Sorting out confusion
How does monetary policy impact asset inflation? This is a question that is not fully answered in economics, although many practical men of affairs have a pretty shrewd idea that it does. This paper suggests a simple theoretical framework that is crying out for empirical testing.*
Defining monetary policy and the stance of monetary policy.
Is monetary policy best defined as money growth (Friedman), cash rates (Taylor) or Chairman's discretion (Greenspan and Bernanke). I add the latter definition, because both these eminent US Fed Chairmen drove cash rates well below any plausible equilibrium after the bursting of asset bubbles they felt unable to influence.
Imagine a simple diagram with a (downward sloping) demand curve for money and an (upward sloping) supply curve. Cash rate is on the vertical axis and the quantity of money is on the horizontal axis. One would naturally define equilibrium in the money market when the cash rate, set by the central bank, is equal to the equilibrium set by the intersection of the supply and demand curves. When the economy is in a state Minsky called 'tranquility' this equilibrium might be described as the 'natural' rate of interest, or varying only by an 'equilibrium' risk premium.
If the central bank sets cash rates below the equilibrium level, this will require an excess supply of money at the original (equilibrium) rate of interest, and if it is set above the equilibrium there will be an excess demand for money. The natural definition of easy, neutral or tight monetary policy follows. If cash rates are set by the central bank at zero, or near zero, it is reasonable to describe monetary policy as very easy. The key point, however, is the defining monetary policy requires information about both demand and supply curves as well as the degree of deviation from an equilibrium interest rate and the amount of excess or deficient money.
This does not seem to be current practice. Most common, perhaps, comes from comparing current cash rates with past averages of cash rates, or with the rate of inflation, or growth of nominal GDP.
I should add (in response to a request) that the dynamics of adjustment to excess or deficient money at the original (equilibrium) interest rate are complicated. It is simplest if the demand for money is based on 'permanant income' and the equilibrium or 'natural' interest rate (neither of which change as a result of the policy induced shift in the supply curve for money). In such a world the price level may be assumed to change instantaneously to pull (or push) the new supply curve back to its original position.
In reality, the demand curve for money may depend wholly or partly on current income and current interest rate in which case more realistic adjustment will take place. Provided the model has the first set of assumptions as its equilibriun solution, the instantaneous adjustment will eventually return to the initial position as temporary income and interest rate effects fade.
Such a model may be generalised to allow growth in money, income and prices. Hysteresis effects due to a changed equilibrium rate of inflation produced by a different rate of monetary growth can be introduced. But except in simulation of high inflation regimes with serious hysteresis effects this should produce results more or less like the static expositional model
The definition of monetary policy will perhaps best be represented by monetary growth, but all the other moving parts will be relevant to a full analysis. In effect, the only general description of monetary policy requires a full 'narrative', which modern central banks provide even when they have a specific inflation target.
The transmission mechanism for monetary policy.
When this writer was a student macroeconomist 'interest rates' were regarded as the main transmission mechanism. It was believed, probably correctly, that cash rates influenced other rates of interest, and the spectrum of interest rates fed into decisions about household demand for goods and services and business demand for investment goods and services. In times of the so-called 'liquidity trap' this sequence broke down.
Some economists also saw a role for 'money' (as part of or a proxy for) wealth in defining the demand for goods and services by households. Both interest rate and wealth effects were usually difficult to determine in the 'single equation' studies that were the norm, especially in the large models that were all the go in the 1970s.
Other economists also saw growth of money as likely to influence inflationary expectations but, again, this hypothesised effect was not easy to find in simple econometric studies. (See Jonson and Mahoney, Economic Record, 1973).
In studies with others at the London School of Economics and at the Reserve Bank of Australia in the 1970s involving small but conceptually complete empirical models, estimated correctly, and with sensible steady state constraints imposed, this writer found powerful excess money effects on household demand and inflation. This 'monetary disequilibrium' effect (as we called it) was dismissed by economists of the (then dominant) school who believed with almost religious fervour that all markets are always and everywhere in equilibrium. So far as I know the concept has not caught on. (See Jonson, JPE, 1976, Jonson,Foundations, 1976 plus papers on the the RBA's RBII model.)
Stockbrokers and other men (and women) of affairs generally believe that excess money spills over into asset markets. Indeed, when the US Fed has interest rates set at zero, with little effect on demand for goods and services, (is this the famous 'liquidity trap'?), it seems intuitively obvious that asset markets will be influenced. Anyone who agrees with this intuitive judgment may stop reading at this point.
The market for assets.
Now we invite readers to imagine a supply curve for assets, for example shares. It is steep, as assets take time to create and to acquire, but upward sloping. The demand curve slopes down, since presumably the cheaper are assets the more of them the average householder or business will wish to acquire. The vertical axis shows the price of assets and the horizontal axis shows the volume of assets.
So what might be the determinants of the supply of and demand for assets? We know from people who write about 'Animal Spirits', and the history of booms and busts, that powerful confidence is sometimes a prime driver of asset demand. 'Animal spirits' themselves are possibly beyond simple explanation but, looking at the history of booms and busts, several factors come to mind. Gold discoveries, peace rather than war, cheap money, powerful innovation, expansion of new lands, are all factors that history suggests may drive the demand curve for assets to the right. And when asset prices begin to rise sharply, there is likely to be a momentum effect that drives demand further, to a point where the process of price hikes becomes self-reinforcing. Then we have a bubble, which might be described as powerful price hikes that seem far too strong for any rational explanation. This process disturbs Minsky's 'Tranquility' and often ends badly.
So my (perhaps obvious) hypothesis is that some rational situation, discovery or time of excess money kicks off an asset boom and, unless it stops when the rational factor or factors stops acting, irrational expectations are likely to keep the boom going and often turn it in to a bubble.
Under the gold standard, gold discoveries in the 1850s and again in the 1890s, produced share booms, although rarely did they reach the heights of madness typical of the Tulip boom in the seventeenth century or the Mississippi and South sea bubbles of the eighteenth century. Clearly excess money was a factor in the share booms of the nineteenth century, but everyone believed that the Bank of England would eventually end the boom so it was hard for irrational expectations to develop. While the demand curve shifted out after larger than normal gold discoveries, it was eventually reined in when the expansionary process led to loss of gold in the United Kingdom and hikes in interest rates by the Bank of England.
The two world wars of the twentieth century had a clear negative effect on share prices, despite strong increases in money to finance the wars. US experience in the second world war is deeply interesting. There was strong monetary growth throughout the war. In the first half, there was strong goods inflation and weak (falling) share prices. In the second half the US government applied effective goods price controls and there was powerful share price inflation, consistent with powerful money growth but also a growing confidence that the war would end with defeat of the axis powers.
Other even more interesting examples are evident when one considers what this writer calls 'aberrant episodes'. In the 1920s, the 1950 and 1960s and the 1990s, America experienced three episodes of powerful share price inflation while monetary growth was contained by what one must call 'firm' settings. These episodes were all times of powerful innovation and strong growth, a rational reason for share inflation.
'Easy money’ often is due to 'easy credit', especially during the era of financial deregulation from the 1980s. But there can be easy credit without easy money, and indeed this is the most obvious way to explain what I call ‘aberrant episodes’ in USA in the 1920s, 1950s and 1960s and 1990s, and Japan in the 1980s.
(In the 1920s, for example, money growth was under control but the New York banks were borrowing from the Fed at 5 or 6 % and lending to speculators at 10%, 12%, at one point 20 %. No impact on measures of money, but a lovely way for speculators to get rich while share prices rose and to lose their shirts after the crash. This is explained in detail in Galbraith’s book The Great Crash.)
However, in the first and third of these 'aberrant episodes', and in the Japanese asset boom of the 1980s, there was clear anecdotal evidence of the development of 'irrational exuberance' pushing the demand curve for shares and other assets to the right for a substantial period. One might expect monetary disequilibrium to show up in both goods and asset markets, but what needs some thought is why, in some occasions - late 1920s USA, Japan in the 1980s, almost everywhere from 2007. goods prices don't seem to respond. This might be due to strong growth and innovation stimulating demand for money (restraining goods inflation or, in the late 2000s severe recession restraining goods and labor markets, leaving excess money to prime asset markets).
Clearly, however, asset inflation or asset deflation depends on a lot of developments other than monetary policy, including sometimes mysterious fluctuations in 'Animal Spirits'. The absence of wars, provision of plentiful cheap credit, powerful innovation, expansion of new lands, are all factors that history suggests may drive the demand curve for assets to the right. If manifestly 'firm' monetary policy is restraining goods and services inflation, (as it was in the 1950s and 1960s USA) this may be an independant reason for asset demand to be boosted.
Strongly expansionary monetary policy will cause asset inflation unless there is some countervailing force - like a serious war or deep depression in the labor and goods markets. Strongly deflationary monetary policy will prevent asset inflation, or turn entrenched asset inflation into asset deflation.
But if monetary policy is more or less neutral non-monetary events, including credit expansion that does not much impact measures of 'money' and such developments may cause asset inflation, possibly turning into bubbles, or bring on asset deflation. Asset inflation or deflation depends on many things other than monetary policy.
Monetary policy should therefore focus on overall economic stability and control of goods and services inflation. This means that there is a role, even a requirement, for other ways to deal with asset inflation or deflation - 'macroprudential policy' in the latest jargon.
Or, as Milton Friedman put it: 'Monetary policy cannot serve two masters'.
Linking monetary policy and asset inflation.
The link between the two sections is that a number of rational things can kick off an asset boom but, when monetary policy or ‘credit policy’ allows, this can and often does, turn into speculative excess.
‘Money’ is not the only accommodating or inhibiting factor, which is why asset inflation is not as closely correlated with monetary policy as is ordinary (goods and services) inflation.
This perhaps overly long, perhaps not sufficiently analytical paper, is the pudding from which I pulled the approach set out here.
There is a consequence of overly easy money that is vital for the stability of modern capitalism. Zero interest rates and massive 'quantitative easing has massively inflated asset prices. This has made rich people richer, and has little benefit for ordinary people. The great thinkers, such as Keynes and Marx, have seen excessive inequality as likely to damage capitalism. So even without factoring in the likely economic consequences of withdrawing the excessive monetary stimulus, there is a serious issue awaiting resolution.
Comment by Professor David Laidler, 28 July, 2014
Not much to disagree with here, other than one or two points of historical detail too trivial and complicated for an email.
Two points, though. At one stage you seem to argue that interest rate close to zero can usually be read as signifying easy money. I don't think that this follows from your basic argument that what matters is Ms-Md. It's not too difficult to imagine situations - the current Euro zone? - where interest rates are very low, but where Ms growth is far below what is needed to the time path that Md would take were the economy on a path to recovery.
You're right about the distaste for "disequilibrium" ideas among the mainstream. If the economy is always in equilibrium, Ms is always equal to Md, and Say's Law guarantees that there is no information in the behaviour of money that cannot also be derived from the rest of the economy. This, I think, is what Woodford's insistence on the irrelevance of money amounts to. On the other hand, things did come apart in the '80s, because to get a reasonable estimate of Ms - Md, you need a reasonably accurate estimate of Md, and that was really hard to come by at that juncture. The trouble was that central banks ran into this problem, and instead of stopping to cope with it, ran away from it altogether - the baby went out with the bathwater. The BUBA and SNB hung on, but were eventually persuaded to join the crowd around 2006, at just the wrong time.
One point that you probably need to pay more attention to for your readers sake: - mostly when there is monetary disequilibrium it shows up in both goods and asset markets (often first in the latter), but what needs some thought is why, on some occasions - late 20s US, Japan ca. 1990, almost everywhere ca 2007. goods prices don't seem to respond. I know you've thought about this, so I won't go on about it, but lots of people haven't, and they need to.