Last time I had Japan under the loop I asked whether there was no end in sight for Japan's economy and as I wind up for another close look, I must say that it is still very difficult to find good news if any at all. However, and for the sake of argument I thought that we might begin with some recent arguments in the context of the global economy which suggest that we may be past the worst of our travails. The first observation comes from the Economist's ever eloquent financial markets pundit, Buttonwood, who recently made the neat point that while we are still stuck in the mire, the second derivative might be turning positive. This suggests that while indicators are still on the decline they are now declining less rapidly. In Tokyo, Cassandra voices a similar sentiment as she takes stock of the number presented earlier last week by the Asian Development Bank that as much as USD 50 trillion, so far, has vanished into thin air during this crisis. Hovering between the "half full, half empty" metaphor Cass notes;
For the moment, I'll take the middle ground and venture that having shed an awful lot of aggregate value (an entire year of global GDP according to the ADB!), we're a lot closer to where we're going than we were.
It may come to a surprise to my readers given the traditional very bearish sentiment expressed at this space, but I actually agree with Buttonwood and Cassandra here. However, I would simply add the important qualifier that there will be a significant asymmetry in terms of where individual economies are going as a function of where they are and were. No where is this more true in the case of Japan and as we progress to the data and analysis it should be abundantly clear that for all the talk of second derivatives and glasses being half full, Japan still look to be in an extraordinarily bad shape. Initial evidence of this comes from the headline GDP figures which don't seem to be blessed with any second derivative effects.
Final estimates from Q4 2008 suggested that Japan contracted at an annualized 12.1% which puts Japan in the dubious pole position of biggest GDP declines among industrialised economies. Q1 estimates are yet to be released, but no-one expects, I think, an improvement as the incoming data so far has been nothing but extraordinarily poor. Sociétè Generale expects Japan to contract sharply in 2009 with Q1 as an forecast bottom. I am not sure about the bottom in the sense that while it may be a bottom in the sense of the second derivative noted above, it won't likely mark a return to sustained growth.
In this note, I will provide an overview of the recent developments in the Japanese economy. Since we last convened some interesting points have emerged. For one, Japan is back in deflation measured on the US style core price index and for the first time in a very long time Japan is now running a current account deficit. This last point will be studied in some detail since it marks a very important issue for the export dependent Japanese economy both in a historical and a current perspective. Before we begin I should note that this post is very big with a lot of graphs and even an econometric model to boot. I understand full well if this deters some of my readers; I shall not hold it against you.
Prices and Consumption, where art thou?
If there is one thing which has been stable in Japan throughout this crisis it has been the persistent sluggish trend in domestic demand measured by top line household consumption expenditures as well as prices on the other hand. These two data points consequently tell an important part of the story of the lack of domestic demand in Japan or more specifically the lack of visible momentum to pull Japan out of the doldrums. One persistent feature of the initial phases of the crisis where markets and global policy makers primarily looked towards the risk of stagflation was that inflation in Japan exclusively was driven by cost-push factors in the form of headline inflation and not demand pull factors. This idea is a well established one at this point, and materialised itself in the fact that as headline inflation shot through the roof core inflation only budged slightly. It is important to point out that this inelasticity cuts both ways and as headline inflation has abated (for now), so has the spread between the two indices narrowed significantly. The underlying point here is thus two-fold. One the one hand it is dangerous to assume that inflation driven by domestic demand conditions will correlate with external headline inflation pressures which, due to global capacity constraints and global demand conditions, look set to shoot higher the minute we move even slightly beyond the current malaise. On the other hand however, we can clearly see, in Japan, that whatever trend we see for headline inflation domestically induced price pressures in Japan are virtually non-existing and now that the crisis is seriously biting Japan is set once again to retrench into deflation despite the central bank's most ardent efforts to apply measures of quantitative easing.
As I have argued before, I believe a large part of Japan's problem with deflation is demographic. In particular, I think that because Japan is basically unable to achieve growth based on domestic momentum a growth scenario strictly based on domestic activity as the one we are seeing at the moment will be de-facto deflationary. However, since Japan is largely dependent on energy imports in so far as goes its consumption of fossil fuels (i.e. a high passthrough effect) the overall inflation indice will diverge from the core of core index which, in Japan's case, is a good proxy for domestically induced price pressures. Now, I realize that my readers will be skeptical of the demographic link here, but let me at least present results that show the broken link between the general price index and core of core prices (which exclude energy and food). Thanks to a novel data set from Japan' statistical office giving us monthly inflation rates (y-o-y) for all three recorded inflation indices since 1971 we have plenty of ammunition on our hands to proceed. In the following all numbers will be based on de-trended time series which in this case simply means that I am using the first difference.
Consider then the very simple representation below which shows the correlation between the general index and core of core index over the entire sample, from 1971-1996 and from 1997-2008.
The emerging picture should be quite straightforward to interpret even for the untrained eye. Consequently, and in so far as we can consider the simple correlation coefficient a credible measure of the strenght of the connection between two variables, then this relationship has clearly deteriorated. In graphical terms we can get an impression of this by looking at the three year rolling average of the correlation between the variables.
Now, the volatility is considerable here and in fact we can see that the correlation has hit rock bottom once before , but the accumulated trend is still one of a decline in relationship between the two variables. If we want to be even more specific we can express this in the form of single linear regression where we let the general inflation index explain the core-of-core index. As is visible below, this also shows a marked decline in explanatory relationship. However, this may not be an adequate conceptualization of the issue at hand. Consequently, let us try to narrate the problem as one of headline inflation leading core-of-core inflation. This potentially brings us into the deep murky vaults of time series econometrics and I shall not belabour my readers with techniques on how to choose optimal lags here (I tried with both a quarterly and monthly). What we end up with is the following small model.
The fit is not perfect and in terms of actual prediction tool I would be weary in using this expression alone although in a standard ARMA framework one could perhaps play around with the lags of other variables. Yet, the picture is now firmly solidified as we observe a secular decline in the model's ability to model the core-of core index.
So, what the heck is this all for then? Clearly, it is difficult to show initially that demographics represent an important underlying explanatory variable in this framework. Yet, it does corresponds with the overall point expressed above that when domestic demand is unable to generate inflation exogenous energy shocks won't necessarily lead to underlying inflation dynamics. On a general note, it is thus difficult to see how Japan can avoid to enter a serious bout of deflation during the course of 2009 especially since, at this point, deflation is being pencilled in across a wide batch of economies across the globe. As will be showed below the BOJ is already coming up with ever more spectacular measures to ward off a lingering fall into deflation. There are two forward looking issues to watch out for when it comes to the comeback of deflation in Japan. One is the point that since everybody is facing deflation, and thus engaging in different forms of QE will Japan then be less of an odd man out? A second a highly related point is what will happen to the JPY in relation to the whole collective edifice of QE among OECD central banks?
Turning briefly to the consumption expenditures and thus the state of the Japanese consumer it really is (un)steady as she goes. Some analysts have expressed the opinion that the Japanese consumer has held up alright up until this point in the crisis. I am not sure what data these analysts are looking at. All I know is that the headline figure for consumption expenditures is still clocking in one negative number after the other and in this light it is difficult to see from where the much awaited boost in domestic demand is going to come from; note for example here that autosales dropped a healthy 27.9% in January. Add to this that retail sales dropped 5.8% on an annual basis with the subcomponent and you have firm evidence of a slump.
