"A Rising Population ...
...makes competition for land fiercer, which in turn leads to an increase in land rent everywhere and pushes the urban fringe outward. This corresponds to a well documented fact stressed by economic historians. Examples include the growth of cities in Europe in the 12th and 19th centuries as well as in North America and Japan in the 20th century or since the 1960s in Third World countries."
(From page 83, section 3.3.2: Economics of Agglomeration:... by Fujita, Thisse).
(link to this post for sharing:
https://meetingthetwain.blogspot.com/2019/03/lao-on-housing.html )
A 2 page condensed version in Word is available here:
https://drive.google.com/file/d/1vMS2rTnFmL3DCf-Sgrbb9VPYwpPW1sw6/view?usp=sharing
Precis:
California's oft-stated 3.5 million housing "shortage" does not exist. The plans of Governor Newsom and others to fill that non-existent "shortage" will reallocate needed housing funds from the poorest groups and areas to the more affluent population.
That 3.5 million number is a
mis-interpretation of the Legislative Analyst Office's (LAO) 2015 report on housing. The LAO argues that
if there had been 3.5 million more housing units built over 1980-2015 then housing costs
would have been lower.
If housing costs were lower, they
hypothesize, more people would have moved to California to fill those 3.5 million housing units. Those who
hypothetically might have moved to California did not move here then so they are
not here now - hence,
no shortage.
They might not have moved to California anyway because during those 35 years there were recessions, lay-offs, etc. in the US and California. A California company goes on a hiring spree for a while and people move to California for the new jobs. Housing prices rise during the hiring spree. Then as markets cool, it stops hiring so people stop moving in. The LAO is reversing "cause and effect" assuming people would always move to California at a steady pace regardless of all other economic conditions.
If (
magically) 3.5 million new homes (a 25% increase of the existing stock) were to appear in California they would be empty and remain so. California has seen a net loss of native population through domestic out-migration for many years. In-migration from other countries has masked this - until now.
Perhaps through a misunderstanding of HCD's "Regional Housing Needs Allocation" (RHNA) numbers there is a view that there is a "housing crisis". RHNA numbers are planning for potential growth. If the growth does not happen, those numbers mean nothing. They are not requirements. Thirteen rural CA counties lost population. They did not “make their RHNA numbers” because there
was no need for new housing. That does not constitute a “crisis”. We are currently in a housing
bubble similar to that of 2008 which will not last much longer. Cf,
https://meetingthetwain.blogspot.com/2018/09/ree-diculous-ree-na-part-i.html
The LAO advocates increasing density in the already densest, most geographically constrained job centers. This ignores the most essential tenet of "Urban Economics" going back centuries -
"increased density increases the cost of housing". If increased density lowered the cost of housing, then Lower Manhattan and Hong Kong would be the cheapest places in the world. They aren't.
The LAO's prescription for lowering housing costs by building more densely in already dense areas will actually
raise housing costs. That, in fact, is what is happening and has been happening for decades. More density, more traffic, and housing prices increase.
 |
Silicon Valley has been increasing housing since it's beginning and the prices keep rising.
So we should build more housing to make prices go down! |
We see it everywhere - gentrification and displacement. It has caused increasing out-migration of companies and population who cannot afford the increased prices or cannot stand the increased traffic.
The obvious alternative is to direct business expansion (and thus population increase) to the less dense metro areas in California with more buildable land such as Sacramento and Riverside. The population growth is naturally happening there, anyway.
Introduction:
Governor Newsom has proposed some re-allocations of state funds for housing. The Legislative Analyst Office's (LAO) looks at those proposals in "The 2019-2020 Budget: Considerations for the Governor’s Housing Plan" (CGHP) by Gabriel Patek dated February 2019 available at:
https://lao.ca.gov/Publications/Detail/3941
As CGHP refers to the 2015 LAO document "California’s High Housing Costs: Causes and Consequences" (CHHC) we also address the references and conclusions from that document available at:
https://lao.ca.gov/reports/2015/finance/housing-costs/housing-costs.aspx
This post is an analysis of the above two documents.
Summary of LAO Documents:
The CGHP (2019) focuses on Governor Newsom's stated strategic concepts on housing .
In particular, discussion centers on the proposed allocation of more state funds for middle class residents. The LAO is essentially arguing that if more state money is allocated to housing for middle income housing then there will be less for the lowest income. They appear to argue that since it is the lowest income group that suffers most from high housing costs, funds should then continue to be directed to those groups. In addition, I would add that the highest housing costs are in the highest income areas (not unrelated facts). Sending more state money to the highest cost areas would be a transfer of wealth from all areas - including the poorest - to the richest areas.
However, the LAO continues on with an argument they have advanced for years that the cause of California's high housing costs (relative to the US average) is a failure to build enough housing in dense urban areas along the coasts. The LAO also argues that it is necessary to increase density to lower housing costs. This contradicts basic theories of "Urban Economics" going back hundreds of years as taught in popular Urban Economics texts both undergraduate and graduate as well as countless academic papers. No specialist in Urban Economics disputes that increased density increases housing prices.
Analysis:
There are 3 points:
Point 1. A "Shortage" of housing units is not the prime cause of high housing-cost-burdens.
Point 2: More "Housing Units per Household" Does NOT Correlate with Affordability.
Point 3: LAO's statistical model is flawed.
Point 1. A "Shortage" of housing units is not the prime cause of high housing-cost-burdens.
Figure 1 in CGHP (below) is comparing California - a state dominated by a few highly urbanized metro areas - with the entire US.
Figure 1
Comparing California with the entire country includes North Dakota, West Virginia and many other rural and semi-rural states. Such a comparison renders the graph almost meaningless. Comparing urban areas in California to other major urban areas on a housing-cost-stressed metric is far better. This is done in the following graph from U. of Southern California (via
Slate). This graph shows that it is not the cost of housing but the incomes which are the determining factors in cost burdens.
Figure 2: (click to enlarge)
This is important in examining LAO's thesis that California communities have restricted housing and that is the reason prices are high. In figure 2, above, the US Census' "SF-Oakland" metro area is near the low end of "cost-burdened" metro areas in the "West" group even though housing prices there are the highest in the US in absolute dollar terms - as are salaries (
hint, hint).
Continuing with figure 2, renters in the communities in other metro areas such as Miami, New Orleans, and Orlando are cost-burdened as much as, or even more than, the California coastal cities (left bars under "West"). It is hard to argue that Miami and Orlando have restricted housing construction. One of Florida's main industries is constructing housing for new Floridians.
Figure 2, above, shows Cleveland as having the most cost-burdened renters in the Midwest yet Cleveland's housing costs are some of the lowest in the country in absolute dollar terms. Cleveland's population during the period 2000-2017 actually
declined 4.2%. Cleveland thus has a
surplus of housing.
A Harvard study considers "residual income" - what is left over after the rent is paid. For example, a couple making $300,000/year could pay 50% of their income for housing and have a $150,000 "residual" - more than enough for other expenses. On the other hand, a couple paying 30% of a $30,000/year income would struggle to get by with a "residual" of $20,000/year. The standard 30% criterion for housing costs does not consider that and should only be used for the lowest income quintiles. (Harvard paper here:
https://www.jchs.harvard.edu/sites/default/files/Harvard_JCHS_Herbert_Hermann_McCue_measuring_housing_affordability.pdf)
When you correct for "residual income", as the Harvard study shows, the difference between LA and Cleveland for lower income households diminishes to only 4% for a family of four.
The following bar chart (Figure 3) shows California is one of the top five states for housing costs as a percentage of income. Florida and Hawaii are slightly higher. Louisiana and New Jersey are slightly worse.
Figure 3: (click to enlarge)
SF-Oakland metro area less "Cost-Burdened" than Orlando or Las Vegas? California less cost-burdened than Florida? Clearly the LAO is
wrong in their assertion that the reason for "California’s high housing costs, ...
is the significant shortage of housing..." page 1 of CGHP (2019). We cannot argue against their evidence for that statement since they offer none. Building more housing is
not valid as a way to help cost-burdened households. Other factors need to be considered.
Point 2: More "Housing per Household" Does NOT Correlate with Affordability.
From the US Census' "Quick Facts" (link below) we see that the US has over 137.4 million housing units for 118.8 million households for an average =
1.16 housing units per household (HupH). This includes vacation homes, housing in transition, apartment vacancies, etc. That is 11.6 housing units for every 10 households.
California has 14.2 million housing units for 12.9 million households =
1.10 HupH. That is 11 housing units for every 10 households. For California to reach the national average of 1.16 it would need only about 5.5% more housing units or 773,600 additional housing units.
Building 3.5 million more homes as Governor Newsom has pledged to do is an astonishing prescription. If an additional 3.5 million housing units were added, the result would be a ratio of 1.37 HupH - almost 14 housing units for each 10 households - far higher than any other state. The following graph, figure 4, shows where California is now and would be under other scenarios:
Figure 4: (click to enlarge)
California (green bar) is the lowest in HupH at 1.10 but not by much. A 5.5% increase in housing would bring it to the national average (purple bar) while 3.5M more homes would put it way above all other states - the top orange bar. None of this matters because when we compare these housing-to-household ratios in figure 4 (above) with housing prices in figure 5 (below) we see no correlation between "housing units per population" and home price.
Figure 5 (click image to enlarge):
Affordability (as measured by "rent-burdened households") has no obvious correlation with "Housing Units Per Household". Florida has low cost housing and
very high Housing Units Per Household ratios. Nonetheless, Florida's population is the most rent-burdened in the country. Hawaii has more expensive housing than California yet has a
higher HupH ratio than the national average. It is hard to argue, as the LAO does, that housing supply is the principal factor in cost burdens.
Point 3: LAO's statistical model is flawed.
The LAO model is described in the "Technical Appendix" at the end of CHPC (2015). In the PDF version it starts on page 36. The LAO's model concludes that if an additional 100,000 homes per year had been built, that would have kept housing costs lower and more people would have come to California, enough to fill 3.5 million homes.
We see this in the CGHP (2019) report in the grey box on the bottom of page 2:
---
"This suggests an aggregate of about 3 million additional housing units would have been needed between 1980 and 2010 to keep California’s housing cost growth in line with cost escalations elsewhere in the nation."
---
This is further elaborated on in CHPC (2015, top of page 4) stating that an additional 100,000 units per year would need to be added to mitigate housing costs. This would add another 500,000 units during 2011-2015 for a total of 3.5 million by 2015.
There are several problems with their model and interpretation.
Problem 1: The most glaring is that it starts with the assumption that people will come to California at the same rate limited solely by housing costs. What we see in actual data is that California's growth rate is highly variable and is declining as is the US's as shown in the graph below:
This assumption of constant growth rate is also part of their conclusion - that more people would have come to California if it were not for the high housing costs. Basing your model on the assumption you wish to prove will always prove your assumption.
Problem 2: Even
if that model were correct, it is still incorrect to say (as many have been saying), that there is currently a shortage of 3.5 million housing units. Those that (hypothetically) would have come to fill the 3.5 million housing units (that didn't get built) would have since gone elsewhere -
which they might have done anyway! Housing is currently in approximate balance with population as shown previously.
Problem 3: As mentioned in the beginning, the field of "Urban Economics" has shown repeatedly that increased density increases housing costs. The figure below illu
A paper (2015) by Alain Bertaud at the Marron Institute of Urban Management at NYU discusses this. He previously held the position of principal urban planner at the World Bank. His paper is available at:
http://marroninstitute.nyu.edu/content/working-papers/the-spatial-distribution-of-land-prices-and-densities
Much more on the costs of housing vs. density here:
https://meetingthetwain.blogspot.com/2017/01/live-work-commute-2.html
Development in the SF Bay Area is limited by the decisions (which I support) to not fill in the Bay and to protect open areas surrounding the populated areas. Without these restrictions, more land would have been available to build on and prices and density would have been lower. But, the SF Bay Area would then be immeasurably less attractive and, without the cooling effect of the bay, much hotter.
Excluding surrounding open space from development also limits the amount of land that can be built on. The cost of remaining land is made more valuable because of it's resulting limited supply due to the restrictions on development in and around the bay. This is seen in figure 6 below:
Figure 6:
The curves below show that as the distance, X, from the center of a metro area
decreases the cost of land
increases.
Reducing the available land raises the cost curve at all distances.
Cost vs. Density Curve
 |
Cost vs. Density Curve from page 74 of "Lectures on Urban Economics" by Brueckner, 2011, MIT Press
Due to urban growth boundaries, "the resulting city has smaller dwellings and is more expensive to live in."
People who want larger dwellings at less cost will leave such areas. |
This same cost vs. density "isoquant" curve is seen throughout basic Urban Economics. A standard undergraduate textbook used at UC-Berkeley is "Lectures on Urban Economics" by Brueckner. He points this out very well as shown in figure 6 above.
Brueckner is examining the effect of politically instituted "Urban Growth Boundaries" but the conclusion is the same whether the boundaries are imposed by laws or by geography. This is stated by Brueckner as "in the presence of an open-space amenity, the socially optimal city is spatially smaller than the city generated by the free-market equilibrium." ("Socially optimal" because the unfilled-in bay and the undeveloped open space are "social" goods making life "optimal".)
 |
"Socially Optimal"
Golden Gate Park in San Francisco |
 |
| Not "Socially Optimal" |
Since less space is available, the space that
is available becomes more valuable. Dwellings become smaller to fit people into the remaining space. Compare housing in Lower Manhattan with that in Peoria, Illinois.
This goes back to the "Law of Rent" from David Ricardo (1772 - 1823), and "Spatial Economics and Economic Geography" of Johann Heinrich von Thünen (1783 - 1850) as described in many, many,
many economics texts, papers and (of course) Wikipedia.
Problem 4: The implicit assumption of the LAO is that urban areas can expand indefinitely. This flies in the face of many decades of economic theory and thousands of academic studies on "Optimal City Size" in Urban Economics. Optimal city size is described as:
----------
"Theories on the optimal city size indicate that, when a city has an optimal population size, the forces of agglomeration economies are offset by the forces of disagglomeration economies, resulting in locally constant returns to scale". (from "The Optimal Size of German Cities An Efficiency Analysis Perspective" by Stephan Hitzschke -
https://www.researchgate.net/publication/254403140_The_Optimal_Size_of_German_Cities_An_Efficiency_Analysis_Perspective )
----------
In simple terms, the above states that a city reaches its optimal size when the advantages of a big city equal the disadvantages. Further growth after that is not possible. A standard Urban Economics textbook (used at UCLA) shows this in the following graph:
Figure 7:
 |
From "Urban Economics", 8th edition, by O'Sullivan
The cost of density rises slowly at first but at a certain point overcomes the advantages of density.
This is the equilibrium point when a city cannot grow without life deteriorating. |
Where the lines of "cost" vs. "return" cross is the "equilibrium point". At that point, the number of people who come for the advantages of big city life (better paying jobs, more opportunity, more cultural events) equals the number of people who leave because they can't stand it any more - traffic is too bad, taxes are too high, housing costs too much - life is too stressful.
The cost of more bridges, highways, subways, etc., etc., starts to exceed the economic benefit of being in a big city. The City has then reached it's "optimal size" and further growth relative to the population of the nation as a whole is very difficult.
Hence LA metro area grew 8% from 2000-2017 while Charlotte grew 90%, Austin grew 69%, and Minneapolis grew 21% in the same time period. See figure 7.
Figure 8: Click image to enlarge.
 |
Top 50 Metro Areas population over 1 million grew 21% on average.
Only less dense areas in California exceeded that.
Data from US Census year 2000 and ACS 2017 |
More housing in the LA or the SF Bay areas will drive up rents, increase traffic, increase urban sprawl, require massive investment in new bridges, tunnels and freeways, requiring increased taxes.
For example - the Golden Gate Bridge has been at capacity for over a decade. Further development in Marin County will either make the bridge impassable or require a second bridge or tunnel costing billions of dollars and increased taxes. Similarly for other parts of the SF Bay Area.
This will continue to drive corporations and individuals to other states. Other people will come to replace those that left and at some point there will be equilibrium when the number of people entering California equals the number leaving - probably preceded by net population loss.
 |
SF Bay Area with 50 Mile Radius around Silicon Valley
Geographically Constrained - Therefore Expensive |
 |
Houston with 50 Mile Radius around Center
NOT Geographically Constrained - Therefore Cheaper |
If there is to be further growth in California it must come where there is space to expand and build the single family homes that most people desire. Sacramento and Riverside are seeing that growth already. Further state efforts at development, in order to succeed, must go where people desire to move.
What Not to Do:
As mentioned earlier, there is a net outflow of domestic residents from California. See color-coded map from Harvard's "Joint Center for Housing Studies" (Figure 6) below:
Figure 9:
Net Flow of Domestic Moves Between States.
Red is "net Outflow". Blue is "net Inflow".
It is only foreign immigration that keeps California's population growing. If that immigration slows down, stops, or reverses (as it has for Latino immigrants) that outflow will lead to a surplus of housing. State expenditures for middle class housing will be wasted.
Governor Newsom apparently campaigned on the LAO model of a 3.5 million housing deficit. It may guide his thinking on housing budgets. Other politicians have used that same 3.5 million to describe a shortage of crisis proportions. But there are not 3.5 million households without housing.
This imagined "housing shortage" is being used to justify the anti-democratic over-riding of local control. It is also being used to justify increased taxes, and building on open space and parks. This increases congestion, and a feeling of being overwhelmed. It is not healthy. This lowered quality of life is increasing out-migration.
At $500,000 per housing unit, 3.5 million housing units is $1.75 Trillion! For comparison, California's GDP is $2.9 Trillion. That is a tremendous amount of money to satisfy a demand which does not exist now and may never have existed outside of the LAO statistical assumptions.
In thanking academics and others who helped them with their model, the authors of CHPC (2015) note that not everyone agreed with their conclusions. It would have been nice to see the dissenting views in the report.
LAO Documentation Standards
(or lack thereof)
Throughout these two LAO documents and similar ones on housing, there are no sources given for any of the data. No calculations are shown. No references for economic theory justifying the assertions.
From the University of Colorado, Boulder:
"You must acknowledge the sources of all your information and any ideas or interpretations you have taken from other works...references are usually placed into notes, with a bibliography at the end of the paper that lists all works used"
Above from Part D of:
The LAO document authors follow none of the above guidelines.
We are not talking about academic peer-reviewed research papers - just an ordinary paper attempting to summarize complex situations for executives with busy schedules looking for informative guidance. Such documents need sources, references and supporting economic theory so they can be verified and serve as a starting point for further investigation. Without reference to those sources and dissenting opinions those topics cannot be further examined for nuance or error.
Conclusion:
The LAO has produced documents with claims deriving from fundamentally flawed models that ignore the long established tenets and observed results of centuries of economic research. As a result politicians have taken these hypothetical and counter-factual theses to advocate and budget for seriously misguided policies - policies which will only make things worse for Californians. This will cause more people to leave for states which actually listen to what residents want rather than what developers want.
For now this is...
