The Myth of IMF Macroeconomics

“… the Fund obviously has program expertise, expertise of dealing with crises, an expertise in seeing how the things add up and fit together… I do think that program design overall has benefited significantly from us working as a Troika and I know this view also shared by our two partners.”

Poul Thomsen, Greece Mission Chief, International Monetary Fund, April, 2012.

“…in retrospect, it was a very good thing that the IMF came, because they had the technology, they brought money, but first of all they had the technology of dealing with this kind of situation that the Europeans didn’t have…”

Jean Pisani-Ferry, Bruegel, 2012.

How many lies and how much gibberish are we expected to tolerate?

Announcing earlier this year a new “integrated policy framework” approach to emerging markets, International Monetary Fund (IMF) Managing Director Kristalina Georgieva suggested in a Financial Times Op-ed that the IMF’s current macroeconomic framework is inadequate as it is “grounded in more conventional economic thinking.” Thus, the new framework would unshackle the understanding of macroeconomic policy for EMs unlike anything employed before.

However, there is a problem. Georgieva’s understanding of IMF macroeconomics is nonsense—though she cannot be expected to know this. In fact, the IMF’s financial programming framework for thinking about macroeconomic challenges facing members is anything but “traditional.” Indeed, were it adequately understood, the IMF already has a framework for understanding EM policy challenges, and it’s anything but traditional or mainstream. In fact, it is closer to the Post- than New Keynesian tradition. That no one on staff corrected Georgieva on this point reflects how ignorant they remain about the macroeconomic tradition that they inherited—now lost to posterity. And for this same reason, the IMF’s World Economic Outlook forecasts continue to be based on nonsense. Hence there is no way of understanding global macroeconomic tendencies from IMF forecasts.

But of perhaps greater concern is the general belief, expressed in the two opening quotes, that the IMF indeed employs some particular and special macroeconomic framework at all. In fact, IMF macroeconomics is a myth. It is a widely held belief that the Fund employs a “financial programming framework” for assessing and calibrating macroeconomic outcomes—when in fact the IMF’s traditional framework has been all but forgotten. This myth therefore creates macroeconomic car crashes—as in Latvia, Greece, and Argentina to name but three over the past decade. And these accidents matter. Livelihoods—and lives—are lost because of them. Moreover, in the process the IMF is failing to deliver on the Articles of Agreement carved out at Bretton Woods.

***

Let’s go back to the beginning.

For the first decade following her formal opening in 1946, the IMF functioned without a coherent framework for the analysis of members’ balance of payments and macroeconomic challenges. This was particularly problematic in lending arrangements, since the Fund was charged, under her Purposes in the Articles of Agreement, with giving “confidence to members by making the general resources of the Fund temporarily available to them under adequate safeguards, thus providing them with opportunity to correct maladjustments in their balance of payments without resorting to measures destructive of national or international prosperity.”

Absent a framework for understanding open economy macroeconomic outcomes, how could the Fund discharge her key purposes? What safeguards in lending were there? In any case, what are the basic relations between policy levers—fiscal policy, central bank credit and bank lending—and macroeconomic outcomes and objectives? Put in simpler terms, what determines nominal (and real) GDP in a monetary economy open to trade and external financial flows.

It was during a visit to Mexico in the mid-1950s that Jacques Polak—later Economic Counsellor to the Fund—took the initial steps needed to address these shortcomings. With this, IMF financial programming was born—which weaved with conditionality to form the basis for IMF lending in the upper credit tranches.

Central to Polak’s programming approach was the close link between an economy’s monetary accounts and the balance of payments—a relation first elaborated by David Hume about 250 years before. As Polak explained: “The opportunities [to resolve analytical shortcomings in IMF analysis] were provided by the monetary statistics that, from the very start, the Fund had been collecting and had encouraged member countries to compile. From its first printed issue in 1948, International Financial Statistics published these data in the form of a ‘monetary survey’, showing foreign and domestic assets of the consolidated banking system as the main counterpart to the money supply. This presentation was in line with the work by Robert Triffin (who was one of the early staff members of the Research Department) on ‘money of foreign and domestic origin.’”

And so, the Fund leveraged the economy’s monetary-financial accounts—and central bank balance sheet in particular, as the fulcrum of analysis—which, when closely integrated with the other sectors of the economy, generate a general equilibrium understanding of sector linkages, macroeconomic tendencies, and balance of payments.

It’s like a crossword puzzle. While each word (sector) might be “solved” separately, only once the cross-checks on each word are applied—the “cross” letters—can the puzzle be completed. For Polak each macroeconomic sector was initially analyzed in isolation. But in the final analysis the sectors are brought together—the sector-linkages made consistent—to gain an overall understanding of likely macroeconomic outcomes and in setting program targets (on central bank and overall credit to the government and private sector, as well as international reserve assets, for example.)

Crucial to this approach was the “monetary survey” or consolidated accounts of the central bank and deposit-taking institutions. The accounts were consolidated in the sense that central bank credit to the banking sector (liquidity) was netted out, but central bank liquidity remained a crucial drain on reserve assets—hence the central bank balance sheet was presented separately. Net foreign assets of the consolidated banking system—or their change at least—including the central bank, provided the link between the monetary sector and the balance of payments. And so financial structure and flows became the defining feature—the financial sector was not an after-thought, but a foundation stone.

And “money” or financial liabilities more generally could originate from two sources—either “externally” reflecting a current account surplus (income account) or net capital inflows from the rest of the world, or “domestically” meaning the provision of credit to the economy by either the central bank, banking system, or financial sector. And money could be “high powered” in the sense of reflecting an expansion of the central bank’s balance sheet, or “broad” in the sense of reflecting lending decisions of private deposit taking institutions.

For Polak, bank credit became “the ultimate variable” as in the key driver of macroeconomic outcomes—while the stock of money was endogenous in the sense of reflecting a portfolio choice.

Finally, as Pentti Kouri of MIT has suggested, the “balance of payments”—in the sense of the change in central bank reserve assets—becomes the open-economy analogue to liquidity preference. In the open economy, the central bank balance sheet reconciles aggregate portfolio choices through the liquidation of reserve assets—or in the eurosystem through the creation of liabilities to the system through the TARGET2 system.

The monetary survey and central bank balance sheet thus became central to IMF financial programming as reflecting portfolio choices. And once these are further consolidated with the remaining sectors of the economy—fiscal, non-financial corporate, households, and other financial intermediaries—the balance of payments, or transactions with the rest of the world, becomes the residual outcome of monetary actions and aggregated saving-investment choices of the combined sectors of the economy.

This approach evolved through the 1960s and 1970s to become the IMF’s “iterative approach” where sector analysis was combined to uncover and correct internal inconsistencies; where projected macroeconomic outcomes were endogenous to fiscal-financial flows. To wit, IMF macroeconomics relied upon an “iterative calculation carried to the point at which sufficient consistency is obtained between the estimated changes in output and prices, on the one hand, and the calculated value of credit creation, on the other hand.” Or, as Polak recalled in 1997:

“Projected numerical values for the relevant [macroeconomic] variables are found not by solving a set of equations but by making iterative calculations. A particularly valuable by-product of this programming approach is that it forces the analyst to use (which frequently means: to construct) a set of consistent data on the balance of payments, government finance and the accounts of the banking system.”

And these technical aspects of IMF lending evolved alongside the IMF’s policy on conditionality—that being, as Joseph Gold has explained, “the policies the Fund expects a member to follow in order to be able to use the Fund’s general resources.” Indeed, “conditionality attached to the use of Fund resources has been an evolving concept which grew and developed through the practice of the IMF, rather than by explicit definition, agreement, or description in the Articles of Agreement.”

Perhaps the high watermark of IMF conditionality and programming was the late-1970s, when the IMF’s book on the “monetary approach to the balance of payments” competed with a publication by economists at the University of Chicago with the same name.

But the IMF’s iterative approach can perhaps be best summarized with the words of Robert Mundell—the Nobel Prize winning economist—who described the Fund’s programming approach as combining: “Hume’s monetary equilibrium concept, banking theory, Polak’s analysis of the role of money in balance of payments analysis, and an application of general equilibrium techniques to the monetary analysis of an open economy.”

***

While this IMF tradition is widely accepted, it is little understood—including inside the Fund. As a result, it has been overlooked that over the past several decades financial programming within the IMF has fallen into a state of disarray and disrepair. It is difficult to precisely trace this intellectual atrophy. But we can say the following.

The changing nature of balance of payments crises encouraged intellectual atrophy during the 2000s. The IMF’s traditional approach was concerned with balance of payments crises emerging due to, typically, monetization of fiscal deficits by the central bank resulting in an external drain—and loss of international reserve assets. This is equivalent to the “first generation” balance of payments crisis in the academic literature—leading to the quip that I.M.F stands for “It’s Mostly Fiscal.”

But by the 1990s, first in Mexico and more prominently in the Asia Crisis, the nature of the balance of payments crisis changed. Instead of central bank reserve loss being an external sign of an internal imbalance, the crises of the 1990s were led by non-resident flight from domestic assets.

Thus the “second generation” of balance of payments crises made reserve pressure an outward sign of an external loss of confidence. In fact, the correct policy response to a loss of confidence by non-residents should be an increase in central bank credit. However, the initial reaction by economists at the Fund following the Asia Crisis was to tighten domestic credit. And when this failed, they would throw out the baby with the bathwater. “It’s not clear that our theory works…” is how one Mission Chief put it to Paul Blustein in his book The Chastening.

More generally, while there was some important work trying to reconstruct the links between balance sheet structure, sector risks, and currency mismatches—which might be thought of as a “stock-focused” elaboration of Polak’s sector-linkages—the iterative tradition of IMF macroeconomics was quietly forgotten.

By the time I arrived at the IMF in 2005, just prior to GFC, while there was indeed a great deal of lip-service dedicated to the notions of “financial programming” and “macro-financial linkages.” But the majority of country-level Fund work had entirely dropped monetary-financial analysis to underpin macroeconomic tendencies, and often failed to be concerned with financial flows entirely. Unspoken, finance was assumed to respond to the needs to the real economy, and not the other way around. When I started working on Latvia in 2008, for example, I asked the desk economist why the team didn’t build a central bank balance sheet or monetary survey—central to Polak’s vision—to be told it wasn’t needed because Latvia was de facto a currency board! This was also true of countries inside the euroarea—all monetary and financial stocks and flows were relegated to after thoughts. Just as the euroarea crisis was beginning the staff of the IMF were completely incompetent and analytically blind.

In short, there was no analysis of the monetary nature of macroeconomic outcomes whatsoever. None of the pinch points of a monetary-financial economy, the role of credit flows and financial structure, could be teased out—none of the reasons monetary macroeconomics is worth contemplating!

Instead, macroeconomic forecasts were driven by two concepts: trend growth and the inflation target. Small deviations from each were tolerated, of course. But in general, within a few years, the evolution of nominal GDP would hinge on the assumption that both these concepts were realized with complete disregard for financial correlaries. There was no concept of the principle of effective demand, of monetary aggregates or velocity of money; current account deficits would always be financed—just because growth requires it! —; money markets and the “balance of payments” as liquidity preference became irrelevant concepts.

The result, of course, has been a series of catastrophic policy mistakes—policy mistakes driven by analytical incompetence by IMF staff. In the case of Latvia, while it’s a long story, there was no effort to walk Polak’s road to iterative consistency—each member of the team produced their own sector analysis, while there was no effort to ensure overall consistency of the crossword puzzle. Rather, the IMF team created a “financial program” with inconsistencies greater than 10% of GDP—but presented it to the IMF’s Executive Board as some kind of holy script. Doing so, the Fund produced a series of documents that essentially created a currency crisis—even though staff have lied unchallenged about this ever since!

In Greece, of which Poul Thomsen is apparently so proud, again the IMF’s financial program was riddled with nonsense, incompetence and inconsistencies—meaning it was impossible for the authorities to achieve their objectives, for which they would receive full blame, while the Greek economy slipped into a coma. But Thomson, supported by Pisani-Ferry, straight-faced presents this as a reflection of the IMF’s macroeconomic tradition! Meanwhile, in Argentina the IMF provided full-throated support to a monetary Ponzi scheme which likewise did not add up and could not possibly succeed.

Which brings us back to the IMF’s WEO forecasts.

Most of what the IMF’s Research Department produces today is entirely irrelevant to supporting the Fund’s Purposes. But twice a year they have the opportunity to undertake at the global level the iterative process advertised, per Polak, to underpin the IMF’s forecasts at country-level. Yet just as each country team—each Mission Chief—is supposed to undertake consistency checks to ensure sectoral congruence within their economy, to make sure the financial structure and flows underpins the expected macroeconomic outcomes, so ought the IMF’s Research Department undertake that there is consistency of overall saving-investment balances at the global level.

Thus, the WEO iterative process is for the global economy what Polak envisaged of his iterative process for sectors of a single economy. Such iteration would ensure that the aggregate of spending would deliver on the expectations of income across the world. It’s the essence of macroeconomics. Overall spending  for the global economy is equivalent to total spending within individual country forecasts. This iterative process ensures that the overall financial structure and flows globally remain congruent—and help steer policy towards a better outcomes for all.

However, today’s WEO, instead of delivering iterative consistency, is simply an aggregation exercise based on some pre-cooked notion of the correct path for nominal GDP. Lessons learnt? Never. Global consistency? Forget about it. Whether China’s current account surplus is consistent with global deficits is not a primary concern; producing a sense of analytical consistency is. So it doesn’t add up. But everyone is hunky dory because everyone is getting paid handsomely.

I’m not having a go at the latest Economic Counsellor. She has no idea of the analytical tradition she inherited—there is no-one inside the Fund willing or able to pass on this tradition. Hence the latest ridiculous WEO forecasts. And this is equally true of all her esteemed (male) predecessors since at least Michael Mussa. A bunch of Ivy League professors provide cover for IMF staff to produce nonsense macroeconomic forecasts—convenient at the political level, but inconvenient for everyone who lives in the real world who has to bear the burden of this bullshit.

END.