The California Report

Summary

So far, 2006 has unfolded about like we predicted.  The housing market has continued to soften, and real estate related employment has moved from a major engine of growth in 2005 to a drag on growth in 2006.  High home prices continue to eat away at income growth – real per capita incomes in California have yet to recover even 5 years after the recession.

Looking forward, the forecast calls for a similar picture.  Real estate sectors will continue to decline, but without significant declines in another sector, the net result will be a slowdown, not a recession.  In the absence of recession, statewide home price are unlikely to experience significant declines.  However, since builders are much more willing to lower home prices than owners, the handful of areas where new homes account for an above-average share of total sales activity could see some price declines.

Employment Trends in 2006

2006 has seen a moderate slowdown in the California economy, with growth in non-farm payroll employment falling from an average rate of 1.8% in 2005 to just below 1% in the first half of 2006.  Unemployment remains at lows not seen since the end of the tech boom, but has mostly moved laterally in 2006. 

CA Non-Farm Payroll Employment (1,000s, SA)

CA Unemployment Rate (SA)

California’s slowing job market in 2006 has been shaped by three forces: a weakening real estate sector, a slightly less weak manufacturing sector, and steady-as-she-goes growth in most of the service sectors.  The 61,000 jobs created by the Construction sector in 2005 were the biggest single source of job growth in California; however, 2006 has been much weaker.  In raw terms, the Construction industry added 7,000 jobs in the first half of 2006.  However, this increase is far below the usual first half seasonal surge that we would expect to see – the substantial seasonal swings in Construction employment make it difficult to differentiate trends from seasonal swings.  There are two ways around the seasonality: we can compare this July to last July by calculating a year-over-year growth rate, or we can calculate the average percentage surge we’d expect to see in the first half of the year, and compare the actual number of jobs to the expected number – a simplified description of seasonal adjustment.  By either measure, the Construction sector is experiencing a major slowdown: the year-over-year growth in Construction employment has slowed from 6-9% to only 2% in July 2006.  Similarly, Construction employment is about 10,000 jobs lower than we would expect given the usual seasonal patterns. 

New Payroll Jobs by Sector (1,000s, SA)

CA Construction Employment (1000s; Bold is SA)

Year-over-Year Growth in Construction Employment

While every region has seen slower growth in Construction employment, the regional pattern has varied widely.  In terms of year-over-year growth, Southern California and the Central Valley have seen similar slowdowns, but in terms of seasonally-adjusted data, the Central Valley has lost jobs, while Southern California has held steady – a 2% year-over-year growth rate is about the average for this time of year in Southern California, but is significantly below average for Sacramento County, which accounts for the bulk of the slowdown in the Central Valley.

Growth in Financial Activities employment has also slowed considerably as the real estate relate components of this sector have felt the pinch from slowing housing markets and higher interest rates.  Layoffs at major mortgage industry players like Ameriquest and Countrywide have made Orange County the primary casualty, but this slowing has occurred throughout Southern California.  Growth in Non-Depository Credit Intermediation (which includes mortgage lenders) and Real Estate have essentially flattened out, while Activities Related to Credit Intermediation (mortgage brokers) has lost about 1800 jobs in Orange County since the middle of 2005.  In fact, the high share of independent contractors in this industry means that the payroll employment data may very well understate the scope of these job losses; nevertheless, even the payroll survey shows significant slowing in the financial side of the real estate sector.

Outside of the real estate portion of the economy, 2006 has been mostly more of the same for the state.  In Southern California, accelerating growth in Leisure/Hospitality and Professional/Business Services has been able to offset some of the real estate weakness, but these local trends have not translated through to the state level.  With service sector growth either about the same or slightly slower throughout the State as a whole, the only offset to the housing-related drag has been slightly less bad news from the manufacturing sector, as the slow but steady job losses in both durable and non-durable manufacturing have moderated in recent months.

Incomes and Inflation

While California’s nominal per capita income of $37,036 in 2005 was over $4500 higher than in 2000, adjusting for local inflation, real per capita incomes have still not recovered from the 2001 recession.  Comparing the national experience with California, it is clear that part of the explanation is that the run-up in incomes at the end of the 1990s was in itself abnormal – funny money dot.com compensation led to a burst of above average growth, leading to a steeper drop as the tech boom busted.  But an equal part of the explanation has been the increased cost of living in California, especially compared to the national experience.  Since 1999, the Consumer Price Index for the U.S. as a whole has risen about 20%.  The L.A. CPI has risen 26% over the period, and San Diego’s has risen 31%.  The Bay Area CPI has risen about the same as the national average, but this was more the result of a few years of high inflation followed by a few years of next to no inflation.

Real Per Capita Income (CPI, 2000 $)

Consumer Price Index: All Items (1999=100)

Consumer Price Index: Shelter (1999=100)

It should come as no surprise that the bulk of this divergence comes from California’s red-hot housing markets.  The shelter component of the CPI has three major components: rent of primary residence, owner’s equivalent rent (roughly how much the Bureau of Labor Statistics estimates that it would cost a home owner to rent an equivalent home), and lodging away from home.  Together, these shelter items receive about 40% of the total weight in the CPI index (the exact amount varies by region) – by far the biggest single item in the index.  Overall, the nationwide housing boom has led to a 22% increase in the CPI’s shelter index since 1999.  Of course, L.A.’s increase was 38%, and San Diego’s was 45%.  If anything, these numbers are a bit conservative: the BLS goes to great lengths to separate the cost of housing services from the investment component of housing; thus, the over 150% increase in home prices in L.A. County from 1999 – 2005 dwarfs the 32% increase in owner’s equivalent rent.  While there is continuing academic debate about which of these numbers more accurately reflects the impact of higher home prices on standards of living, even the conservative estimate imply that the housing boom has taken a serious bite out of real incomes in California.

Housing Markets

The first half of 2006 has continued the slowing trend from the second half of 2005.  Sales activity in both the Bay Area and Southern California has fallen over 30% from recent peaks, and year-over-year growth in median sales prices in the two regions has slowed to single digits.  But within this broader trend of flattening prices, San Diego and several counties in the Sacramento region have actually seen declines in median sales prices.

Home Sales Index (2000=100)

Year-over-Year Chg in Med. Sales Price

For almost a year, the Forecast’s California report has been arguing that we are unlikely to see significant declines in state home prices without a recession.  So how do we reconcile this argument with the empirical fact that a handful of California counties have seen median sales prices fall in 2006?

Our answer comes in two parts.  First, just when we would like an unambiguous read on what’s happening with housing prices, different data sources are telling different stories.  The table below compares the percentage change in OFHEO’s Home Price Index from 2005Q4 to 2006Q2 to the change in Dataquick’s median sales price over the same period.  While the broad trends may agree, there are substantial differences.  While the seasonally-adjusted median sales price in L.A. County grew less than 1%, the HPI for L.A. County was up over 6%.  Several metro areas show positive appreciation in the HPI, but show declines in median sales prices.  The discrepancy comes from a fundamental difference between the two measures.  Median sales price is fairly straightforward: it is the average sales price of all the homes sold in that period.  The calculation of the HPI is much more involved calculation that tries to nail down the change in the price of a “constant quality” home by matching the sales price of the same home over two different periods across thousands of transactions.  Generally speaking, economists tend to prefer to use repeat-sales measures like the HPI to gauge changes in home prices, since it controls for the changes in home mix and quality that could alter average sales prices even if the price of an “average” home was unchanged.  In addition to changes in mix and quality, there are two other important reasons why these two measures might differ.  The HPI focuses exclusively on detached resale homes, while the Dataquick averages include both resale condos and the sales price of new construction (new homes and new condos are averaged together).  Last but not least, the HPI sample is derived from conforming Fannie Mae / Freddie Mac loans, so the absence of the higher end of the home price distribution may skew the index a little.  However, this problem is not as bad as you might expect: although the conforming limit in 2005 was only $359,650, 67% of the homes in the 2005 sample for California were valued at higher than this limit.  Bottom line?  In markets where the median sales price is falling but the HPI is rising, it’s hard to sort out exactly what’s going on with home prices – but we lean towards trusting the HPI.

Home Price Appreciation, Median Sales Price vs. OFHEO Home Price Index

The second part of our answer focuses on the differing behavior of new homes and resale homes in a soft market – something that the Dataquick numbers can illuminate.  The graphs below show the median sales price for resale houses and new homes in Southern California from 1988 to 1995.  The path of these median prices during the recession is markedly different: the median price of new homes fell sharply, but bounced back sooner, while the median price of resale homes fell much more gradually, but for a more prolonged period of time.  This could be another side effect of severity of the local recession; however, Sacramento County shows a similar pattern in home prices even though the recession was much less severe.

Southern California Median Sales Price of New and Resale Homes (SA)

Sacramento County Median Sales Price of New and Resale Homes (SA)

Why would new prices be flexible downward in a soft market when resale prices aren’t?  While an owner could easily respond to a softening market by simply waiting to sell or renting out the house, builders can neither carry the inventory nor take the depreciation hit on renting a new home.  Thus, builders will respond to a fall in demand by lowering the price, while resellers may simply retreat from the market.  We have often explained the connection between recessions and falling aggregate home prices by arguing that job loss creates motivated sellers who need to raise cash quickly, and are therefore willing to trade a quicker sale for a lower purchase price.  Our story about builders is just an extension of this logic: builders are always motivated sellers.

Let’s apply this logic to the current market.  Lack of any significant economic distress so far in 2006 has limited the number of motivated resale sellers, but the slowdown in sales should motivate builders to lower their prices.  This argument offers several predictions.  Since the bulk of home transactions are resale transactions, aggregate home prices should be mostly flat.  However, regions where the sale of new homes makes a significant share of total sales should see slower or even negative growth in median sales prices.  So far in 2006, both of these predictions have been true.  The broad aggregates for home prices have flattened out, but haven’t shown significant declines – which is consistent with the historical relationship that statewide home prices have only fallen in recessions.  At a county level, we find that the counties with higher shares of new homes are also the counties which have experienced the weakest price growth.

California Home Sales, 2005-2006

So does this change our forecast for California housing markets in the next two years?  Not really.  We are still firmly convinced that the economy is the primary driver at the state level: statewide home prices are unlikely to decline significantly unless there is a recession.  However, regions where new building is a substantial portion of the housing market may experience some price declines even if there is no recession in California, since builders and their willingness to lower prices make up a bigger chunk of the local market.  So far, San Diego and the Sacramento region have followed this pattern.  On the other hand, new homes made up 40% of new sales in Riverside County in 2005, yet prices are still rising.  Whether this means that the other shoe has yet to drop in Riverside, or simply that there are demographic factors that continue to favor the housing markets in Riverside relative to other counties is still an open question. 

Forecast and Conclusions

The main themes of last quarter’s California forecast have carried over into this quarter.  Building permits will continue to decline, bottoming out in 2008 as activity returns to levels seen in 2000.  This decline in building activity will continue to weaken Construction, which will lose around 100,000 jobs over the life of the forecast.  The weakness in real estate related Financial Activities will also continue, slowing overall growth in that sector.  On the other side, a flattening out in Manufacturing employment and steady but not spectacular growth in the rest of the service sector will offset some of this weakness, keeping overall payroll growth around 1% through 2008.  Slowing growth in nominal personal income and taxable sales will be offset somewhat by declining inflation: growth in real personal income and taxable sales will remain weak in 2007, but pick up some lost ground in 2008.  Since we are not predicting a recession, we look for nominal home prices in the state to stay flat through 2008, though real prices will fall.  However, a handful of regions with above average levels of new construction relative to total sales could see moderate price declines.