Below is a commentary originally available to subscribers at Piscataqua Research, Inc. on 16^th September 2005.

Is M-2 the Canary in the Coal Mine?

Summary: This paper shows how the real estate market effects consumer money supply. We bring together the results of our research in the "Interest Rate Conundrum", "The Housing Bubble-An Update", and "The Sources and Uses of Household Cash Flow" papers.

Home equity loans have become the primary source of M-1 growth. Home resales have become an important source of non-M-1 M-2 growth. This situation is under-recognized.

M-1 is already under liquidity pressure. As M-1 increasingly draws funds from non-M-1 M-2, non-M-1 M-2 will also face liquidity pressure. With home resales an important source of non-M-1 M-2 growth, slowing resales could limit M-2 liquidity.

We wonder: Has the Federal Reserve created a consumer liquidity trap? If lower home resales causes M-2 to decline, we could have an answer.

We never expected our research to lead us where it has. Our observations have led us to the doorsteps of our homes. This analysis brings together the information in our "Consumer Cash Flow", "The Housing Bubble-An Update", and "The Interest Rate Conundrum" papers also available on our website at www.piscataquaresearch.com.

In this analysis, we integrate the effects of consumer cash flow and the real estate market with consumer money supply. The result is as close to an economic "smoking gun" as we can imagine. The fingerprints on the gun are those of the Federal Reserve.

The diagram on page 2 shows that consumer money supply now flows backward. We have a "vacuum" condition where consumption and debt service drains money out of the M-1 money supply faster than the consumer replaces it with income. Consumers must now resort to increasingly large real estate debts to manage their money needs.

This condition started to develop in 1990/1991 as the New England/CA property bubble burst. The deterioration gained momentum after the Federal Reserve set low interest rates in 1992/1993. Money supply flow finally reversed in 1999 after the Federal Reserve again resorted to low interest rates in 1998. When the Federal Reserve went to the low interest rate well from 2001 to 2004 yet again, the backwards flow intensified.

The key to consumer money supply growth is now home resales. The home resale market is a tremendous source of consumer debt and helps fuel the current growth of non-M-1 M-2. If M-2 starts to decline, it could occur because home resales slow substantially and eliminate an important source of consumer money supply creation.

Housing-Related Money Effects

In the first section of this paper, we present a simplified cash flow model for households. In the remainder of the paper, we will show real estate interactions with the money supply. The housing market has five separate money supply effects:

• Debt service requirements;
• Home equity lines;
• Mortgage refinancings;
• Resale of existing homes; and
• Purchases of new homes.

The first observable real estate effects of current Federal Reserve policy should be faltering M-1 growth. The second effects of Federal Reserve policy will be slowing non-M-1 M-2 growth as slowing home resales reduce a substantial source of money supply.

Conclusion

The 1990 recession showed us that reduced home resales influences the growth of M-2. With money supply now so sensitive to housing and consumers draining M-1, interest rate increases may provide a more spectacular result than expected. Lower home resales may lead to an unusual post-WWII decline in M-2.

Household Cash Flow Model

We have restructured the information that we gathered in our previous papers to show the interaction of consumer cash flow and consumer money supply. Historically, money supply flows from left to right in this model. Consumer daily life usually produces surplus funds that can be used for consumer investing.

In 2005, money is flowing backward; or from right to left. The consumer needs to constantly replenish M-1 with funds from non-M-1 M-2 in order to avoid a liquidity crunch. This means that long-term investing is the source of the funds for daily life instead of daily life being the source of funds for long-term investing.

*Estimated as of September 9, 2005

In this chart, we have simplified the flow of debt. It helps identify the debt and income interactions supporting the consumer investing process. The left side of the chart is where everyday activities occur: income production, consumption choices and debt service payments. The right-side of the chart shows long-term activities: savings; investing in assets; and debt increases.

We have made the following adjustments to items developed by the Federal Reserve and the Bureau of Economic Analysis:

• Personal income is disposable personal income plus mortgage interest;
• Consumption equals Personal Consumption Expenditures plus Personal Transfers minus net non-cash consumption including depreciation on houses;
• Debt service is based on the Federal Reserve debt service information; and
• Money flow between M-1 and M-2 is the amount needed to balance M-1 changes.

Otherwise, the rest of the chart is based on the information in our paper on household cash flows.

The chart is astonishing. Daily life is causing depletion of M-1 money supply. M-1 money supply is only $1.35 trillion, or less than 2 months of payment needs. Households now annually withdraw from non-M-1 M-2 for use in supporting M-1 an amount that equals nearly 20% of M-1. Funds in non-M-1 M-2 usually support long-term investing activities; not daily life.

Overall, the chart shows that reducing debt growth will cause some surprisingly big effects. In the past, the consumer had positive cash flow from M-1 as a shock-absorber to cushion lower debt accumulation.

Now, M-1 is reducing non-M-1 M-2 even as the Federal Reserve is trying to reduce debt growth. This money supply flow should cause a larger than normal percentage reduction in investment-related activities.

The remainder of this paper will identify different real estate side effects and determine how they will effect the money supply. We discover that M-1 liquidity will decline first as home equity credit will be the first debt source constrained by Federal Reserve interest rate policy.

After M-1 growth slows, non-M-1 M-2 growth should also decline. Non-M-1 M-2 will suffer attrition as the consumer needs to continue raising liquidity to support M-1. As the competition for non-M-1 M-2 funds becomes more severe, home resales could decline and accelerate the loss of liquidity.

Real Estate and Money Supply

In the introduction to this paper, we identified the different real estate effects that relate to debt and money generation. We will analyze the individual effects to show how they effect the money supply.

M-1 and non-M-1 M-2 are traditionally considered consumer money supply. M-1 contains transactional balances and currency. M-2 contains consumer savings accounts and other interest-bearing consumer accounts. Non-M-1 M-2 is the savings portion of the consumer money supply.

Debt Service

Debt service amounts are generally payable monthly and are paid out of available funds. The monthly debt payment is subtracted from demand deposits near the beginning of the month. This effect causes M-1 to drop. M-1 is replenished over the rest of the month.

We discussed this effect in detail in our paper entitled, "The Interest Rate Conundrum". As debt service increases, income generation needs to increase faster than consumption or demand deposits will tend to diminish. The chart below shows the level of demand deposits in the last few years as a percent of total consumer money supply.

Demand deposits as a percent of consumer money supply have tended to trend down as debt service levels have built. During 2005, they are trending down, again. We appear to have achieved "just-in-time" money supply. This result is a predictable outcome of increasing reliance on floating rate home equity loans as back-up liquidity.

Home Equity

Home equity debt is consumer credit that functions as back-up liquidity for M-1. The advent of home equity accounts and their broad expansion during the 1990s enabled M-1 to decline during the mid-1990s even as economic activity expanded. Effectively, home equity debt now serves the same purpose as an extra cushion of checking deposits did prior to the mid-1990s. It is simply a substitute for M-1 liquidity.

Home equity debt is tied to a short-term interest rate that can be reset monthly. It will be more or less attractive depending on the level of short-term interest rates and the likely future direction of short-term interest rates. When the Federal Reserve raises short-term interest rates, home equity will be the first piece of real estate-related liquidity affected. The cost of this debt will rise first and fastest.

As interest rates rise, two things will happen. First, the household will try to avoid accessing the back-up liquidity: i.e., M-1 growth will be allowed to drop. Second, refinancing will tend to pick up as interest rates reach a level where short-term financing can be converted to long-term financing at cheaper rates.

The chart above shows that M-1 growth is decelerating quickly. M-1 growth declined quickly starting in late 2004 as Federal Reserve interest rate increases finally achieved a high enough level to make home equity lines less attractive.

Refinancing of Debt

Mortgage refinancing restructures debt and provides funds for other activities. As documented in our paper entitled, "The Interest Rate Conundrum", these funds tend to show up in non-M-1 M-2 interest bearing accounts.

The cash-out refinancing supports the purchase of cars or other high-priced consumption items; the payment of short-term floating rate debt; or the purchase of additional property or other investments. It shows up in the money supply as savings until it is used for one of these activities.

For 2005, Piscataqua Research estimates that about $280 billion of non-M-1 M-2 funds will be diverted to M-1 just to support current consumption and debt service. Consequently, any reductions in refinancing opportunities should show up as a reduction in the growth of non-M-1 M-2 and in the overall growth of M-2.

The next chart shows the annual growth in M-2.

This chart, too, shows the effect of real estate-based financing. The level of mortgage refinancing has dropped significantly from its peak in 2003. It has negatively correllated with long-term interest-rate increases and declines in 2004 and 2005.

Resale of Existing Homes

The resale of existing homes has an unexpected effect on money supply. In fact, it has two effects that merge and become a net contribution to savings accounts in M-2.

A resale of an existing home involves a household purchaser and a household seller. From the perspective of counting consumer money supply, only the consolidation of the two households matters. This example shows a potential $500,000 transaction.

A resale of an existing house has a direct effect on non-M-1 M-2 money supply and outstanding household debt. The entire change in the value of household assets is explained by a change in household debt and supports the growth of non-M-1 M-2 money supply. Households experience no increase in equity.

Author's Note: Some Wall Street economists claim that the "income" produced for the seller in this transaction provides a significant boost to personal income. However, a proper consolidation of household accounts shows no change in equity. This activity produces equal amounts of cash and debt. It does not produce any "income" for households in aggregate. It seems to us that Wall Street tries to sell us this same accounting consolidation error over and over and over: Enron's accounting, pension accounting, and national income accounting. Unfortunately, we appear dumb enough to fall for the same line over and over and over, too.

In the current environment, almost every home resale will provide a net increase in non-M-1 M-2 money supply. The increase in non-M-1 M-2 money supply will equal the net increase in debt from the transaction. Though these assets can be used for another investment or for consumption, they boost the flow of debt to non-M-1 M-2 consumer money supply.

With approximately 7.2 million existing home sales expected in 2005, this should cause a substantial increase in non-M-1 M-2. An assumption of an average $200,000 sales price with a 20% extraction rate means that about $290 billion could be added to non-M-1 M-2 through the resale of existing housing.

In our household cash flow model diagram, we expect that non-M1 M-2 will grow by about $180 billion in 2005. The resale of homes may be the only thing that stands between an increasing and a decreasing non-M1 M-2 during 2005.

As a result, we would expect the growth of M-2 to fall if the level of home resales fell. This is exactly what happened from 1990 through 1992 as the level of home resales and then the level of relative prices fell after the last recession. This effect is important, under-investigated and could lead to an unexpected decline in consumer liquidity. It will be discussed in the next section of this paper.

Residential Construction

The effect of residential construction on money supply is clear. As an investment, it subtracts money from non-M-1 M-2. The investment capital (down payment) for the purchase of a new home comes from non-M-1 M-2.

Finally, residential construction increases the flow of personal income. The building of the asset reduces current non-M-1 M-2 by drawing off investment funds, adds to current M-1 through the creation of one-time personal income, and increases the future flow of debt service requirements that will draw from M-1 every year.

The 1990 Recession and Its Aftermath

The cash flow history of the 1990 recession shows us the progression of monetary effects through the cash flow statement. We will overlay our real estate monetary process and compare it to cash flow.

The 1990 recession progressed through consumer money supply in a nearly text book fashion until 1992. Early in the recession, consumers behaved in a normal fashion. They increased their M-1 liquidity to meet economic uncertainty and reduced their use of debt. They also reduced consumption growth and maintained a strong flow of excess funds from M-1 to non-M-1 M-2. Home equity loans were not common and did not effect M-1.

The table below shows that new debt, M-1 contributions to savings account growth, residential construction, and flows of cash to other investments appeared normal. As time progressed, non-M-1 M-2 growth declined precipitously and became negative.

The unusual behavior of non-M-1 M-2 coincides with what should have been falling equity extraction from home resales. The next chart shows the level of home resales since 1989.

Annual home resales bottomed in 1990. However, from 1989 until the middle of 1993, home resales were essentially flat. Real estate activity picked up in late 1992 as significant relative price and interest rate declines changed the investment attractiveness of existing homes. The next chart shows the pattern of relative price declines.

*Federal Reserve information as of 3/31/2005

The relative price declines started in 1990 with the start of the recession. However, they did not accelerate until late 1992 into 1993 and 1994. Since overall real estate sales declined in 1990 and 1991 with the recession, the equity extraction effect began limiting non-M-1 M2 growth in 1990 and 1991.

When relative prices began to decline rapidly in 1992, equity extraction from resales of existing homes bottomed. In 1993 and 1994, selling prices for existing homes increased slowly, but not as rapidly as the cost of hew homes. Non-M-1 M-2 growth began to improve.

The 1990 recession shows that the secondary effects of a popping housing investment bubble can help to limit the expansion of consumer money supply for years after the original recession. The 1985 to 1988 housing investment bubble was much more limited than the current nationwide housing investment bubble. Yet, it appears to have had money supply effects that stretched from 1990 into 1993.

If you lived in Southern California or New England during that period, you remember how thoroughly the real estate markets effected local economic activity. Many banks were liquidated. Loans were difficult to obtain. The value of homes declined. Jobs were difficult to find. The local money supply effects were obvious.

Summary

This paper demonstrates the effects of real estate on the money supply. M-1 back-up liquidity has been replaced by home equity loans and non-M-1 M-2 growth is dependent on debt-based equity extraction from home resales and from mortgage refinancings. When coupled with net negative M-1 growth from combined consumption and debt service in excess of cash income, interest rate policy could severely restrict the growth of consumer money supply.

A consumer liquidity crisis will probably start in M-1. It will be caused by the Federal Reserve raising the cost of M-1's back-up liquidity source: home equity debt. It will spread to non-M1 M2 as investment-related money supply is diverted to M-1.

When non-M-1 M-2 becomes the only source of M-1 liquidity, the consumer will be forced to make choices about the use of funds in non-M-1 M-2. It is likely that consumers will reduce investments in the financial markets and home resales.

The information in Appendix I shows that money supply in 2000 and 2001 behaved similar to our envisioned process. Though not shown in the Appendix, M-1 declined in 2000. The Appendix shows that funds were then extracted during 2001 from the financial markets and home resale markets in aggregate. The process was stopped by significant interest rate cuts that again made debt-funded investments attractive.

Piscataqua Research expects that the home resale market will follow a similar pattern as the 1990 recession. It will slow as prices stabilize. Slowing home resales will reduce the debt-based equity extraction that expands consumer money supply.

Effectively, the consumer will create a lower debt flow by lowering the level of home resales. The lower debt flow will leave a smaller amount of net funds available to support M-1 and consumer investment-related activities.

With M-1 drawing off a larger percentage of the reduced flow of newly created debt, consumers will reduce investing activities supported by non-M-1 M-2. A nightmare scenario could occur if consumers reduce consumption and investing to protect liquidity.

It is possible that the consumer might be unable to maintain money supply levels either in M-1 or in M-2. Since total M-2 did not fall on a rolling 1-year basis even in the early 1990s, falling M-2 could constrain the economy in a way that leads directly to deflation.

If this were to unfold, we could discover that the missing ingredient for preventing a nightmare liquidity scenario is a high national savings rate. With consumer money supply growth now highly dependent on real estate-driven debt activities, it might be foolish for the Federal Reserve to restrain the current source of consumer liquidity.

Appendix I

Sources and Uses of Consumer Investment Funds