Morgan Stanley's Stephen Jen and the Importance of Demographics

In keeping with the tone set in Edward's recent post you might ask what the h'll we are doing commenting on something coming from one of the biggest investment bank's chief currency strategist. I mean, have we gone completely mad? Well, not exactly and as we progress I am sure you will concede that this is worthwhile. What are we talking about then?

In so many words we are quite simply taking about the impact of demographics on financial markets and more specifically the way demographics are affecting (and will affect in the future) cross-country capital flows and global liquidity conditions. This is of course a topic which, contrary to many other issues dealing with demographics, has been widely debated in the literature and amongst financial practitioners and economists. As far as our own contributions go within the DM team I have authored two notes on the subject; Ageing and Financial Markets - Going for Yield? and Ageing and Financial Markets ... a revisit. As you will see by visiting these notes I have already been somewhat tuned in to Stephen Jen's work and in that light the headline above shouldn't come as a complete surprise. My immediate impetus for this post is consequently Stephen Jen's recent note over at Morgan Stanley's Global Economics Forum entitled: Demographic Trends and the Financial Markets. So, what does Jen has to say and do I agree with him? Well, I would like to begin with the simple point that Jen really does seem to be poking the right places and in fact if we include MS GEF's Robert Alan Feldman we can clearly see that the Morgan Stanley analyst team is pretty well cued up on the fact that demographic trends actually do exert a strong influence on global economic and financial markets. In this light alone I have been very impressed in general with MS GEF team's analysis on the topic. However, I still (of course) have a few and important quibbles.

First off we have the notion of dissaving which leads Jen to the following argument in the context of Japan;

The permanent income hypothesis suggests that very ‘young’ and very ‘old’ countries tend to dis-save, while those with low dependency ratios should be saving. The constellations of the C/A imbalances in the world are not consistent with this pattern. What this may imply is that, as Japan ages, its savings rate should decline, as retirees start to draw down their savings. This may very well start to happen. With such a high C/A surplus position (4% of GDP), this prospective trend should not pose a problem for Japan.

There are a lot of important points here but let us start with the first one which is of a theoretical nature. In the context of the theoretical framework as such there is no doubt that the general point is 'true' whether you take Friedman's permanent income hypothesis or Modigliani's life cycle theory of consumption and saving. Yet, we need to understand two things here I think. Firstly, there is an important distinction between the micro- and macrolevel here. In this way and while I have no problems with the argument from the point of view of the individual household viewed as a black-box I think we need to attach some qualifiers once we venture towards the macroeconomic level. Secondly, we need to realize I think that whatever characteristics we apply to the process of (rapid) dissaving we are essentially talking about a hypothetical endpoint in time which is not, at this point, possible to identify. Moreover, we also must remember the utter unattractiveness of such a process from the macroeconomic point of view since if such a process were to occur in the context of lingering lowest-low fertility we would effectively be talking about a rapid hollowing-out of whole societies which is not exactly feasible and desirable; I hope! This then brings us to those qualifiers mentioned above and in the context of Japan (and potentially other ageing societies). They can be summarised in three main points of which Stephen Jen in fact himself has inspired me to one.

1) A change in the consumption/saving schedule which is likely to emerge in the context of the relatively smaller working cohorts in the sense that they will be prone to save more of their disposable income in order to compensate for the rising dependency ratio as well as to secure an adequate level of life in retirement, particularly given the rising life expectancy. This will have as an important side-effect the tendency to further depress domestic consumption/demand.

2) This relates to the Japanese export surplus which Jen narrates as a 'buffer' from which to dissave. In the context of the general argument of dissaving I can understand this point but once again I think it is important to point out that the Japanese external surplus is a little more than a buffer at this point. In fact, as Edward argued recently over at Japan Economy Watch Japanese growth is deeply and structurally dependant on external demand. This once again serves to substantiate my point on dissaving in the sense that a country such as a Japan will most likely fight long and hard to avoid bringing itself into a situation of (rapid) dissaving since this essentially would mean 'game over'; I mean, how on earth would Japan for example be able to finance an external deficit?

3) This point is actually taken from Stephen Jen's own work and analyses and relates to the decline in home bias of Japanese investors. As he puts it in the note in question;

Fixed retirement age, coupled with ever improving life expectancy, has created a ‘longevity risk’, whereby retirees can no longer be confident of their ability to defend their lifestyle if they end up living much longer than they expect at the time of their retirement. In the case of Japan, this has led to more risk-taking, not less, as retirees try to enhance their expected investment returns by diversifying away from assets with low credit risk. In contrast to the first hypothesis, this alternative hypothesis suggests that retirees should have a bigger appetite for equities.

Please note that I am in no ways trying to catch Jen off guard here but, at least to me, the rising risk appetite of Japanese retail investors (don't forget those housewives) and the subsequent decline in home bias indicate yet another level through which dissaving can be (is being) postponed since in stead of actually 'dissaving' as such we are witnessing the attempt to earn a return on those massive poolings of savings.

The second and final niggle I have concerns Jen's perception of 'ageing' and essentially his conceptualization of what constitutes an old society/economy. This might seem as a trivial question of calibration since in the end we are all ageing but as we will see, an initial calibration mistake/difference can end up making the whole argument scoot off towards quite differing conclusions. In short, I am at odds with Jen's pooling of e.g. Japan and the US in the same bracket as 'ageing'/old economies. There is no way around this one I feel and quite simply this is not 'possible' given the different age structures of these two countries. In order to substantiate my point I think that we should take a look at the two graphs below which demonstrate the difference between two countries in terms of ageing; the countries in question are the US and Germany where the latter is as useful proxy for Japan as any in connection to age structure (at this point I have not yet completed my median age calculations on Japan).

As we know ageing occurs as a result of a joint process of falling fertility and rising life expectancy where migration of course provides an important potential factor in smoothing out what otherwise seems to be a pretty linear process of ageing. Regarding life expectancy and when we speak of the biggest OECD economies we can safely say I think that the divergence in ageing does not decisively come from differing degrees of longevity. Rather we need to look at fertility and to some extent also migration where the latter's effect of course tend to work with almost no lag relative to changes in fertility which tend to take longer time in showing their effect but also then demand a much longer backdrop of action in order to swing the situation back into some kind of 'balance'. If we look at the US and Germany above and the evolution of fertility and median age we can, quite unsurprisingly, see a very strong correlation between the two. There are two things to note in particular regarding the evolution of fertility. The first is that both countries seem to have entered a decisive stage of the demographic transition which began in the beginning of the 1970s (we don't know whether this in fact is also the final stage). Now, the interesting thing here is that the US and Germany started off from different starting points (taking 1970 as year '0') with the US 'starting' the transition from a TFR at about 2.5 and Germany from about 2. This had the obvious effect that while both countries' TFR rate went below replacement level the German TFR stayed persistently lower (at about 1.5) than is the case in the US (at about 1.7). If we then factor out the difference in starting points we can see that during the 1980s the relative decline in TFR was no greater in Germany than it was in the US. However, from the 1990s (1987 in the US) and onwards a striking difference emerges by which the US TFR begins to climb towards replacement level (a bit below) whereas Germany's TFR falls below the dreaded mark of 1.5 where it has lingered ever since (the fertility trap?). Now, both the absolute as well as relative difference in evolution of TFR rates quite dramatically translate into the median age where we can see above that Germany is substantially older than the US. In fact, the time series above tend to underestimate the difference a bit since the REAL impact on German median age (and Japan's, Italy's, etc) is going to come in the decades to come. So, the US is younger than Germany (and thus Japan); that much is pretty certain I think. But how can we extrapolate this to my general point in terms of calibration? This brings me to my second point on fertility.

As such, we can see that for both Germany and the US, the evolution of median age seems to exhibit a very strong linear trend over the course of the last half of the 20th century and into the 21st century. This is most likely due to a steady upward movement in life expectancy as well as of course a steady downward path of fertility, especially once we get into the 1970s and beyond. However, and here comes the important point, we need to realize I think that although ageing occurs with a steady linearity (migration flows can distort this though) the effects of ageing exhibit notable non-linearities; this at least is our hypothesis. In short, there seems to be a median age range (let us say <40 years) years where conventional economic theory seem to cope 'pretty well' with understanding the processes taking place; i.e. developed economies in this range seem to respond strongly to low real interest rates/credit expansion, are able to sustain asset price booms, can run large external deficits, have a solid domestic demand component, etc. Yet once we transist into the 40s regarding median age there seems to be a change in all of this as we have tried so many times to point out here at DM in the context of e.g. Japan, Germany, and Italy(who IS running an external deficit mind you). Now, in terms of Jen's remark this all shores up at his point on how demographics will have a tendency to be tantamount to Dollar weakness and thus a misconception in my view regarding the point where the US would 'hypothetically' start to dissave.

In Summary

In order to sum up why don't we have a look at the way Jen summarizes his main points;

Demographic trends have important economic and financial implications. Without remedial action, global ageing in the developed world tends to raise the level of real interest rates, flatten the yield curves, benefit equities at the expense of bonds, and lower the value of the dollar.

By the reading the note above you could be tempted to believe that I disagree quite strongly with Stephen Jen. This would be a mistake. However, I do want to emphasise the two points noted above . Firstly, it does not serve us well for analytical purposes to narrate ageing in the context of dissaving since this takes us away from what happens in the transition towards this situation which we after all might not want to reach. Secondly, I think that we need to be rather careful with which economies we denote as 'old'. Of course all OECD economies are ageing but they are doing so in different tempi and moreover as I have argued above the real effects of ageing seem to materialize with notable non-linearities relative to the process of ageing itself which is fairly linear. Finishing of with Stephen Jen's work in general much of my own work on ageing and financial markets has transpired from Jen's continuous digging on the topic in the context of MS' GEF. Especially, his hypothesis on a decline in home bias in the context of Japan is a very important contribution to the big picture. Also, I am very sympathetic to the way Stephen Jen has been working his way towards this. Basically, he started digging on Sovereign Wealth Funds went on to Sovereign Pension Funds and then ended up with this important contribution to the debate which I am sure will be modified and refined as we go along. And all this over the course of a year in the context of MS' GEF; who the h'eck ever said that blogging couldn't add value?