Hedging is the attempt to reduce exposure to risk from fluctuations in the market. Unlike many securities and futures, physical real estate assets cannot be perfectly hedged. Property derivatives are based on indices that tract specific types of property values for broad regions. The location and structural attributes of a specific property may not correlate well with the index. If the property is located far from the index area, there may be no correlation.

In physical real estate, it may not be economically practical to hedge the entire value of the property. Therefore, we will need to ascertain how much of the property's value is at risk and how much of that risk we are comfortable with carrying.

Hedging Real Estate

As mentioned, we can hedge a portion of the property's value that we feel is at risk. Larger institutional investors (also speculators) would utilize swaps and forwards to limit some of their market exposure. For the small investor/speculator, we can utilize the real estate futures with CME Group. Utilizing the future contracts, we can attempt to limit our loss in property values if the market drops. A future can also be used for hedging against an upswing in market values. We will look at hedging against a down market, and follow with hedging against a rising market.

We are looking flip a house in Winter Garden, Florida. The house costs $110,000 in a neighborhood were similar homes had sold for $130,000. Instead of purchasing an existing home, we could also be looking to sell a new "spec" home for $130,000 with a total construction cost estimate of $110,000. Either way we are looking to capitalize on a property value gain between our costs and the sales price. Additionally, we are interested in hedging any unexpected drops in property value. To hedge real estate does require us to do some analysis.

Winter Garden is located near Orlando, but the closest index that trades is Miami. Although both cities are in Florida, there is no guarantee that property values in one city will have anything to do with the other. We should probably limit our index comparisons to the MSA's if we are in the same state and the Composite-10. If we are not in the same state as one of the traded MSAs, we should stick with the Composite-10. Since we are looking to hedge market movement, we need to check if our market correlates with the selected indices and choose the "best fit". A correlation implies that there is a relationship between the price movement of our house and the index. The analysis is simple and can be done on spreadsheet software. First, obtain some monthly data on our property's zip code or city for average home values. Today many real estate websites have regional transaction data that can be downloaded directly into a spreadsheet. Next, download the S&P Case-Shiller Home Price Index from the Standard & Poor's website. They will give you all the monthly historical data for twenty-two indices. Since there are futures on only eleven of the indices, we must seek out the best fit from the smaller traded group. If you are lucky enough to be in one of the ten MSAs, you may still want to check how well the neighborhood prices correlate with the index.

When comparing the data, we are going to run into some technical issues. First, we are comparing residential transaction prices with a filtered repeat sale single-family price index. We are not comparing "apples to apples", but our purpose to find an index that is a reasonable representation of local home value movement. Next, none of the dates the values were published will likely match. Here we will need to make a judgment on how to line each set of data up against the other. For example, the Case-Shiller indices are published at the end of the month. Transactional data for Winter Garden is available, but is based on the 15th of the month - in the middle. Should we stagger the transaction data to match with the end of the month, or match it to the beginning of the month? Since the indices do not move too drastically each month it probably does not matter, but each case could be run to see if anything changes drastically in the correlation. Now we should plot our data on a chart to "eyeball" our property values compared to the indices of interest. The property values can be put on a secondary vertical axis so the data overlays correctly (See Figure 1 below).

Figure 1, Chart comparing actual housing values to indices.
A chart to compare local home value changes in Winter Garden with the Miami and Composite 10 CSI indices.

The chart plots the Winter Garden transactional property values against the Miami and Composite 10 home price indices. Looking at our cart, we can see that data tend to move together reasonably well. The Miami index seems to be a better fit than the Composite 10 index, granted that Orlando and Miami are similar with heavily tourism based economies. Before we decide to utilize Miami index futures, one last step needs to be done. We are interested in comparing the monthly changes to each other to see how well the data matches for month-to-month changes. To do so, the following steps need to be taken with the data. First, in a separate column we will compute the log-returns. This is simply taking natural log (LN) of the current value divided by the prior value.

LR = LN[ (Value t)/(Value t-1) ]

Compute the log-return for each index considered and our local property value data. What we end up with is equivalently equal to the percentage change between the monthly values. If we place our returns on a chart, we can see how well each month compares to each data set (See Figure 2 below).

Figure 1, Chart comparing Log returns for housing prices and indices.
A chart to compare the Log return home value changes in Winter Garden with the Miami and Composite 10 CSI indices.

It is not as easy to "eyeball" a winner in the returns chart. Each index and our property values seem to agree in degree for major price movements, but looking close they seem to be all over the place. It is the correlation in the monthly log-return data that we are interested. Therefore, we can run a correlation calculation between our Winter Garden log-returns and each of the indices log-returns.

The correlation coefficient ranges in value from +1.0 to (-1.0). A coefficient of one would be a perfect correlation of 100%, whereas, a coefficient of negative one would indicate the data is perfectly mirrored (when one chart moves up a point, the other moves down a point). The closer to one a coefficient is, the better the linear relationship between the data sets. A coefficient of zero means that no relationship is present. For real estate, we would be looking for correlation coefficients from zero to +1.0. Although there may be a situation where a negative correlation could exist between a property and an index, it is unlikely. So what does coefficient value need to be? Here is the "rules of thumb". A weak relationship gives a value between zero and 0.3. Between 0.3 and 0.7, the relationship is moderate. From 0.7 to 1.0 the correlation coefficient indicates a high level of relationship between the data.

How did we do? Looking at the table with the correlation results (See Table 1 below), we can see that the Winter Garden log-returns have a high level of relationship with both the Miami and Composite-10 Case-Shiller Indices log-returns. The Miami index, however, is better correlated to Winter Garden than the Composite index. Therefore, we will look to utilize the Miami futures to hedge based upon the Miami index. Typically, another analysis would be done to test the hedge efficiency of the derivatives to the underlying, however; the property derivative markets are illiquid compared to other markets and the index cannot be replicated. We will need to run some numbers to figure out our optimal hedge ratio of the project costs based upon our appetite of risk.

Table 1, Log Return Correlation Coefficients
Winter Garden Miami Index Composite 10
Winter Garden 1.000 0.895 0.787
Miami Index 0.895 1.000 0.900
Composite 10 0.787 0.900 1.000

Going back to our Miami index log-returns, we will make a histogram with our spreadsheet to see the frequency of price movements in the data.

Figure 2, Miami S&P Case-Shiller Index histogram of log returns.
A chart to compare the Log return home value changes in Winter Garden with the Miami and Composite 10 CSI indices.

The histogram shows us some interesting information (See Figure 2 above). First, most of the price movement in the past has been positive, with 2% being the most common price increase. The highest monthly gains were at 3%. The losses seem to have their own "typical" value around (-2%), with the highest losses at (-4%). We know the over speculation during the real estate bubble has played havoc with our chart, and the true distribution of prices is probably not as skewed as our histogram shows. It may be reasonable to remove the bubble, but this is the exact behavior we are trying to hedge with/against - probably better to consider. On a more technical note, the log-return data is not distributed normally and is skewed to the left. The distribution may be a stable distribution. Briefly, this means is that if we utilize probability theories based on a normal distribution we will underestimate the probability of the higher market movements. To make life easy on ourselves, we will look at simply computing the worst, middle, and best cases. From scanning the results of our scenarios, we should get a good feel for what we are willing to pay for our hedge.

Which price movements should be considered? The percentage change per month is not too wide a range, from (-4%) to 3%. Therefore, to keep our calculations relatively simple we will look at four scenarios were the Miami housing price index moves by (-4%), (-2%), 0%, and 2% per month over our project life. The (-4%) monthly change is our apocalyptic scenario, while the ranges between -/+2% seem to be where we could expect to end up.

If we wish to be technical, we can compute some statistics for the Miami log returns. The mean computes to 0.003, which is effectively zero. With a standard deviation of 0.017, our returns would generally be between +/- 1.7%. Usually around two thirds of the data falls within the standard deviation. We could continue on assigning probabilities to index movements to determine what we wish to model. However, since the distribution is non-normal we would need to apply sophisticated analysis to the data to get reasonable results. Either approach of "eyeballing" or analysis can be equally effective. Even with the most rigorous analysis of past events, we still cannot predict the future any better.

It is time for us to research the prices of the Miami S&P Case-Shiller Home Price Index (Miami) futures to see what is available. Before we begin, we are estimating our project to take no more than six months - eight months worst case. Therefore, we should be looking for contracts that will match our project timeline as best possible. It is the market risk during the project where we are most vulnerable. Looking back at the Miami futures, we find that there are some contracts available six months (May) out and nine months (August) out. The Miami index is currently at 149.09, the May contract is priced at 154 (ask) for a long position and 140 (bid) for a short position. The August contract is priced at 154 for a long position and 138 for a short position. If the prices were not to our liking we could always post a limit order at the price we are willing to accept. The biggest risk with the limit order is that it could go unfilled. In any event, we are interested in a short position since we are trying to protect ourselves from a downward movement in the market. Remember that a short position gains value as the market prices fall - similar to a put option. Upon reviewing the bid prices on the futures, it would seem the short positions are anticipating the index to decrease in value two basis points over a three month period, about 1.3% or 0.45% per month. Being optimistic about our project management prowess, we will settle on the six month contract (May). If the project begins to drag on, we can always take another position with the Miami futures. Additionally, we are picking up two basis points of index movement.

Now we need to run some calculations to see how much we should be hedging our property to get to a level of risk that we are willing to accept. We will investigate three different levels to hedge: the required equity, the mortgage, and the expected sale value. Because the property is considered an "investment" property by the financial institution, we will be required to put 20% down. Again, the calculations are easily handled on a spreadsheet.

Table 2, Hedged property senarios
Hedge Model
Total Cost $110,000 Est. Sale $130,000
Equity $ 22,000 Future Bid (Short) 140.00
Mortgage $ 88,000 Current Index 149.09
Hedge Positions
Value Contracts Cost Notes
$ - 0 $ - No Hedge
$ 22,000 1 $ 2,025 Equity Only
$ 88,000 3 $ 6,075 Mortgage Only
$130,000 4 $ 8,100 Total Est. Value
Monthly 6 Months Home Sale Future Settle
2% 12% $145,600 166.98
0% 0% $130,000 149.09
-2% -12% $114,400 131.20
-4% -24% $ 98,000 113.31
*Returns based on change in property price and number of future contracts.
Contracts 12% 0% -12% -24%
0 $35,600 $20,000 $ 4,400 ($11,200)
1 $28,855 $17,728 $ 6,600 ($ 4,527)
3 $15,364 $13,183 $11,001 $ 8,819
4 $ 8,619 $10,910 $13,201 $15,492
% Returns
Contracts 12% 0% -12% -24%
0 32% 18% 4% -10%
1 26% 16% 6% -4%
3 14% 12% 10% 8%
4 8% 10% 12% 14%

Table 2 shows the projected monthly index movements that we had ascertained from the log-return histogram. Continuing the index movement out for six months, the index change ranges from a 12% increase down to a 24% decrease. The value of the hedge represents either no hedge, the 20% equity, the mortgage, or the expected home sale value. To determine the number of futures contracts to purchase for the hedge, the "value to hedge" is divided by the future contract value based on the bid price (140 x $250). The fit is not perfect, so we end up rounding the number of future contracts up to cover the required value. Our estimated property sales price is based upon our expectations of the local market at each level of expected index change. If home values did not change, or changed very little, we would expect to get our sales price of $130,000; 0% index movement. Since the index is measuring housing value, any gain or loss in the index level would be reflected in the estimated sales price. The property value gain/loss is the "estimated property sale" minus the total cost of the project, $110,000. To determine the estimated future value of the index, the six-month index change was accounted for in the current index level at 149.09. The gains or losses on the future contracts would be the value difference between the bid price at 140 and the estimated future value. The value difference is then multiplied by $250. It is assumed that the future contracts would be held to expiration. Finally, the total gain/loss is determined by adding the property and future contracts gains/losses together. To make it easier to compare the results the "total gain/loss" was divided by the total costs of $110,000 to get a quick gage of profit or loss.

The example has been simplified a bit. We do not account for brokerage fees, closing costs, insurance, and finance payments. All of which would need to be modeled in the projects initial feasibility study that can now include hedging calculations. Albeit our example is simplified, it shows how hedging can work for us. The return on the house will probably fall between the +/-12% range. The implied housing return on the index expects our August contract in six-months to be 2 points below the current index level, which would equate to a (-1.4%) change. If the market is correct, we will end up closer to the zero percent change.

Looking at the "profit/loss" column we see that without hedging we span the return range from a positive 32.4% down to a (-10.2%) loss. Bracketing our concerns to the +/-12% range, the profit/loss is between 32.4% and 4%. The futures market is putting us closer to our 0% projected index movement for an 18.2% profit - or just below. In all of the above cases, we still make a profit, except if housing values in Miami drop by around 16% (calculated - not in the chart). Below 16%, we begin to lose money. Remember, all of our calculations are approximate since the index is in Miami and the house is in Winter Garden. There is nothing to prevent our home in Winter Garden from dropping in value while the Miami index increases.

In every case, the hedged positions will reduce the gains if housing prices increase. There is a price to pay for protection - there are no free lunches. The concept behind a hedge is to limit losses for unexpected market moves, a downward price movement in our case. Scanning across the "profit/loss" column, we can see that the hedged house projects do indeed limit our losses when the index loses value. The single futures contract hedging our equity does lose money in the extreme case, but at half the magnitude of not having a hedge. Hedging against the mortgage value, with two contracts, returns a profit in every case. When hedging the entire sales value with four futures contracts, we also profit in every case. However, hedging the entire sale reduces our profits by a significant percentage when the index remains the same or moves upward. Which level of hedging is best?

The hedging cost cannot be financed in the house purchase; it is a speculative investment, not real estate. Cash must be available to place the futures performance bond, plus a bit to spare in case of a margin call. In each hedged position, a margin call for every contract would have been made if the index rose 12%. If we are tight on available capital, one contract may be our only option. If capital were available, hedging the mortgage value with three contracts would seem to make the most sense. Although we are giving up 6% of profits if the index remains flat, having a reasonable profit in almost every case we modeled is appealing. Far from being a guaranteed, "locking" in a profit may make more sense than taking the risk additional return.

Another point needs to be made. The contracts are settled at expiration unless we have closed our position early. If we will owe money on our contract positions, we will have to pay at settlement in cash. If we are expecting to get the cash to settle from the sale of the house, we should probably give ourselves a bit of wiggle room for the sale of our house and opt for a contract further out.

Hedging against increasing value

Hedging a project against a downward index movement makes sense for physical assets. To hedge real estate against rising values would be beneficial for speculators looking to enter a market in the near future. Supposed we would like to purchase a "fixer-upper" in San Francisco, but cannot make the purchase for a few months. To protect our purchasing power we can purchase forward positions in the San Francisco futures. How large a position we take depends on our financial ability, but it would probably make the most since to cover the anticipated down payment. If property values continue up, the index should reflect it with rising values. Had we purchased the futures at an appropriate price, we should also see an increase in the value of the contracts. Ideally, we would have offset the increase in property value with the increase in value of the contracts. Our down payment should have grown in value to match the increases in the market. If property values decrease, our contracts should lose value in approximately the same proportion. Although we have lost the amount of down payment available, the properties of interest should have also lost value. The down payment value should still be sufficient.

Prefer to flip pages instead of scrolling?

Synthetic Real Estate Book Image
Buy the book, Synthetic Real Estate Investment for the Small Investor.