Tuesday, September 9, 2014

Speculative Position in Housing, Part 2

In last week's post, I forecasted a 75% - 100% growth in homebuilder revenues over a 3 year time frame.  Firstly, I'm not sure that that forecast is really much outside the typical forecasts for some of the individual builders.  Also, it's not outside the range of outcomes from the previous expansion cycle.  Here is a graph of 3 year revenue growth rates for the homebuilders as a group, and another graph of 3 year revenue growth rates for the individual firms.


The entire industry experienced that rate of growth for a decade.  My hunch is that bottlenecks in credit creation, home price trends, and builder capacity create a sort of growth ceiling in that range, and that the homebuilders will plateau at that range.  The question will mainly be how long will it persist.

In the 2000's, there was a fairly wide range of growth levels.  Converting back from logarithmic to a geometric scale, except for a couple of outliers, the range was generally from about 60% to 120%.  The relative position in this range is obviously an important consideration for valuing each firm, but this post is simply intended as a framing device for creating a starting point for speculative positions.


Annual Growth Rates
Here are a couple of graphs comparing industry revenue growth with growth in total home equity.  This uses a more thorough sum of revenues than the one I used last week.  The first graph shows the fairly tight correlation over a three year period between industry revenue growth and home equity growth.  Also, when we look at trailing 3 year time frames, the revenue growth for homebuilders from 1999 to 2004 was very consistently around 93% (shown in the graph as about 66% logarithmic growth).

In the next graph, we can see how equity and total real estate market value tended to grow together until the monetary crisis caused homes to be overleveraged.  Also, we can see that a homebuilder revenue forecast (on a YOY basis) based only on home equity levels has pretty accurately tracked actual revenue, especially over 2 to 3 year periods.

Further, homebuilder valuations track revenues fairly well.  Here is a graph of homebuilder revenues, enterprise value, and market capitalization.  (Note, I am using a kind of lazy enterprise value here that consists of total liabilities plus the market capitalization of equity).  Here, we can see that there has been roughly a floor of 1 on the EV/Rev ratio.  Currently, the market capitalization also reflects about $6 billion of tax assets from the downturn, many of which are still off-balance sheet (as of the end of FY2013).  We can see here that the market is anticipating growth.  So, there already is a large deviation between Enterprise Values and current Revenues.  Much of this, surely, is in anticipation of some future growth.  In fact, for the entire industry, Enterprise Value already reflects a near doubling of Revenues.  Again, this suggests to me that my home market value and homebuilder revenue projections are already being anticipated, at least within the homebuilder equity market.

So, as a broad brush framing for speculative opportunities among homebuilders, I will compare the gains to equity holders, given certain increases in revenues, with forecast valuations based on EV/Rev=1.  In the 2000's, firms tended to converge at a Market Cap / Enterprise Value ratio of about 50%, so in this broad comparison, I will assume that operational liabilities will eventually limit market cap to this ratio.

Below is a table of homebuilders comparing several metrics:


The first column compares the potential equity returns of the various major homebuilders, given a growth of 90% in revenues.

The second column compares the growth required in each builder's revenue in order to justify today's share price, at an EV/Rev. ratio of 1.

The third column compares consensus 2 year revenue growth estimates for 2014-2015.

I think it is interesting that the weighted average 2 year growth rate is under 40%, with a standard deviation of about 16%.  This is very low compared to any year since 1996, outside of the crisis years.  The 120% break even growth rate suggests that the market has already priced in high expectations for future homebuilder revenue growth.  But, this isn't reflected in the conservative analyst growth expectations.

It is a common dilemma in forecasting that forecasts tend to have less variance than actual outcomes.  So, I think it is likely that homebuilder prices reflect expected growth that isn't being reflected in published analyst estimates.  This could reflect a consideration of highly negative outcomes in the analyst estimate figures, bringing the expected values down, or it could arise from anchoring effects where, in highly volatile contexts, the most accurate forecasts will seem unreasonable ex ante.  The market seems to be generally pricing in a bullish revenue expectation.

In the fourth column, I have simply tripled the consensus 2 year growth rate, to arrive at a sloppy estimate for 3 year growth rates in a bullish home market.  This is a broad attempt to capture the relative expectations for each builder while adjusting the consensus to my bullish expectations.

When I compare the expected returns to equity for each homebuilder to each homebuilder's current financial leverage (estimated with MC/EV), I find a systematic relationship where the less leveraged firms have the lowest expected returns and the most leveraged firms have the highest expected returns.

The expected returns of the safest homebuilders (in terms of leverage) show here as negative because I am using a static valuation forecast of EV/Rev.=1.  But, for the least leveraged firms, lower equity risk premiums due to the lower leverage would lift their static relative valuation levels.  So, while I haven't engaged in this adjustment here, one can imagine the right hand of this relationship being pushed up by this factor, so that the returns of the least leveraged firms will tend to be higher.  My broad measure here does not account for the accumulation of profits during the 3 years elapsed, so if we imagine the healthiest homebuilders earning reasonable profits during this time, as they certainly would, then a firm showing a negative return in my estimate could still provide investors with reasonable expected returns, due to both earned profits and to the higher terminal relative valuation.

In effect, this has simply been a long exercise in finding high forward-beta equities.  The forward-betas of the most leveraged homebuilders should be extremely high, since a bear market would probably kill them and a bull market will lead to positive results from operating and financial leverage, including additional leverage through options on land.  The hypothetical exposure of these builders to fluctuations correlating with the broader market are probably well in excess of 2.

So, this really is just a regular, crude beta play.  Not that there is anything wrong with that.  If you are confident in a bullish forecast, especially in a high equity risk premium environment, grab some beta - as long as you know the risks.  But, I think when betas get this high, there is an added kick that is available to expected returns.  Because, with such highly variable securities, where the terminal valuations will almost certainly be very different than the beginning valuation, the position becomes more of a play on the first derivative of the beta.  If these positions go south, beta will be irrelevant.  The equities will have little or no value in any case.  But, if conditions evolve such that these positions accrue gains, leverage will decrease, operations will normalize, and these firms will start to look normal.  I'm not sure that in extreme contexts, this potential gain from changing beta is efficiently priced.  And, in extreme contexts, this can be a significant source of gains.

One last point from the scatterplot graph.  The most leveraged firms also tend to still have the highest levels of tax assets remaining from the losses they recorded during the downturn.  The red series in the graph reflects market cap with tax assets (both on- and off-balance sheet) subtracted.  Beazer and Hovnanian had market caps at the end of 2013 barely as large as their tax assets, and both have seen falling market caps since then.  And tax assets reflected about half of KBH's equity value.  This adds another layer of leverage.  To a certain extent, with these three builders, there is a discount being applied by the market to their tax assets, based on uncertainty about whether they will be able to utilize them.  This also is a factor which might not be efficiently priced, due to the highly variable, bilateral distribution of probable outcomes.  And, this should be another source of additional gains, to the extent that a bullish forecast is accurate.

It might be the case that the market is generally pricing in a bullish expectation for the industry, but is discounting the lowest quality equities using less bullish expectations.  Or, it could be that the prices of the safer builders reflect the potential gains that would come to them in bad scenarios from the exit of the less stable builders from the market.  This seems unlikely, though, since the riskier homebuilders represent a fairly small portion of the industry.

Here are some individual comparisons of Enterprise Value and Revenues over time.  Beazer, Hovnanian, and KBH are the three riskiest positions highlighted by this analysis.  And, while each of them had Enterprise Values above Annual Revenues at the end of 2013, the difference could be mostly attributed to tax assets.  (My measure of Enterprise Value is total liabilities plus market cap.  Normally, Enterprise Value would be calculated by deducting cash from debt and adding market cap.  I think the lazy version of EV that I am using here gives a relatively stable indication of firm capital, except where firms have large holdings of tax assets.  These have some cash value which will be reflected in market capitalization but should probably be deducted from Enterprise Value in order to create a stable relative valuation measure, compared to the pre-crisis firms.)

Meritage and Ryland are safer options that, at first glance, may have some upside potential in a bullish market.  Note that these firms were priced well above EV/Rev.=1 at the end of 2013, even though they had relatively few tax assets remaining.  So, their higher valuations reflect higher growth expectations and higher terminal valuations.

Note that in all cases, in the pre-crisis period, EV/Rev.=1 provided a fairly stable valuation metric across firms, with all firms approaching or exceeding that level at some point, and several of the firms following that level fairly closely.  Keep in mind that this is a comparison of annual revenues (a flow) to Enterprise Value (a value), so that during the year, prices fluctuate tremendously, and over a period of months the Enterprise Value would tend to fluctuate above and below the revenue level.  Enterprise Value here is a combination of the year end balance sheet and a rough, informal estimate of market cap near the end of the calendar year.





No comments:

Post a Comment