I just finished reading the book "When Genius Failed" by Lowenstein, an extraordinary account of the story of Long Term Capital Management (LTCM) a hedge fund created in America in the second part of the 90s by formers Salomon Brothers bond traders and 2 Nobel Prize Professors, Robert Merton and Byron Scholes (Nobel Prize Winner 1997 for the BlackScholes Option Formula).
The book analyses the boom and bust story of LTCM, that in 5 years of existence was managing at one point over 125 billion dollars but at the end was responsible for the financial crisis of 1998 that almost blow completely several emerging market economies (the Federal Reserve Bank of New York took the unprecedented step of facilitating a bailout of the LTCM, out of fear that a forced liquidation might ravage world markets. A first example of the Too Big To Fail Theory).
The guys of LTCM were extremely good in spot arbitrage opportunities in the bond market; in practice, as soon the correlation between the price of 2 bonds (example T Bond vs. France Bond) was different from its historical average calculated by a complex LTCM model, LTCM entered in the market shorting one bond and buying the other and gaining on the price difference.
LTCM was extremely reliant of its own VAR Risk Model and for every trade was able to calculate not only the probability of a loss but also the amount of the expected loss.
What went wrong? 2 main reasons:
The firm was victim of its own success. Strong money inflow makes it is difficult for LTCM to find arbitrage opportunities in the Bond Market and started to diversify in untested water as the trading of equity and emerging market currencies. The team feels so smart and with the advice of 2 Nobel Professors what could have gone wrong?
The VAR Risk Model was obviously incomplete and didn’t consider nonrational factors such as the behaviour of the political system of emerging countries (as Russia) and the probability that other companies were following the same investment strategies and were forces to sell in the same moment of LTCG causing the collapse of the asset value. TLCM VAR model forecasted a 40% probability for a monthly loss of max $40m. LTCM lost over £4bln in the space of 5 months in 1998….
What can LTCM teach us about the risk assessment preinvestment in a start up?
Mostly investors assume a very high risk of a failure of their seed stage investment. A common valuation methodology is the Harvard Venture Capital Model that assume a very high Discount Rate (WACC) to discount the expected exit proceeds.
The methods assume that 50% is on average the right WACC for the valuation of a startup.
The WACC could fall to maybe 40% if the company is post revenues, with some tractions and not far from breakeven or rise to 60% if the company is still developing the idea (Beta phase).
As the above show the fluctuations in the calculation of the WACC is not a really scientific theory and is mainly based on the assumption of the Investor.
What if could improve the WACC calculation with some more proper assumptions?
Let’s start from the 50% WACC, assuming that 50% is associate for a start up with the following features:
2 Founders (with one previous start up experience)
Monthly Revenues >£20,000
Month on Month Revenues Growth: >25%
Monthly Loss between £10,000 and £20,000
TAM: >£500m
The following table illustrates how differences in the above parameters could increase or reduce the risk associate with the startup and consequently the WACC.
The fictional Start Up Pall Mall has a much higher risk in 4 or the 5 parameters than the average start up (50% WACC) and so a higher WACC.
Standard Case 

 2 Founders (with one previous start up experience)  100% 
 Monthly Revenues >£20,000  100% 
 Month on Month Revenues Growth: >25%  100% 
 Monthly Loss between £10,000 and £20,000  100% 
 TAM: >£500m  100% 
Average  100.0% 
WACC  50.0% 




Start Up Pall Mall 

One Founder (no previous start up experience)  75% 
Monthly Revenues: £50,000  120% 
Month on Month Revenues Growth: 15%  90% 
Monthly Loss: 50,000  75% 
TAM: £250m  80% 
Average  88.0% 
WACC  56.82% 
The investor has still strong discretion in the calculation of the correlation vs the standard parameters but is also able to identify the additional risk factors and calculate a specific WACC for the company under analysis.