VoxEU Column EU institutions Europe's nations and regions

When the IMF evaluates the IMF

The IMF has just released its self-evaluation of its Greek lending, in which it admits to many mistakes. This column argues that the report misses one important error – reliance on the Debt Sustainability Analysis – but notes that the IMF’s candour should be a model for the other participants in the lending, namely, the European Commission and the ECB.

The IMF has now released its self-evaluation report on the programme for Greece between 2012 and 2016 (IMF 2017). This report admits most, if not all, of the glaring mistakes and calls for significant changes. Unfortunately, it does not always get to the bottom of why these mistakes were made.

The requirement that the IMF self-assesses and publishes its interventions in the case of programmes with exceptional access was adopted in the wake of the highly controversial East Asian crisis. Exceptional access occurs when the amounts lent by the Fund exceed the normal ceiling of 145% of the country’s quota per year, or a total of 435%. (At the time the limits were 200% and 600%, respectively). A first programme provided for 3200% of the Greek quota. As it was going astray in 2012, it was interrupted and replaced with a second programme worth 2159% of the quota. These are numbers never seen before.

What’s good

The report is candid on a significant number of mistakes. It acknowledges that its forecasts were “overly optimistic”, which justified the front-loaded and historically deep budget deficit reduction condition that created one of the longest and deepest contractions ever recorded. This, in turn, led to a rising debt to GDP ratio, the opposite of the stated goal. The IMF has already admitted this failure, based on low fiscal multipliers (Blanchard and Leigh 2012). The report further notes that export price elasticities were expected to be large and that the banking system was presumed to be healthy throughout.  The report goes further by listing the implications of this mistake, including political turmoil, the rise in non-performing loans (NPL) that undid banks several times over, and failure to meet a number of programme conditions that appear ex post as unrealistic.

Exceptional access can only be granted if the staff certifies that the debt has a ‘high probability’ of being fully serviced. With considerable internal misgiving, the staff obliged. It did so not once, by twice, for each programme. Its assessment was based on its highly optimistic forecasts, which is open to the interpretation that optimism was endogenous to the need to certify debt sustainability. At the same time, well aware that the debt situation was dicey, the IMF supported the deep budgetary measures that it now regrets. The IMF offers two explanations. The first one is that a deep debt-restructuring would have hurt some systemic foreign banks. So soon after the Lehman Brothers collapse, it was feared that the Global Crisis could be reignited. This led the staff to surreptitiously add global systematic stability as a new clause to provide exceptional access. The precedent thus created was explicitly repealed in 2015, effectively acknowledging that it had been a mistake. The second explanation is that the EU and the ECB were staunchly opposed to any form of debt restructuring. This raises even deeper questions of cooperation with a monetary union. The report candidly recognises that it had made no preparation for such an event and that this issue still remains largely untreated.

Several other mistakes are acknowledged. They cannot all be presented here, but a few can be mentioned. One is the failure to recognise early on that the Greek polity did not ‘own’ the programmes, which indeed resulted in backtracking. Another is that the request that a great many structural reforms be designed and implemented exceeded both administrative capacity and political feasibility. Yet another one is a failure to properly monitor bank fragility. In addition, structural reforms were assumed to immediately boost the supply side, in contrast with much accumulated evidence that the benefits come in very slowly. Finally, the report mentions the failure to reform the goods market, in contrast with the labour market – deep labour costs did not lead to significant gains in export competitiveness. More generally, it notes acute problems of governance in both corporations and the government, which have remained largely untreated.

What’s lacking

The frequent references to the co-management of the Greek crisis with the European Commission and the ECB – the Troika arrangement – suggest that this was a complicating factor that led to conflicts. Rather lamely, the report calls for an in-depth search for procedures to deal with a monetary union. Reading between the lines, one can guess that the IMF considers that the biggest mistakes (optimistic forecasts, no debt restructuring, excessively tight budgetary conditions, etc.) were not due to internal analytical failures, but to its position as junior partner in the Troika. The Troika was a historical precedent. Normally, when it lacks sufficient resources, the IMF single-handedly manages its programmes while calling upon ‘friends’ of the treated country to provide the needed additional resources. With the exception of Latvia in 2008, never before had the IMF accepted a junior role. So far, there has been no explicit analysis of this momentous step. Since it goes beyond the Greek case, the report may not be the place to tackle an issue that is politically hot, as the continuing disagreements between the IMF and the EU show.  

The report emphasises the unusual amount of uncertainty affecting the programme. It considers that the programme was extraordinarily risky and that its chances of success were limited from the start. The lesson it draws is that, in such a situation, the conditions required from a country must be gradual and spaced out over a long period. Neither the observation, nor the implications are explained. Of course, many weaknesses pre-existed in Greece, but this is standard fare in most countries that need IMF support. Two factors made Greece special. First, it could not devalue to counterbalance budgetary stabilisation, and second, it did not have access to lending in last resort. Devaluations, however, are not a panacea. In this case, not only would it have aggravated indebtedness, already excessive, but Greece would had had to leave the Eurozone, a momentous move for which neither Greece nor the Eurozone had made the slightest preparation. The proper conclusion, therefore, would indeed have been to be gradual with budgetary stabilisation – but not for ‘risk’ reasons – and to encourage the ECB to fulfil its (implicit) lending in last resort duty. Diplomatically, the report does not touch upon the role of the ECB, its fellow Troika member.

What’s wrong

Two critical mistakes are not mentioned. The first one is technical. In view of the exceptional, but not unique, access situation, debt sustainability was a crucial issue that remains at the heart of current debates. The IMF has developed a Debt Sustainability Analysis (DSA) procedure, which goes as follows. It starts with assumptions about the path of future primary budget balances, interest and growth rates over some 30 years or more. A straightforward simulation then delivers the future path of the debt ratio. The IMF then compares this path with some benchmarks. As with all compounding exercises over long horizons, the results are extremely sensitive to small changes in the assumptions. The debt path, therefore, is no more than the unstable representation of assumptions, as are the benchmarks (Wyplosz 2011). Assumptions over long horizons are highly uncertain, however. It would seem natural to associate the simulation results with confidence intervals. Zettelmeyer et al. (2017) show that the confidence intervals are very wide, casting doubt on whether any policy conclusion can be drawn from DSA. Sadly, there is no hint in the report that the Fund is ready to question this procedure.

It may be true, as US and European policymakers forcefully argued at the time, that a Greek default would have sparked global systemic effects. Preventing the crisis was therefore in the interest of a large number of important countries, which raises the issue of burden sharing. Fairness and feasibility argue in favour of sharing the burden. Instead, as we know, Greece was instructed to borrow, which means that the burden has fallen entirely on its taxpayers. The burden is so heavy that the IMF now calls for a debt reduction, which would be ex post burden sharing. As the world’s benevolent referee, it should have refused ex ante to be complicit and part of such an imbalanced approach. The issue of burden sharing is not even mentioned in the report. The implicit answer is that political realities had closed that door, but then it raises grave concerns about the Fund’s independence. It also reminds us of Keynes’ two famous lost battles. First, after WWI, he had opposed German reparations. Second, in Bretton Woods, he had raised the issue of symmetry. Greece represents his posthumous third loss.


The IMF must be commended for imposing self-evaluation reports upon itself. They sometimes come on top of reports by the IMF’s Independent Evaluation Office (for the report on Greece, see Wyplosz and Sgherri 2016). It is about speaking truth to yourself, which can be delicate because the programme’s actors, most of whom are active in the building, have skin in the game.

These reports can fulfil an extremely important role if they identify mistakes that should not be repeated in the future. Does it happen? A previous self-evaluation took place after the first Greek programme. Many of its observations are the same as those of the second report, which is disheartening. The Fund argues that, because the first report was published after the start of the second programme, its conclusions could not be taken on board. It calls for a faster production of the self-evaluation reports. Would that be enough? Scepticism is warranted when we observe that a number of the mistakes reported in this report were already mentioned after the East Asian crisis.

With all its limitations, the fact that self-evaluation occurs and that the report is made public deserves to be commended. The procedure should be a model for the two other Troika institutions, the European Commission and the ECB. Most regrettably, self-evaluation is not part of their institutional culture. They seem to follow the prescription attributed to Napoleon: “In politics never retreat, never retract, never admit a mistake”. 


Blanchard, O J, and D Leigh (2013), “Growth Forecast Errors and Fiscal Multipliers”, IMF Working Paper 13/1.

IMF (2017), “Greece : Ex-Post Evaluation of Exceptional Access Under the 2012 Extended Arrangement”, IMF Staff Country Reports, February.

Wyplosz, C (2011), “Debt Sustainability Analysis: Mission Impossible”, Review of Economics and Institutions 3(2): 1-37.

Wyplosz, C, and S Sgherri (2015), “The IMF’s Role in Greece in the Context of the 2010 Stand-By Arrangement” in: The IMF and the Crises in Greece, Ireland, and Portugal: An Evaluation by the Independent Evaluation Office, IMF.

Zettelmeyer, J, E Kreplin, and U Panizza (2017), “How high for how long? Examining the “realism” of primary surplus assumptions for Greece”, unpublished, Peterson Institute for International Economics. 

5,354 Reads