Thursday, May 26, 2011

Forecasts: What and How Do Business Economists Think?


The WSJ and Philadelphia Fed surveys of economists were released last week. It’s of interest to consider what they imply for the macro outlook, and additionally, how they believe inflation will evolve as a function of other variables.

The Macro Outlook
Because the WSJ and SPF forecast mean are essentially the same for GDP, I’ll focus on the WSJ forecasts. Figure 1 depicts the forecast mean, and trimmed high and low forecasts (where trimming is based on the five quarter growth rates).
wsji1 economy
Figure 1: GDP (blue), WSJ forecast mean (red), and trimmed high (Lavorgna/Deutsche Bank) and trimmed low (Leamer/UCLA) (gray), all in bn Ch.2005$, SAAR. Trimming removes top and bottom five respondents. NBER defined recession dates shaded gray. Source: BEA, 2011Q1 advance release, WSJ May 2011 survey, NBER, and author’s calculations.

Forecasters predict continued growth. However, there is some dispersion of forecasts. Moreover, while growth is predicted to continue, it will not be at such a pace to quickly close the output gap.
wsji2 economy
Figure 2: Log GDP (blue), WSJ forecast mean (red), and trimmed high (Lavorgna/Deutsche Bank) and trimmed low (Leamer/UCLA) (gray), and potential GDP (CBO January 2011), all in bn Ch.2005$, SAAR. Trimming removes top and bottom five respondents. NBER defined recession dates shaded gray. Source: BEA, 2011Q1 advance release, CBO, Budget and Economic Outlook (January 2011) data, WSJ May 2011 survey, NBER, and author’s calculations.

Figure 2 indicates that by 2012Q2, forecasters are projecting output at 3.8% below CBO projected potential GDP (in log terms). The trimmed high is 3% below, while the trimmed low is 4.5% below. Even a 3% output gap by mid 2012 is substantial, and suggests to me that policymakers need to be extremely circumspect about tightening policy over-rapidly.

The graph is useful in reminding us of the cost of the recession, which started in 2007Q4. As of 2011Q1, the cumulative output shortfall relative to potential GDP was 2.1 trillion Ch.2005$. Using the WSJ mean forecast, as of 2012Q2, the cumulative output shortfall will be 2.9 trillion Ch.2005$ — and the output gap will still be 3.8%!

These forecasts are conditional upon certain policy measures. One of those is monetary policy; here it is of interest to note what monetary policy is assumed to do.
wsji3 economy
Figure 3: Fed Funds (red), WSJ forecast mean (red squares), ten year constant maturity (blue), and ten year note yield (blue triangles), all in percentage points. NBER defined recession dates shaded gray. Source: St. Louis FREDII for interest rates, WSJ May 2011 survey, and NBER.

What is interesting to me is the fairly gradual upward trajectory for the ten year interest rate — and how those projected interest rates compare against those earlier in the decade.

I can understand how some people might ask how interest rates can be so low with such a large budget deficit. But in a loanable funds framework, saving and demand for total credit determines the price of bonds, and as long as private demand for credit is depressed (consistent with a 3% output gap), real rates should remain relatively low. Shocks to risk appetite could also induce flight to US Treasurys.

An alternative interpretation of these rising interest rates is that inflation is expected to rise, despite the fact that Treasury-TIPS spreads and other measures of expected inflation [0] exhibit muted pressures. The short to medium term inflation expectations are also muted in the WSJ survey:
wsji4 economy
Figure 4: Actual CPI y/y inflation (blue), WSJ forecast mean (red squares), and trimmed high (Riding,DeQuadros/RDQ) and trimmed low (Harris/UBS) (gray +), all in percentage points. NBER defined recession dates shaded gray. Source: St. Louis FREDII for interest rates, WSJ May 2011 survey, and NBER.

Inflation Dynamics

One interesting question, given the pervasive (among some circles) belief that hyperinflation is just around the corner, is what determines inflation. The Phillips curve posits current inflation is a function of expected inflation, the output gap, and input price shocks. 

π t = π et + f(yt-y*t) + Z t

Where π is inflation, the e superscript denotes expected, (yt-y*t) is the output gap, and Z is a function of the growth rate of input prices. For average inflation rates close to zero, the expected inflation term can be approximated by zero.

The WSJ survey does not contain an estimate of the output gap, but one can take a look at how forecasted inflation rate over the next year correlates against the forecasted growth rate of GDP:
wsji5 economy
Figure 5: Average forecasted y/y inflation versus average q/q annualized growth, all in percentage points, excluding James Smith/Parsec Financial Management, n=53. Nearest neighbor fit (bandwidth=0.3). Source: WSJ May 2011 survey, and author’s calculations.

Running a regression of average forecasted inflation against average forecasted growth, and the change in the average forecast oil price from $100 leads to the following estimates.

π t = 1.05 + 0.53 × Δ yt + 0.176 × ΔPoilt
Adj-R2 = 0.32, SER = 0.052, d.f. = 52.

Where Δ yt is the average q/q growth rate, and ΔPoilt is the change in the average price relative to $100, and bold coefficients are statistically significant at the 10% msl. In words, most business economists believe more rapid growth is associated with higher inflation. It’s possible that it’s monetary policy that is believed to drive both growth and inflation jointly (of course, that would be inconsistent with what has often been characterized as a Keynesian view of the world, but consistency and familiarity with data is not a strong point amongst those who are most worried about hyperinflation [1]).

Proxying the looseness of Fed policy by average Fed funds rate in 2011, one finds there is no link of looseness with higher inflation in 2011-2012. In fact, it is the reverse; in a OLS regression, higher average expected inflation in 2012 is associated with tighter policy in 2011, significant at the 5% msl (and with higher increase in oil prices). (The adj-R2 = 0.12.)
wsji6 economy
Figure 6: Average forecasted y/y inflation in 2012 versus average average Fed funds rate in 2011, all in percentage points, excluding James Smith/Parsec Financial Management, n=53. Nearest neighbor fit (bandwidth=0.3). Source: WSJ May 2011 survey, and author’s calculations.

In other words, expected inflation does not appear to be primarily a function of loose monetary policy. Rather, it appears to be driven by factors consistent with a Phillips curve relationship obtaining.

Views on Policy

Economists in the business sector tend to have a view of the world consistent with an expectations and supply augmented Phillips curve, and inconsistent with a strict monetarist/Quantity theory view (or a strict real business cycle view).

It’s of interest, then, to consider what their views on the policy outlook are. Figure 7 highlights the monetary policy outlook.
wsji7 economy
Figure 7: Quarter in which Fed begins raising the Fed funds rate (blue bars) and when the Fed begins exiting quantitative easing by allowing “mortgage-backed securities to mature without being reinvested.” Source: WSJ May 2011 survey.

The modal quarter for QE exit is 2011Q4, while that for Fed funds tightening in 2012Q1. I can see how these dates can be rationalized within the context of a sustained, albeit moderate, recovery in GDP. However, I worry about overly rapid tightening against a backdrop of ill-advised over rapid tightening of fiscal policy.

This is particularly important to recall, in this time of fears of debt accumulation, that much of the accumulation of debt as a share of GDP occurs because of Bush era fiscal policies and the economic downturn, as highlighted by the CBPP:
wsji8 economy
Source: CBPP.

One can see that a large chunk of the debt accumulation is attributable to the 2001 and 2003 tax cuts. The economic downturn is another key contributor.

As Aizenman and Pasricha observed, the fiscal stimulus merely offset the Contractionary effect emanating from the state and local government spending cuts and tax increases. The proposals to cut spending out of the next fiscal year’s budget, without addressing out-year spending and revenue, will merely increase the dark blue component (“economic downturn”) in the above graph.

The WSJ economists (not a notably liberal group, when it comes to economics) also do not appear to be strong adherents of the “expansionary fiscal contraction” view (see my views here and here). In the March survey, the response to the question “Will cutting the federal budget by an annualized $100 billion this year help or hurt economic growth over the next two years?”, was roughly 50-50. My favorite quote was “Claims that cuts are stimulative in the short run are nonsense.” I think we should take this comment to heart, as we wonder if oil prices and other shocks might push us below “stall speed”.

See the original article >>

No comments:

Post a Comment

Follow Us