The crisis from an AD perspective

(Note: This piece was written in November 2009 as “notes to myself” when I decided to write up a story based on NGDP targeting and see if it fitted the “facts”. I believe the illustrations help in understanding the story. Judge the results for yourselves)

Rarely a day goes by without a new analysis on the “cause” of the crisis. A lot of the discussion on this topic is centered on the importance of the role that the elements on the “most wanted” list of suspects played. Those include:

The excess savings in emerging markets (savings glut),

Expansionary monetary and fiscal policies in the advanced economies,

The securitization of finance,

The underestimation of aggregate risks,

The fall in loan underwriting standards,

The errors made by rating agencies,

Regulatory failure,

The aggressive tactics behind mortgage loans,

The public policies related to homeownership.

Most likely, all the “suspects” share some degree of guilt. However, this is not of much help if the goal of the “finger pointing game” is to provide lessons for the future. For example, if the public policies related to the mortgage market and home ownership had not been adopted, maybe the aggressive mortgage loan tactics, the fall in loan standards, the degree of securitization and even the underestimation of risks would have evolved differently.

From another angle, and this enticed a defense by Bernanke in the recent AEA meetings in Atlanta, many analysts attach a high weight to monetary policy “errors” that would have occurred between 2002 and 2004, when “interest rates were kept too low for too long”. This conclusion follows from examining the Taylor Rule for calibrating the FF rate by the Fed illustrated in figure 1.  Since the FF (Fed Funds) rate remained significantly below the rate required by the “Rule” (FF-TR), monetary policy is interpreted as being “too easy”, contributing to the increase in the demand for mortgage loans and the consequent unsustainable (“bubble”) rise in house prices.

Apparently, as shown in figure 2, this view has merit since there appears to be a close correlation between the FF rate and the 30 year mortgage rate. During 2002 – 2004 the fall and subsequent stability at the 1% level of the FF rate “pulled” the mortgage rate downwards, a fact that may have contributed to the increase in house demand (which was already being driven by the public policies – homeownership incentives – adopted for the sector). 

As figure 3 indicates, the “contribution” of the “easy” monetary policy to the “bubble” in house prices may not have been as significant as envisaged by many. Looking again at figure 2 we observe that the mortgage rate remains low even after the FF rate begins to rise in mid 2004. On the other hand, we note from figure 3 that house prices had been rising for some years prior to the rate reduction and continued to climb after the FF rate began to rise, with the peak in prices being reached two years later in mid 2006.  This fact (FF rates increasing and long rates stable) was codified by Greenspan as a conundrum. One “solution” to the conundrum that became famous was provided by Bernanke in 2005, going by the name of savings glut (meaning the “excess” of savings in emerging markets whose positive current account balance – reserve accumulation – was reinvested in US Treasury Bills), in practice blocking one of the transmission channels of MP.

If figures 1 and 2 constitute evidence in favor of those that point to failures in monetary policy for the development of the “bubble” in house prices, the bursting of which is a root cause of the crisis, figure 3 throws some doubts on this narrative since the behavior of house prices doesn´t change during the period of “low” rates.

Critics accuse Greenspan of being a “bubble blower”. Since this was certainly not his explicit objective, what motivated the Fed to adopt an “easy” monetary policy in the first half of the last decade? To understand the FOMC´s actions we have to appeal to the Fed´s dual mandate: “Price stability (interpreted as a “low” inflation (2%)) and maximum employment”.

Figures 4 and 5, illustrate what the Fed was “seeing” at that moment. Between 2001 and 2003, following the 2001 recession, inflation was below its “target” and falling while the rate of unemployment was in a clear rising trend. In November 2002 Bernanke, at the time a Fed governor, delivered his famous speech: “Deflation: Making sure “it” doesn´t happen here”.

A passage in the speech clearly indicates what was going on in the minds of those responsible for monetary policy: “That this concern is not purely hypothetical is brought home to us whenever we read newspaper reports about Japan, where what seems to be a relatively moderate deflation–a decline in consumer prices of about 1 percent per year–has been associated with years of painfully slow growth, rising joblessness, and apparently intractable financial problems in the banking and corporate sectors”.

Note on the figures that FF only began to rise after inflation reached the “target” and unemployment was on clear down trend.

Figures 4 and 5 sort of exculpate Greenspan and the Fed since the policy implemented was the correct one given the Fed´s dual mandate.

These two opposing views – one based on Taylor Rules, that indicates that monetary policy was “easy”; the other, given the Fed´s mandates, indicates that policy was correct – lead me to think, especially in light of the crisis that manifested itself further down the road, if the objective (inflation target, even if only implicit in the case of the US) and the instrument (FF rate) of monetary policy are robust.

There are calculations of the Taylor Rule indicating that the FF rate should now be -5%. Having reached its zero lower bound, however, the FF rate cannot be manipulated any longer. There remains the perception that at the time it is most needed, monetary policy becomes powerless with “fiscal stimulus” remaining the only alternative to get the economy back on its feet.

On the other hand, the use of an interest rate as the instrument of policy creates confusion regarding the stance of policy. “Low” rates are confused with an “easy” monetary policy (MP) and “high” rates with a tight MP stance. Japan, for example, has kept its “FF” rate close to zero for many years, but Japanese MP cannot be called “easy”. Between April and September 2008 the FF rate was kept at 2%. As experience has shown, far from indicating that MP was “easy”, this “low” rate reflected a “tight” MP stance!

The Inflation Targeting (IT) regime also gives rise to some problems. By definition the IT regime requires that MP be symmetric. If inflation is below target MP has to be manipulated to bring inflation back on target. Conversely, MP has to constrain inflation when it is above the target.

However, while we may think of inflation above target as being “bad”, a fall in inflation below target (or even some deflation) is not necessarily something “malign”. The problem, as perceived from the Bernanke quote above, is that even a moderate deflation can be damaging. This happens due to the fact that he, as well as many others, has in mind the events surrounding the Great Depression when MP errors resulted in deflation and a steep fall in real output (RGDP).

That deflation is not necessarily harmful is illustrated in figure 6 which shows what was happening in the US economy in the last quarter of the 19th century.  While RGDP increased at an average rate of 3.8%, prices fell at 1.3% on average. Note that this is something quite different from the Japanese experience over the last 20 years.

The arguments above may be indicating that an IT regime may not be ideal. In the second half of the 1980´s several researchers, notably Bennett McCallum, with the purpose of devising “rules” for MP, discussed the idea that MP should be geared to maintaining stability in aggregate nominal expenditures (aggregate demand (AD)). This “search for rules” ended up being won by the proponents of the IT regime, with the interest rate (the FF rate in the US case) being the “instrument” of policy.

One of the effects of the ongoing crisis, characterized by an abrupt and steep fall in AD as will be seen below, was to bring back the discussion of what should be the target of MP. To the proponents of the view that the Central Bank should have as its objective the stability of AD, this is the “key” to macroeconomic stability – national or global.

For this group, shocks to AD are the true source of macroeconomic volatility while inflation is just a symptom of those shocks. Just like in medicine, symptoms may be related to different causes, so that prescribing the correct treatment is not trivial. In this sense inflation (a symptom) may be difficult to interpret: is inflation high (low) due to a positive (negative) shock to AD or to a negative (positive) supply (or real) shock? In situations of this kind, the prescribed “medicine” may not be the right one.

Figure 7 provides an illustration. The figure depicts the essence of a basic Dynamic Aggregate Demand/Aggregate Supply (AD/AS) model. The vertical axis denotes inflation while RGDP growth is depicted on the horizontal axis. The AD growth curve AD1 represents nominal expenditures growing at the rate of 5%. Along the AD1 curve the product of inflation and RGDP growth is 5%. The long run AS line LAS1 represents the growth of potential RGDP, which has been about 3% for the US over the last 50 years. The long run AS is vertical to reflect the fact that in the long run AS is determined by real factors. In the short run, however, the AS curve SAS1 is positively sloped due to temporary rigidity of prices and wages.   The dynamics of the level of RGDP and prices is illustrated in figure 8.  In the graphical model (fig. 7) the initial point is at a to which is associated an RGDP growth of 3% and an inflation rate of 2%. Let us suppose that the economy experiences a real, or productivity shock that permanently raises the level of potential RGDP (fig. 8). In this case the economy goes from point a to point b where RGDP growth  increases to 4% and inflation falls to 1% with AD growth remaining at 5%.

To associate the model to real world events, some may recall that in the second half of the 1990´s there was a heated discussion about if the Fed was “behind the curve”. Several analysts, among them Paul Krugman, accused the Fed of being complacent with inflation by not raising the FF rate when RGDP growth was “clearly” above “potential’.

Greenspan would reply that the US economy was going through a period of higher productivity growth (a positive supply shock), that had temporarily increased the sustainable rate of output growth, something that did not call for an increase in the FF rate.

Note that in time the economy returns to point a with growth falling back to 3% and inflation moving back to 2%. The level of RGDP is permanently higher and the price level will be permanently lower.

This will be the outcome if the Fed keeps the growth of AD at 5%. However, if it strives to keep inflation at 2% after it falls to 1% as a result of the positive supply shock, increasing AD growth to AD2 the economy will move from point b to point c. Since at point c the economy will be growing above its potential, inflation pressures will appear so that AD will have to be constrained (FF will increase) in order to keep inflation on target. In this case RGDP and employment volatility will certainly be higher than if the Fed had operated to keep AD growth stable.

Panel 1 below depicts the theory described in the simple model (figures 7 & 8) above. While growth in the US economy was higher than what was considered “normal” by analysts, being  the result of a higher rate of productivity growth, a fact that few other than Greenspan noted, the FF rate was stable and AD growth remained close to its average of a little more than 5%. After 1997, however, when inflation fell below the “target”, the Fed reduced the FF rate. AD went on to grow more than 7%.

At this point RGDP growth climbs to close to 5% while inflation remains below “target”. Note that the trend growth of productivity is maintained over the period, indicating that the fall in inflation was the predictable outcome of positive supply shocks.

From reacting to an inflation below “target” reducing the FF rate in 1998, the Fed than reacts to an RGDP growth considered “exuberant”. So, when the unemployment rate falls below 4% in 1999 the Fed caved in to “popular demand” and raised the FF rate.

Panel 2 describes what happened next.

The main point to note is the strong reversal of AD growth. Inflation remains below the target and productivity growth remains strong. RGDP growth falls towards zero with 2001 being officially defined as a recession year.

AD instability after 1998 is associated with Fed actions. First the Fed reacted to a below target rate of inflation and then, in the opposite direction, reacted to an exuberant RGDP growth rate. This is indicative that an IT regime has an important flaw that manifests itself when the economy experiences supply shocks (also called real or productivity shocks).

Furthermore, we cannot discard the fact that a Fed induced AD instability also has consequences for asset prices. This is shown in figures 9 & 10 where we observe the strong positive correlation between AD growth and the NASDAQ and S&P stock indices.

Note that the NASDAQ more than doubled between mid 1998 and early 2000 while the S&P rose by a little over 30%. This difference in the behavior of the two market indices is related to the fact that the NASDAQ, composed of technology stocks, was also influenced by the “productivity stories” of the time. Thus, what became known as a “bubble” was to a significant degree a reflection of AD instability.

So, if MP were to focus on stabilizing AD growth (level targeting), in addition to preserving macroeconomic stability it would also contribute to avoiding the appearance of “bubbles” in asset prices!

As was observed in figures 4 and 5, following the 2001 recession inflation remained low and unemployment high and rising with both reflecting the robust growth in productivity shown in Panel 2. Panel 3 describes how things unfolded in the 2002-05 period.

Expansion in AD only stabilizes (at the high rate of 7%) when inflation is brought back to “target”, the rate of unemployment is consistently falling and RGDP growth resumes its long run rate of about 3%.

As observed in figure 7 which illustrates the simple Dynamic AD/AS model, if MP is successful in keeping AD growth stable, the system adjusts while volatility remains low. If the Fed manipulates AD, influenced in turn by an inflation below target and then influenced by RGDP growth above potential, the system also adjusts – in the sense that inflation returns to target and growth to potential – but the shocks to AD resulting from these manipulations impart an undesired high level of volatility to the system, also affecting asset prices with spillovers (through wealth effects) to the real economy.

Given its size, events in the US economy have significant impact on the global economy. Figure 11 shows the behavior of AD (nominal expenditure growth) in the US over the last 50 years. It is relatively straightforward to associate this behavior to the periods that became known as “The Great Inflation” and “The Great Moderation”.

Figure 12 shows the similarity of the AD behavior in a group of 24 countries in the OECD. In the 1960´s and 1970´s the focal variable was the unemployment rate. During the Kennedy-Johnson administrations AD shocks, from MP and Fiscal Policy (FP), parted a positive bias to AD growth.

In the 1970´s, characterized by negative supply shocks (notably oil prices), increases in AD to compensate for the negative effects of the shocks on unemployment and RGDP growth resulted in high and rising inflation.

Panel 4 illustrates the “final sprint” to the crisis. As indicated in figure 3, in mid 2006 house prices began to fall. From this moment the delinquency rate began to grow affecting the health of important financial institutions. A few (New Century Financial, for example) went broke already at the beginning of 2007.

Notable, however, is the fact that between 2006 and mid 2008 AD growth was kept relatively stable and close to the trend path. Therefore, in spite of the crisis, the adjustments in the economy were able to proceed in an “orderly fashion”.

Observe in figures 13 and 14 that between 2006.II and 2008.II employment change was mostly positive and the unemployment rate remained stable. Something very different happens after 2008.II when AD “dives”, becoming negative for the first time since WWII. The change in employment is strongly negative and the unemployment rate shoots up.

Figures 15 and 16 help explain why it so happens that between mid 2006 and mid 2008 employment didn´t “dive” and the unemployment rate didn´t “soar”. With nominal spending growth (AD) relatively stable, resources flowing out of the residential construction sector were, among other, relocated to the export and non residential construction sectors.

Only when AD “melts” is that a significant cyclical problem manifests itself. From this moment on the drop in economic activity is generalized and no “compensations” are possible.

It is interesting to note how many analysts see things in a completely different light. In a recent column in the Washington Post, for example, Alan Blinder who was a governor and vice chairman of the Fed (1994-96) and like Bernanke is a Princeton University professor writes that while between mid 2007 and mid 2008 Fed actions fell short, following the Lehman blow-up “the Fed deserves extremely high marks for its work since then. It has hit the bull’s-eye regularly under very trying circumstances”.

When it is said that the Fed regularly got it right at the same time that AD “melts down” something must be very wrong with people’s perceptions.

What is lacking is an explanation for the steep drop in AD. Once again the fact can be connected to the IT regime. Since early 2006 commodity prices were rising. In early 2007 oil prices joined in the fray, with price more than doubling in the next twelve months. Figures 17 and 18 illustrate.

Panel 4 showed that core inflation remained stable (even if slightly above target). The Fed feared that commodity and oil price increases would “infiltrate” the price system and push core inflation upwards.

The crisis gained the headlines in early August 2007 when three funds in Bank Paribas were closed for redemption. In the September FOMC meeting, in a gesture of “appeasement”, the Fed began a process of FF reduction. Nevertheless in every meeting after that the Fed always showed that it was very concerned with the risk of a rise in inflation. In the April 2008 meeting the FF rate was set at 2% and remained at that level until early October, when the “panic” set in.

In the August 2008 FOMC meeting, when commodity and oil prices had already turned down the Fed showed a very hawkish stance. One of the participants voted for an immediate increase in the FF rate and the release after the meeting indicated that the next move in the rate would likely be up! In the September 16 meeting, 24 hours after the collapse of Lehman and a few hours after receiving the news that industrial production had dropped by 1% in August, the Fed maintained the FF rate at 2% and reaffirmed its concern with inflation.

The Fed´s inflation fighting credibility – and the manifestations in the sequence of releases after meetings – indicated that the growth of nominal spending (AD) would be reduced in the future. Just like inflation expectations are an important determinant of present inflation, the same happens with AD growth.

The negative supply shock (from oil and commodity prices) and an IT regime meant that if the Fed strives to keep inflation at the “target” level it will contract AD growth. Given the Fed´s credibility, nobody doubted that that would happen. No wonder it did…

The financial crisis was certainly made worse by the expectation of a contraction in AD, and contributed  to increase the economy´s “melt-down” rate.

I wonder in what kind of world we would be living today if 20 years ago the group that argued for the stabilization of AD growth (level targeting) as the rule to be followed by MP had won the debate over those that proposed IT with the interest rate as the policy “instrument”. Quite likely we would be experiencing much less “fiscal stimulus” with all its attendant risks.

As an end-note, Bernanke knows very well how to conduct MP at the ZLB. In a paper presented in 1999 (check Japanese Monetary Policy – A case of self-induced paralysis in the “Suggested Reading” list) he gave very useful advice to Japan (including “price level targeting”, a variant of AD targeting). One wonders why he did (and does) not take his own advice.

14 thoughts on “The crisis from an AD perspective

  1. Marcus,

    What are yours views on the soaring productivity growth during the Great Recession?

    BTW, you shold try to publish that somewhere else! Spread the word.

  2. Pingback: TheMoneyIllusion » Confusing “history” with numbers

  3. Pingback: NGDP Targeting >> Inflation Targeting « Retronomics Blog

  4. Pingback: Confusing “history” with numbers-Economic News | Coffee At Joe's

  5. Pingback: TheMoneyIllusion » About the GDP numbers

  6. Pingback: About the GDP numbers-Economic News | Coffee At Joe's

  7. Pingback: "It´s insufficient aggregate demand, stupid" – A primer « Economics Info

  8. Hi, Marcus. I would like to respond here to your comment on my post.

    You quoted me: “Before Volcker we had 15 years of worsening inflation. After Volcker we had 30 years of worsening economic growth, attenuated by acts of supply-side desperation and government deficits, leading nonetheless to the system shutdown we call ‘the crisis’.”

    You said “This paragraph is WAY OVER THE TOP” and “it is just VERY WRONG!”

    But you left me wondering exactly what you objected to. First I thought maybe it was when I said we had 30 years of worsening growth after Volcker. (And maybe that is part of it.) But after reading the post that you recommended (your post, above) now I think you disagree mostly that the crisis was the natural result of anything.

    I do think it has been all downhill since the post-WWII “golden age”, and that the crisis was the natural conclusion to that decline. I was only surprised it took so long in coming. (I guess, if you don’t see a decline there, then it would be difficult to see the crisis as its natural conclusion!)

    I liked your post very much; I evaluate (what to me is) the main idea of it in mine tomorrow ( 4 June ). Now let me see if I can defend my view by pointing out what I see as flaws in your presentation …the hobbyist said boldly.

    You open with thoughts on “the ’cause’ of the crisis”, looking for possible “lessons for the future”. You consider “monetary policy ‘errors'” but ultimately question whether “the objective” and “the instrument” of monetary policy are “robust”. Thereafter you point out differences in policy decisions that must arise from “Inflation Targeting” versus “NGDP Targeting”; this I found extremely interesting.

    Following your Figure 8, you discuss the trends of the second half of the 1990s, a time Robert J. Gordon has called a “macroeconomic miracle”. You write that Greenspan said “that the US economy was going through a period of higher productivity growth”. You write that economic growth “was higher than what was considered “normal” by analysts, being the result of a higher rate of productivity growth, a fact that few other than Greenspan noted”.

    You evaluate how the “shock” of this improved productivity was handled by policy and how it might have been handled differently. That’s good, I guess. I just think it might be more important to figure out why that “miracle” happened in the first place.

    After your Panel 2 you write: “The main point to note is the strong reversal of AD growth.” A few pages later you say, “What is lacking is an explanation for the steep drop in AD. Once again the fact can be connected to the IT regime.”

    You do make the connection, but I thought the argument was weak. Responding to Bank Paribas, you write, the Fed lowered rates. But it was still talking tough about inflation. Here is your whole explanation of the crisis and steep drop of AD:

    “The negative supply shock (from oil and commodity prices) and an IT regime meant that if the Fed strives to keep inflation at the “target” level it will contract AD growth. Given the Fed´s credibility, nobody doubted that that would happen. No wonder it did…
    The financial crisis was certainly made worse by the expectation of a contraction in AD…”

    Expectation is the worst excuse. I’m with Keynes, who said people generally expect existing conditions to continue. Your relevant existing condition was that the Fed would raise or lower interest rates to fight inflation. That had not created a crisis for 80 years or something. Why would people expect a sudden collapse? I don’t buy it.

    The other relevant condition is that actual conditions have been deteriorating for a long time. This is the problem. The “expectation” that things will get worse was created by the fact that things were getting worse. So I dismiss “expectation” as an explanation.

    That leaves policy. The up-and-down of interest rates. And “nobody doubted” the Fed’s credibility. But nobody thought the Fed wanted to collapse AD! Again, the problem I think is actual conditions, not how we anticipate those conditions. If you want people to think better of the economy, the productive thing would be to make the economy better. To depend on expectations to make the economy better is circular logic. Just because economists have no better explanation than expectations, does not make expectations a good explanation!

    • “Just because economists have no better explanation than expectations, does not make expectations a good explanation!”
      But they do! I send you to Krugman:
      “Nonetheless, the ECB raised rates — and that was when the euro really began falling apart. The direct effects of the rate increase can’t explain that unraveling, but the effect on expectations — aha, so they really are that fanatical about price stability! — can.”

      • As economic conditions deteriorate, more people grow concerned about the economy. Their “expectations” come to matter more. If expectations matter, it is because things are so bad.

  9. A technical question: How is productivity measured? Wikipedia defines productivity as “output per unit of input” more or less. So my question is: How is the input measured? Is it measured in dollars? For example, suppose we wake up tomorrow and everybody’s income is 10% less than before. So the cost of input is 10% less. Would this be calculated as an increase in productivity?

  10. Pingback: Mais monetarismo de mercado « De Gustibus Non Est Disputandum

  11. Pingback: Monetarismo de Mercado e nós « De Gustibus Non Est Disputandum

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s

This site uses Akismet to reduce spam. Learn how your comment data is processed.