Facebook Blogging

Edward Hugh has a lively and enjoyable Facebook community where he publishes frequent breaking news economics links and short updates. If you would like to receive these updates on a regular basis and join the debate please invite Edward as a friend by clicking the Facebook link at the top of the right sidebar.

Tuesday, September 06, 2005

The Oecd Ageing Model

This is - I hope - a relatively brief follow-up to my last post on fertility. This is a quick account of a paper - THE IMPACT OF AGEING ON DEMAND, FACTOR MARKETS AND GROWTH, by OECD economists Joaquim Oliveira Martins, Frédéric Gonand, Pablo Antolin, Christine de la Maisonneuve and Kwang-Yeol Yoo which really is mainly interesting for the fact that it is the economic support for the above-mentioned fertility paper.

What really amazes me in all this is that the OECD work seesm to go on as if the IMF didn't exist. There is only really one paper that they seem to mention in the bibliography that is common to the two approaches - Yaari, M. (1965), “Uncertain Lifetime, Life Insurance, and the Theory of Consumer”, Review of Economic Studies, 32(2) - and this is really so old that this fact in itself is curious. Obviously the IMF economists are much further advanced in their work, and I don't understand why the OECD economists don't simply try and stand on their shoulders (or perhaps I do, and that's what worries me).

Be that as it may, I will do my best. The majority of this paper seems to consist of covering already well-known facts about the situation, eg that there is both increased life expectancy and reduced fertility.

One finding they note which is perhaps worthwhile is:

Moving the old-age threshold in line with longevity gains would only affect old-age dependency if aged workers participate in the labour force, are employed and remain in good health. Labour force participation and employability of older workers will be discussed in Section 5. With regard to healthy ageing, the evidence is scattered and points to many unknowns. No clear pattern emerges from the comparison of trends in life expectancy (LE) and disability-free life expectancy (DFLE) based on available cross-country data (Table 2.4). In some countries, there is a balanced increase, in others the DFLE progresses faster than LE and in a few the reverse is observed.

It's interesting that they find no clear pattern, it's also interesting that they don't mention along with the other factors important to the effectivity of moving the old-age threshold "if aged workers maintain the level of aggregate productivity". This, I would have thought is the key question.

On the consumption impact, which is obviously one of the key areas, they note the following:

"Household survey data suggest that total consumption displays a hump-shaped profile across agegroups (Figure 3.1). These profiles are subject to uncertainties and are not equivalent to saying that the consumption profile is hump-shaped over the life cycle, mainly due to the existence of cohort and time effects.8 Nonetheless, they would suggest that the pure consumption-smoothing hypothesis is only partly supported by the micro data."

This is all broadly in line with other findings, including the consumption smoothing, but is subject to the same caveats as I offered in the Bryant case.

On consumption generally they reach this hardly earth shattering conclusion: "With these orders of magnitude, ageing-induced changes in consumption shares are not expected to generate major structural changes in the economy."

Which is fine, but hardly helps us understand what we are already seeing in Germany and Japan.

As far as capital markets goes they use an Overlapping Generations model The specification of the model draws principally on Börsch-Supan, A., F. Heiss, A. Ludwig and J. Winter (2002), “Pension Reform, Capital Markets, and the Rate of Return". This paper will now need looking at. In particular they inform us that to simplify, "each country is viewed as a closed economy and the labour market is exogenous".This is obviously at variance with the Bryant and McKibbin models where the international dimension is paramount.

They then test three reform scenarious against a baseline 'doing nothing' approach. As they inform us:

"In the two reform scenarios, policy changes are implemented from 2005 onwards. After the reform is announced, households revise their optimal saving paths in order to smooth future consumption levels. This implies that before 2005 consumption for each cohort remains equal to its level in the noreform 'rising contribution rate' scenario. The economies return to a stable population level and structure by 2080, when a steady state equilibrium is reached. In this situation, GDP per capita growth is exclusively determined by the (exogenous) growth in total factor productivity (TFP) and capital deepening grows in line with TFP, at 0.45% per annum.17 By contrast, during the demographic transition, the dynamic equilibrium is driven by ageing trends and pension reforms."

Now its not clear how realistic this is in the first place, since it isn't evident where we will find a stable population structure by 2080, but this is so far out it may be beside the point. We need to understand here the difference between stationary and stable population. A stationary population maintains the same numeric level, but it's stucture may change due to below replacement fertility and increasing longevity (setting each other off). A structurally stable population is one which maintains the shape of the pyramid. It is this latter that I am doubting.

This finding anyway many will find re-assuring.

"The results give little support to the so-called "asset meltdown hypothesis", according to which the massive pension withdrawals during the second phase of the ageing process could induce a large decline in asset prices (see Box 2). Other quantitative studies using general equilibrium models also tend to reject this hypothesis. This is because the asset meltdown hypothesis reflects a partial equilibrium view where changes in economic agents’ behaviour spurred by interest rates movements are not taken into account. Also, the hypothesis may also be based on an exclusive focus on the impacts of ageing on savings, ignoring the impacts via labour supply on investment. In a general equilibrium setting, accounting for these mechanisms largely offsets the asset meltdown effect."

This result is broadly similar to the findings of Börsch Supan et al, but again this isn't so surprising, since they use the same model. I'm not sure what weighting they are giving to house prices in private wealth, and what evolution they see for these. I am aslo not clear whether they do not rely excessively on capital deepening, since this has well known diminishing returns induced limits.

Of course none of these studies seem to take into account the existence of a zero-bound downward limit on interest rates. It is OK to speculate that interest rates may rise later, but the stylised fact we are getting from the 'first wave agers' is that savings first go up, and global interest rates come down.


Now they get onto productivity:

"In the most optimistic scenario, assuming a flat productivity profile for old-age workers, the resulting aggregate productivity levels would increase over the next two decades and stabilise thereafter (Figure 5.3). In the other two scenarios, productivity would uniformly decrease. However, in all three scenarios the order of magnitude of the change is small (+/-2.5%) and this level effect would stabilise after a certain time. This result is related to the evolution of the average age of the labour force (Figure 5.4), which is increasing but will stabilise over the next decades if retirement patterns remain unchanged. This explains why, despite contrasting assumptions about individual age-productivity profiles, ageing per se does not have a major impact on the aggregate productivity level. In this context, concerns about the current "greying" of the labour force have to be seen as a rebound from a previous sharp decrease in the average age of the labour force, the "rejuvenating" shock that took place during the 1970s, from which OECD economies are just recovering. In fact, in the 1990s Japan already experienced the average age of the labour force that is projected to occur in Europe and in North America by 2050."

I remain very sceptical of all this. I think the principal think to say is that we really don't know. We don't know the pace of technical change going out into the future, and we don't know how a greying workforce will adapt to that change. The mention of Japan is tendentious, since it highlights the fact that *some* EU countries face relatively benign ageing (UK, France), while others (Germany, Italy, Spain) face a much more severe problem. Germany eg isn't that different from Japan now, while the UK won't be where Germany is now for at least another twenty years.

Before leaving this perhaps its worth mentioning that alot of the attention is given to the impact of PAYGO pension systems. Obviously these complicate matters enormously, and lead coutries like France with much more moderate birthrate issues to have serious sustainability issues at govt finance level. But these are relatively easy to comprehend. So I don't think they are where the main thrust of our effort should be going right now. What we need to be able to do is model the impact of the demographic changes without the incidentals like PAYGO. We are still some way from doing this, and I think it the pressing issue here. Without adequate models we will not be able to orient our assessments at all.

I suppose it goes without saying: I don't think that this paper is a very useful or important contribution to our knowlege bank.

However, having been so critical, I would like to say that the technical annex 2 which outlines the overlapping generations model is precise and clear. Whoever did this bit knew their sausages.

Low Fertilty and Public Policy

Many thanks to New Economist who has just sent me a link to a paper from OECD staff - "Trends and Determinants of Fertility Rates: The Role of Policies" by Anna Cristina d’Addio and Marco Mira d’Ercole.

What follows is an on-the-fly review of the paper. Perhaps the first thing I can say is that the research in this paper seems to depend on the macro economic work of another group of OECD staff, headed by Joaquim Oliveira Martins - The Impact of Ageing on Demand, Factor Markets and Growth - and I will look at that work directly after I consider this paper.



My initial impression is that this work does not represent any major advance in our understanding and that the work of macroeconomists like Bryant and company, or demographers like Lutz and Thomas Sobotka (who has conducted the most extensive recent study of low fertility) don't seem to get much in the way of a look in. In fact none of these authors even figure in the bibliography, which I consider lamentable, (nor does Faruqee, or any of the other IMF staff economists whose work I have been looking at lately). This apparent inter-institutional rivalry is to be deplored in an issue as important as this. Let e be clear: in general the IMF work is much better.

Another serious and astounding omission is the failure to look at the US as a very distinctive separate case. If I were talking about OECD fertility, the US 'outlier' would form a central part of what I wanted to talk about.

One demographer whose work does feature is Australia's Peter McDonald, and while I am generally sympathetic to what he is saying (and see especially this recent paper - again not in the OECD work), it is important to be careful with this work, in particular since it runs counter to the previous - more quietist - consensus.(See this influential paper by the UNs Paul Demeny).

So why does a paper like this need to be 'handled with care'? Well basically they don't address head-on the idea that there might be a fertility trap below the 1.5TFR level. They thus don't distinguish clearly enough those countries suffering from extremely low fertility from others which are settling down to what appears to be long term TFRs in the 1.7 - 1.9 range. These are the countries which have applied most the child-friendly policies that the OECD authors are correctly advocating, but they don't, in my opinion, consider sufficiently the limitations and possible ceilings to the proposed pro-fertility policies.

Also they certainly don't take into account the possible negative feedback of the macroeconomic environment, and how emerging fiscal and employment issues may make it difficult to raise fertility beyond certain levels, but here it may have been the work of Oliveira Martins which has mislead them.

Let's take some examples:

"Simulations of the possible effects of various reforms show that a package of policies relaxing some of the constraints to childbearing may significantly raise total fertility rates and, thereby, the size of the population and of the labour force. For example, in the case of Japan, such policies may raise the total fertility rate to a level of 2.0, which would leave the population in 2050 at 94% of the 2000
level, as compared to 79% projected by national authorities.
"

This is really astounding. If they are getting Japan fertility rates up to 2.0 out of their simulations they really need to re-examine their parameters. Japan is currently at TFR 1.3, and has been for some time, but a jump to 2.0 would be huge, and it is hard to believe this would be attainable at any point within the significant time horizon. Particularly since Japan with estimated government debts of around 160% of GDP is about to enter a period of severe fiscal restriction *and* rising health and pension costs for the elderly, at the same time as it tries to raise labour force participation rates among women. I have to ask: do these authors live in the same world where I am living?


Postponement of the first childbirth is probably the most important event of what has been labelled as the "second demographic transition" that is characterizing most OECD countries (van de Kaa, 1987). Postponement results in the rise in the mother’s age at childbirth (see Gustaffson and Wetzels, 2000). An indicator that is often used to describe this phenomenon is the mean age of mothers at first childbirth. For the seventeen countries depicted in Figure 2, this mean age at first childbirth has increased, on average, from 23.8 to 27.2 years over the period 1970-2000, an increase of over 1 year per decade.

"If successive cohorts have the same average number of children per woman, but delay their childbearing until later in life, this will lead to a temporary reduction in the period fertility rate; the opposite would occur if each cohort of women advanced the timing of their childbirths. Changes in the mean age at first childbirth for different cohorts of women can therefore generate cyclical swings in the period fertility rate (a decline, followed by recuperation) even when the cohort fertility rate is unchanged. The use of total fertility rates, when postponement of childbirth is occurring, will thus overestimate the short-run effect of the decline in fertility rates."

Yes, but they really are confusing the question here, since we are not dealing with cyclical swings, but with a long term upward rise in the timing of childbirth. There is no evidence that this is reversing anywhere, and there is no indication whatsoever of a 'swing' phenomenon.

"The decomposition presented in Figure 3 helps understanding whether lower fertility rates in one period, due to reductions in fertility rates at younger ages, are compensated by increased fertility rates in subsequent periods. However, it does not allow determining whether full recuperation is occurring. This reflects two factors. First, postponement alters the contribution of the two groups of women to the total fertility rates: even an increase in fertility rates of older women in one period that exactly matches the decline of younger ones in the previous decade may leave the total fertility rate below the level that prevailed before the onset of postponement. Second, period indicators give only a cross-sectional view of what is unfolding at the cohort level."

The childbearing patterns observed for these four birth cohorts confirm that, in all countries, recent generations of women have fewer children at early stages of their reproductive cycle and more children at later ages. In general, however, the higher number of children that women have when old does not fully compensate for the lower number of children that women have when young: for example, in the case of Australia, the age-specific fertility rates of different cohorts decline by significant amounts at age 25 to 29, when moving from the oldest to the youngest cohorts, while they increase by small amounts at age 35 to 39. Sweden and other Nordic countries are exceptions, as the increase of age-specific fertility rates at age 30 to 34 is much larger than in other countries. In Sweden, in addition, the median value of the distribution shifts to the right, reaching a higher level than that recorded in the previous period; in other words, by the time Swedish women reach age 25 to 29, the cohort born in the years from 1961 to 1965 had a higher fertility rate than the maximum attained by the cohort born from 1941 to 1945.


This paragraph is important:


Bearing these caveats in mind, Table 2 suggests that women born in 1956-60 are likely to experience a further significant reduction in their total fertility rate relative to the level realised by the previous cohort in Austria, Italy, Japan and Spain, while possibly recording a recovery in France, Norway, Sweden and the United States. The average decline in completed fertility rates increases for later cohorts (as "projected childbearing" over their full reproductive cycle declines from a level of 1.89, for womenb born in 1961-65, to 1.81 for those born in 1966-70 and to 1.77 for those born in 1971-75), before recovering slightly for the cohorts born in 1976-80. For women born in this later period, fertility rates are slightly above those needed to ensure replacement of the population in France, Netherlands and the United States, while Austria, Japan, Poland and Sweden have fertility rates of around 1.5 or lower. In these and other countries, the size of recuperation in fertility rates required to bring birth rates back to the levels achieved by the cohorts born in 1951-55 is very large: on average for the cohort born in 1971-75, the extent of recovery required is 54%, and this rises to 65% or more in Italy, Japan and Spain. While still being "feasible" in biological terms, the pace of such recovery would be without historical precedents. This suggests that the decline in fertility rates observed over the past three decades on the basis of period and cohort indicators is likely to be lasting.


The authors do note that:

"Postponement of childbearing has important consequences on both the number of children women have over their life and on the mother's and child's health. Close to half of all children are growing up without siblings in several OECD countries; the share of women that remain childless at ages 30 and 40 has increased strongly over time; and the risk of some health problems for mothers and their children has also risen."

Which is a point I was highlighting here, and they fail to assimilate the important point that Lutz is making that the postponment of childbirth has long term structural consequences for the population (regardless of the final cohort TFR values) since in the next generation there will be succesively less young adults capable of having children.


The second chapter gives an assessment of the societial and individual changes which have lead to the changed childbirth patterns. This is all largely uncontroversial, and is a worthwhile read for anyone not familiar with the situation.

One of the items they have a section on is 'marital status'. This may not appear at first sight to be self-evidently important. Then consider this: one of the factors which influence fertility is te rate of new household formation, and one of the factors which affects age on childbirth is the age of setting up the household. Now in Northern Europe and the US people may well go and live together and maybe marry later, whereas in Southern Europe or Japan people may 'hang on' until they are able to formally marry. This obviously affects household formation, and hence fertility.

The key to the whole issue comes with the discussion of the widening gap between desired and observed fertility rates:

While changes in structural conditions and life styles are contributing to delay and decline of birth rates, the effects of these changes on the number of children that women will have over their reproductive life have been exacerbated by the constraints that individuals and couples face in everyday life, by the emergence of new risk factors confronting them (labour market insecurity, difficulties in finding suitable housing, unaffordable childcare) and by the failure of social policies to provide adequate support. Indications about the potential role of these constraints on women’s childbearing decisions can be derived from answers to questions about the "desired" or "ideal" numbers of children provided from opinion surveys."

In fact McDonald suggests that the two *key* phenomena are changing gender identities on the part of women (including attitudes to education and labour market participation) and the lack of similar changes in the male mindset. Where the first occurs rapidly, and the male minset either doesn't change, or only slowly (Southern Europe) then it is very easy to fall into the low-fertility trap.

Chapter 3 deasl with government policies, which of course is the other key parameter which may affect the childbirth decision. As the authors indicate there are direct, and indirect costs associated with having children, and the indirect (or opportunity) costs are probably much the more significant ones:

Beyond direct costs are the indirect (or opportunity) costs associated to childbearing. If direct costs may be shared among parents, indirect costs fall almost exclusively on mothers. While the difficulties of estimating their size are even larger than in the case of direct costs, it is very likely that the size of these indirect costs rises alongside the higher employment opportunities available to mothers. For women that — because of their education and preferences for financial autonomy — can get a foothold in the labour market, the decision of having a child will affect their career opportunities. They may have to withdraw from the labour market, at least temporarily, shortly before and after childbirth; they may not be able to return to work after childbirth, or may have to work part-time or under atypical schedules; or they may find that, in the longer term, their career prospects have worsened relative to childless women and to men. Often, one immediate consequence of these changes in work arrangements following childbirth is a loss of income. Further, as the longer a mother stays out of the labour market the more difficult it becomes for her to re-enter it, indirect costs of children increase as the mother ages.

The paper considers a number of alternative policies - tax policy, child care provision, parental leave from employment, - and then carry out a number of simulation tests to assess possible outcomes. As I said, my impression is that they are way to optimistic here, but that doesn't mean that the policies they advocate should not be given urgent and serious consideration. Even a much smaller impact would be a strong positive factor. Their conclusions:

"Despite the limitations of the analysis undertaken in this chapter – which reflect the difficulty in considering the full range of factors that may contribute to cross-country differences in the levels and changes of fertility rates in OECD countries – the evidence presented suggests that changes in a range of policies may prove to be helpful in removing obstacles to childbearing faced by individuals and couples. Childcare arrangements, transfers to families that reduce the direct cost of children, as well as provisions that allow mothers to better cope with their family and career responsibilities all can help in removing obstacles to childbearing decisions. Illustrative simulations of the possible impact of a range of policy changes also point to significant increases in fertility rates of several OECD countries, with significant effects on population size (and structure) and with smaller but still large effects on employment levels."

Monday, September 05, 2005

Welcome To The Menagerie IV

The next paper in the Bryant/McKibbin saga is Incorporating Demographic Change in Multi-Country Macroeconomic Models: Some Preliminary Results (2003). As in the 1998 paper they suggest:

Demographic shifts may profoundly influence the world economy, directly in the countries experiencing the demographic change and indirectly through changes in global trade, capital markets, and exchange rates. Though that point is now widely acknowledged, it is much less widely understood that existing analytical tools are inadequate for assessing the generalequilibrium and cross-border consequences of demographic change.

Unsurprisingly, they propose to remedy that:The research reported in this paper takes preliminary steps to improve the required analytical tools.

They inform us that they build on earlier theoretical work by Blanchard, P. Weil, Faruqee, Laxton, and Symansky, who use a revised life cycle approach to consumption and saving behaviour. (Actually, the MIT opencourseware, just coincidentally, has some notes on the OLG model, they are a bit technical, but here they are).


The Blanchard model could be called 'forever young'.

Bryant and Mckibbin allow changes in birth and mortality rates to be combined with an approximation of age-earning profiles in orderto allow demographic shifts to influence human wealth, consumption, and asset accumulation. They do this using two simplified empirical models: (i) a two region abridgement of the IMF’s MULTIMOD model and (ii) a two-region abridgement of the McKibbin MSG3 multi-country model. As they indicate "the stylized shock on which we initially focus is an unanticipated and transitory demographic bulge, analogous to the "baby boom" experienced by some industrial nations several decades ago."

So this is a very partial and specific shock. What this model effectively looks at it an 'ageing' scenario, where one very large cohort distorts the age pyramid as it moves upwards.They use a two economy (home and away) model to test for this shock both on an equally distributed and on a lop-sided (restricted to one country) basis.

They report that their research "strongly confirms the hypothesis that differences across countries in the timing and intensity of demographic shifts can have significant effects on exchange rates and cross-border trade and capital flows".

Parts of the paper reads - even word for word - as very similar to the 1998 one.

This part, which I didn't go into in any detail before is important:

The tradeoff facing researchers about the two routes is primarily a matter of time horizon. Multi-cohort OLG approaches that explicitly keep track of different cohorts, their saving decisions, and their wealth stocks can be more rigorous theoretically. Other things being equal, a multi-cohort OLG approach thus may appear more attractive. But other things are not equal. Multi-cohort OLG models are more difficult and demanding than models that use analytical shortcuts to get demographic effects into the consumption-saving specifications in macroeconomic models. The OLG
models, moreover, are likely to take much longer to advance to the stage where the models can deliver interesting empirical conclusions. The requirements of a multi-cohort specification are of course especially demanding in a model with numerous separate national economies and national currencies.

Another disadvantage of the multi-cohort OLG specification is that it might, if calibrated only to partial-equilibrium relationships derived from micro-level evidence, deliver misleading inferences about aggregative macroeconomic relationships. As Hamid Faruqee emphasized to us in a comment made after the international workshops that launched this project, it is both the virtue and the vice of a micro-level specification for individual agents that variables such as goods prices and interest rates are taken as exogenously given. Goods prices and interest rates self-evidently cannot be modeled as exogenous at an aggregative, economy-wide level. It is thus unclear whether a disaggregated OLG model, based on partial-equilibrium relationships estimated from micro-level data, will yield correct inferences and predictions for general-equilibrium, macroeconomic behavior (in the sense of being able to replicate the moments of actual macroeconomic data).



Forever Young

Blanchard showed that a simplifying assumption can expedite aggregation and modeling and thereby allow a researcher to avoid the adoption of a more complex and analytically difficult multicohort OLG approach. Blanchard’s key assumption was that each individual, throughout life and regardless of age, faces a constant probability of death, p. The expected life of an individual is thus 1/p . With this assumption, researchers can choose a value for p anywhere between zero and a large number. If p is put at the limiting case of zero,individuals live forever and the model yields the infinite-horizon results familiar from still simpler models; values of p in the range .03 to .01 yield model "life expectancies" in the range of 33 to 100 years.

The constant-probability-of-death assumption can be combined with an assumption, based on Yaari (1965), that the economy contains life insurance companies permitting agents to costlessly make annuities contracts contingent on their deaths.18 The two assumptions together permit the derivation of an aggregate consumption function without keeping explicit track of the consumption and wealth of multiple cohorts. Aggregate consumption turns out to be a relatively simple linear function of human and non-human (financial) wealth, with the marginal propensity to consume dependent on the age-invariant probability of death and individuals’ rate of time preference.


The authors inform us that in order to facilitate a clearer comparison of models and alternative treatments of economic behavior they constructed stylized and simplified versions of two existing macroeconomic models. The first a stylized, two-region
abridgement of the IMF staff’s MULTIMOD model as modified by Bryant and Long Zhang, and secondly a stylized, two-country abridgement of the McKibbin and Wilcoxen G-Cubed models.


As I have been indicating the study looks at the impact of a 'baby bulge' generation passing through one or both economies, and its impact on consumption and saving. Since exactly this process is taking place in the US at this moment, the following is not without some significance:

The cyclical movements in per capita human wealth and consumption during the years 10 to 30 and the associated cyclical behavior of other domestic macroeconomic variables are partly due to the fact that the new members of the population resulting from the (adult) baby bulge are at first relatively low savers. Speaking loosely, these younger adults are ascending the left side of the hump of their age-earning profiles (Figures 1 and 2). As the bulge cohorts reach their years of peak earnings and high savings, another inflection point is reached. Per capita human wealth and consumption in the fourth decade of the shock begin a long decline relative to baseline. Eventually, as the baby boomers become elderly, their labor income and human wealth decline and they begin to consume out of their financial wealth. Per capita consumption in the long run returns to the baseline level of the initial steady state.

Obviously there will be an impact on US consumption and saving as this generation pass through the earnings peak.

Finally, and as with so many of these dense technical papers, perhaps the most interesting part comes with the diagrammatic figures which start on page 59 as viewed with the adobe acrobat reader.

Watching Indonesia

Last week at New Economist I posted about the apparent fragility of Indonesia's financial situation. I was partly put in mind of this by a post by the New Economist himself on Buttonwoods interest in emerging market risk.

Basically the idea is that demography is a context, economic theory explains the details. To see what I am getting at, lets go back to the idea (which isn't correct, but some state) that I am 'putting it all down to demography'. Obviously you can't do this, and this isn't really my intention. What I do suggest is that we can get onto the economics later when we understand the demographic settings better.

What do I mean? Well the thing is we could take the difference between stable and unstable equilibrium. Now when economies move from one 'local state' to another, then they move from one more or less stable equilibrium to another, and it is during this process that the danger of 'crisis' is high, and policy is terribly important.

Look at the median ages of some of the societies which have had notorious crises in the last 15 years: Mexico, Argentina, Thailand, Turkey.(For the record they are currently: 24.93, 29.42, 30.88, 27.77 respectively).

What is interesting is that all these countries have had crises as they moved from being child societies to becoming young adult ones. Once they have firmly made the transition they become a lot more stable. (China will, of course, be the big test here). My guess is that Turkey is firmly ontrack, Argentina would be if someone would lend them some money and Thailand is now also following a much more solid trajectory. Mexico is still rather on the young side, and prehaps still vulnerable.

So what about Indonesia? Well Indonesia's median age is 26.8, whioch means it is making the transition now, and is in the high risk area: which is why this piece in the Morgan Stanley GEF again caught my eye:


We have been quite concerned about the Indonesian situation as the rupiah and the market lost substantial ground over the past month (Between a Rock and a Hard Place, August 26, 2005). The move by Indonesia on August 30 in raising short-term interest rates and reserve requirements for banks, and tightening foreign exchange market regulations in my view will not fundamentally address the current woes confronted by Indonesia.

In the face of a structurally weak economy, high debt, a lack of fiscal latitude, a weak rupiah and soaring oil prices, we think difficult policy trade-offs will be required. The government has chosen to raise short-term rates and impose restrictions on the banking system and the foreign exchange market. As the banking system and consumers are quite interest-rate sensitive, a substantial rise could torpedo the economy. Yet, raising short term interest rates and putting restrictions on the banking system and the foreign exchange market is not a permanent policy solution. Tackling the root of the problem — the inefficient fuel subsidies — is the ultimate solution. However, such a subsidy removal would also pose considerable social-political risks as a significant segment of the population relies on this subsidy and it could also torpedo private consumption. Indeed, it is an ugly policy trade-off as the current administration may risk social and political-economic instability, but hopefully a rational policy move would win the support of the Indonesian people.


I would say this is the quite a clear statement of the classic child to young adult policy dilemna. Be careful Indonesia.

Moving up to the next level, Japan's 1989 crash came as she was moving from mature adult to elderly dynamics. Does that tell us why the shock was so severe, and the results so enduring? I've not quite got there yet, but this type of argument is definitely getting me someway along the road.

The G-Cubed Model

Interestingly enough Warren Mckibbin and Peter Wilcoxen's G-cubed model (original explanatory paper here) which is considered by some to be so useful for modelling demographic impacts , seems to have been originally developed, in part at least, to handle global warming issues:


The G-Cubed model has been constructed to contribute to the current policy debate on environmental policy and international trade with a focus on global warming policies, but it has many features that will make it useful for answering a range of issues in environmental regulation,microeconomic and macroeconomic policy questions.

The model combines the dynamic macroeconomic modelling approach taken in the MSG2 model of McKibbin and Sachs (1991) with the disaggregated, econometrically-estimated, intertemporal general equilibrium model of the U.S. economy by Jorgenson and Wilcoxen (1989). As the authors explain:

The Jorgenson-Wilcoxen model breaks the economy down into 35 separate industries, each of which is represented by an econometrically estimated cost function. The G-Cubed model has only 12 sectors but each sector is based on econometrically estimated cost functions.

We are informed that G-Cubed is still in the process of development but it is already a large model. In its current form it contains over 5,000 equations and 110 intertemporal costate variables. Nonetheless, it can be solved using software developed for a personal computer. Wow!

Sunday, September 04, 2005

Welcome To The Menagerie III

Well I'm back from my guest blogging stint at New Economist. Afoe is also temporarily down for maintenace repairs, so poor readers I have no excuse to not come home and carry on with this unusual series of posts.

Right now I am steadily working my way through the papers of Ralph Bryant and Warren McKibbin trying to see what can be learnt from the trajectory they have followed.

The first significant paper they seem to have written in the field of demographic consequences seems to have been "Issues in Modeling the Global Dimensions of Demographic Change" (1998).

The paper is a kind of state of the art summary - back in 1998. It is really a write-up of feedback obtained at workshops held in Washington, Paris and Tokyo during July 1998, and as the preliminary blurb says:

"As a first step in developing a better framework for dealing with global demographic change, this paper surveys the areas in which we need to improve existing global analytical frameworks to deal with the range of important policy issues that will emerge as a part of the demographic shifts. The paper attempts to summarize what is now known, identifies areas where important unresolved debates still exist, and explores theoretical and empirical issues on which more research needs to be undertaken."

As the authors say:

"The goal of the paper is to take stock of what the economics profession knows about modeling the macroeconomic consequences of population aging. We focus particular attention on the likely consequences for changes in saving-investment-current-account balances and in asset prices -- for individual nations, and for the world economy as a whole."


Speaking of the ambition of their work the authors state:

"The objectives of this project, with its emphasis on the cross-border consequences of demographic change, cannot be achieved without use of one or another multi-country, generalequilibrium macroeconomic model."

However:

The most notable inadequacy in existing multi-country macro economic models is their failure explicitly to incorporate the effects of demographic changes. As preconditions for achieving the major objectives of the project, therefore, researchers working with the multicountry models must improve their equation specifications in at least three respects. An improved analysis is required of the effects of demographic changes on:

(1) consumption (including possibly patterns of consumption across different goods and services), saving, and wealth accumulation, with appropriate allowance for the
openness of national economies;

(2) the production/supply sides of national economies, again with appropriate allowance for openness; and

(3) expenditures, transfers, and revenues in government budgets.


As they also note:

"Most industrial countries have experienced two important behavioral trends. Labor force participation rates for male workers, especially in the last years of normal working life, have been declining. And labor force participation rates for female workers have been secularly rising. The first of these behavioral trends exacerbates the adverse effects of population aging on the effective labor force. The second has worked in an offsetting direction."

As both James Hamilton (Econbrowser) and Dave Altig (MacroBlog) have recently indicated these kinds of demographically associated behavioural trends may be pretty much at the heart of recent US labour market phenomena - this however is marginal to my present preoccupations.

Perhaps it is important to note that all the above sets down a pretty substantial agenda for macro economics, so in this sense it could possibly be regarded as the starting point.

The paper also contains this idea:

"Population aging will occur at differing paces and with differing degrees of intensity in the industrialized countries of the world."

So it is the relatively different pace of ageing which needs to be part of our focus. One thing is becoming clearer. Actually, if you look at it from this point of view, the whole demographic transition is an *ageing* one, after the initial kick-start. What do I mean? Well, initially you get a substantial rise in the birth rate and fall in infant mortality, which presumeably shoots the median age right down to all-time historic lows, and from there on, as they say, it's uphill all the way. Japan is currently just over 42 years old, and rising, and we don't know where the limit is.

Of course, this clears up one other key relation: the important thing is surely how rapidly we move upwards through the ages our active capacities in relation to the rate of increase in median age.

The Lifetime Model

There is another thread worth picking up in this paper, and that is that of the assesment of the work of Friedman and Modigliani in relation to savings and consumption behaviour over the life-cycle:

During the July 1998 workshop discussions for this project, a large part of the dialogue was devoted to analytical methods for studying the consequences of demographic changes for consumption, saving, and wealth accumulation. Given the wide variety of analytical views held in the profession, this area is one of a few central topics on which substantial further research is warranted.

"At a very general level, since the work of Modigliani-Brumberg (1954, 1979) and Friedman (1957), economists have accepted in broad terms the idea that many households will wish to smooth their consumption across time."

I think we get off to a bad start here, and 'll explain why. I don't think the best way to go about this is to specualate about what households (or even individuals) may wish, or not wish to do (the explanation), where we need to be is with the outcomes, what the data can tell us about what they actually do. I stress this particularly since I think far too many macroeconomists spend their time worrying about the implications of their work for the microeconomic foundations of modern macro, and I tyhink this is an overly 'theorised' approach. We need to have models of the macro economy which tell us something aout the important things going on, and one way to achieve this may be to 'black box' the micro issues. This is what I am doing. I am simply agnostic in this regard.I simply don't have the time, and this seems to me to be the 'cleanest cut'.All of this becomes important when you look at the continuing sentence:

"The degree of this intertemporal smoothing and the main factors driving it, however, continue to be much in dispute."


This is something you will see repeated on numerous occassions, it has become part of the 'economists folk culture'. But actually how valid is it? And if it isn't valid, aren't most macro theorists, when they base themselves on the view, really puttting the telescope over Nelson's bad eye?

"One set of issues concerns the manner in and degree to which households or individuals voluntarily want to act as intertemporal smoothers"

This is the set of issues that I am suggesting we should neatly side-step.

"A second set pertains to whether constraints external to the household inhibit or prevent agents from acting as intertemporal smoothers."

This looks much more promising.

"Explicit life-cycle approaches hypothesize that consumers save little in their early years, save most in their middle-to-late working years, and then spend down their wealth accumulation after retirement."

Now this is not a statement about intentions. It is one about outcomes. This can be tested. This is the way to go.But somehow it doesn't seem quite right.

Basically, following the work of swedish demographer Bo Malmberg, and drawing on Dvaid Bloom and others, we can break down the demographic transition as it affects age structure into stages.

First of all there is a child dominated society (Niger, Bangladesh, Nigeria etc) where the burden of having so many non-productive children normally produces negative saving and indebtedness (especially at the state finance level).

Then comes the young adult stage, where saving predominates over investment, China is, of course, the classic example of that right now. Then you have the mature adult society - with one more time indebtedness and negative saving, but this time producing consumer driven booms - here we might think about the UK and the US.

Finally old age arrives, and once more people begin to save appreciable portions of their income, and consumer demand is flat, or even falling slightly. I say finally, but maybe I spoke too soon, since we still don't really know what comes next. On the standard version of the life cycle theory people should start at some point to draw down their savings, but to date there is little evidence of this (and this is one of the objections that is being raised to the life cycle theory). However, if we look at things in terms of Ricardian equivalence, where we take the aggregate of private and state borrowing into account, then clearly looking at Germany, Italy and Japan there is growing evidence of some kind of symmetry with the debt problems of the child dominated societies, certainly in terms of the move towards unsustainable government debt dynamics. At the same time we may yet see the savings 'drawdown' predicted by the young Modigliani, we need to wait and see.

Anyway this highlights two issues: the mature adult and the old age societies, maybe early versions of the life cycle model put this much too simplistically.


Some researchers interpret the empirical evidence as broadly supporting the lifecycle view, including the hypothesis that consumers are patient enough to begin saving for their retirement early in their working lifetimes. Examples of researchers sympathetic to the life-cycleview include Attanasio-Browning (1995) and Meredith (1995). Other researchers, however, read the empirical evidence as providing only weak support for the life-cycle view of saving and instead supporting the hypothesis that consumers save and accumulate wealth primarily to insulate consumption against uncertainty about fluctuations in income. In this latter view, the saving for retirement motive is much less important than the precautionary saving for uncertainty motive. (Precautionary saving for uncertainty leads to the accumulation of "buffer-stock" assets.)

As I keep highlighting the preoccupying detail here is we keep talking about *motives*, when what we should be looking at are outcomes.


"Several participants in the July workshops emphasized the point that the life-cycle hypothesis, as studied in the context of microeconomic panel data, appears unable to account for the most prominent observed changes in several countries' saving behavior. For example, the lifecycle hypothesis does not do a good job of explaining the pronounced decline in the saving ratio in the United States in the last several decades. Nor can it explain the pronounced increases in saving ratios in several Asian countries (such as China, Indonesia, South Korea, Singapore, and Thailand)."

Well, I'm sorry gentlemen, as I am explaining, I think a version of the life cycle model, suitably calibrated with the data, can fit these developments quite well.

The US has, remember experienced an immigration induced demographic shock since the mid 80s which has transformed its age structure, and obviously the net impact (at the aggregate level) can be negative for saving. The isssue with the asian tigers needs less comment, as I think Bloom and Williamson's findings have now achieved a widespread acceptance.

"Most of the households in the U.S. economy seem to have cut their saving at the same time, and most of the households in the relevant Asian economies seem to have increased their saving at the same time."

Of course understanding this seems to depend on understanding the meaning of the term 'most'. These movements are perfectly compatible with substantial demographic shifts. There is also the issue of individual vs household, but this can wait for another day.

"Yet the microeconomic analyses of saving behavior are not throwing up a consensus explanation for these time trends. In particular, the demographic components of the simpler versions of the life-cycle hypothesis are certainly not providing the explanation. "

Exactly, the simpler versions of the life cycle hypothesis need modifying.


Angus Deaton in a recent and highly readable assesment of the Modigliani thesis has this to say:

In these extreme precautionary or “liquidity constrained” accounts of saving, consumption is smoothed, not over the whole life-cycle, but over much shorted periods of a few years at a time, see again Carroll (1997) and Deaton (1991). In the literature, this is often referred to as “highfrequency” smoothing of income, as opposed to the “low frequency” or “life-cycle frequency” smoothing that was postulated by Modigliani and Brumberg.

These results are not inconsistent with Modigliani’s argument about precautionary saving for (possibly the majority of) consumers who, under the original life-cycle model, wish to save, not borrow. These will be people who are sufficiently patient so as to be happy to postpone their consumption, or who do not expect much income growth in the future. Indeed, it is possible to reconcile the traditional and the new view by noting that it is the youngest families who are likely to want to borrow, but either cannot or are too prudent to do so, and are therefore more or less constrained by their current earnings, while those in middle-age behave in the traditional lifecycle way. That such a formulation is consistent both with expected utility maximization and with the survey data has been shown in an important paper by Pierre-Olivier Gourinchas and Jonathan Parker (2002).


OK, now back to the Bryant and McKibbin paper:

"A further difference of view in the saving-consumption literature concerns agents that in practice may not be able to borrow and lend freely as the simplified intertemporal smoothing models presume they can. The theoretical treatment and empirical importance of such liquidity constrained households is controversial."

Naturally the paper notes that "It seems natural to suppose that large changes in the demographic structure of populations will have significant effects on consumption and saving. The effects should be larger, the more pervasive are life-cycle elements in the behavior determining saving."

However:

"Most aggregative macroeconomic models have failed to allow for such changes in the demographic structure of the population. The seminal contributions to growth theory in the 1950s and 1960s abstracted from such demographic factors."

So this is a clue to what has happened.

In effect, those models contained no children and no elderly, with the result that a faster (steady-state) rate of growth of population caused the saving rate unambiguously to rise in response to higher requirements for investment and a higher capital stock. Tobin (1967) developed a simulation model differentiating workers from retirees (but still without children). Tobin's model also predicted that faster population growth would raise the private saving rate, because the faster growth caused the population distribution to become younger (more working, saving households relative to retirees who were older and dissaving).Even as theoretical growth models matured, most empirical macro models still abstracted from shifts in the demographic composition of the population. This omission persisted despite the major emphasis in consumption theory of life-cycle considerations in the microeconomic behavior of households.

Now comes the clue:

A strand in the literature on development economics was sensitive to these issues; see, for example, Coale-Hoover (1958), Mason (1987, 1988), and Taylor-Williamson (1994). The builders of empirical macroeconomic models, however, did not try to incorporate this work. To be sure, the development-economics literature for the most part did not directly address the details of how to adapt the consumption-wealth specifications in general-equilibrium macro models. The development-economics literature on population dynamics and saving is reviewed in recent papers by Higgins-Williamson (1997) and Bloom-Williamson (1997). Deaton-Paxson (1997, 1998, 1998) and Paxson (1996) are recent studies.

And now we get to the rub:

"When specifying the consumption-saving-wealth relationships in macroeconomic models,
issues of aggregation are very important. In particular, the aggregation issues are central for getting an adequate macro specification of the demographic influences."

However:

"Macro models built up from a micro theory positing a single representative agent are not easily adapted so as to incorporate demographic changes."

Oh. And why would that be?

"By definition, changes in the demographic composition of the population require analysis to acknowledge the heterogeneity of agents -- at the very least heterogeneity in age."

Ah yes. Quite.

As they then go on to explain, there is another class of models, the overlapping generations one, which attempts to grapple directly with one or more dimensions of heterogeneity across agents.

So the OG models could be more promising, but, there are no free lunches. They are much more difficult to construct and work with:

An explicit multi-cohort approach, at least seen from some perspectives, appears more rigorous theoretically. But such an approach may also be more difficult and demanding, and will probably take much longer to advance to the stage where the models can deliver interesting empirical conclusions. The requirements of a multi-cohort specification are of course especially demanding in a model with numerous separate national economies and national currencies. Another possible disadvantage of the multi-cohort specification is that it might, if calibrated to partial-equilibrium relationships derived from micro-level evidence, deliver misleading inferences about aggregative macroeconomic relationships.

The authors however remain agnostic:

As of the time this paper is written, the economics profession has no sound basis for judging whether an explicit, multi-cohort approach will over the longer run provide the most reliable empirical conclusions and projections about macroeconomic behavior.

As they suggest, despite the limitations, adapting existing models - like the IMF Staff's MULTIMOD - may prove to be the most direct and effective way of getting at some orientative results. Well, this has been long. It has certainly been useful for me. If you have arrived this far I hope it has been for you too. I'm afraid there *is* more to come.