Tuesday, October 22, 2013

United States Housing Index Compared with Several Measurements

Nick Edmonds has a post that I found very interesting.

UK Housing a Look at some Ratios

I decided to take a look at United States housing through the lens of Government Provided Money Supply.

The Federal Reserve provides three measures of money supply, M1, M2, and MZM.  So far as I know, all three are crafted to aid the Federal Reserve in it's job of managing the economy.  There is no simple scaling factor that relates these three measures, making them difficult to use for macroeconomic analysis. 

On-the-other-hand, the concept of Government Provided Money Supply provides a common reference standard for macroeconomic purpose.  The components are recognized as either money supply or near-money, depending upon the reader's perception of "moneyness".  Step-changes in these components are accepted indications that step-change purchase events have occurred or will occur.  Bank loans is one component of Government Provided Money Supply that conventional wisdom would accept as the best measure of house pricing.

I recognize that calling Government Debt "money supply" is unconventional. Calling bank loans "money supply" is also unconventional but a little more acceptable; and it is in keeping with the MMT "loans create deposits" slogan.  

We will use data from the Federal Reserve FRED series to plot data series and compare trends.

Figure 1.  Three components of Government Provided Money Supply.  The top thin line is Government Debt held by the public.  The diamond line is Total Bank Loans reported as FRED series TOTLL.  The bottom box line is bank deposits less bank loans.  It can be considered as backing for bank loans in excess of the deposit created by the loan.  The bottom thin line is Government Debt held by the Federal Reserve. It is included only as reference as to scale.  All the components are divided by GDP to provide relief from the inflation caused distortion which magnifies the effect of current data.

Figure 1 is a comparison of Government Provided Money Supply components. The important line in Figure 1 is the top dark line which is bank loans reported as series TOTLL.  This line will be the base line of comparison in several of the following graphs.  The bottom dark line in Figure 1 is total bank deposits less total bank loans.  It is left as a reader exercise to use bank deposits as a money supply standard.

Figure 2 compares house prices with bank loans.  All Transactions House Prices are reported as an index of 100 based on 1980.  Bank loans are scaled to 100 to create an index by simply applying the proper factor based on 1980's value.  The result is two lines reflecting the relative change in value between data points.


Figure 2.  Compare bank loans and house prices.  The heavy line is bank loans.

The most interesting revelation from Figure 2 is that the growth rate of bank loans is much greater than the growth rate of house prices.  I had some expectation that the two lines might be parallel but they are certainly not.  Money supply, whether measured by the bank loan component or by Government Debt is certainly not driving house prices as the sole factor.  Other factors in the United States that influence house prices would include large tracts of undeveloped land, great improvement in tools used to build houses, competing uses of disposable income such as cars, and government policy.

Figure 3 compares house prices with personal income.  Similar to Figure 2, personal income is indexed to 1980 to provide a relative comparison with house prices.  Here we see that house prices have not increased as fast as has personal income.  This observation would reinforce the notion that competing uses for personal income is impacting the share of income spent on residential property.
Figure 3.  Personal income compared to house prices.  Personal income is the top thin line.  House prices are the lower dark line.
Figure 4 will be the final figure.  Here we compare personal income with bank loans.  Using the same indexing technique used in previous figures, we can see that bank loans have increased faster than personal income.  Again, I had an expectation that personal income might increase parallel to bank loans, which would be parallel to the increased money supply. 


Figure 4.  Personal income compared to bank loans.  The lower thin line is personal income.  The upper heavy line is bank loans.
Not shown here, when personal income is indexed to money supply measure M2 using the 1980 index point, the match is MUCH better than we see in Figure2 with bank loans.

The failure of personal income to increase at the rate of bank loan increase indicates that additional factors are in play.  One possibility is that imports are being purchased with borrowed money.  Stated another way, borrowing is employing foreign workers whose income is not reported, not United States workers with reported income.

Another possibility is that borrowed money is financing increased valuation of existing fixed assets.  

Yet another possibility is that government spending is becoming a very large part of national income.  Government spending would not be expected to follow bank lending closely.

Perhaps all possibilities are at work.

I will close the post with a word of caution: 
Inflation over the period of record has skewed data in a logarithmic fashion.  This skewing makes indexing very sensitive to the date of reference.  The general trend between lines remains valid but be very careful in how exact differences are described.

No comments:

Post a Comment