Cyclical companies: ways to evaluate

Hello guys,

This is my first post here, and i’m looking for some practical ideas on how to evaluate a cyclical company. To be more accurate, the company is based in Brazil and all of its revenues come from domestic market. They produce vehicles and trailers (focused on towed vehicles for the agriculture market).

The company business is highly cyclical and dependent on credit supply, it doesn’t have a very big barrier to entry as the products are not high in technology, but they do have a considerable economy of scale and brand recognition. In the last 10 years the ROIC has been fluctuating from as low as 15% to as high as 60% (their WACC is something around 13%). They have been losing market share to smaller competitors, but not profitability.

The manager has a very long and succesfull track record in the business.

Their current P/E is 7,3 and P/TBV is 1,4. There are no comparable companies listed.

Problem: If we try to use DCF, the assumptions for the next 5 years will vary widely and the perpetuity will have a huge impact depending on the FCFF of the last year.

The P/E appears to be attractive compared to other industries in Brazil, but it could be due to the market forecasting terrible years to come (the question is whether it’s being “too pessimistic”).

The P/TBV is on its lowest point ever, but a mean reversion may not happened if company loses its pricing power (ROIC dropping as there are no barriers to entry).

My idea so far: i’m looking for the “normal” year of the company. Even though the company was able to generate a lot of excess value in prior years, i don’t think it will be able to keep that up. For that, I’m setting a “normalized” ROIC of 20% and considering the Invested Capital I get to the “normalized” NOPAT.

From here, I have assumed that the company is maintaning its market share, thus growing the same thing as the country conomy in perpetuity. And now i consider that in perpetuity the EV should be:

EV = (Normalized NOPAT)*(1-Reinvestment Rate)/(WACC-growth)

With the reinvestment rate being calculated as:

Reinvestment Rate = (growth)*(1-efective tax rate)/(Normalized ROIC)

From this, I’ll get the EV subtract the Net Debt and arrive on the Market Cap.

I would like to hear ideas and sugestions from other researchers.

Determine the length of one business cycle for the industry. Take two business cycles and normalize them (usually around 10-15 years). Take only the historicals of one business cycle starting today if it is more relevant to current operations.

Normalize the IS and BS for these years with OLS, proforma statments if nessecary. Use the TREND function in excel. Project one more business cycle into the future using the normalized growth rates, or any additional assumption you might have. Project another extension cycle with their respective assumptions. Then determine the contiuing value using RONIC=WACC.

You could make it more accurate with detailed segmental revenue, then project volume of each product based on the trend of market share going forward (gaining, holding, losing? How fast?), then multiply that by the estimated price of each year. Then model the rest of the IS as a % of Rev (except for deprc).

Seems to me this modelling thing is very hocus pocus. For one thing, no two business cycles are alike. Duration and amplitude are quite different. Then you have the company fundamentals to deal with. A cycle of its own. Lots of assumptions here.

You model based on all the available information. Even if that doesn’t reflect the imminenet future, you are still correctly valuing the security at hand. While appropriately adjusting as new information becomes evident.

That is the whole concept of modeling. Who are you to assume that the next business cycle will not resemble a normalized previous cycle? That doesn’t make you more right than I am, nor it does me either. But the unfolding of information will converge the security price to it’s intrinsic valuation assumed beforehand.

You cannot assume things when you operate with your own money. With someone else’s money, maybe. But not with your own. There is a fundamental difference between this coming cycle and the previous. You cannot assume. You have to know or know with a high degree of certainty what will eventually take place.

I am intrigued by the model thing in AM. Seems like everyone is doing modeling. Just like everyone is trying to get a CFA. When everyone wants something, it usually means it is in the state of a BUBLLE. But my question is: do you let the model dictate what you think? Or the models just confirm or support your thinking process?

You have to assume at some level in your model. If you know exactly what value said security should be priced at, then we’d all be rich. Well actually, the market will become efficent faster than we can make any money in that case.

You cannot know with a high degree of certainty, modeling is a science just as much as it’s an art. It is critical that you do not make common valuation mistakes, but you will reach a point where there is simply not enough information to project with. Leaving you digressing, hopefully not too much, on some variables, especially as your implicit period lengthens and goes forward.

Fundamental analysis gives you good investment selections, modeling tries to draw intrinsic values to support purchases for your portfolio, checking if the security is not too expensive, and drawing conclusions on the present value given an approximate framework as a going concern.

The models don’t dictate what I think, nor do they confirm or support my thinking process. Modeling should be as unbiased and complete as possible. If the valuations conclude that the price of an investment that interests me is, at most, fairly valued. Then I should consider buying it. The same principle goes for shorting or selling.