The piece “Reverse Elasticity Drives Up Deposit Balances” has several errors in basic economics.

First, consumers do not demand deposits—they supply them.  Demand curves slope downward: a falling price is associated with a rising quantity demanded, other things being equal.  Supply curves slope upward: a falling price is associated with a smaller quantity supplied, other things being equal.  In the context of the market for deposits, other things being equal, a lower interest rate on a deposit account is associated with depositors reducing their balances in the account—they supply less deposits.  When looking at depositors, price and quantity desired move in the same direction, so we’re talking about supply.

In contrast, from the perspective of banks or credit unions, we’d want smaller balances if deposit interest rates rise, other things being equal.  Because price and quantity desired move in opposite directions, financial institutions are the demanders of deposits.

Second, “elasticity” does not refer to whether a reduction in price is associated an increase in quantity versus a decrease in quantity.  The concept of elasticity refers to whether the change in quantity is large relative to the change in price (high elasticity) versus small relative to the change in price (low elasticity).  For example, if the price dropped by 10% (let’s say from a deposit rate of 1.00% to 0.90%) and consumers supply 15% less balances (they reduce balances from $100 to $85), we would say supply is elastic, because 15% is greater than 10%.  If instead they reduced balances by only 5% (balances shrink from $100 to $95) we would say supply is inelastic, because 5% is less than 10%.  There's no such thing as “reverse elasticity” in economics (although there are rare cases of upward-sloping demand, such as when consumers perceive a lower price as indicating a lower quality).

Third, the observation that deposit rates are falling while deposit balances are rising is a sure sign that the “other things being equal” qualification isn’t holding.  When talking about a deposit supply curve, we look at changes in the deposit interest rate but hold other things constant.  If other things aren’t constant, we need to look at a new supply curve, rather than look at movements along the original supply curve.  The article talks about things like changes to stock market returns and consumer sentiment, which are classic examples of factors that move you to a new supply curve.  When other things aren’t held equal, we can’t draw any conclusions about a given supply curve or its elasticity.

Fourth, the article has no discussion of how a financial institution should go about evaluating the tradeoffs in setting deposit rates.  Ordinarily, that involves comparing the incremental impact of a rate change on net interest income—how does the change affect total interest costs, and also how does it affect total interest income through its impact on total investable assets.  For example, if we reduce rates from 1.00% to 0.90% and estimate depositors will reduce their balances from $100,000 to $85,000, we would estimate our annual interest expense will drop from $1,000 to $765.  But in order to evaluate that change, we’d need to know the interest income we would have had on that other $15,000 in balances and compares it to the $235 decrease in annual interest expense.

Moreover, through Mr. Peterson’s comments I realized that sometimes it takes more than just an abstract to present a new concept, and therefore, in this write-up, I will provide specific examples to further clarify this concept. 

Let’s start with the basic question of whether deposits are on the demand or supply side of elasticity.  The answer is - deposits are on the demand side because deposits are the enabler of the supply of money to the market though lending and credit.  The main source of confusion, in this context, is that supply, demand and price in the banking industry are a mirror image of supply, demand and price in consumerism.  For example, in consumer economics, money in the hands of a company is an asset, and loans are a liability – exactly the opposite of their classification in the banking industry.  The same applies to price – higher price in consumer economics is the equivalent of lower APY in deposits – both have an adverse impact on consumers’ earning or spending power. 

Now to the main concept presented in the article.  Conventional economics is very rigid – especially the principles of supply, demand and elasticity.  According to the conventional theory of price elasticity of demand, elasticity is one-dimensional because it measures only the ratio between the percentage change in quantity (i.e. balance) and the percentage change in price (i.e. APY).  Nothing else is taken into consideration.  Moreover, since elasticity is expressed in absolute figures (disregard for negative outcome), it is limited in its ability to point to changes in direction, such as we have in the case before us. 

Now let’s see what happens when we apply conventional elasticity theory to an actual scenario of consumer-deposit behavior, and how its limitations prevent us from understanding current events and behavior  Between August 2003 and November 2007 (pre recession), domestic deposits increased from $5.2 trillion to $6.8 trillion – an increase of 30.8%.  At the same time, the national average APY for deposits increased from 1.91% to 4.23% – an increase of 121.5%.  Hence, elasticity is 0.25 – i.e. inelastic.  On the other hand, from December 2007 to December 2010 (during and after the last recession), domestic deposits increased from $6.9 trillion to $9.4 trillion – an increase of 36.2%., however,  the national average APY on deposits decreased from 4.15%  to 0.79% - a decrease of 81% or elasticity of 0.45 -  i.e. inelastic.  So, according to conventional elasticity, nothing changed between these two periods – they are both inelastic.  But that’s no the case – a major event occurred in the pivotal point of December 2007 – a complete reversal of the independent variable (APY), white the dependent variable (deposits) continued to grow.  In other words, conventional elasticity doesn’t recognize nor captures this reversal in the direction of the independent variable.  

Since conventional elasticity measures only two variables, which are indicators, it ignores factors that impact the relationship between the two variables.  Just for the record, in one of his comments, Mr. Peterson mentions stock market returns as a “classic example of a factor” – it is not; it is an indicator.  Factors can’t be observed nor measured.  With the help of factor analysis, we can capture the impact factors have on the relationship between APY and deposits in the example I noted above.  Let’s examine what happened.  In the pre-recession period, the relationship between APY and balances was inelastic, but more importantly, it was “normal”, which means that the two variables related to each other in accordance the principle of price elasticity of demand.  However, during and post recession period, there was a complete reversal in the behavior of the APY (price), which is abnormal, and does not “fit” within the conventional theory of elasticity. 

In lieu of going into detailed explanation of Latent Class Analysis and Structural Equation Modeling, let’s examine a verbatim statement made by a deposits customer when commenting on an article in mainstreet.com on the $1 trillion added to deposits since the start of the last recession in late 2007. 

“We aren't saving so much as not spending (not exactly the same) because we are scared. I might lose my job in a week, so, even though I make good money, I am not going to spend NOR will I invest in traditional vehicles because I need the money liquid and I cannot afford to gamble and lose. I am going to let money pile up in the bank in case I need it, to pay rent and buy food.”Friday, December 03, 2010 

There are two major implications to the new state-of-mind of deposit consumers.  The first is the safety and security of an insured deposit.  In times of uncertainty, people gravitate towards safety.  In the context of deposits, the assurance that the money is insured provides a high comfort level to many customers.  This means that we should expect to see deposit balances grow, despite decreasing interest rates, as long as the economic uncertainty persists. The second implication is the preference of liquid accounts over term accounts.  The knowledge that deposited money can be withdrawn right away without penalty is comforting to people, who feel unsure about their sources of income.  Thus, in a case of a job lose or other economic ill, deposit money can be immediately available. 

In summary, conventional elasticity models were not designed and aren’t capable of capturing these types of factors, and as such, they are limited in their ability to provide us with a realistic assessment of the situation.  Hence, caution should be exercised in using elasticity models to price deposits because the output can be misleading and, needless to say, costly.

Dr. Dan Geller – March 30.2011