economies are involved.  

Further studies should attempt to develop this idea further since it provides a useful venue of analysis as an alternative to the traditional idea drafted from life cycle theory that global ageing will entail dis-saving on an aggregate level.


List of References


David M. Cutler & James M. Poterba & Louise M. Sheiner & Lawrence H. Summers (1990)
"An Aging Society: Opportunity or Challenge?," Brookings Papers on Economic Activity, Economic Studies Program, The Brookings Institution, vol. 21(1990-1), pages 1-74.

Henriksen, Esben (2002)A Demographic Explanation of U.S. and Japanese Current Account Behavior, Graduate School of Industrial Administration, Carnegie Mellon University

Higgins, Matthew (1998)Demography, National Savings, and International Capital Flows, International Economic Review, Volume 39 (1998) Issue (Month): 2 (May) pp 343-69

Bryant, Ralph C (2006) – Asymmetric Demography and Macroeconomic Interactions Across National Borders, Brookings Institute, the paper was presented at a conference hosted by the Reserve Bank of Australia in 2006 (

  Borsch-Supan, Axel H; Alexander, Ludwig; and Krüger Dirk (2007) Demographic Change, Relative Factor Prices, International Capital Flows and their Differential Effects on the Welfare of Generations, NBER Working Paper No W13185


[1] With the German surplus mainly materializing itself in an intra-European imbalance.

[2] Paul Krugman (2009) – The Eschatology of Lost Decades, NYT blog post

[3] Martin Wolf (2008) – Global Imbalances Threatens the Survival of Free Trade

[4] Which is assumed to be exogenously determined for all involved economies through the equilibrium in this system.

[5] Remember that I am assuming that quantity is fixed and that the entire adjustment takes place on the price. In a more realistic representation the adjustment would of course take place on both the price and quantity.