Suppose the Bank of Canada pegged the price of gold. If an increased demand for gold caused a recession, would macroeconomists be told they needed to pay more attention to the theory of the demand for gold? Or would the Bank of Canada be told to change its policy?
Suppose the Bank of Canada pegged the exchange rate to the US Dollar. If a fall in the demand for Canadian-produced goods relative to US-produced goods caused a recession in Canada, would macroeconomists be told they needed to pay more attention to the theory of international trade? Or would the Bank of Canada be told to change its policy?
Suppose the Bank of Canada pegged the money supply growth rate. If an increased demand for money caused a recession, would macroeconomists be told they needed to pay more attention to the theory of the demand for money? Or would the Bank of Canada be told to change its policy?
In each of those three cases, the macroeconomic importance of some particular sector of the economy is merely an artefact of a particular monetary policy.
Is the macroeconomic importance of finance, likewise, merely an artefact of a particular monetary policy?
What lesson should we learn from the recent recession? Is it that macroeconomists should pay more attention to finance? Or is it that we should change monetary policy?
In the past, the lesson we learned is that we should change monetary policy. When fluctuations in the demand and supply of gold became macroeconomically important, we dropped the gold standard. When fluctuations in the trade balance became macroeconomically important, we dropped fixed exchange rates. When fluctuations in the demand for money became macroeconomically important, we dropped money growth targeting. When fluctuations in finance became macroeconomically important, what should we do? Why should this time be different?
Now, the analogy is not exact. The Bank of Canada does not peg a nominal interest rate, except for 6-week periods. It adjusts that nominal interest rate peg eight times a year, as needed, to try to peg the CPI inflation rate. But interest rates, in particular the gap between market rates of interest and their corresponding "natural" rates of interest, play a key role in monetary policy tactics. And the monetary policy strategy, targeting 2% inflation, can be seen as pegging the real rate of interest on currency at minus 2%, which is the interest rate differential between holding currency and holding the CPI basket of goods. Current monetary policy uses one interest rate differential as an instrument to target a second interest rate differential. Current monetary policy, in both tactics and strategy, is all about interest rate differentials. And the theory of interest rate differentials is finance. A financial crisis is a big sudden change in interest rate differentials.
It doesn't have to be this way.
In the past, we changed monetary policy to make the demand for gold, the trade balance, and the demand for money, macroeconomically unimportant.
We should now change monetary policy to make interest rate differentials macroeconomically unimportant.
Worthwhile Canadian Initiative 
JKH said: “And M1 money supply targeting used an interest rate instrument to control money supply growth.”
Canada has a 0% reserve requirement. Let’s say monetary base stays the same, but M1 goes up.
That’s mostly about interest rates, right?
Nick,
I think the option exists to switch from strategic interest rate differential targeting (money rate/CPI) to something different (money supply, gold, FX, land) but I don’t think there’s an option to switch from tactical interest rate differential targeting (policy/natural) to something different – because central banks require the latter for balance sheet management purposes. When they hit the zero bound, they extend term on tactical targeting via forward guidance. I think QE is a form of tactical adjustment, but it doesn’t replace tactical differential targeting at the core. I think we know this because the CB still sets the policy rate. The CB adds QE to the zero bound policy decision instead of going negative on the policy rate.
Peter N,
You seem to argue that the risks involved in lending can never be foisted entirely onto the private sector. I don’t agree. The private sector carries the entire risk when it comes to the shares and bonds issued by corporations. That is, if the stock market collapsed to a quarter it’s present value, I sincerely hope government would not come to the rescue of the asset rich who hold those shares.
Of course if that collapse was part of a wider recession, then I accept it would be government’s duty to maintain aggregate demand using monetary and/or fiscal means, but that’s a separate point.
As to banks (i.e. “corporations that specialise in lending”), I see nothing wrong with private sector shareholders carrying the entire risk. Indeed if they don’t, then taxpayers do carry part of the risk, and that’s a subsidy of banking, and subsidies misallocate resources.
Nick:
I disagree that inflation targeting is about finance because it is possible to characterize it as targeting the real interest rate on hand-to-hand currency at minus the targeted inflation rate.
Now, the modern Friedman rule, where the inflation (deflation) rate has to make the real interest rate on currency equal to the natural rate does tie the target back to finance. But the other way around, where the real interest rate on currency is the negative of the inflation rate–no. It is just an artifact of hand-to-hand currency with a zero nominal yield.
Viadas’ argument about collateral doesn’t apply to more recent proposal for index futures targeting–at least not exactly.
If the only way the monetary base can change is buy speculators taking positions on a futures contract, then disruptions in collateral markets might prevent the quantity of money changing.
But if index futures trading is a constraint only, and the quantity of money can change without any trading of the futures contract, then the effect of problems is just less volume. If the monetary authority is making an error, the futures market won’t signal it as well.
At the very least since Sumner has been blogging, he has been proposing that the central bank set base money at the level it expects to keep nominal GDP on target and then adjust that tentative target according to trades of the future. If there are difficulties in obtaining the necessary collateral, then there is just less trading of the future.
I think the same problem occurs with “second moment” issues, but I don’t really understand that very well. With the actual proposals, changes in the risk just impact volume for any expected deviation of nominal GDP from target. The effect is symmetrical. The correction method is less effective. Low (or no) volume leaves base money where the monetary authority expects it to leave nominal GDP on target.
I don’t think the wrong interest rate on bank reserves is relevant to index futures targeting. So what if a central bank sets its interest rates on reserves wrong and takes a loss? What does that have to do with anything important?
Bill:
There is a possibility that disruptions in collateral markets distorts the signal futures are sending.
Regarding “second moment”, consider a scenario that is unrealistic, but serves well as an illustration. Suppose NGDP futures are always on target, but the distribution of probabilities is perverse. For example, there is a 90% probability of overshooting the NGDP target by 0.5%, and 10% probability of undershooting it by 5%.
“So what if a central bank sets its interest rates on reserves wrong and takes a loss?”
Ex-post losses may not be a big problem if the central bank is recapitalized, it may become a problem otherwise. Ex-ante expectation of losses means that the allocation of resources in the economy is not optimal, and the real supply of money is too high. However, the opposite problem is more important, it is a very bad situation when during the financial panic the real supply of money is too low (as a result of interest on reserves that is too low, even though NGDP is expected to be on target).
JKH,
With FX targeting there is no zero lower bound – an FX targeting CB eases by printing cash and buying foreign currency. This can be done without limit (see Zimbabwe). Liquidity can be increased without bound. For what I’ll call a “very open economy” (VOE), easing makes imports become more expensive (raising inflation) and exports become more competitive (export boom leads to more inflation). So where an interest rate targeting CB loses one tool for easing at ZLB, an FX targeting CB never loses its primary easing tool.
On the other hand, an FX targeting CB has a tightening bound. An FX targeting CB must expend foreign reserves to tighten. Since foreign reserves are limited, there is a limit to the amount of tightening a FX targeting CB can perform with its primary tool. Once reserves are limiting, other tools – taxes, retirement accounts, capital controls – must be substituted. Those are rather severe things, so foreign reserves are very important for an FX targeting CB.
Note that due to covered interest rate parity (which does hold in practice), forward points and swap offer rates can go negative. If there are any domestic bank loan agents offering loans based on swap offer rates, they might have their heads explode when they realize their adjustable rate loans now have a negative interest rate. This happened recently in Singapore where SOR based mortgages are commonly available. In practice, the regular banks put in a zero-lower-bound. But this is of no concern to the CB – who merrily continues to tighten.
Ralph,
I’m just opposed to covert subsidies. If we need low cost 30 year fixed loans and cheap college education, we should subsidize them explicitly and not just leave a mess for future taxpayers to clean up.
“Is the macroeconomic importance of finance an artefact of current monetary policy?”
If finance can cause a recession without there being a current monetary policy, then the answer is no.
Vaidas, targeting the mean value of NGDP realizations (which is what NGDP futures give you) is in fact optimal if you have a quadratic loss function. If your preferred loss function is different, you can build a synthetic security that has the appropriate mean, and target that instead. It’s even easier if you care little about small deviations, as then you can just target a fixed corridor instead.
Peter, “However, as the lender of last resort (ie. last man standing), the government is always the final bearer of extreme tail risk.”
I see it like this:
Banks take in demand deposits and loan most of them out long term like 10 or 20 years
At some point banks get into trouble because there is always some amount of withdrawls from the demand deposit accounts that they can not handle. This is called a “Banking Crisis”.
Government bails out the banks.
Government gets too much debt and gets into trouble = “sovereign debt crisis”.
Central bank bails out the government = monetizing debt = printing money.
Currency Crisis
So the currency gets it in the end. To me it looks like much of the world is about to get to that last step.
Vincent, to truly be at this step:
“Central bank bails out the government”
Don’t you think it must be the case that the central bank is obviously overpaying for government debt (or any other asset)? If they’re not overpaying, then I don’t see OMOs as a troubling sign. For example, if the Fed were to stop buying Tsy debt tomorrow, do you think the price on Tsy debt would collapse, or even move much at all?