Other Alpha Sources

I have been enjoying myself in the Austrian Alpes last week and hence the lower output. Here is my look though, of a number of notable news stories and contributions. 


Global Liquidity

Benoît Cœuré, Member of the Executive Board of the ECB has penned a speech (and argument) on global (excess) liquidity. Izabella likes it and I agree with her that it is a good piece. I am not sure though that it is that much different than the Savings Glut argument put forward by Bernanke, but I may be missing the fine print (i.e. need to read it more carefully). The biggest problem I have is that he assumes that the lack of safe government (i.e. AAA rated assets) is cyclical and due to market failure or other "temporary" factors. Izabella interprets it as follows, 

What’s the solution to this vicious liquidity circle? Simple, says Cœuré. The euro area needs to regain its role as a global supplier of safe assets. Something which could be achieved by a) ensuring that Eurozone countries have become fiscally sound and b) diverting excess liquidity from other zones back into “programme countries” by way of the IMF.

I disagree. The failure of euro zone economies and indeed large parts of the OECD edifice in general to provide "safe haven" assets is deeply structural and tied to population ageing. Unfortunately, there is little prospect that the euro zone economies will be able to supply AAA rated securities for a long time and herin lies the rub. Of course, if we are talking euro bonds, but then again. I will believe it when I see it. 


Japan and the currency wars 

A recent Bloomberg article suggested that Japan has been "secretly" selling JPY to try to stem the tide and force through depreciation of the Yen.

Japan used so-called stealth intervention in November as the government sought to stem yen gains that hammered earnings at makers of exports ranging from cars to electronics.Finance Ministry data released today showed Japan conducted 1.02 trillion yen ($13.3 billion) worth of unannounced intervention during the first four days of November, after selling a record 8.07 trillion yen on Oct. 31, when the yen climbed to a post World War II high of 75.35 against the dollar. The currency’s strength has eroded profits at exporters such as Sharp Corp. and Honda Motor Co., just as faltering global growth undermines demand. 

Open market operations to sell domestic currency are so old school. Didn't they get the memo in Japan? In a world where all major central banks are either at or very close to the zero bound, it is central bank balance sheet expansion (quantitative easing) that matters. On this note, both Japan and the Fed are being left decisively behind by the ECB and BOE (at least in the past six months). Of course, even the usage of "standard" measures in Japan is being contested and as long as this is the case, the Yen will continue to strengthen.


Don't bet on deflation with the current team of global central bankers 

Elsewhere, I am wondering where all the deflation, let alone disinflation, is. I am a sworn deflationist and I believe in the main thesis of the deleveraging/depression/deflation crowd. However, I have the utmost respect for the inflationist bias of global central banks and with the current batch of policy makers at the helm, deflation is a very remote risk.

The latest data show that inflation in China recently quickened as well as producer prices in the UK increased in the week that the BOE announced another round of QE. Of course, this is not all clear cut. Chinese real M1 (YoY) recently moved into negative territory for the first time since 1996 and in the UK, it is noteworthy that core inflation (ex food, beverages, tobacco and petroleum) came in noticeably lower in January. 

I will change my views on the basis of changing data, but I am beginning to think that the bout of global headline disinflation we are expecting as a result of the global slowdown will reverse itself much, much quicker than many (including me) have expected. Arguably, we still need decisive easing in emerging markets and QE3 from the Fed, but it is more a matter of when and not if this happens and as such, global central bankers remain fully committed to creating inflation.

The main problem so far for those arguing for strong central bank action (including me) is the absence of nominal growth in output in excess of consistently rising headline inflation. Could this be a result of doing too little, perhaps, but at the moment stagflation remains the best way to describe our current economic situation and thus inflation in all forms is a drag on growth. Should the genie finally come out of the bottle in the form of consistent wage increases central bankers may find that they got more than they bargained for even if the alternative is equally painful.  


The Greek experiment is about to end

Greece remains the main talking point and also the only thing that appears to prevent equity markets ripping to new highs. Greece is bankrupt and while I understand that the patience of the rescue committee will run out at some point, I am astounded that anyone expects this hideous experiment to end well. Greece will see its fifth year of contraction this year and for what? A membership of a currency union that does not work anyway?

We are told by the Troika, the EU and the IMF that failure to reach a deal would be catastrophic and thus that Greece has no way out but to take the medicine. However, Greece has a real choice and the stronger she is pushed the more obvious the end result is. Internal devaluation and decades of austerity don't work; not in Greece and not elsewhere. This remains the KEY issue that the euro area politicians and the ECB have not understood. The social fabrics of society won't stand the pressure and strain. Textbooks tell us that the cure is simple when you can't devalue, but practical experience have now shown otherwise. 

I am neither on the Greeks' nor the IMF/Troika's side, but I simply point out the obvious destiny of current events; failure! Even if Greece manages to appease its creditors with austerity, the end result in terms of Greek macroeconomic balances is still unsustainable and thus the underlying problems will not have been solved.

The ECB and the IMF will likely face significant drawdowns on their Greek bondholdings regardless of whether they use such drawdowns as  "carrot" for Greece to push through austerity measures. This is what the establishment has not yet understood.  


MF Global investigation fails to uncover illegal activity?

Megan McArdle has an amazing article suggesting that the investigation on the failure of MF Global is finding it difficult to uncover anything illegal. 

Megan quotes a piece from Reuters (no link available)

Lawyers and people familiar with the MF Global investigation of the firm that was run by former Goldman Sachs head Jon Corzine say that even though the hunt is still on to find out whether or not officials at MF Global intended to pilfer customer money in a desperate bid to keep the brokerage from failing, the trail at this point is growing cold.

This seems very odd to me even if I have not followed the aftermath in detail. I completely agree with the sentiment expressed by Megan

I don't understand how this could be true. To be clear, I am not saying that it couldn't be true-only that I don't understand how such a thing could have happened. There is more than a billion dollars missing from supposedly segregated client accounts. I understand that it was chaotic, but what kind of chaos causes you to accidentally move money out of money that any moderately sophisticated compliance system should have automatically flagged for approval?

While my professional responsibilities are confined to the smooth running of a macro research product I sit in an office, and work, with asset managers and ever since the failure of MF global I would imagine that their general level of concern has increased. This is understandable. If your main counterparty as an asset manager (i.e. your prime broker) essentially decides to steal your deposits and/or allocate them to losing trades against the principle of segregated accounts, it really does not matter what you do. No matter the tightness of the shop run on the asset managers' end, he will face significant and perhaps even fatal losses. 

Obviously counterparty risk is as old as finance itself and any decent asset manager today will deal with more than one broker and even have a strategy on how to manage counterparty risk. Ultimately though, mutual trust between asset managers and their prime brokers is a commodity which has been severely impaired by the MF Global failure and this is an issue for all players in financial markets. 


Dealing with vintage data in economic forecasts using instrument variables (wonkish!)

A recent note from the George Washington University points to an interesting study from Warwick University on the forecasting of data vintages in the context of US output and inflation forecasts. The problem is as follows; 

Consider a simple benchmark autoregressive model that a forecaster might use to forecast an economic variable yt. In order to estimate the parameters to be used for the forecast, typically the forecaster will obtain the most recently updated data on yt (i.e. the vintage of yt available at that time) and estimate the model using those data. However, the data in this single time series may in fact be coming from different data generating processes. The data some time back in the series have gone through monthly revisions, annual revisions, and perhaps several benchmark revisions. The most recent data, however, have been only “lightly revised,” as Clements and Galvão term it. Therefore, Clements and Galvão argue that the data in a single vintage are of“different maturities.” Forecasters may want to forecast future revisions to data as well as exploit any forecast ability of data revisions to improve forecasts of future observations. In their article, Clements and Galvão suggest that a multiple-vintage vector autoregressive model (VAR) is a useful approach for forecasters working with data subject torevisions. This comment discusses the importance of taking revisions into consideration and compares the multiple-vintage VAR approach of Clements and Galvão to a state-space approach.

This is a significant issue but remember; if the following holds, we need not worry too much about it. 

If the revisions are unpredictable and the early data are efficient estimates of future data, then we may not need to be concerned about the different vintages. 

Most economists assume that the statement above is true and simply force through their model. Being a great believer in practical usability when it comes to empirical economics, I would argue that in most cases this will not cause too many problems in most cases. However, a growing body of evidence suggest two important issues to consider. Firstly, revisions are predictable and thus provide important ex-ante information which should be incorporated into the the forecast. Secondly, even if revisions are unpredictable, the manner in which data is revised may itself provide important information on future data readings. 

I agree, but the problem is potentially much more severe. Another issue then concerns that situation where you try to forecast Y(t) as a function of X(t) where both variables may be subject to revisions. Normally, we would solve this issue by restricting X(t) to variables where revisions are minimal (or absent alltogether). One way to do this is to use market based data (market prices, closing values of securities etc) which are, by definition, not revised. However, in the context of the e.g the classical leading indicators framework pioneered by Geoffrey H Moore, this issue re-emerges X(t) is cast in the form of real economic variables (themselves potentially subject to revision).

We have replicated and refined many of the LEIs described by Moore et al and applied it to various economic data series with specific fitting of a time series regression in each case. However, such an approach may still suffer from vintage data issues (as described above. One solution that I been thinking about is to imagine two forms of right hand variables. X(t, economic) and X(t, market based); if the latter is unrevised it might be possible to find an instrument for X(t, economic) (final revision!) using a variation of X(t, market based). This would, in my opinion, constitute an elegant way to solve the issue of data revisions in your explanatory variables.

In practice, you could also try to replace Y(t, economic) with Y(t, market based), but this is probably too a-theoretical and ad-hoc.