Models vs. Markets (Reply to DTM)


DTM remains unconvinced of my macro narrative, citing a few specific areas where he disagrees with me. From his comments:

I don’t think it is possible to get an answer to a question about the future effect of the stimulus in this way. Too many other factors are confounding market prices, and the stimulus was in the works well in advance of passage, so how does he know what the appropriate baseline market price absent the stimulus would have been? And of course markets can only process information known at the time–they don’t have crystal balls. So that’s just not a real measure.

This is fair enough. However, saying markets “were not impressed” is not the same as saying “markets plummeted”. Ostensibly, markets did not plummet. The DJIA broke 6% during the week the stimulus passed, and commodity prices remained steady. Interest rates did fall, however, on everything from short to long term assets (30-yr fixed); and the TIPS spread didn’t move, suggesting that the ARRA had little to no impact on inflation expectations.


Note: While inflation expectations were rising under the implementation of the TALF, these levels are way too low to intimate recovery under any model.

Remember, I am not claiming that the ARRA didn’t “save or create jobs”, as this is not a fruitful argument. What I am concerned about is whether the ARRA had an impact on forward-looking inflation expectations, which should be priced nearly immediately. In any model with rational expectations, a fiscal stimulus should raise current inflation expectations — and thus should raise output. I prefer market forecasts to model forecasts. I’m assuming DTM prefers the latter. I see the aggregate price levels of markets (percent change from previous period) as signalling that the ARRA did not make a dent in inflation expectations — one could argue that this is because it “wasn’t big enough”…I argue that if a central bank has in mind a specific target, even at the zero bound it will sterilize the impact of fiscal stimulus. Fiscal stimulus just cannot credible commit to a sustained extreme level of spending indefinitely.

Secondly, he comments on my views about monetary policy:

Blanchard’s main point seems to be that setting a higher inflation target and not paying interest on reserves would have been a vastly superior policy. I think he is completely wrong about interest on reserves: the Fed can set reserves levels directly, and that was actually about making sure excess reserves went into it long-term asset buying program, instead of sitting around somewhere.

The bit about setting a higher inflation target is potentially more promising, but at least in that post, Blanchard makes no attempt to estimate the likely magnitude of that effect (a problem Yglesias has had as well). There is every reason to believe it wouldn’t have much effect–that target has to be credible, and that target can’t be arbitrarily higher and remain credible.

While a higher inflation target would have been nice going into the recession (and by recession, I mean 2008:Q3), and an explicit inflation target (3, 4,5 percent take your pick, anything higher than the trend rate of 1.5-2%) would have helped out immensely (a point which I don’t think is controversial among both Keynesians and monetarists). I do not advocate targeting inflation as a monetary policy. I favor targeting a trend rate of nominal output (NGDP), with level targeting.

Turning for a second to his criticism of my analysis of the interest rate on excess reserves…he is correct to say that the Fed can set reserve levels directly. This was actually one of the main problems in 1937-38, when the Fed arbitrarily raised the reserve requirement 1/4%. What is the payment on excess reserves today? 1/4%! Excess reserves didn’t go anywhere, and that is the exact effect that the Fed intended…which is the wrong intent! Hall and Woodford noted this, and Robert Hetzel noted the contractionary impact of policy which increases the demand to hold reserves:

As banks attempted to offset their loss of excess reserves, the money stock stopped growing. Money growth declined after 1936:Q3. Thereafter the level of money fell moderately from 1937:Q1 through 1937:Q4. The level of money remained basically unchanged in the first half of 1938. Money began to rise when banks restored the pre-reserve-requirement level of excess reserves in 1938:Q2. Money then began to rise steadily, basically coincident with the cyclical trough in June 1938 when recession replaced recovery. A chastened Fed retreated from its attempt to again become an active central bank and continued to freeze its holdings of government securities. Monetary base and money growth resumed with gold inflows and the end of Treasury sterilization (Friedman and Schwartz 1963a, Chart 40, and Friedman and Schwartz 1970, Table 1). Because inflation (CPI) turned to deflation in 1937:Q4, the trough in real M1 occurred in 1937:Q4. The return of growth after the business cycle trough in June 1938 is consistent with the increase in real M1 stimulating expenditure through portfolio rebalancing, that is, through a stimulative real-balance effect (Patinkin 1948, 1965).

As you can see in the chart above, the amount of excess reserves skyrocketed in late 2008, coinciding nearly perfectly with the Fed’s decision (October 8th, 2008) to pay interest on excess reserve balances. Ceterus paribus, doubling the money supply will devalue currency by half. Indeed, this is exactly what FDR did in 1933 when he devalued the dollar against gold. Reserves sitting idle are not doing anyone any good (as noted by David Hume in the 18th century). Under recession conditions (like the fabled “liquidity trap”), exchanging reserve balances for risk-free government debt is what Keynes famously termed “pushing on a string”. However, if you eliminate transaction costs of actually purchasing government debt, and instead pay interest on excess reserve balances directly, you incentivize hoarding reserves. Checking on this theory from recent evidence proves strong correlation, and (perhaps weak?) causation.

Scythia asks a question:

2. What do you think about his idea that the energy bubble, and its subsequent impact on the Fed’s inflation policies, is the true cause of the recession, rather than the sub-prime fiasco?

I think this is a problem of wording. To start, I “believe” that the financial crisis was definitely caused by underlying real factors (markets and policy) which led to the first part of the recession (2007:Q4-2008:Q3). After this time, I “believe” causation shifted from financial problems causing growth to slow to tight monetary policy exacerbating financial problems. The Fed’s use of backward-looking measures of inflation was merely the mechanism that caused the Fed to allow monetary policy to become inadvertently to tight for the relative needs of the economy at that point. So yes, I “believe” that tight money caused what we term “the recession” today (everything after 2008:Q3). Of course this is not some crackpot idea. Milton Friedman and Anna Schwartz showed how tight money was the cause of the Great Contraction (1930-1932).

I hope that cleared some things up!

P.S. I apologize for not being able to do any macro modeling. I simply don’t have the time, and (unlike other bloggers like Menzie Chinn or James Hamilton) macro modeling doesn’t intersect with my “real life” (although I’m always free to start *hint!* ;]) .

Advertisements

2 thoughts on “Models vs. Markets (Reply to DTM)

  1. His criticism about reserves is a complete joke. If you want money (aka credit) to flow why in the hell would you pay interest on excess reserves? Seems to me you’d want to do the exact opposite.

Leave a Reply

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

WordPress.com Logo

You are commenting using your WordPress.com 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 )

Google+ photo

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

Connecting to %s