Wednesday 22 November 2017

Greek governments and GDP forecasting: not a good match

It was state-budget submission time in Greece yesterday and as it is natural I guess, out of the whole budget, most people paid attention to a few figures, namely the primary surplus envisaged and the GDP growth rate projection. In this post I'd like to zero in on the latter. 

The GDP growth rate projected by the respective Greek government is regarded with heavy scepticism (bordering on irony) by both the press and the wider public. Hence, I believe that it would be interesting to look at what the actual data tell us.

The data I'll look at here are those concerning those budgets that are available in the Hellenic Finance Minstry's website, i.e. those from 2007 onwards.  

Here's how these projections stack up against actual growth rates.

source: Eurostat, Hellenic Finance Ministry

And here is the same chart but in a slightly altered form. A positive reading means that, GDP-wise, things turned out better than envisaged while a negative one that they turned out worse.


source: Eurostat, Hellenic Finance Ministry, own calculations

If we look at the chart above and break it up in sub-periods according to which government was in office so as to evaluate their GDP-forecasting skills (or should I amend this to "the Finance Ministry's forecasting skills while they were in office"? hmm) we can see that: 

- the 2007 - 2009 ND government (not in office when the 2010 budget was voted though) didn't do very well, as far as GDP forecasting is concerned.
- the 2009 - 2011 PASOK government (not in office when the 2012 budget was voted upon, but only just) didn't do that well either.
- the Nov 2011 - May 2012 PASOK - ND - LAOS coalition government under Loukas Papademos' projections proved to be too optimistic as well (of course the time between them coming into office and voting on the budget was too short).
- the 2012 - 2014 ND - PASOK government has the most stellar record among those commented upon here since in both of the years that they were in office GDP turned out better than projected.
- finally the Syriza - ANEL - Greens coalition government's performance is a mixed bag since 2015 turned out far worse than predicted, while 2016 turned out better and now 2017 is en route for a miss (since for the 2,7% projection to come into fruition Greece's GDP has to grow by ~4,85% on average in the two remaining quarters of 2017).

There's a number of reasons that those forecasting misses can be blamed upon, e.g. the change in trend which is when projection misses usually occur, the depth of the recession that was hard to forecast etc. but the fact remains that Greek governments and GDP-forecasting are definitely not a match made in heaven.

I'll wrap this up by way of posting a chart showing what Greece's GDP would be today if government projections were proved to be correct vs. where it actually stands. 


source: Eurostat, Hellenic Finance Ministry, own calcualations

Makes you wish they'd got it right, eh? If they had, Greek GDP would be ~25% higher today...


Friday 20 October 2017

The Odyssey of the Greek Middle Class

It is said that a country's middle class is the bedrock of its democracy. Hence, saying that its prosperity should be considered tantamount by anyone who believes that democracy is the optimum form of government, cannot be dismissed as  an exaggeration. But what happens in a country such as Greece where household incomes have been clobbered by a depression that saw GDP decrease by as much as 28% (peak-to-trough)? How did the Greek middle class fare throughout the depression?

First, one has to define middle class. ELSTAT's household budgets survey divides households into eight income classes. My intention is to deem the lower two income classes (0 - 1100 EUR) as the "lower class", the middle four ( 1101 - 2800 EUR) as the "middle class" and the two highest (>2801) as the "upper class".

So, what happened to the Greek middle class? Did it suffer the same fate that Greek GDP did?

Well, as counter-intuitive as it is, it actually didn't dwindle. On the contrary, it grew. In 2016 the aforementioned definition of Middle Class accounted for ~54% of Greek households compared to 49,5% in 2008.

source: ELSTAT, own calculations

Of course this can be attributed in its entirety to the fact that what was defined here as the "Upper Class" lost 2/3 of its pre-depression mass. Most likely the households that left the two upmost income brackets migrated to those "right" below (although one can hear plenty of anecdotal evidence of much more extreme moves across brackets). At the same time, the "Lower Class" doubled in size so that in 2016 it accounted for ~33% of Greek households compared to ~15% in 2008.

The next graph charts the change in the number of households in each income class.

source: ELSTAT, own calculations
If we combine data from this chart with those from the next one (which depicts what chunk of middle class' total, households belonging to each of its four income classes account for) we can see that contrary to 2008 when the lower brackets represented about 45% of total they now correspond to a bit more than 55%. So, despite growing slightly in size, it becomes evident that Greece's middle class has, on a whole, become poorer. Now that is definitely more intuitive.

source: ELSTAT, own calculations
  
All in all, contrary to what someone would expect Greece's middle class didn't dwindle in number but it certainly hollowed out. Despite the move downwards already in place, insecurity and low morale due to the threat of even further downward mobility are hanging above Greek Middle Class' collective head like Damocles' sword. Here's to hoping that Greek middle class will hold.

Friday 22 September 2017

The Greek manufacturing rebound charted.

Anyone casting even a furtive glance at Greek macro data must have noticed that Greek manufacturing is currently staging a much needed bounce. A bounce that still belongs to the "too little too late" category if one compares its magnitude to that of the massive contraction that the sector underwent during the 2008 - 2012 period. Let's drill down a bit more and look at the specifics of this bounce and the underlying situation in individual sub-sectors.

source: ELSTAT, own calculations

In case someone wants some perpective, it should be remarked that the lowest the overall production index got was 30,3% off its highs and that the current bounce brings it 10,4% off its lows, i.e. it currently stands 23,1% off its late-2007 highs.  

source: Eurostat, own calculations

I thought that it would be very interesting to see the overall index's moves de-composed into these of its main components (unfortunately the narrowest de-composistion ELSTAT offers, includes 24 sub-sectors and since I couldn't find a similar methodological explainer for Eurostat so I went for that) so I re-constructed the headline index from scratch. The resultant index wasn't an exact replica (for reasons that obviously elude me) of the original but it was decent enough I think. Anyway, here is how it stacks up compared to the real thing.


source: ELSTAT, own calculations

And here is the decomposition. I know one gets woozy by all the tiny coloured bars but I did my best to give somewhat flashier colours to the index's main components (i.e. Food ~ 19,87%, Oil Products ~ 15,08%, Basic Metals ~ 7%, Beverages ~ 8%, Non-metallic Minerals ~ 8% of total).


source: ELSTAT, own calculations

The current rebound's catalysts are mostly oil products, basic metals and pharmaceuticals while beverages and non-metallic minerals contributed too to the previous big production spike in mid-2016.

Headline figures show that momentum is currently waning but this could very well be due to one or two big sectors slowing while, beneath the hood, growth remains broad-based and buyoant. One way to see if this is the case here is to count the sub-sectors that are currently in expansion mode. 


source: ELSTAT, own calculations

It seems that, below the hood too, the rebound is indeed moderating since out of the 24 sub-sectors charted here, just 8 were expanding in July compared to a high of 21 in July 2016.  

Taking a look at individual sub-sectors, it seems that the only one going through a phase of - seemingly - unhindered growth is pharmaceuticals whose production is climbing to new highs compared to all the other sectors that still try to reach their previous highs. What's more, secondary, mostly inward-oriented sectors, have been all but annihilated by the crisis, with their current production levels hovering at about 20% of their pre-crisis highs.


source: ELSTAT, own calculations


To wrap this up, the current rebound in Greek manufacturing is decent in magnitude but it currently appears to be slowing. It remains to be seen whether it will gain steam again of the current policy mix of over-taxation, high energy costs and cost-of-capital will make it succumb to the pressure.

Saturday 12 August 2017

Greece's Consumer Price Index components analysis for 2010 - 2017

For quite some time now I wanted to take a look at Greece's Consumer Price Index in order to gauge each of its components' contribution to the headline figure. I somehow though never got round to it. Now, during a viciously hot August in Athens that makes sleeping at nights "a tad" harder, I finally managed to sit down and do it.

I used the 12 sub-indices that ELSTAT publishes along with their respective weights to translate their monthly, year-on-year, changes into their contributions to the headline figure. If those are used to re-construct the overall index, this is how the constructed index compares to the actual one. Good enough, I think. 

source: ELSTAT, own calculations

Since 2014 ELSTAT is supposed to revise the sub-indices' weights every year according to the Household Budget Survey, something that undoubtely improves the National CPI's quality and representativeness. That said, I haven't managed to find any revised weights for 2017... *sigh*

So, here's the relevant chart. 

source: ELSTAT, own calculations
As it is plain to see, the drivers of the 2010-2011 inflation spike were none other than the sub-indices covering the sectors for which indirect taxes were hiked (i.e. housing, transportation, food, alcoholic beverages and tobacco) while the rebound in crude prices in 2010 and the first half of 2011 also played a role.

Also interesting, that the main drivers of the second leg of the infamous Greek deflation were the sub-indices that where affected by the fall in crude's prices (i.e. housing due to heating oil and transportation). On the contrary, its first leg was significantly more broad-based but, as mentioned in an older post, since the turning of headline inflation to negative territory was preceded by an analogous move of the import price index, the catalyst also seems to be not domestic.

The signing of the 3rd Greek MoU in July 2015, which foresaw a hike in VAT for food items and the tourism sector, translated into inflationary pressures stemming almost exclusively from these two areas.

Finally, the move from deflation to inflation in early 2017 was driven from those sub-indices that were affected by the positive year-on-year change in crude prices, i.e. housing and transportation, while the tourism sector also contributed.

To wrap this up, it seems that some of the policies pursued due to the Greek MoU were not deflationary at all. On the contrary, they proved to be rather inflationary since their result was the highest inflation reading for Greece post-2000, in 2010. On the other hand, the notorious Greek deflation was mostly due to import prices falling, while the catalyst behind Greece's swing to positive inflation in 2017 was also a change in import prices, this time a surge.


P.S. In case someone's interested, here's the evolution of CPI sub-indices' individual weights. The increased importance for food items and housing will probably work towards making Greek CPI more volatile.

source: ELSTAT

Sunday 11 June 2017

One feature of the current emigration wave out of Greece left unsaid.

The current emigration wave out of Greece is certainly well documented and remarked-upon. It is often compared and constrasted with past emigration waves, mostly with that of the '60s that saw droves of Greeks flocking to western Europe (mostly Germany), the US etc. There is one feature of the current wave though that sets it appart from past ones. Current Greek emigrants don't seem to be sending money back home. 

Quite a lot of countries, mostly EMs, depend on worker remittances to shore up their current account balances. Greece did, in part, during the '60s and '70s. Although, detailed data regarding remittances flowing back into Greece for those decades do not exist, a simple look at secondary income balance shows that this was quite prevalent during the 1960-1970 period (flows from the EU after the country's accession in 1981 are easily distinguishable, since the secondary income balance deteriorated as the economic situation in Greece was improving and emigrants started coming back and exploded upwards in the 80s).

source: AMECO, own calculations

If we look at worker remittances data for the present period we come across a paradox. As the emigration wave out of Greece gained steam, workers' remmitances declined and as of now, haven't picked up.

source: Eurostat

Now, one can spin a ton of different explainers on this (new emigrants are young and don't have families of their own to support back home / due to EU-induced freedom of movement people take their families with them / since a lot of recent emigrants are highly educated their families are mostly well-off hence not in dire need of support / they barely make ends meet in their new home country so there' nothing left to send back home etc.) but in order to draw data-based conclusions  one needs detailed data which I don't have.

It is worth remarking that this is not the case in other countries of Southern Europe, which witnessed similar emigration waves these past few years. Both in Italy and Portugal, worker remittances picked up along with emigration flows.

source: Eurostat

source: Eurostat


Given the dire economic situation in Greece and the relevant historical patterns of emigration, the lack of remittances sent back home from the fresh wave of Greek emigrants seems peculiar but it might be explained by the characteristics of the households sending emigrants abroad as well as the households that the emigrants set up in their destination countries.



Sunday 4 June 2017

More proof that Greek deflation was not due to anemic demand

These past few years we were showered with superficial analysis by celebrity economists (academic or not) who claimed that the occurrence of deflation in Greece can be blamed on the policies implemented because of the MoU.

This particular line of thinking is something found in economics textbooks and it could have very well been the case in another country where product and services markets functioned properly. But not in Greece. 

I had written another post about this in 2016 but I choose to revisit this now because of the fact that inflation in Greece turned positive again.

As I had claimed back then, inflation peaked almost 4 years after final consumption did so how can someone, in all seriousness, claim that these two incidents were directly connected? Now as far as the deflationary nature of the policies implemented is concerned I beg to differ and say that, on the contrary, some of them (like indirect taxes hikes for example) were clearly inflationary and contributed to the incidence of stagflation (and also to the higher rate of inflation that Greece experienced post-EMU accession) during the early years of the Greek adjustment program. 

Now, to stop blabbing, the reason that deflation left the building is the same that brought it on back in early 2013, a change in import prices. 

source: ELSTAT

As the chart makes rather obvious, both of the times that the inflation rate changed from positive to negative territory (and vice versa) were preceded by analogous changes in import prices. The biggest component of the import price index by far is crude oil, which accounts for 21,3% of the total, so this is the component that mostly drives headline figures. 

So, instead of generic analysis derived from first-year economic textbooks maybe, for a change, we should pay attention to facts and to the special features that make every economy different.

Friday 12 May 2017

How to bring on Hyperinflation: the Argentine experience as a cautionary tale for Greece

We have lately seen a lot of comparisons between the current Greek depression and various sudden stop episodes in different countries. All those pieces though have missed the case most apt for such a comparison. The Latin American Debt Crisis of the 80s. Out of all the Latin American countries that went through extremely hard times back then, I'll focus on Argentina due to greater data availability. I won't go into comparing the performance of the two economies in this post but I'll focus instead on the occurrence of hyperinflation towards the end of the decade and the causes for that occurrence.

In both cases there was a cluster of sudden stop episodes encapsulating the whole region (in Argentina's case Latin America, in Greece's case the European periphery), making the countries' access to capital markets rather pricey if not outright impossible. 

In Argentina's case, which had what many Grexit-advocates prescribe as a way for Greece to overcome its plight, i.e. the ability to print money at will, the lack of access to capital markets was plugged the usual way. By putting the printing presses to work.

The next chart makes rather obvious that when government revenues dropped, inflation in Argentina spiked something that implies a jump in seignorage. It is worth noting that throughout the 80s inflation dropped below three-digits only once, in 1986 (during the Austral plan), when it stood at ~90%.


source: IMF, inflacionverdadera.com

If we zoom in on the period right before and after the hyperinflation episode, it becomes crystal clear that hyperinflation was preceded by a momentous surge in monetary financing for the government by Banco Central de la Republica Argentina, the country's central bank.


source: BCRA, inflacionverdadera.com

Given that many Grexit proponents envisage that the country can e.g. raise its pension spending by printing money, the Argentine experience becomes rather relevant for Greece as well. Moreover, in case of a messy default and unilateral Grexit wouldn't tax revenues drop significantly? How would government finance the resultant deficit? Money-printing again seems to be the path of least resistance. If we add international trade disruptions and the subsequent shortages in the above mix, a "(much) more money chasing much fewer goods environment" emerges.

Many outsiders dismiss the possibility of hyperinflation in case of a messy Grexit as too far-fetched and unlikely. I, on the other hand, think that takes like this are overly optimistic and tend to ignore how a rather large part of the Greek political systems thinks and acts. I would advise them to take a look at the cluster of hyperinflationary episodes in Latin America during the 80s which came on as a result of the cluster of sudden stop incidents in the region. Suddenly, the possibility of hyperinflation making an appearance  in case of a messy Grexit doesn't seem that low, does it?

Thursday 9 February 2017

Is Greece's underperformance compared to sudden stop episodes in other EMs really due to the Euro ?

Yesterday a post from Matthew C Klein under the title "Greece has done much worse with the euro than EM basket cases did with their own currencies" was published in FT Alphaville and caused quite a stir. I will try to pen a response piece since I believe that the said post overlooks some pretty essential points that, in a large part, explain the reason why Greece underperformed the EMs mentioned above. (hint: that reason is not each country's currency of choise)

A sudden stop episode is an abrupt halt/slowdown in private capital inflows into an economy that causes an abrupt current account reversal. Hence, I find it peculiar that there was no mention of the magnitute of the current account deficits involved in the episodes used by Mr. Klein in his analysis in order to draw any conclusions.
The magnitude of the pre-crisis imbalances involved defines the adjustment needed to be undertaken so that the economy in question acquires a more sustainable footing and also pinpoints the liquidity that will be taken out of each one of these economies because of the sudden stop.   

In the current post I only look at the respective sudden stop episodes of Argentina, Indonesia and Thailand because the size of the adjustment appears to be larger and as a result more easily comparable to that of Greece. In the charts posted later on in the post I define the pre-crisis year as the one before the current account balance exploded into positive territory. As far as the Asian crisis is concerned, this is not consistent with normal crisis timelines who define the pre-crisis year to be 1996.
First of all, the level of the current account deficit with which Greece came into the crisis makes the ones of the other three countries appear miniscule. 
source: IMF
The difference between the three EMs and Greece is that the sudden stop episode in their cases was really abrupt and within one year their current account balances swung from varying levels of deficit into hefty surpluses. Since Greece's current account deficit was humongous back in the pre-crisis year, if Greece chose to withdraw from the Eurozone, the amount of liquidity withdrawn from the economy would be monumental and so would be the ensuing depression. If Argentina's swing from a ~ 1,3% deficit to an 8% surplus meant a ~11% drop in GDP what would the resultant drop in Greek GDP be if the curent account swung from a 15% deficit to an unknown, but undoubtely large, surplus?  

One further point overlooked in Mr. Klein's post is the countries' fiscal imbalances, which played a really important role in their respective performance.
source: IMF
source: World Bank, own calculations



Here too, Greece's imbalances were huge. It's more than sufficient to say that if one summed up the respective deficits of Argentina, Indonesia and Thailand in the pre-crisis year, the resulting deficit would still be smaller than that of Greece. What's more, Greece's fiscal deficit grew about 50% larger the following year, magnifying further the level of the fiscal adjustment needed as well as the negative drag on the country's GDP.

Moreover, due to the composition of its economy Greece was not as good positioned as the three other EMs to substitute internal growth with external growth. During the 1960 - 2015 period, Greece posted a balanced trade account only once and that was in 2015. Calculating the cumulative trade balances for all four countries for the 1960 - 2015 period makes obvious that, in this respect too, Greece is an outlier. 
source: World Bank, own calculations


If one wants more evidence that an actual currency devaluation didn't induce Greek exports to perform she/he need look no further than the 80s. During that particular decade the Drachma depreciated by about 66%. According to plain vanilla economic arguments, Greek exports should have posted some stellar performance. Reality though was a bit different. Greek exports during the 80s performed worse than they did during the 00s (until 2008 that is, before the Great Trade Collapse) under the "hard" Euro. And this goes some way to show how simplistic arguments that equate currency devaluations with exports' overperformance are.


source: AMECO, own calculations
source: AMECO, own calculations






To wrap this up, I think that all the arguments outlined above prove that Greece was not only an outlier as far as its performance after the sudden stop episode is concerned but also and more importantly, in its fundamentals. Greece's imbalances were significantly larger than those of the other three EMs examined in this post and these imbalances would also take significant more economic pain in order to be ironed out. So even if Greece had reverted back to the Drachma in 2010, the ensuing contraction would still be much larger than that observed in other sudden stop episodes.