Tuesday, May 6, 2014

Efficient Markets, Rational Agents and Asset Bubbles (Micro-founded Endogenous Money IS-LM)

I first took a stab at Efficient Markets when I attempted to reconcile the opposing views of two Nobel Laureates (Why Both Fama and Shiller are Correct).  Shortly thereafter I worked on Endogenous money IS-LM, which revealed the enormous pro-cyclical influence of money.  A logical next step is to marry the two frameworks and develop micro-founded IS-LM.  In layman terms, this is an attempt to reconcile economic booms and busts with the idea of rational agents.

Asset bubbles represent a fundamental challenge to Efficient Market Hypothesis.  How can markets be always right if asset prices are subject to tremendous swings - rising to unsustainable levels only to drop precipitously when bubbles burst?  Endogenous money IS-LM provides the macro answer - in addition to rational fundamentals, asset prices reflect relative preference for money.  Under a gold standard or central bank with credible anti-inflation stance, excess supply of exogenous monetary base (MB) over endogenous asset money demand (Ma) will fuel an asset bubble. In this post, I will attempt to put forward the mechanism that describes this relationship on micro-level.  The building blocks of this micro-founded model are as follows:

  • Rational agents do not necessarily produce rational macro outcomes when acting as a group.  There is a feedback loop between a decision by an economic agent and the macro economy as a whole.  In other words, our economic decisions have an impact on the macro economy, which exerts an impact back on us.  However, this feedback loop is simply unknowable to individual agents.  Even if a third party were to attempt to close the feedback loop, such information will be ignored because it is pitted against powerful self-interests (i.e warnings of housing bubble went unheeded by real-estate investors, home buyers and banks because there were still plenty of profits to be made in real estate).
  • When it comes to markets, this feedback loop creates a dependency between the decision to take a risk (invest or not to invest) and the return associated with such risk.  In other words, if you choose to invest, you are driving down returns for everyone.  If you choose not to invest, you are driving up returns for everyone else.  The same is true of consumption.  If you choose to increase consumption, you are driving up inflation expectations causing other people to consume more.  If you choose to consume less, you're driving down inflation expectations causing other people to consume less.
  • The two points above can explain micro imbalances. On macro-level such imbalances would cancel out - a bubble in one asset class will depress prices in other asset classes.  An increase in consumption by some individuals will be offset by decrease in consumption by others.  In order to explain macro fluctuations, we have to take into account the supply and demand for money.  Since money is denominated in itself, changes in the value of money affect prices in other markets. If excess money is directed toward asset markets, it will lead to an asset bubble.  On the other hand, if such excess is directed toward consumption markets, it will lead to inflation.  Money is the link between the micro-world and the macro-economy.  As I referred to earlier, Endogenous money IS-LM reveals the true source of macro economic fluctuations as the mismatch between exogenous monetary base (MB) and endogenous asset money demand (Ma).  Chart 1 below is as a vivid demonstration of this relationship. 
Chart 1 (US Data from FRED)

If you are not familiar with any term, please refer to my post on Endogenous money IS-LM.

Real return (R) is equal to growth expectations (g) plus expected asset appreciation (a) less inflation (i).  Time preference (T) is equal to the expected return (R) unadjusted for inflation (for more on time preferences please refer to this post). Time preference and the interest rate (r) are two distinct measures of the same thing - expected future returns.  Time preferences measure those expectations as expressed in the real economy in terms of actual spending and investment decisions.  Interest rates measure those same expectations as expressed in financial markets in terms of the supply and demand for money.

The other components of expected return - asset appreciation (a) and inflation (i) - are both functions of the factor (T-r).  As I've observed repeatedly, because of the existence of exogenous monetary base, time preferences in the real economy do not have to match interest rates in financial markets.  If (T-r) is positive, assets will appreciate.  This relationship is the embodiment of the fact that money is denominated in itself. Changes in the relative value of money are reflected in the prices of other assets.  The same is true for inflation (i) which measures the increase in prices of consumption goods.

The extent to which the factor (T-r) drives asset appreciation as opposed to inflation depends on a number of things.  Income and wealth distribution play an important part with the wealthy having lower propensity to consume.  Also, under gold regime or central bank with credible anti-inflation stance, the factor (T-r) will exert greater influence on asset prices.  Under fiat money regime with no credible central bank, most of the impact will be directed toward inflation.

Change in investment (I) depends on the difference between actual growth (G) experienced in the current period and the expected future growth (g).  However, the factor (T-r) also exerts an influence on the level of investment.  To the extent that this factor causes assets to appreciate, it will also attracts more investment.  Conversely, if (T-r) is negative, asset price declines will also lead to reduction in investment.

Change in consumption (C) depends on the actual growth (G) less the investment factor (g-G). This represents the idea that in order to invest, you have to reduce consumption. However, there is a third factor that affects consumption. If you refer again to the post on time preferences, consumption growth over the last 55 years is greatly influenced by the factor (T-r) adjusted for the difference between future inflation expectations and inflation observed in current period.

The inverse relationship between changes in actual growth (G) and changes in investment stems from the law of diminishing returns.  As investment increases, growth begins to decline. The factor (E) stands for actual technological and entrepreneurial innovation with lower-case (e) representing future expectations for such improvements.  For ease of use, I assume that this factor is fixed (E=e=0%)

Expected growth (g) stands for the rational component of expectations.  It corresponds to the expected productive growth of the economy.  If this model is applied to a specific asset class, g will represent expected dividend yield or income stream.

Next, let's consider the driving force behind the factor (T-r).  Endogenous money IS-LM reveals the significance of asset money demand (Ma) and the exogenous monetary base (MB).  The mismatch between these two monetary aggregates is the reason why time preferences in the real economy (T) do not match interest rates in financial markets (r).

Even more important is the inverse relationship between asset money demand (Ma) and time preference (T), which represents the feedback loop between the macro economy and individual decisions by micro agents.  If return expectations improve due to technological innovation, time preference will increase as well, driving down the demand for asset money (Ma).  Unless the monetary base (MB) declines to match the decline in Ma, time preference (T) will rise above the interest rate (r).  Positive factor (T-r) initiates asset appreciation driving returns even higher and setting the stage for an economic boom.  The reverse is also true.  If investments fail to produce expected return, actual growth declines dragging down return expectations (R). This lowers the time preference (T) as well.  Lower time preference leads to an increase in asset money demand (Ma).  However, if the exogenous monetary base (MB) does not grow to match the increase in money demand, time preference (T) falls below the interest rate (r). Negative factor (T-r) sets the stage for the economic downturn characterized by asset price declines and reductions to spending and investment.

Now, let's consider a couple of scenarios assuming the same starting condition where level of investment (I) produces growth (G) and the technological factor is fixed (E=e=0%).

Scenario 1:  Money Demand Targeting (Monetary Base always matches endogenous asset money demand: MB=Ma). 

Under this scenario, interest rates in financial markets always match time preferences in the real economy, which in turn are all equal to actual growth (G) and growth expectations (g). 

Such condition is associated with stable growth without boom and bust cycles.  Appreciation of specific asset classes is driven by their relative performance compared to the growth in the overall economy.  As a result, asset bubbles on macro-level are impossible - lofty expectations in one asset class will be offset by lower expectations in other asset classes.  Any change in growth expectations (g) will be offset by an equal change in interest rates.  If new technologies require different level of investment, such change will be offset by change in consumption.

Scenario 2. Inflation targeting by Central Banks (MB > Ma)

Scenario 1 above is only hypothetical because central banks do not target monetary aggregates.  Instead, scenario 2 is the true representation of the real world where central banks target 2% inflation. 

For a somewhat jovial explanation of why central banks target 2% inflation please refer to this post.  Basically, low levels of inflation act as cushion against the zero lower bound, which preserves central banks' capacity to influence the business cycle.  This is quite ironic - central banks are using money to fix a problem that was caused by money in the first place.

In order to achieve their 2% inflation target, central banks increase the monetary base (MB) above the level demanded (Ma).  This lowers interest rates below time preferences in the economy.  As a result, the positive factor (T-r) causes assets to appreciate (a>0).  This initiates a self-fulfilling cycle of positive expectations.  Ever higher returns due to positive asset appreciation lead to more investment and consumption which begins to drive inflation expectations higher.  The incremental investment puts downward pressure on growth (G) which drags down growth expectations (g).  At some point lower growth expectations (g) and higher inflation (i) overwhelm the impact of asset appreciation (a), which is the trigger for the bust.

It is notable that under gold standard or central bank with credible anti-inflation stance, excess money supply (MB>Ma) leads to asset appreciation first. The reason for that is simple.  Central banks conduct monetary policy by manipulating money balances held by the public.  Such balances are predominantly concentrated among the wealthy who have low propensity to consume and high propensity to invest in existing assets.  As a result, excess money balances are likely to churn in asset bubbles before having an effect on the real economy.  Only as a result of the wealth effect, would excess money balances begin to flow into new investments and consumption resulting in higher inflation.  The bigger the asset bubble, the longer the time lag between the wealth effect and inflation.  The second irony here is that central banks do not concern themselves with asset prices but rather take their cues from inflation.  This gives bubbles plenty of time to grow as we experienced during the "great moderation" of 1990's and 2000's.

This micro-model reveals the feedback loop between rational micro-agents and the macro-economy, a dependency embodied by the inverse relationship between time preference (T) and asset money demand (Ma).  Endogenous money IS-LM shows that this relationship is at the heart of macro fluctuations because central banks do not manage the monetary base (MB) aggregate, but rather target inflation.  Because inflation signals are subject to a time lag as I explain above, inflation targeting by central banks is the direct cause for asset booms and busts.  The message is clear - we need a paradigm shift.  Central banks should strive for a balance between the demand and supply of base money (scenario 1).  This is a new market ideal without inflation and asset booms and busts, two distortions which are both  direct result of government policy.

No comments:

Post a Comment