Monday, May 2, 2011

Using comparables for the valuation of internet startups at emerging markets

PART #2. Country Risks

As it was mentioned in the previous post we deal with a theoretical Russian Company which is getting compared to the set of mature rivals In the USA. Obviously markets are different. There are higher political and economic risks in Russia. We should take this fact into account. Let’s have a try.

There are several different approaches. To make a cross-border adjustment we can:
  • Calculate Sovereign Bonds Yield;
  • Find a correction coefficient through the market multiples ratio;
  • Use smart multiples;
  • Do something more adequate.
You can easily find description of those three approaches in those articles which have already been mentioned in the previous post. I just want to make several comments on them.

Sovereign Bonds Yield
Making adjustments of market multiples according to the non-market variable seems quite strange. Nevertheless nobody can forbid us from trying to do that on our own. Let’s take Russian Eurobonds and U.S. treasuries with the maturity in five years.

Correction multiple is 0.675. As we see the yield of Russian Eurobond is higher. We can make an assumption that the spread between euro-5 and tn-5 was provoked by the higher level of risk in Russia. On the other hand, I don’t see any logic in taking this method as a fundamental one.

Market multiples ratio
Nothing bad can be told about this kind of approach. To do a cross-border adjustment according to the Market Multiples Ratio method you have to gather information about multiples from both of the markets. In our case this markets are ones in Russia and USA. To get a result you need to conduct a calculation in accordance of the following formula:

Calculating the market multiples ratio can be proved with a simple sense of logic, what can’t be told about the Sovereign Bonds Yield Ratio. But as for me I don’t have a lot of time to collect everything needed for the calculation. That’s why this method seems to me unreasonably resourceful.

Smart multiples
The farther into the forest, the thicker the trees. In few words smart multiples aggregate information about market and specific risks. I can describe it in more details according to your request. To find out more about them I recommend you to read [Bhojraj, Sanjeev, Charles M.C. Lee, and David T. Ng. 2003. International Valuation Using Smart Multiples. Working Paper (March 14)].

These were the methods I had a chance to find the information about. I am sure that there are many others. And now I’d like to share with you my own approach based on calculating the ration between market indexes.

This method seems to me quite useful for several reasons. Among them:
  • Doesn’t need to gather a lot of information.
  • It has a good theoretical underpinning proved by classical statistic analysis.
  • Based on market variables but non-financial somehow related staff.
The idea is that we can make cross-border adjustments according to stock market indexes. We will calculate the ratio between S&P 500 and RTS Classic. Though there are plans to build an international financial center in Moscow we should face a reality - the liquidity of the Stock Exchange in Russia is dramatically lower in comparison with one in the US. Moreover there is no good alternative for NASDAQ in Russia, though it’s well everywhere that Russia has an extraordinary technological potential. It's a pity but let's get straight to the point.

where t is a set of months in the period from September 2009 to the April of 2011.

Let’s make a cross-border estimation with this method. The information needed is presented in the table below. All calculations were made in Stata.

There is nothing easier to get the estimation of correction multiple needed:

Let’s get to the statistical proof of this method. It seems reasonable to say that if there is a significant linear regression with RTSI and GSPC as variables we can say for sure that our approach is statistically correct.
The correlation pole for these two indexes looks in the following way:

The correlation is evident. Now it’s high time to see at the results of Stata analysis.

We can observe that the probability of obtaining a test statistic for the model is almost 0, the same is with the RTSI as a factor. As a result we can conclude that our model is significant. Bingo!

In conclusion I place histograms for RTSI and GSPC for the reason that they show the way how variables are distributed.

Summarising everything I can say that cross-boarder adjustments are always subjective. Different approaches have a different combination of complexity and accuracy. It's up to you to decide which kind of method to chose. What methods do you use? Would you like to implement mine?

No comments:

Post a Comment