Not for the Faint-Hearted: Econometrics as a Valuation Method

As a recent college graduate, I try to involve the material I learned during my college career in my professional pursuits. I took econometrics during my sophomore year, and I have been applying it left and right. Suffice it to say, economists will stop at almost nothing to construct and test an econometric model to most data they can get their hands on.

Currently, I am looking for openings in finance. By default, I looked for jobs where I could apply econometrics. To my surprise, I found little to no application of econometrics in finance. That came to me as a shock for a number of reasons, but mostly because my intuition tells me that econometrics is best fit for finance. As a disclaimer, I would like to state that I do not have an all-encompassing knowledge of all the different applications of econometrics in finance.

I know of one line of work that applies econometrics within finance: trading. My question is the following: how can econometrics be applied to corporate finance, valuation practices in general, and forecasting models used in portfolio management?

I will start with one potential way of implementing econometrics: in an application (that I never submitted, sigh…) to Soros Fund Management for a portfolio generalist position, I produced a short company valuation to substantiate the claim that the company’s share price was undervalued at the time (and it turns out it was, now that it has increased significantly--almost exactly $40--since I quoted its share price about three months ago). Here is how I did it:

- I collected what I thought was relevant financial quotes on the company in question
- I collected the same information for the leading 15 other companies within the industry
- From the data, I constructed a multiple regression that best predicted the price of the company in question
- I added time trend variables
- I added relevant variables (these variables will differ depending on which company, industry, country, and region the company is in)

The model works, but I cannot say with certainty if I got lucky or not; I tested it only once with this instance. I also do not know how to use python, so the task of cross-checking the reliability of this model across different industries is a huge undertaking and time consuming; python really helps in scraping the web for data and could save one plenty of time. I would love to hear your opinions on this method, and I am happy to share with you some of the details of my model. I also would love to hear if you have applied econometrics in any unconventional way and have found success. (Please message me if you are interested in helping me/know how to scrape data using python).

 

Why don't you keep doing this with your PA and see how it goes? Sounds interesting and I am sure some hedge fund or firm would find what you are doing interesting, if not the initiative that you have taken. Write up a short investment thesis for every trade you make (short on powerpoint), have it ready with the data and your P/L, available from your broker statement. Then as you network and meet people, you can show this to them.

I used to do Asia-Pacific PE (kind of like FoF). Now I do something else but happy to try and answer questions on that stuff.
 

Jamoldo, thank you for your advice. It sounds exactly like what I had in mind, so it is affirming to hear this from you. I also was hoping this would give the analysis some more definition as well as attention.

 

I'm not trying to dumb down your analysis, but it looks like it is just an industry comparison. For example take the ratios and FS items (not sure what you mean by quotes) and compare them to the industry averages. If the P/E is low and everything else looks similar, hey its undervalued.

Am I way off?

Big 4 Accounting Recruiting Guide Interview Questions and Answers, Networking Guide and more - Complete 50 page guide.
 

As I mention in one of my replies, this is indeed based off of the comp analysis model. I just tried to leverage it with statistical analysis.

I also tried to address issues of intrinsic values, and for that I generated multiple variables that take into consideration most of what is taken into consideration in a normal comps. My analysis is not any different apart from the use of regressions instead of just comparison without an underlying formula. I hope that helps.

 

Not at all dumbing it down. If anything, thank you for challenging it.

You are also correct (and so are you, itherosuper); this is based on a basic comps model; however, I was trying to develop it a bit further by deriving the regression formula based on statistical significance levels, then conduct the comps analysis. In other words, I was trying to leverage the type of analysis into something more dynamic and fluid (as opposed to static).

 

so the issue with this kind of analysis is that correlation doesn't imply causation - past trends in data you look at don't actually cause a stock price to go up. That is why a good investment analysis requires a catalyst that details a causal mechanism for the stock price. However, establishing causality is extremely tough.

it doesn't make make sense to perform regress for a single stock - a regression is a conditional probability, and if you believe in the frequentist interpretation, investment decisions only on a regression would be a recipe for failure. a frequentist will argue your forecast of what happens only if you could roll the dice many times, but with investing, you typically have one shot.

lastly, there are many types of regressions, depending on what purpose you need it for - forecasting isn't the only thing you do with regression. How you diagnose the fit etc really depends on your end goal and i'd argue a place like soros would use regression for puropses other than forecasting.

 

Thank you for your comprehensive response, I appreciate it.

"so the issue with this kind of analysis is that correlation doesn't imply causation"

I sort of agree, but the wording is getting to me. I would argue that statistical analysis implies causation (but does not conclusively determine causation, its direction, or its presence).

"past trends in data you look at don't actually cause a stock price to go up"

They do in some instances, but past prices often soak up a majority of the variation in the sample, so I avoided using them. What i did instead is I looked at a cross-sectional set of observations.

"That is why a good investment analysis requires a catalyst that details a causal mechanism for the stock price. However, establishing causality is extremely tough."

I agree. I could further explain what I did to pinpoint that catalyst if you would like me to.

"it doesn't make make sense to perform regress for a single stock - a regression is a conditional probability, and if you believe in the frequentist interpretation, investment decisions only on a regression would be a recipe for failure. a frequentist will argue your forecast of what happens only if you could roll the dice many times, but with investing, you typically have one shot."

I might have not explained my methodology thoroughly enough, but I agree with you for the most part. What I did was come up with a number of regressions that are slight variations of each other. I then ran regressions on each of the companies' prices, and just observed which regression formula best described the price fluctuations for each of the companies. I think took that formula and fit it for the company initially in question.

"lastly, there are many types of regressions, depending on what purpose you need it for - forecasting isn't the only thing you do with regression. How you diagnose the fit etc really depends on your end goal and i'd argue a place like soros would use regression for puropses other than forecasting."

Again, I agree. My take on this is that regression analysis provides a limited scope in the sense that it cannot account for what it does not include as a variable. That short-coming leaves most forecasting regression formulas open to what Nassim Talib calls a black swan (or a highly unlikely event that does occur). Moreover, it is tough to account for all variables, ensure the absence of a lurking variable that is distorting your sample, and even coming up with an efficient instrumental variable that could clean up the distortion.

 

I see what you are trying to accomplish but this is by no means a thorough analysis for determining the intrinsic value of a stock. I’d also like to start off by saying I am by no means an expert in this field.

First off, this is only looking at a relative valuation and ignores any DCF application. Also, there are reasons firms may trade at discounts/premiums to peers that you may not be capturing as a variable within your model (which is very likely). You are also assuming past data predicts future data which is also another issue. Also, testing a model for reliability is a difficult task.

It is nice to see you trying to apply what you learned in school but you should probably see why professionals aren’t using models like yours to predict intrinsic value. I suggest you do some research on the types of models used and why, then incorporate any further knowledge you may have to improve what is currently used.

Buy fear, sell cheer
 

Many thanks for your response.

"First off, this is only looking at a relative valuation and ignores any DCF application."

There is no particular reason why DCF is more sound than the example I provided; DCF is simply a method that is used more often than the one I devised, but it is not more correct.

"Also, there are reasons firms may trade at discounts/premiums to peers that you may not be capturing as a variable within your model (which is very likely)."

I do not have access to that information, so I have to assume that the effect of such trading is uniform across all companies (especially since this is a regulatory and compliance issue; there are laws that ensure the uniformity of this effect on all companies).

"You are also assuming past data predicts future data which is also another issue. Also, testing a model for reliability is a difficult task."

In fact, I am not. My data set is a cross-sectional one. I do not have 'past' prices per se; I do, however, have a cross section of industry prices (or prices of different companies) that, in aggregate, would play a role in determining the price of one company (the company in question) within the same industry.

"I suggest you do some research on the types of models used and why, then incorporate any further knowledge you may have to improve what is currently used."

I agree; more research is due. Thank you for your advice.

 

I am not saying a DCF is necessarily superior, but a relative valuation model does rely on current market prices. By only looking at relative values you make the assumption that the market is pricing things correctly which know is not always true (i.e., 2007/2008). Also, discounting cash flows is literally the foundation of finance and the true value of any asset is its discounted cash flows at the required rate of return. I am sure there are plenty of investors that only use relative valuations to make decisions and some may prosper and some may fail, but try presenting an investment thesis without running a DCF first and brace yourself for laughter. Market mispricing is only half the equation.

' "Also, there are reasons firms may trade at discounts/premiums to peers that you may not be capturing as a variable within your model (which is very likely)." I do not have access to that information, so I have to assume that the effect of such trading is uniform across all companies (especially since this is a regulatory and compliance issue; there are laws that ensure the uniformity of this effect on all companies). '

Actually this data is available, run a regression of company’s A (insert ratio here[P/E, P/B, P/S, etc.) against the industry average overtime and there is your answer. And what do you mean when you say “(especially since this is a regulatory and compliance issue; there are laws that ensure the uniformity of this effect on all companies).” There is no compliance issue for me to tell you there is such thing as a conglomerate discount…

Even with cross-sectional data, same issue I raised earlier exists; you are assuming the market is correctly pricing assets. Just because company x looks cheap to peers does not mean that the entire industry isn’t xxx% overvalued (think tech bubble).

Stay interested, most people in finance do not use or understand statistics to its fullest power so keep at it. Just because this model may not be perfect could just mean it isn’t perfect yet. Good luck.

Buy fear, sell cheer
 

As somebody mentioned above, I think you’re making a classic mistake of extrapolating causation from correlation.

A company’s value is determined by its operations, financial results, and myriad of other issues that are not represented in the financial statements. And a rigorous analysis of the past by looking at the company’s book only mirrors the past, which is only half of the analysis. The rest is determined by one’s thoughts about the future, and this surely is not factored into the current and past information.

I think some time-series analysis and stuffs could be useful in predicting macroeconomic variables, but when it comes to individual stocks, I don’t think it adds much merit.

 

Thanks, sanjose04, for your input.

"I think you’re making a classic mistake of extrapolating causation from correlation."

I have not attributed causation; I simply selected a best fit formula for a number of companies within an industry. The formula predicts price ranges (but does not determine causality).

"A company’s value is determined by its operations, financial results, and myriad of other issues that are not represented in the financial statements."

Correct, and I agree. I took into account operations, financial results, and other issues as well. However, this comes off as an unfair criticism because I cannot prove to you that I am not being thorough enough. On the other hand, this model is to be continuously developed and revised for every purpose. I am definitely not putting this model out as 'the model' (except for the one price/company it targets).

 
Best Response

my brother would hate me for saying this because he loves econometrics, but if you're trying to find something that's undervalued, you've got it wrong by applying scientific methods like econometrics to accomplish that. what you don't realize (or maybe you do, you didn't indicate it here though) is that finance & econ are not hard sciences like biology, chemistry, or physics. by that I mean if you drop 2 objects from the same height and they have the same wind resistance/aerodynamic properties, they will reach the same velocity at impact and impact the ground at the same time. if you take an atom with 6 electrons and 6 protons, you will have Carbon. if you add an electron, you'll have an ion (I think, I took chemistry in middle school and never looked back so my memory is shaky, but you get the point). if you combine 2 Hydrogen atoms with 1 Oxygen atom, you have a water molecule. these are not emotional decisions, these are facts. these are laws of nature, without any emotion attached to them.

Economics has tried for centuries to be a science, with supply and demand curves, laws about utility, the words "ceteris paribus," etc etc etc., but it will never be so, because of PEOPLE. People make up the world in which economics exists, so future behaviors cannot be predicted with pinpoint accuracy regardless of how good your math is, because economics and by extension econometrics are not natural sciences. Carbon atoms don't give a damn who's around, they're going to keep spinning their 6 electrons until another force changes them (in which case we can predict the result). this is the reason Graham came up with the margin of safety. you can write equations and add variables and be as exact mathematically as you can when trying to find undervalued companies, but the thing is, the market is not a series of formulas that are predictable over long periods of time, it's an aggregation of humans & humans operating machines.

there is no multiple regression that can tell you what a company's intrinsic value is, you have to guess. sure, you can use DCF and other methods depending on your school of thought and if there's a wide enough margin between what the company's worth is and what the current price is, you buy the stock. to say that you can use math to predict stock price movements is asinine, it cannot be done with consistency over meaningfully long periods of time, unless you are Claude Shannon mixed with Warren Buffett mixed with Ed Thorp mixed with Howard Marks mixed with Seth Klarman mixed with George Soros. it cannot be done. what you can do is invest wisely and be patient so that eventually (if your thesis is right) the market recognizes its errors and realizes a company's intrinsic value. an alternative is that the company keeps creating value over time (like KO), and you never truly realize intrinsic value because it keeps growing.

bottom line is yes, you should have a process by which you analyze companies, and there should be some math in it (in my opinion nothing more complicated than DCF). if you are trying to predict future price movements based on past price movements and saying that you're finding a company that's undervalued, in my mind you've got it wrong. you're doing technical analysis, which has some merit I'll admit (once a trading decision has been made, it's important to look at 50 day moving averages and the like to determine whether we go all in at once or average in or something else), but if you look at the data, it's virtually impossible to make more money with TA in the long term net of transaction costs versus being a fundamental investor.

test this model over a 5-8 year time period so you can experience a full market and business cycle. then, and only then, will you have determined if it works. 3 months is short enough for it to be luck or skill. to take it one step further, I'd argue methods like this don't have any merit unless they can make you money for your own personal life cycle (decades), the downside to that is if you're wrong, you're out of money and can't make it back!

sidebar: if you're talking options, commodities, forex, interest rate derivatives, forwards, swaps, futures on any of that, etc., I think advanced math might have a place. that is not my area of expertise, but I know that those markets are much more technical and might have a place for highly advanced math, but even still, your "edge" is limited.

 

thebrofessor makes an awesome post and I cannot agree more with most of his sentiments (regarding Econ etc as trying to be a hard science - I have seen it and still do on a daily basis) etc, except his last point of testing things out forever, because frankly, it sounds like you want a job right now, rather than work things out over years (things are always changing, what works today may not tomorrow etc)... However, I think you are being discounted too highly on this board, though there is some good discussion of how to maybe refine your processes or thoughts or things you should look into (I've sort of half skimmed a number of those responses). Regardless of how simple or flawed your methods might be, it sounds like you are trying something somewhat different to many at your stage in your career.

Like I said, take some of those results, refine your methods a bit, put them on a few slides, network and try to find an in with people and present your ideas and findings. Even if you are wrong, hopefully you'll be seen as hungry, eager, cheap and someone with potential and get your way in. That's my holistic 50000 foot view at least.

Good Luck

I used to do Asia-Pacific PE (kind of like FoF). Now I do something else but happy to try and answer questions on that stuff.
 

This is a good point, my statement about trying it out was based on this systems merits as an investment strategy. I think if op showed this to someone they would take away that you're strong in math and modeling, both desirable skills, so long as you disclaim it like "I know this might not be the investment strategy this fund uses, but..."

 

Jamoldo, you guessed it. I was desperately trying to devise a way to prove my ability to do valuation work. Nonetheless, I am finding myself still interested in such topics and hope to continue this kind of work on my own. Fortunately, this is not the only project that I worked on to prove my ability.

Thank you for your input!

 

Thanks for the great response, thebrofessor (and sick penname!)

My intention behind setting up this model was the following: how can I capture the market and the current conditions' bearing on it? I was very careful in not trying to predict the price and instead providing a model that agrees, so to speak, with the rest of the companies in the same industry. However, I do think the model would require huge adjustments over time.

In a sentence, I was trying to see where the market was pointing, whether it was where the market itself (according to my model) expected to be, and whether it uniformly suggested the same trajectory for the different companies given similar conditions. I hope that explains my reasoning.

 

It helps, but it doesn't change the fact that I disagree with forecasting something that doesn't behave the same way a pendulum does, unless someone has a manic depressive pendulum. I realize this is your field of study and econometrics is never going away so it's not my intention to bash it, what I'm saying is that once you take forecasting principles in Econ and try to make money off of it in the stock market, you've gone too far, and it won't work.

 

"if you are trying to predict future price movements based on past price movements and saying that you're finding a company that's undervalued, in my mind you've got it wrong." This would be violation of Efficient Market Hypothesis Weakest form... sorry I had to throw this in there...but I enjoyed your detailed explanation. Good stuff thebrofessor.

 

I was just about to type a long tirade but you took the words right out of my mouth. +1

PZ87:

You are describing a multiple regression analysis. In order to do this, you need to first specify a model which you are test (e.g. Price = F(Earnings, Growth, Beta, Margin...) + error). And in order for this to be really useful you need to believe that.....

+1 for you too, finally someone who understands you can't just apply equations blindly to situations and draw hasty conclusions.

@OP: out of curiosity, where did you attend undergrad?

Currently: future neurologist, current psychotherapist Previously: investor relations (top consulting firm), M&A consulting (Big 4), M&A banking (MM)
 

Econometrics is really just statistics with economic variables.. Idk, you make it sound like it's alchemy

As other people have said, seems like you're just describing industry analysis, comparing stock prices against a peer group..

I interned at a non-profit econ research institute, we used econometrics to forecast GDP growth for a bunch of different countries, one of the variables was just lagged GDP data as a kind of all-encompassing lagged variable, similar idea

 

From what I've seen in my 'metrics and time series classes, the assumptions holding your models together aren't really realistic, specially when approaching these assumptions through heuristic understanding. That's my biggest issue with trying to quantify investing like this.

 

Fair, but what would you adjust in my assumptions? I also would like to think that my assumptions would be far more well-thought-out had I had the time to dedicate to this. Unfortunately, I put this model together during my senior spring of college with an above average course load (at a really tough school) and simultaneously doing the job search (plus some partying).

In other words, how would you set up the assumptions so that this model is more realistic?

 

Econometrics and equities don't mix, simply because equities (in my opinion) is an asset class most prone to irrationality, especially when you deal with specific stocks. I mean I COULD see a potential use for say equity indicies (i.e. try comparing SP500 to Dow jones to Nasdaq and see what that yields).

Application of econometrics on corporate finance will fail - simply because it's too 'black box' for a board to accept, if there's a shift in culture perhaps you'd be able to make an inroads.

I think broadly speaking econometrics is useful as an ancillary risk measurement tool, not a primary risk-taking tool (i.e. stock x underpriced, therefore buy it).

 

Hello, greetings from Indonesia. A soon-to-be graduate too. High five.

The Chief Economist of my previous employer use econometrics and regression all the time for macroeconomic analysis, i.e. inflation, GDP, exchange rate. Never for other analysis though. So.

Fortes fortuna adiuvat.
 

Have you already considered "financial econometrics" literature? You can apply fin metrics to portfolio and risk management, statistical arbitrage, empirical asset pricing, derivatives, real estate markets, etc.

Have a look at some fin metrics papers from Andrew Lo (MIT).

 

So Im an analytics consultant and more into the stats side than the finance side. But from what I'm hearing there are 2 components to an equity price.

Intrinsic price based on book value, current cash flows from operations. Perceived price from expected future cash flows taking into account risk (industry, country, legal, economic, etc) and intangibles (brand, smart employees, core competences, etc.).

Intrinsic price I can see as being able to model since its book value, cash flow and discount rates which are pretty easy. But, the problem is perceived price since that is prone to irrationality and belief about what the future will hold. Some variables you can put into your model like expected industry or country growth and maybe even the impact of law changes (Obamacare estimates). This makes statistical techniques probably easier at the macro level. However, the intangibles, particularity at the individual firm level, are by far the hardest to model.

For example, what is the cost of Steve Jobs' death on Apple? Or loss of popularity of myspace.

That said, some equities are more prone to the news than others since while everyone knows Apple and a few of the top 100, there are many firms no one knows anything about and the are probably less responsive to the news. I dont have time to test this but this would be a good project for you, add sentiment and topic models to your regression model. Basically, learn some text mining and natural language processing using scrapped data. Twitter is really good for this and there are many tools already, especially for python. It might not overcome some of the inherent issues of pricing the unexpected, but i have a feeling you can track the number of stories and sentiment which will affect brand and/or market perception of the future cash flows.

And to test the model in sample and out sample as the poster above said. Figure out the mean squared error and/or mean absolute error. Basically, how far does your predicted price deviate.

Also even with the pure unknowns like a Steve Job's death, you can at least guess some scenario ranges and your model hopefully should provide the rest of expected price within a reasonable margin of error.

 

If you use this method to find a large number (dozens) of supposedly undervalued assets and invest in them collectively as a portfolio, it should theoretically be more efficient then just investing in a few assets.

You should also screen the set that you've found for assets that are "undervalued" for a reason and remove them. Also remove every asset for which a reason not to buy it can be figured out by some other means. The remaining number of assets should still be fairly large. This way you get a diversified portfolio of seemingly neutral assets that are supposedly undervalued according to your analysis.

Worst case scenario, you invest in a diversified portfolio that will perform more or less at the market level. Best case, the majority of the portfolio will perform at the market level, while some of the assets were actually correctly identified as undervalued and will outperform. You can also hedge by shorting S&P (or whatever is appropriate if you portfolio turns out less diversified than that).

 

I think this thread shows how many people know some analytical and statistical methods without having really thought about their fundamental use/meaning.

You are describing a multiple regression analysis. In order to do this, you need to first specify a model which you are test (e.g. Price = F(Earnings, Growth, Beta, Margin...) + error). And in order for this to be really useful you need to believe that:

1) You have specified the right functional form i.e. the true model is linear and not quadratic etc. 2) The parameters are fixed over time e.g. if 1% extra growth gets you a turn of multiple today, that'll also be true tomorrow 3) The expected value of the error variable is 0 i.e. there are no other persistent factors affecting the stock price other than those captured by your model and so any deviations will correct over time.

This is pretty ambitious, no?

We will often screen companies using simple regressions i.e. look at P/E multiple versus expected earnings growth. But when you find something that looks "cheap", the difficult task is understanding what that might be i.e. what factors are not in your model that is driving price - it can be any of an almost infinite number of thing e.g. it is a B2B business and the "cheap" company is overexposed to the SME sector which long term is under threat from new technology, the company is getting some good earnings growth in the medium term by cutting costs and investment but people think that'll leave it exposed down the road, the mgmt has a reputation for doing value destroying M&A.

Regression analysis is not useless, but identifying a deviation from the model is only the starting point. Discussing why there is a deviation and what this means is where the real value is.

 

Interesting thread. I think econometrics/statistics is one tool in your toolbox, that can be used to identify and quantify relationships that are not readily apparent. It's based on hard data and repeatable well established analytical methods. Of course these methods won't tell you why a relationship exists or what is driving it to persist. Fundamental/macro analysis involves a lot of assumptions and guessing...it might be more conceptually sound, but it won't necessarily pan out just because you believe it to be so. There is no perfect method, in some sense it is always a roll of the dice.

 

Voluptas accusantium soluta sed perspiciatis iure. Deleniti minus illo dolores tenetur sint et recusandae. Incidunt ab minus deleniti natus rerum.

Perspiciatis assumenda harum omnis autem quod non. Nesciunt laudantium consequatur sapiente tempore. Necessitatibus et repellat ea maiores sapiente sint rerum. Dignissimos ut tenetur aliquid eius quisquam autem inventore.

 

Velit quo consequatur ex doloribus quisquam. Aperiam qui aliquam nobis maxime facere unde itaque. Rerum ducimus soluta excepturi excepturi amet. Consequatur quae reprehenderit eos enim et aspernatur est. Id deserunt possimus rem totam occaecati aut.

Ut quidem accusamus odit in. Sit quo nam exercitationem possimus et nesciunt exercitationem. Quibusdam voluptates quisquam cum quae sed.

Eum possimus quo officiis vero eligendi saepe eum. Ab fugiat eos omnis molestiae. Neque quidem amet officia et dolores.

Nostrum minus magni illo ea explicabo enim sunt sit. Sequi id sit quasi.

 

Et ab reprehenderit aut velit sunt facilis. Quis nihil et quo exercitationem est. Autem earum minus debitis voluptates doloribus pariatur. Et ea sed mollitia praesentium sed. Rerum et molestias porro consequatur aut.

Molestiae sapiente nihil similique iusto qui. Consequatur repudiandae ut deserunt explicabo aliquid possimus vero. Aliquid asperiores odit enim asperiores aliquam modi magnam.

Dignissimos aut sed libero porro quos sed. Porro est repellat harum in. Repellat facere non sed deleniti labore vero maiores qui. Veniam numquam aspernatur aut nobis accusamus. Vel qui repellat possimus commodi rerum. Vitae et laboriosam voluptatibus officia est.

Blanditiis minus aut harum cupiditate maiores est. Et explicabo a tempore numquam ullam. Laudantium nisi quaerat totam veritatis dolore expedita doloribus nihil.

Career Advancement Opportunities

March 2024 Investment Banking

  • Jefferies & Company 02 99.4%
  • Goldman Sachs 19 98.8%
  • Harris Williams & Co. (++) 98.3%
  • Lazard Freres 02 97.7%
  • JPMorgan Chase 03 97.1%

Overall Employee Satisfaction

March 2024 Investment Banking

  • Harris Williams & Co. 18 99.4%
  • JPMorgan Chase 10 98.8%
  • Lazard Freres 05 98.3%
  • Morgan Stanley 07 97.7%
  • William Blair 03 97.1%

Professional Growth Opportunities

March 2024 Investment Banking

  • Lazard Freres 01 99.4%
  • Jefferies & Company 02 98.8%
  • Goldman Sachs 17 98.3%
  • Moelis & Company 07 97.7%
  • JPMorgan Chase 05 97.1%

Total Avg Compensation

March 2024 Investment Banking

  • Director/MD (5) $648
  • Vice President (19) $385
  • Associates (86) $261
  • 3rd+ Year Analyst (13) $181
  • Intern/Summer Associate (33) $170
  • 2nd Year Analyst (66) $168
  • 1st Year Analyst (202) $159
  • Intern/Summer Analyst (144) $101
notes
16 IB Interviews Notes

“... there’s no excuse to not take advantage of the resources out there available to you. Best value for your $ are the...”

Leaderboard

1
redever's picture
redever
99.2
2
Secyh62's picture
Secyh62
99.0
3
Betsy Massar's picture
Betsy Massar
99.0
4
BankonBanking's picture
BankonBanking
99.0
5
dosk17's picture
dosk17
98.9
6
DrApeman's picture
DrApeman
98.9
7
kanon's picture
kanon
98.9
8
CompBanker's picture
CompBanker
98.9
9
GameTheory's picture
GameTheory
98.9
10
Jamoldo's picture
Jamoldo
98.8
success
From 10 rejections to 1 dream investment banking internship

“... I believe it was the single biggest reason why I ended up with an offer...”