### What is the Dividend Discount Model?

The Dividend Discount Model (DDM) is another popular valuation methods to value a company. The theory is that the worth of a company TODAY is equivalent to ALL future dividend payments discounted back to present value.

### Why use Dividend Discount Model in REITs?

I find DDM to be particularly applicable and relevant to REITs because REITs are mandated to pay out 90% of their distributable income as dividends. Unlike stocks, you don’t have to worry about certain years that the company don’t pay dividends. If DDM is used on a company that pays IRREGULAR distribution of dividend payout, the valuation outcome would not be useful. Since REITs are guaranteed to pay out dividends every year to get tax exemptions, we do not have to worry about this issue.

### Different types of Dividend Discount Models

#### 1. Zero-Growth Dividend Model

There are several types and variations of DDM models out there. The first is zero-growth DDM. But we know there are no companies out there that is growing at 0%. This is more applicable for perpetuity bonds, perpetual securities where a fixed coupon payment is paid annually without any maturity date. Since not many REITs out there issue perpetuities, I would just skip this.

#### 2. Gordon Growth Model

The next type of DDM model is known as the Gordon Growth Model. In this model, we add in an additional assumption: **growth rate**. Here we expect that dividend payouts cannot possibly remain constant forever, there must be some kind of growth in dividend payouts over the years. However in Gordon Growth Model, the growth rate is ASSUMED to be constant.

So what we are trying to do with Gordon Growth Model is to ask if this company was to pay out dividends growing at x% every year for the entire life, how much would this company be valued at? The formula is:

There are 3 simple things to find out. **First** is the dividend for next period which is calculated by current dividend x expected growth rate. **Second** is discount rate which we will use Capital Asset Pricing Model (CAPM) to calculate. And **Third**, is the growth rate of the REIT’s dividend payouts. I will use Parkway Life REIT as an example as I love Parkway but haven’t got the chance to got in yet.

I calculated the dividend growth rate of parkway life using their BASE dividends. Exception dividend payouts are excluded as I see them as bonus which does not reflect the underlying growth of dividends. It’s better to be conservative in valuation too. Here we can see that average growth rate and compounded annual growth rate is relatively the same, I will just take 7% to be Parkway’s growth rate figure.

Now that we have got the growth rate, we can calculate what is the expected dividend payout in the next period. Its simply the dividend payout on the last period FY 2017 ($12.46) x the growth rate (7%) = $13.30.

We have the growth rate, we have the expected dividends for next period, all that’s left is just our discount rate which is calculated using CAPM.

Beta measures the volatility of a stock (REIT in this case) against the market. A beta closer to 1 means it moves exactly in the same tandem as the index. A beta less than 1 means its price are less sensitive to movements in the market. Parkway’s beta of 0.66 is taken from Reuters and its no surprise that its beta is 0.66 since Parkway is a defensive, low-volatile REIT. Market Risk Premium is taken from KPMG’s equity research report and risk-free rate takes the figure of a 10-year Singapore government bond yield quoted by Bloomberg. After plugging in the figures, expected returns from shareholders came up to be 5.83% and this is our Discount Rate. So far so good.

But here comes the problem, our calculated discount rate of 5.83% is lower than the calculated growth rate of 7%. That would give a negative value and a REIT’s price can’t possibly be negative. The reason why growth rate can’t be HIGHER than discount rate is because we are discounting future dividend payouts at perpetuity. If dividend growth is higher than discount rate, theoretically speaking, all the money in this world would invest in this REIT and the value would reach infinity.

### Valuation is about STORY + numbers

This is where the challenge of valuation lies. The challenge is not in understanding the formulas or calculating the figures. The workings of DDM formula can be easily found online from google or YouTube. But the *art of valuation* comes from the **story**. I learnt this from Aswath Damodaran, a corporate finance/valuation prof in Stern School of Business at NYU. This guy is really solid. Check him out if you want to go deep into valuation models.

All along I have always thought valuation is a science, its a quantitative and number game, but I did not pay so much attention on the qualitative story. He mentioned that a good valuation has a good story and I realized that’s the hard part. Alot of times that requires insider knowledge, ground experience, talking to management from within, understanding the industry and we seldom have access to all these resources. **The output of one’s valuation is ONLY AS GOOD as the input of one’s assumptions.** I won’t say my assumptions are good *(far from it)* but let’s try again and see how we can refine the above assumptions to come up with a better valuation model.

### Growth Rate Assumptions

Firstly, estimating the growth rate of one’s company is already not an easy feat to begin with. There are so many possibilities, uncertainties and outcomes of what could happen next and nobody knows. The best we can do is to come up with assumptions that REFLECTS the economic conditions, underlying business model and the story of its business.

This is Parkway’s year-on-year growth rate over the past 10 years. The immediate observation is a declining growth trend. Does it still make sense to simply say Parkway Dividend payouts is going to grow at 7% FOREVER? Probably not… but I have no idea what would be the best way to arrive at a constant growth rate figure. I have tried playing around with moving averages, weighted-moving averages, exponential averages but have not found one that quite makes sense. The idea of using a constant growth rate is already flawed to begin with.

Nobody knows what are the potential acquisitions in the pipeline, what is the yield-accretive % in dividend growth, or the extent of using equity financing in the future. Economists can’t even predict the economic outlook for the next year, what more perpetuity? So you kind of get the idea that its’s almost impossible to predict a perfect arbitrary growth figure as there are too many uncertainties and variables out there.

The way I would do it is to place more emphasis on the recent 5 years and make a conservative estimate of 3%. This is a discretionary approach and by no means require any form of statistical assumptions or “story” behind it. Its not ideal but let me know if you have a better way to come up with an estimated constant “growth figure”. Till then, let’s see what’s the value of Parkway Life is if we use 3%.

### Sensitivity of Growth Rate Assumptions

Hmm… $4.53. Expected dividend is calculated by taking dividends for 2017 12.46 x 3% growth rate which gives us 12.83. Discount rate we will just stick for the current assumptions at 5.83% to keep things simple. Is $4.53 a fair value? What if I decided that growth rate should be 3.5% or 4%, how would it change? Let’s do a sensitive analysis to show how sensitive the outcome would be.

This is another problem of valuation, results are VERY sensitive to changes in the growth rate and discount rate assumptions. All i am changing is just the growth rate by 25 basis points and you can see how the value of Parkway grows exponentially. Even though the difference between growth rate of 3% and 5.5% is relatively small in absolute terms, its valuation outcome ranges significantly from $4.53 to $39.83!

The stark contrast is due to the fact that we are playing with perpetuity. As growth rate approaches cost of equity or the discount rate, the value of Parkway would approach infinity. The reason as mentioned above is because if growth rate exceeds discount rate, every dollar in the world would flock to this asset and thus its value would reach infinity.

### Is Dividend Discount Model still useful?

There are 2 metrics I would use to decide if DDM is applicable. The first is **coefficient variation of growth rate** and second is **comparing earnings growth rate and dividend growth rate**.

### Coefficient of Growth Rate Variation

Dividend Discount Model is more applicable IF the growth rate of the REIT’s distribution history is less volatile, constant and stable. The more predictable one’s growth rate is, the less error there is in approximating the growth figure. One way to measure this is to find out what is the coefficient of variation.

This is calculated by taking the standard deviation divided by the mean. We already got the mean which is calculated to be 7% earlier. As for standard deviation, I used excel’s formula STDEV and it outputs 4.13%. So 4.13% / 7% gives us the RATIO of standard deviation to the mean. The higher the coefficients means data points are highly dispersed away from the mean. Vice versa, the lower the coefficients means data points are closer to the mean.

Parkway Life’s growth rate coefficient of variation is (4.13% / 7%) **0.59**. Value closer to 1 equates to higher volatility and standard deviation while value closer to 0 equates to lower volatility and standard deviation. Since Parkway Life coefficient is above 0.5, attempting to derive a constant growth figure is much harder to do so and there is a higher room for error in our assumptions.

**Earnings Growth Rate vs Dividend Growth Rate**

Comparing earnings growth rate and dividend growth rate can also give us an idea of the soundness of using Dividend Discount Model. Theoretically speaking, earnings growth rate shouldn’t be LOWER than dividend growth rate. Neither can it be HIGHER than dividend growth rate.

To illustrate, let’s say earnings growth rate is 4% and dividend growth rate is 6%. Over perpetuity, dividends would increase exponentially over earnings and that would end up in the firm having insufficient cash flow to pay out dividends.

Conversely, if earnings growth is 6% and dividend growth rate is 4%, The dividend payout ratio would eventually diminish to zero at perpetuity. Firm has more and more cash, but RATIO of dividends payout to cash is getting lesser and lesser. This is going to affect REITs as they are required to pay out 90% of their net income as dividends.

DDM model works better if dividend growth rate is similar to its earning growth rate. This ensures that both earnings and dividends are growing at the exact SAME pace and dividend payout ratio can be maintained at 90%. We have already derived Parkway’s average dividend growth rate which is 7%. Let’s see if Parkway’s average earnings growth rate is also around 7%.

Parkway’s earning growth rate averages out to be 20.17% but its average dividends growth rate is growing at 7%. That’s completely off. However… compounded annual growth rate is 3.98%, which is very similar to our previous assumption of 3% earlier when doing DDM valuation.

### Reasonable assumptions?

The question then is whether is it appropriate to use compounded annual growth rate instead of average. I would think it is appropriate because compounded growth rate is actually measuring the CONSTANT growth rate that brings initial investment to the ending investment. That’s to say, Parkway net income was actually $68.64 million in 2007. If it grows at a constant growth rate of 3.98% with profits reinvested over 10 years, its investment would be $101.46 million in 2017. That’s exactly what we are looking for isn’t it? the “constant” growth rate. The limitation here is just past growth rate doesn’t necessarily predict future growth rate.

### Revised Valuation for Parkway

Since our earlier assumption of dividend growth rate is 3% and earnings growth rate is concluded to be 4%, I think it would be fair to take the average of both which is 3.5% as our constant growth rate. The assumption is that BOTH dividends and earnings have a constant growth rate of 3.5%. This ensures dividend payout ratio remains constant at 90% and it would strengthen the soundness of our model.

Now that we have decided on the constant growth rate to be 3.5%. Let’s see how can we adjust the CAPM assumptions. Instead of just blindly taking 5.5% as our Market Risk Premium from KPMG’s equity report, maybe we can substitute that with a better metric to refine our assumption of investor’s expected rate of return. That metric is the **Singapore’s All Healthcare Index**. I used this because If an investor who wants to get into healthcare REIT or stock would use the healthcare index as a benchmark for market return. It is also possible to use STI return as an assumption. There are no hard rules around valuation and it really is subjective.

The SGX All Healthcare Index comprises of 31 constituents (Including Parkway Life) and it measures the performance of listed healthcare companies in Singapore. Based on SGX’s report, the 5-year average annualised returns from this index is 8.1%. Hence, Market premium would be revised to (8.1% – 2.2% = 5.9%).

As compared to the earlier valuation, at least now we have a better assumptions behind our DDM model. We have the growth rate, which is approximated by taking the average of Parkway’s earnings compounded annual growth rate and our 3% dividend growth rate. We also have the discount rate, where I only changed the market risk premium to put Singapore’s healthcare index into context. With these 2 figures, its sufficient to use DDM to value Parkway and it turns out to be $4.97. The above assumptions we came up with may not be the best, so take it with a pinch of salt.

As always, I always add in another 20% discount to the final value to give a bit of margin for error. Therefore, final fair value of Parkway is calculated to be **$3.98** (80% of $4.97).

### 3. Multi-Stage Dividend Discount Model

The last way of modifying the Dividend Discount Model is to say, companies do not grow at 0%, neither do companies grow at a constant growth rate, but instead, companies grow at variable growth rate cycles. Even though the variable growth rate model simulates a more realistic scenario of a company’s future, estimating the different growth rates itself is also another challenge. Assumptions are always a challenge, regardless of which valuation models.

In the multi-stage dividend discount model, it is generally assumed that every company has a life cycle just like every living thing in this universe. Companies rise and decline, empires rise and decline, humans grow old and die etc. The cycle of a company begins with start up, young growth, high growth, mature growth, mature stable and lastly, decline. If you want to understand more about life-cycle of companies, be sure to check out the video I posted on my FB page. It will definitely be worth it!

Anyways, multi-stage DDM can be broken down into 2 parts. The **two-stage DDM** and **three-stage DDM**. Two-stage DDM simply means a company would experience higher growth rate for the first couple of periods before stabilising to a mature long-term growth rate over perpetuity. Three-stage DDM is a variation to the two-stage DDM. in Three-stage DDM, it added in a transition phase of gradual decline growth rate in between. For example:

2-stage DDM:8%, 8%, 8%, (beginning) 3% (perpetuity)3-stage DDM:8%, 8%, 8%, (beginning) 6%, 5%, 4%, (transition) 3% (perpetuity)

Actually, there are no fixed rules to the variable growth rates in 2-stage, 3-stage or whatever stage. If you have the necessary resources to get an in-depth growth rate assumptions, you can tweak it to the specific figures for each year into the future. For example it could be 25%, 25%, 20%, 20%, 15%, 10%, 5%, 5%, 5%, 5% etc. After which, just discount back the PV value for each year and sum up all future dividends.

The surprising thing about Parkway’s behaviour is while net income has grew over the years, its dividend growth rate is declining. Its understandable if you think about it as Parkway is still a relative young company with only 10 years of experience. Retained profits definitely has to be channelled towards acquisitions to increase its asset portfolio in the long-term. Warren Buffet has also states that paying dividend should be the last resort in corporate management and Berkshire Hathaway does not pay out dividends at all. It uses all its money to re-invest and grow its business.

Keeping that in mind, we know that REITs are unlikely to pay out dividends at “high” growth rate especially when a company is young and starting out on its acquisition spree. Actually, the entire REITs industry in Singapore is still relatively young as it only started out in 2002 when CMT first got listed. But Singapore is also a very small market with limited space, many home-grown REITs have already covered most of the lands around and are looking abroad to grow their asset portfolio. So its really hard to say if we should consider these REIT companies as young or mature.

Nevertheless, it is still hard to put Parkway’s case into a multi-stage growth DDM as its dividend growth rate is not even HIGH to begin with. This is because the money is used to acquire nursing homes and healthcare providers facilities in Japan throughout the decade. Generally, DDM model is more applicable for mature to declining companies that have already stopped searching for growth and is slowly distributing cash back to shareholders. But for now, we don’t really see that in Parkway as its still in the growth stage.

### Summary

If there is one thing to take away after reading this post, it is to know that valuation is NOT just about numbers but also the story of its business and assumptions behind *(even though I did not cover much of the ‘story’ in this post)*.

Dividend Discount Model (DDM) is a simple formula to calculate fair value of an asset. There are 3 main variations of DDM: zero-growth, constant growth and variable growth rate. DDM is applicable to REIT because REITs pay dividends consistently unlike stocks. One way to check whether an asset can use DDM with a higher confidence is to compare earnings growth and dividend growth OR compare coefficient variation of dividend growth rate. Another check of using DDM is to evaluate the life cycle of a company. DDM is more applicable for mature stable companies instead of companies that are in its early growth stage.

As much as I try to come up with a fair value figure for Parkway, I can’t be sure or confident of my valuation as I am well aware that when it comes to valuation, there are way too much uncertainty variables to quantify our assumptions.We can’t be 100% sure in trying to predict or value a company, there are simply way too much uncertainties and variables out there.

Another thing is that assumptions of growth rate and discount rate are EXTREMELY sensitive to the fair value outcome. Most of the time we have a pre-conceived notion and bias on what price range should be. Our assumptions would then “fit” what we thought the price should be. I could just tweak the rates around and come up with a story to value Parkway near its traded price and leave it as it is. But I know that would only be discounting myself and I would be giving biased valuation figures. Remember that the quality of valuation is only AS GOOD AS the input of assumptions.

As I work on the “story” side of macro factors, I would further refine some of the above assumptions, come up with new methods and see if we can get a clearer story behind Parkway or any other counters. Till then, keep a lookout for the next post and keep learning! Learning is key to wealth!

Thanks for the write up! Very comprehensive article and glad to have a takeaway or two from this! š