Consumer Confidence Sheds Some Light on New Home Sales

The Conference Board recently reported that the Consumer Confidence Index declined slightly in September. The present situation index declined by 1.5% from 148.4 to 146.1, while the expectations index rose by 0.5% from 101.7 to 102.2. As a result, the Consumer Confidence Index decreased by 0.5% to 119.8 in September on a seasonally adjusted basis, from 120.4 last month.

The Conference Board also reported the share of respondents planning to buy a new home within six months. In September, there were 1.0% of respondents planning to buy a new home within six months, compared with 1.2% in August. As shown in the figure below, despite the monthly volatility, the underlying trend in the share of respondents planning to buy a new home within six months has been moving upward since 2012.

Figure 2 shows both new single family home sales and the share of respondents planning to buy a new home. The red line represents new single family home sales one-month ahead since 1978. The blue line shows the share of respondents planning to buy a new home within six months over the same time. Reports suggests that a prospective buyer signs a contract within 30-60 days. The dashed line is the six-month moving average of the blue line to smooth the monthly volatility.

As shown in the figure below, new single family home sales have moved closely with the share of respondents planning to buy a new home within six months until 2002 when new single family home sales started to soar. Between 2002 and 2006, new single family home sales jumped from 0.9 million to 1.4 million at a seasonally adjusted annual rate and then declined steeply. Meanwhile, the trend in the share of respondents planning to buy a new home within six months was about flat with some volatility over the same time period. The increases in new single family home sales over the 2002-2006 period widened the gap between the red line and the dashed line, suggesting that new single family home sales deviated from buyers’ expectations during the housing boom.

After 2006, new single family home sales fell to the trough, matching decreases in the share of respondents planning to buy a new home within six months. Since 2012, both new single family home sales one-month ahead and the share of respondents planning to buy a new home within six months have trended up, although new single family home sales one-month ahead are slightly above the trend of consumers’ expectations.

The scatterplot below is a different way of presenting the relationship between the six-month moving average of the share of respondents planning to buy a new home and new single family home sales one-month ahead.

Figure 3 includes all of the data from 1978 to the present and shows the positive and linear relationship between the six-month moving average of the share of respondents planning to buy a new home and new single family home sales one-month ahead. The black dashed line is the linear trendline, indicating that the higher the share of respondents planning to buy a new home, the larger the number of new single family home sales.

An alternative interpretation is that the relationship between the six-month moving average of the share of respondents planning to buy a new home and new single family home sales one-month ahead is non-linear. This is represented by the red dashed line in Figure 3. The dots, located above 1 million new single family home sales (enclosed in the red circle), did not follow the trendline as the others did. These dots recorded between 2002 and 2006 represent the period of the housing boom. During the housing boom, higher new single family home sales were not matched by increases in the share of respondents planning to buy a new home proportionately. It suggests that the housing boom changed the relationship between the six-month moving average of the share of respondents planning to buy a new home and new single family home sales one-month ahead from linear to non-linear one.

This hypothesis is assessed by demonstrating both a relatively well fitting and linear relationship over the periods outside of the housing boom years. Three scatter plot analyses were applied as discussed below, covering the period before the housing boom, after the housing boom, and the entire period excluding the housing boom.

First, the statistical analysis in Figure 4, using the data from 1978 to 2001, suggests the linear relationship between the six-month moving average of the share of respondents planning to buy a new home and new single family home sales one-month ahead before the housing boom. The fit of the trendline over the pre-housing boom years is a bit weaker but still indicates that a linear trend can explain the relationship. Second, using the data from 2007 to the present in Figure 5 shows that after the housing boom the six-month moving average of the share of respondents planning to buy a new home and new single family home sales one-month ahead are highly correlated with each other. Third, combining the data from 1978 to 2001 and from 2007 to the present as shown in Figure 6 indicates both a linear relationship and a good fit, although the goodness of fit over the entire series excluding the housing boom years reflects a stronger one after the housing boom than before it.

Additional analysis confirms the statistical relationship between the six-month moving average of the share of respondents planning to buy a new home and new single family home sales one-month ahead, making the share of respondents planning to buy a new home a valuable predictor of new single family home sales, especially the period from 2007 to the present. In addition, it concludes that the relationship is likely linear under normal housing market conditions, but, housing boom “bent the curve”, making the relationship non-linear seemingly.



Tags: , , , , ,

Leave a Reply

Your email address will not be published. Required fields are marked *