NFS Treasury Blog

Companies’ plans for 2010 and lessons learnt from the past months

2009-11-19T07:05:00.000+01:00

Irrespective of which industry to ask almost the same answer from all. I.e. the crisis seems to hit all industries almost in the same manner and the preventive and swift measurers taken of the corporate treasury are very similar. Here is a brief list of conclusions.

Liquidity is Paramount

Intensify/strengthened banking relationships and to make an overview to decrease the dependencies on banks, not to be too reliant.
Improve working capital management decreasing AR, collections, disbursements etc.
Improve cash flow forecasting and simplifying the model (migrate way from using spreadsheets instead establishing accurate cash flow forecasting systems and processes).

Counterparty Risk Must be Tightly Managed

Pay closer attention to counterparty exposure and to the ownership structure of counterparties (the obligations of the parent can impact the subsidiaries).
Increase the number of the customers (A/R) who are required to pre-pay before delivery.

Collateral Agreements

Implementing more stringent collateral agreements with counterparties. Introduced and change approach to CSA Thresholds.

Risk Management

Further improved credit & risk processes. Focus on risk management. Need for more feasible risk systems (be able of slice the analysis; country, type of counterparty etc.)

Business Units, Treasury and the Executive Board

Treasury should act as business partner to the Executive Board (provision of scenario analysis, stress tests on liquidity development, etc.). There need to be a greater interaction and integration between treasury and business units to overcome barriers.

The corporate plans for 2010 include closely update all sensitive parameters as the general macro parameters and the variables specific to the own business. The corporate treasuries’ plan to reduce bank debt, increase public debt and ensure refinance/repay maturities well ahead of maturities and much sooner than they would have done prior to the crisis.

55% of the respondents stated they performed the actions necessary. They feel confident with their staged efforts. Of course we all get hindsight, the lesson learnt is to fought harder to close some of those gaps and approximations which existed before the crisis hit, accurately set Risk management routines, have more bank rating follow-up and monitoring, spread out maturity of hedges and upgraded systems earlier.

Please contact me at johanna.lenner@nfs-group.com if you want to receive the complete survey report.

Steps of Collateral Management Process

2009-09-10T14:38:00.005+02:00

The past year’s market turbulence has lead to increased attention to counterparty risk exposures. The market circumstances led to downgrades of many counterparties. How can your organisation manage it?

One area is setting limits on each counterpart. Yet another area is managing the risk on financial counterparts by entering a CSA (Credit Support Annex) as part of an ISDA (International Swaps and Derivatives Association) agreement.

The process to implement Collateral Management through a CSA is based on a number of variables, as:

Legal contracting e.g. standardize over counterparts so you have the same terms.

Process setup e.g. ensure optimal thresholds, payment frequency.

Netting of exposures.

Decide which assets shall be eligible as collateral.

It simplifies to use a structure similar to this:

The European Treasurers’ Peer Group performed a survey and saw a great interest in entering CSA agreements, however only 28% of the respondents had actually done it. Read here for more information.

How can companies deliver such solid Q2 results and what does it mean for the US and European economic outlook?

2009-08-01T13:19:00.012+02:00

Until now, about 80% of the SP500 companies have delivered results exceeding forecasts made by analysts and the mean deviation for the profit has been about +14%. The same pattern has been obvious for European companies as well. These figures are the largest ever since the figures was started to be tracked back in 1993. To me, a lot of questions have popped up during the last couple of days related to the fact that the Q2 results so far have been extremely solid compared to the broader economic outlook, i.e.:

How can this be possible during such a big economic downturn?
Have analyst, as a group, become too pessimistic about the economic outlook and about companies’ capacity to deliver good results even in these bad times?
What are the implications for the outlook and business cycle for a longer period?

I will below elaborate on this topic and discuss my view of what has happened and also what I think will be the future related to company results and the US and European economies.

First of all, let me state that companies results are truly impressive! This is the largest downturn since the Great Depression during the 30th. Still companies earn large amounts of money which never before has been the case during certain difficult market conditions. GDP and industrial production decreases by between 4-8% respectively 15-25% in nearly all of the G8 countries! If we take the industrial sector as an example, the main bunch of these has made tremendous results during Q2 2009! How can then companies earn so much money?

The answer to this question is that companies have been very quick in their response to the economic outlook and been really successful in cutting their costs. Companies have also during this downturn been able to cut cost more than during earlier downturns which contributes to their current fairly good results. Today compared to 20 years ago, companies are more built up in blocks where they buy much more from subcontractors. This means that costs in a much larger extent than before are variable and companies can quickly downsize if demand vanish. If we at the same time take a look at Q2 revenues figures for the same companies, these figures are not at all that impressive compared to the results presented. For further details, see the table below describing some companies in the industrial sector in United States and in Europe:

Revenue figures are much more in line with the broader economic outlook and even lower than analysts pessimistic expectations. So my answer to the question if analyst has become too pessimistic about the economic outlook is NO (!) but they have instead underestimated the new “era” where companies quickly can scale down and cut costs because they are built up in a different way than before. One good example of this is the Swedish company SKF which is the world leader for solutions and products related to rolling bearings. If you go back 20 years in time, SKF would have made large losses during the current market conditions.

But who are really the winners and losers in this new era? Or are there only winners now when companies earn much more money during at least the beginning of a very weak period for the World Economy. I probably surprise many people by using the term “beginning” because most people and economists right now start talking about a recovery. There will for sure be smaller upturns in the shorter perspective due to some very obvious circumstances:

Very large amount of money are currently spent by governments (fiscal policy) in different countries by implementing different stimulus packages.
Central banks around the world runs a very expansive monetary policy where they support the economy by reducing interest rates and print a lot of money – which in the end most probably will cause inflation to increase substantially.
Industrial production and industrial orders can’t decrease by 15-25% and 25-40% respectively without popping back up at least temporarily. The reason for this is that at the moment companies in a large scale are reducing their inventories which means that they most probably becomes “underinvested” for some time with the result that their demand will increase temporarily in the future.

The above circumstances lead to a road more bumpy than smooth. But does it also mean that we are heading for a quick and broad recovery? My answer to this question is NO and I will below explain why this not will be possible.

As mentioned before, the fact that companies at the moment earns a lot of money means that they so far have been very successful cutting costs and cutting staff. But isn’t this all good? If you look in a short range of time and only look at the individual company and its shareholders, i.e. not aggregate companies together and let it sum up to the World Economy – Then it for sure is good! But for the World Economy, it’s probably not good and will work in the opposite direction.

The process of downsizing is still an ongoing process and companies at all levels and scales are cutting back making more and more people redundant. The result is that the global spendable income per capita will continue to decrease for a long time and the demand for companies’ products and services will because of this NOT ramp up in a larger scale. I.e. people can’t spend money they don’t have! So the fact that the crisis now have been ongoing for a fairly long time will contra dictionary result in that it will keep on going. Neither the US or European consumers nor the companies themselves are able to help the world economy to get the “kick start” it would need to start ramping up again.

When we have had such a deep economic crisis, a recovery is only possible where it will go gradually and because of this very slowly. What I mean is that the only alternative for a recovery is a gradual recovery where companies will start hiring again and people will start spending money again because their confidence increases. And these two can’t take place individually so this in turn means that a recovery automatically will be in very small steps. Neither companies nor the consumers in US/Europe will at the moment take the first step to ramp up in the shorter perspective!

But what is our hope then? How are we going to get out of this economic crisis? During the last 6-8 years, the Asian economies have been the key driver of the world economy. In this difficult situation, the development in this region will be even more important for the US and European economies during the next couple of years. So my conclusion is that we should cross our fingers that China, India and the other Asian countries can help the world economy to start a broader recovery!

Extreme value statistic approaches for estimating Value at Risk (VaR) – Results from a study covering the FX market

2009-07-11T23:24:00.001+02:00

I have in two earlier contributions discussed the current high volatility in the FX market respectively how you can use extreme value statistics to overcome the drawback that the most commonly used VaR method encounter.

In this contribution I will present results from a study that I and my colleague Peter Bergström has performed where we compare the three different methods mentioned in the former contribution, i.e. results will be presented comparing:

Assumption about that returns are normally distributed
Block Maxima (Generalized Extreme Value Distribution - BM)
Peak Over Threshold (PoT)

If you would like to review the former contribution about the different methods, press here.

The data used in this study is the same as in the former contribution describing the current high volatility in the FX-market, i.e. we are estimating VaR for the currency pairs USD/SEK and EUR/SEK. For USD/SEK we have used time series from 1982 to 2009. For EUR/SEK we have used time series from 1999 to 2009. Below some tables are presented describing the result from our study.

n is a variable describing how many data points you have in each block, i.e. for BM you divide your dataset in blocks. There is no distinct value which you should use for n – In the literature describing the two methods, descriptions are given for how to estimate a proper value for n. When you have found a proper value for n, you can elaborate by changing n around this value to see if the calculated VaR value changes significantly (which not is the case for us). I have left out the more detailed description here for how to estimate n.

As can be seen in the tables, the assumption about normality gives absolutely the lowest estimate of VaR for both USD/SEK and EUR/SEK. Peak Over Threshold is in the middle for both currency pairs and Block Maxima gives the highest estimate of VaR for both currency pairs. This is in line with that the commonly used VaR method do not consider that financial return series have “fat tails”, i.e. they are NOT normally distributed and that the risk for large losses are larger than anticipated by the normal assumption.

The overall result from the above means that if funds and financial institutions estimate their risk using the normal VaR method - which more or less has become a market standard - this means that they probably underestimate the actual risk in their portfolio!

If you would like to have a more detailed description of our study, please send an email to fredrik.tallsten@nfs-group.com or peter.bergstrom@nfs-group.com.

Bonus programs for sure fulfils their purpose!

2009-06-11T22:56:00.003+02:00

The last couple of month there has been an ongoing debate about bonus programs in US as well as in many European countries. It all started when about 400 top executives at the American insurance company AIG where qualified for getting bonuses of around $165 million for 2008 despite that the company had received more than $170 billion in taxpayer bailout money from the Treasury and Federal Reserve during 2008/2009.

My personal view is that the bonus debate has turned in the wrong direction and that it unfortunately has become very political. In for example Sweden, the government has proclaimed that no bonus programs should be allowed in companies controlled by the Swedish state and that Swedish pension funds should work against bonus programs in private companies where they own stocks.

I truly agree with the criticism about the enormous excess bonuses that from time to time have been paid out to top executives but my opinion is at the same time that it should be possible to become rich by working and not only on Sports! No one argues when a sports star earns $100 million! I am convinced that bonus programs for sure can fulfil their purpose depending on how they are set up. My view is that the individual economical compensation needs to be much more closely connected to the individual’s performance. This will benefit both the company as well as the individuals!

In my mind, the debate should instead be focusing on other important questions related to bonus systems and economical compensation for individuals like:

How to set up an adequate bonus programme that is very closely related to the individual’s performance and at the same time also benefits the employer? This is not an easy task but to get people striving, this is very important!
How to set up a bonus programme that actually reflects the individual’s performance and not only to a large extent reflects the current business cycle? Most bonus programmes tend to be very much correlated with the business cycle where top executives get large bonuses during good times and small or no bonus at bad times. But most probably, people’s performance is distributed evenly over good and bad times which means that it should be likely easy to get a high bonus during bad as well as good times! So my opinion is that a good bonus programme needs to as much as possible scale back the influence of the business cycle.
Where in a company’s organisation and hierarchy is bonus programmes most effective? If a top executive earns $50 million/year in base salary, he/she is probably fairly indifferent if he/she gets another $50 million in bonus. You can’t spend that money anyway! At the same time you know for sure that people that works on the assembly line works more effectively if their salary is based on incentives. My point is twofold here; First, bonus programmes to top executives probably are of little value if the person already has a very high base salary. Second, bonus programmes many times probably works better in the lower part of an organization than for the top executives because for these people it really matters if you can influence your own salary by doing a god job! To answer the question where in the organisation a bonus program is most effective, the important thing is that it is possible to very closely relate it to the individuals performance, see point no 1 above.
Bonus programmes seldom reflects both opportunities and risks. What I mean is that if a stock trader at a trading desk as an example earns very much money one year, he/she will get a very high bonus but the downside is limited where he/she will get only the base salary if not successful. This drives people in certain job roles like in the financial sector to take large risks because risk and return is very closely related to each other.

So my opinion is that the bonus debate has turned in the wrong direction and that bonus programs for sure fulfil their purpose. But, -they should be directed to the right people and the conditions for the programs needs to be set up so they really get people to strive to do an excellent job!

Another aspect of the ongoing debate which is closely related to the discussion about bonus programmes is a more general discussion about the need of flexible salary payments. In these bad times when many companies order intake and production decreases by as much as 40-50% it is obvious that companies either need to make people redundant or decrease salaries and the total working time! There has been a very tough debate about this but to me it seems obvious that companies need to be able to link revenue and salaries to a larger extent, at least in these certain market conditions!

Extreme value statistic approaches for estimating Value at Risk (VaR) in a portfolio

2009-05-10T23:13:00.023+02:00

During the ongoing financial crisis it has become more evident than ever how important it is to be able to measure the risk in your portfolio. But how do you appropriately measure the risk in a portfolio? The most common way is to use a Value at Risk (VaR) model where the Variance-Covariance method is the absolutely most used and which most people in the financial sector have heard about.

I will below shortly go through this method and compare it to two alternative methods which have been established during the last couple of years and which are designed to take into consideration the drawbacks found with the Variance-Covariance method, i.e; Peak-Over-Threshold and Block-Maxima.

Variance-Covariance method
This is as mentioned the absolutely most common way of estimating VaR for a portfolio. Historic data is used where:

The Variance for each individual element is estimated
The Covariance between all elements are estimated

These elements can be presented in a vector/matrix which can be used to fairly easily estimate VaR for a portfolio.

The largest drawback with the Variance-Covariance is that it underestimates the probability of large losses. The Variance-Covariance method assumes that the returns are normally distributed which not is true. If you look at the probability function, it has “fat tails” which means that the probability of getting a larger loss actually is higher than the normal distribution predicts. This means that you will - by using this method - underestimate the probability of large losses which of course is big problem with a risk method.

Extreme value statistic methods
I will below present two different methods which have its base in extreme value statistics and which have been developed much more recently than the Variance-Covariance method. These methods have the advantage that they do not in a systematic way underestimate the probability of large losses which was the case for the Variance-Covariance method. Both methods are fairly complicated and I will below only present the most important steps when calculating VaR.

1. Block Maxima (BM)
In this method, you divide your return series into blocks with typically 20-40 data points in each block and where you sort out the most negative value in each block. After that you will fit your new return series to the “Generalized Extreme Value Distribution”. This will give you estimates of the parameters "kappa", "sigma" and "my" which can be used to calculate something commonly denoted as “Block-VaR”.

To calculate daily VaR from “Block-VaR”, you have to estimate something denoted “extremal index”. I will not go into detail here how to estimate the “extremal index” but there are actually three different methods for estimating it; “Block method”, “Block-runs method” and “Runs method”. When you have an estimated value for the “extremal index” it is straight forward to calculate daily VaR.

2. Peak-Over-Threshold (POT)
In this method, you start to estimate a threshold level for your return series where you would like to sort out return series data above (actually below) a certain level. The method for doing this can for example be found at:

http://biblioteket.ehl.lu.se/olle/papers/0002918.pdf.

When you have the threshold level, you should sort out all data in your return series that have losses larger than the threshold level and fit that data to the “Generalised Pareto Distribution”. This will give you estimates of the parameters “sigma” and “gamma” which can be used to fairly easily calculating VaR.

Result from VaR estimates
I will in a later contribution present results from an investigation me and my collegue Peter Bergström have performed comparing the three different methods above. We have for this investigation used the same time series as in the earlier contribution “High volatility in the FX market – A USD/SEK and EUR/SEK” study” presented a couple of weeks ago.

No real effects of the financial crisis so far!

2009-04-27T13:47:00.016+02:00

You have probably during the last 18 month been bombarded through media with information about the current financial crisis which more and more has become a broader economic crisis. Every day, newspapers, television, magazines and other media are full of information about the crisis. But have you personally felt any big effects from the crisis? Most people, at least in Europe, answers NO to this question and I will below explain why this is the case.

If you ask people about the real effects they have seen from the crisis so far, they most probably mention any of these:

Their employers are more wary about spending money, i.e. companies are trying to reduce their costs.
They have heard about that large companies have given notice to a lot of people the last couple of month. Note that this does not mean that people have lost their jobs yet, only that they probably will do that in the near future. If you go back about 6 month, nearly no company had started to make people redundant which means that the trend is fairly new despite that the crisis has been ongoing for over 18 month.

During the same period, i.e. over the last 18 month, most people have got higher wages combined with that the overall inflation has decreased. This means that people’s disposable income has increased substantially, i.e. people can buy more for their money now than they could just a short time ago. These circumstances with higher disposable income combined with low unemployment rates have made most people feel comfortable despite the ongoing economic crisis which probably is the worst since the great depression around 1930.

As mentioned, the crisis so far has not had any big implication for most people but this will change during the next couple of months! We will from now on, very rapidly go over to a new phase in the crisis where:

More and more people actually will loose their jobs making the overall unemployment rate increase.
Disposable income will be flat or go down for many people. I.e. even for the people that keep their job, the economical environment will be tougher with slow salary increases or even salary reductions.
People will be careful spending their money which in turn will lead to that consumption will be slow or fall even further.

This could lead to that companies once again need to reduce their costs with the result that even more people are laid off which in turn means that a negative spiral has started. But for sure the future will be choppy and we will have periods of recovery as well as new downturns. My main point is that this new phase unfortunately will be much more tangible and it is now the crisis actually starts where people will feel the pain!

For more information about unemployment rates for EU27 and US, press at the corresponding link.

High volatility in the FX market – A USD/SEK and EUR/SEK study

2009-04-06T17:21:00.023+02:00

Volatilities in the FX market have not been higher for decades. I have been studying volatilities back to the early 80-ties and the fluctuations we have seen the last 6 months are extreme. Below are two examples of return time series presented [USD/SEK and EUR/SEK].

At the right part of the diagrams, you can see that the span for the return series increases notably where both tops and bottoms have larger magnitudes. The extreme value of about -14% for USD/SEK which occurs between 1980 and 1985 corresponds to when the Swedish Krona was devalued 15.9% against a linked basket of currencies (date: 8th of Oct 1982). Click at the links below to find the corresponding cross currency diagrams for USD/SEK and EUR/SEK for the last 12 months. The Swedish Krona has depreciated remarkably since July 2008. This has of course to do with the current financial crisis and the trend of “flight to quality” during uncertain times. It can for sure be relevant to ask yourself if it really is “flight to quality” to buy dollar and sell the Swedish Krona during this financial crisis – I will not do that judgement here. I will only conclude that the size of the depreciation and also the volatility is remarkable in a historical perspective!

In close connection to this I am at the moment researching more about VaR where I am studying different methods for estimating VaR. I will come back presenting the usual VaR concept as well as some extreme value statistic methods for estimating VaR. Using extreme value statistic methods for estimating VaR is a fairly new area where you can overcome some drawbacks related to the usual VaR concept.