Sunday, August 23, 2015

A Primer on Money Demand Targeting

My work on time preferences revealed that banking and the monetary base may prevent micro-expectations from properly aggregating into macro interest rates. Such condition results in ex-ante disequilibrium between money supply and demand, which gives rise to economic fluctuations, inflation as well as asset booms and busts. This line of reasoning naturally led to the question of monetary policy. According to these findings, central banks should adopt a money-demand-targeting rule which calls for 0% inflation target and a velocity peg. For that purpose, central banks need new tools in the form of consumption and investment tax credits. Such credits can raise or lower time preferences of private agents thus affording central banks direct control over money demand and by extension, the velocity of money. There are several specific benefits to Money Demand Targeting:
  • Central banks can achieve stable economic growth with neither inflation nor asset bubbles.
  • Monetary policy will no longer be constrained by the zero lower bound.
  • Countries experiencing trade deficits will have capacity to direct foreign capital flows into productive investment as opposed to unsustainable consumption and housing bubbles.
  • Countries experiencing trade surpluses will be in position to re-balance their economies toward domestic consumption and investment.
  • Money Demand Targeting can also bring particular benefits to the Eurozone as it can substitute for fiscal union and enable the ECB to customize monetary policy to the specific conditions in each member-state.

Money demand targeting is a new monetary policy regime whereby the central bank seeks to balance the supply of base money with asset money demand. Asset money demand stands for long-term savings held in the form of money by private agents who expect declining or stagnant incomes. As such agents have negative time preferences, they are willing to hold their long-term savings in the form of money regardless of interest rates[1]. Banks are not in position to fully accommodate this component of money demand with bank money as the supply of borrower liabilities, which is the raw-material for money creation by banks, is inversely related to the time preferences of private agents[2]. In other words, when time preferences decline, asset money demand goes up while the demand for borrowing goes down constricting the supply of bank money. 

It is only through the exogenously-supplied monetary base that fluctuations in asset money demand can be fully satisfied. A shortage of base money results in excess desired savings causing the economy to perform below expectations and prices, including those of financial assets, to decline. An oversupply of base money results in excess desired borrowings causing the economy to perform above expectations and prices to increase. Such variance between actual and expected outcomes exerts a reinforcing influence back on the time preferences of private agents thus closing the feedback loop which powers the business cycle.

Money demand targeting ensures that the supply of base money is consistent with private agents’ preferred allocation of their long-term savings between money and financial assets. Under such conditions, the economy performs according to expectations and interest rates match the aggregated time preferences of private agents as of the respective future term point. The general price level is unchanged and financial assets are subject to neither booms nor busts.
The actual implementation of this rule calls for the central bank to target zero inflation [3] and adopt a velocity peg. For that purpose, central banks require new tools in the form of temporary consumption and investment tax credits. Temporary consumption credits increase the utility gains of current consumption while investment credits increase the expected return on investment. Accordingly, both have capacity to raise or lower the time preferences of private agents (negative tax credits would lower time preferences).  Ability to manage time preferences affords central banks direct control over money demand and by extension, the velocity of money.

It is important to note that an economy performing to expectations does not mean full employment. If time preferences are depressed due to ageing demographics, exogenous shocks, mal-investment or unsustainable public or private debts, the economy can experience a persistent output gap and unemployment. In such instances, central banks can deploy the new tax credits to raise time preferences and set the economy on a path to full employment. For example, at the zero lower bound agents with negative time preferences are indifferent between holding their long-term savings in the form of money or financial assets.  Accordingly, monetary base expansion cannot restore growth.  On the other hand, temporary consumption and investment tax credits can lift time preferences back into positive territory and enable the economy to break free from stagnation.

Tax credits have traditionally been the domain of fiscal policy, but clearly they have significant monetary effects in terms of the supply and demand for money.  Accordingly, it is not unreasonable to propose that the body charged with managing the money supply should also control the means for money creation as well as the incentives that determine money demand.  Furthermore, lack of coordination between fiscal and monetary policy can detract significantly from the government's response to macro-economic conditions.  Vesting the central bank with limited counter-cyclical fiscal authority will go along away toward improving the effectiveness of such response and mitigating the associated side effects such as inflation and asset bubbles.

The theoretical framework underlying money demand targeting explores the aggregation of micro-agents’ decisions into macro-outcomes. The critical insight is that banking and the exogenous nature of the monetary base[4] prevent borrowing and saving decisions by micro-agents, be they rational or irrational, from properly aggregating into macro interest rates. The model defines time preference as the expected change in utility gains due to delayed consumption. An agent’s time preference is a function of her expected growth in real income and wealth. Such time preference along with the interest rate as of the respective future term point determine whether the agent will choose to borrow or save. However, as banks can meet the demand for borrowing by creating bank money and savers have the option to save by holding base money[5], there is no requirement that the aggregated time preferences of private agents should equal interest rates and by extension, that ex-ante desired borrowings should equal ex-ante desired savings. Only when the supply of base money equals asset money demand[6], can central banks prevent such ex-ante disequilibrium and ensure that the economy performs according to expectations as measured by the aggregated time preferences of private agents[7].
This framework can provide a consistent explanation of the observed historical record spanning the gold standard including the Great Depression, the dollar standard under Bretton Woods, the period of stagflation in the 1970s, the emergence of the dollar as the dominant global reserve currency in the 1980s, the asset bubbles under inflation-targeting in the 1990s and 2000s and last but not least, the experience at the zero lower bound in the aftermath of the Great Recession.

To receive a copy of the working paper titled “Time Preferences, Interest Rates and Money Demand Targeting”, please contact the author at p.valerius.h at

[1] As interest rates cannot fall below zero, holding money even at 0% interest is guaranteed to produce higher return than the required return of such agents per their negative time preferences.

[2] Banks can meet this component of money demand with bank money to the extent they go long on duration. However, banks’ appetite for duration is very pro-cyclical plus bank regulators require banks to duration-match their assets and liabilities as to mitigate interest rate risk.

[3] Lack of inflation indicates zero aggregation error assuming that domestic agents perceive domestic currency as a monetary reserve capable of satisfying their asset money demand.

[4] Under the gold standard, base money was exogenously determined by the available supplies of gold. Under fiat-money, central banks supply base money according to an exogenously-set interest rate target. Such target may or may not take into account all endogenous processes taking place in the economy.

[5] Savers can save by holding cash or bank deposits resulting in excess reserves at the Central Bank. Both cash and reserves are components of base money.

[6] This assumes that banks are perfectly duration-matched.

[7] This framework can also be considered a re-interpretation of the Quantity of Money Equation whereby velocity is a measure of ex-ante money demand and the causality in the ex-post QP=MV identity flows from MV with prices and employment being the residuals.


  1. I'm not sure I understand all of this but one objection that comes to my mind is that ultimately, time preference of economic agents are based on after tax, after inflation consumption. Distorting time preference and thus after tax yield curves with new taxation tools is not a very good way to proceed if you want to allow agents to efficiently negotiate the term structure of their consumption and investment. It muddles the negotiation with guesses on what different tax levels are going to apply over the future horizon and makes it difficult to achieve optimal decisions.

    Running a monetary policy that allows explicit yields to be negotiated for different terms relative to a predictable and stable level of tax and inflation should allow agents to better optimize the consumption and investment patterns over time. Of course this is subject to having inflation targets kept high enough that interest rates never hit the zero lower bound. Other than that, intertemporal decisions and negotiations between different actors in the economy including allocation decisions about the optimal amount and the optimal time to project maturity of private and public spending and investment should be more transparent and efficient under stable or at least predictable taxes and inflation.

    1. Great points - you are correct that time preferences are based on the projected utility of after-tax, after-inflation consumption (which, by the way, is a function of agents' expectation of future incomes). Let me try to respond to the issues you raise.

      The goal of Money Demand Targeting is that interest rates match time preferences - under such condition desired savings equal desired borrowings and the economy performs according to expectations with neither inflation nor asset booms and busts. The problem is that sometimes time preferences are negative (when agents expect declining incomes). Under current monetary policy even with interest rates at zero, agents are unable to negotiate the term structure of their consumption efficiently because there is an excess number of people who want to save today and consume tomorrow and not enough people who want to consume today and save tomorrow. This results in the savings paradox: as more people want to save, incomes decline and the economy produces less savings. The proposed tax credits can lift time preferences back into the positive such that time preferences equal interest rates, which ensures that for every saver there is a willing borrower and vice versa.

      As to your second point on stable tax/inflation level. Let's assume that under inflation targeting, you have stable tax and inflation level and the observable yield curve allows agents to lock-in their intertemporal choices. However, you will still have uncertainty of future incomes hence you have people who borrowed excessively in the past or others who continuously save despite steady or growing incomes. Having flexibility on taxation can offset fluctuations in incomes thus mitigating the associated uncertainty.

      The ironic thing is that it is precisely the 2% inflation target that introduces income volatility. Prices rise when interest rates are below time preferences. Accordingly, central banks have to lower interest rates below time preferences such that they can meet their 2% inflation target. Problem is that this causes excess borrowing and inflates asset prices. In other words, over the last 25 years central banks eliminated one source of volatility (inflation) but introduced asset booms and busts and excessive debt levels.

  2. It's an interesting approach that could work in theory, and it would be way better than allowing the economy to hit the ZLB like with the way the current western world monetary system is managed. But I still don't see the advantage over allowing sufficiently negative real rates when time preference go low by simply having high enough inflation targets.

    If you do it with tax credits the difficulty becomes determining the right mix of consumption and investments of different term lengths. With an explicit an transparent yield curve relative to stable prices and taxes, the mix negotiates itself.

    For example, if time preferences are negative, with Money Demand Targeting, the correct tax credit mix would probably be skewed towards the future with little or no consumption tax credits and more of investment credits with maturity relatively far in the future. That is, when people want to work now and consume later (such as a cohort near retirement), you want to promote economic activity in projects that can be built now but provide value for a long time in the future. I'm not sure what the exact mix should be however, and how you would go about to figure it out.

    I would argue that high asset prices in times of low or negative time preference are not to be avoided. They don't have to incur excess volatility if we can stay away from the ZLB and they are simply a market reflection of those time preferences. They just mean that people put a lot of value on things that can store value for the future, that allow them to consume the fruit of their labor later rather than now.

    The high prices for assets that maintain value, spur more creation of these types of assets and allow markets to fill the needs for future consumption. Sure you can probably achieve the same result by giving tax credits to projects that create value in the future, but again, how are you going to get the maturity structure right without relying on a market with a transparent yield curve? Why even bother to try to figure it out if you can let the market solve the problem easily and simply with a high enough inflation target?