The Push and Pull of Knowledge!


    In my musings on valuation, I’ve lengthy described myself as extra of a quantity cruncher than a storyteller, however it’s as a result of I really like numbers for their very own sake, fairly than a passion for summary arithmetic. It’s that love for numbers that has led me at first of every 12 months because the Nineties to take publicly obtainable information on particular person firms, each from their monetary statements and from the markets that they’re listed and traded on, and attempt to make sense of that information for a wide range of causes – to achieve perspective, to make use of in my company monetary evaluation and valuations and to separate info from disinformation . As my entry to information has improved, what began as a handful of datasets in my first information replace in 1994 has expanded to cowl a a lot wider array of statistics than I had initially envisioned, and my 2026 information updates are actually prepared. If you’re curious about what they comprise, please learn on.

The Push and Pull of Knowledge

    After a 12 months throughout which we heard extra speak about information and information facilities than ever earlier than in historical past, normally within the context of how AI will change our lives, it’s value contemplating the draw that information has aways had on not simply companies however on people, in addition to the hazards with the proliferation of information and the belief we placed on that information.

    In a world the place we really feel adrift and unsure, the attraction of information is evident. It offers us a way of management, even when it’s only in passing, and gives us with mechanisms for making selections within the face of uncertainty. 

  1. Sign within the noise: Anybody who has to cost/worth a inventory or assess a undertaking at a agency has to make estimates within the face of contradictions, each in viewpoints and in numbers. All the level of excellent information evaluation is to search out the alerts within the noise, permitting for reasoned judgments, albeit with the popularity that you’ll make errors.
  2. Coping mechanism for uncertainty: Traders and companies, when confronted with uncertainty, usually reply in unhealthy methods, with denial and paralysis as frequent responses. Right here once more, information might help in two methods, first by serving to you image the vary of potential outcomes and second by bringing in instruments (simulations, information visualizations) for incorporating uncertainty into your decision-making. 
  3. Prescription in opposition to tunnel imaginative and prescient: It’s straightforward to get slowed down in particulars, when confronted with having to make funding selections, and lose perspective.  One of many benefits of taking a look at information variations over time and throughout corporations is that it might assist you elevate and regain perspective, separating the stuff that issues loads from that which issues little.
  4. Protect from disinformation: On the threat of getting backlash, I discover that individuals make up stuff and current it as truth. Whereas it’s straightforward in charge social media, which has supplied a megaphone for these fabulists, I learn and listen to statements within the media, ostensibly from specialists, politicians and regulators, that trigger me to do double takes since they don’t seem to be simply unsuitable, however simply provable as unsuitable, with the info.

    Whereas information clearly has advantages, as a data-user, I do know that it comes with prices and penalties, and it behooves us all to concentrate on them.

  1. False precision: It’s plain that attaching a quantity to one thing that worries you, whether or not or not it’s your well being or your funds, can present a way of consolation, however there’s the hazard with treating estimates as info. In considered one of my upcoming posts, as an illustration, I’ll take a look at the historic fairness threat premium, measured by taking a look at what shares have earned, on an annual foundation, over treasury bonds for the final century. The estimate that I’ll present is 7.03% (the common over your entire interval), however that quantity comes with an ordinary error of two.05%, leading to a spread from rather less than 4% (7.03% – 2 × 2.05%) to better than 11%. This estimation error performs out again and again in virtually each quantity that we use in company finance and valuation, and whereas there’s little that may be accomplished about it, its presence ought to animate how we use the info.
  2. The Position of Bias: I’ve lengthy argued that we’re all biased, albeit in various levels and in numerous instructions, and that bias will discover its approach into the alternatives we make. With information, this will play out consciously, the place we use information estimates that feed into our biases and keep away from estimates that work in the other way, however extra dangerously, they will additionally play out subconsciously, within the decisions we make. Whereas it’s true that practitioners are extra uncovered to bias, as a result of their rewards and compensation are sometimes tied to the output of their analysis, the notion that lecturers are someway goal as a result of their work is peer-reviewed is laughable, since their incentive methods create their very own biases. 
  3. Lazy imply reversion: In a collection of posts that I wrote about worth investing, at the least as practiced by lots of its old-time practitioners, I argued that it was constructed round imply reversion, the belief that the world (and markets) will revert again to historic norms. Thus, you purchase low PBV shares, assuming (and hoping) that these PBV ratios will revert to market averages, and argue that the market is overpriced as a result of the PE ratio at this time is way larger than it has been traditionally. That technique is engaging to those that use it, as a result of imply reversion works a lot of the time, however it’s breaks down when markets undergo structural shifts that trigger everlasting departures from the previous. 
  4. The info did it: As we put information on a pedestal, treating the numbers from emerge from it as the reality, there’s additionally the hazard that some analysts who use it view themselves as purely information engineers. Whereas they make suggestions primarily based upon the info, additionally they refuse to take possession for their very own prescriptions, arguing that it’s the information that’s accountable. 

    As the info that we accumulate and have entry to will get richer and deeper, and the instruments that we have now to research that information turn into extra highly effective, there are some who see a utopian world the place this information entry and evaluation results in higher selections and coverage as output. Having watched this information revolution play out in investing and markets, I’m not so certain, at the least within the investing house. Many analysts now complain that they’ve an excessive amount of information, not too little, and battle with information overload. On the identical time, a model of Gresham’s legislation appears to be kicking in, the place dangerous information (or misinformation) usually drives out good information, resulting in worse selections and coverage decisions. My recommendation, gingerly provided, is that as you entry information, it’s caveat emptor, and that you need to do the next with any information (together with my very own):

(a) Think about the biases and priors of the info supplier.

(b) Not use information that comes from black containers, the place suppliers refuse to element how they arrived at numbers.

(c) Crosscheck with alternate information suppliers, for consistency.

Knowledge Protection

    As I discussed initially of this publish, I began my information estimation for purely egocentric causes, which is that I wanted these estimates for my company monetary analyses and valuations. Whereas my sharing of the info could seem altruistic, the reality is that there’s little that’s proprietary or particular about my information evaluation, and virtually anybody with the time and entry to information can do the identical. 

    

Knowledge Sources

    On the threat of stating the plain, you can not do information evaluation with out getting access to uncooked information. In 1993, once I did my first estimates, I subscribed to Worth Line and acquired their company-specific information, which about 2000 US firms and included a subset of things on monetary statements, on a compact disc. I used Worth Line’s trade categorizations to compute trade averages on a number of dozen objects, and offered them in a number of datasets, which I shared with my college students. In 2025, my entry to information has widened, particularly as a result of my NYU affiliation offers me entry S&P Capital IQ and a Bloomberg terminal, which I complement with subscriptions (largely free) to on-line information. It’s value noting that these virtually all the info from these suppliers is within the public area, both within the type of firm filings for disclosure or in authorities macroeconomic information, and the first profit (and it’s a massive one) is simple entry. 

    As my information entry has improved, I’ve added variables to my datasets, however the information objects that I report mirror my company finance and valuation wants. The determine under gives a partial itemizing of a few of these variables:

As you’ll be able to see from looking this record, a lot of the info that I report is on the micro stage, and the one macro information that I report is on variables that I want in valuation, similar to default spreads and fairness threat premiums.   In computing these variables, I’ve tried to remain in line with my very own pondering and instructing and clear about my utilization. As an illustration for consistency, I’ve argued for 3 many years that lease commitments must be handled as debt and that R&D expenditures are capital, not working, bills, and my calculations have at all times mirrored these views, even when they had been at odds with the accounting guidelines. In 2019, the accounting guidelines caught up with my views on lease debt, and whereas the numbers that I report on debt ratios and invested capital are actually nearer to the accounting numbers, I proceed to do my very own computations of lease debt and report on divergences with accounting estimates. With R&D, I stay at odds with accountants, and I report on the affected numbers (like margins and accounting return) with and with out my changes. On the transparency entrance, you’ll find the particulars of how I computed every variable at this hyperlink, and it’s solely potential that you could be not agree with my computation, it’s within the open.

    There are a number of ultimate computational particulars which can be value emphasizing, and particularly so should you plan to make use of this information in your analyses:

  1. With the micro information, I report on trade values fairly than on particular person firms, for 2 causes. The primary is that my uncooked information suppliers are understandably protecting of their company-level information and have a dim view of my entry into that house. The second is that in order for you company-level information for a person firm or perhaps a subset, that information is, for essentially the most half, already obtainable within the monetary filings of the corporate. Put merely, you do not want Capital IQ or Bloomberg to get to the annual stories of a person firm. 
  2. For international statistics, the place firms in numerous international locations are included inside every trade, and report their financials in numerous currencies, I obtain the info transformed into US {dollars}. Thus, numbers which can be in absolute worth (like complete market capitalization) are in US {dollars}, however many of the statistics that I report are ratios or fractions, the place foreign money is just not a problem, at the least for measurement. Thus, the PE ratio that I report can be the identical for any firm in my pattern, whether or not I compute it in US greenback or Chilean pesos, and the identical will be mentioned about accounting ratios (margins, accounting returns).
  3. Whereas computing trade averages could seem to be a trivial computational problem, there are two issues you face in massive datasets of numerous firms. The primary is that there will likely be particular person firms the place the info is lacking or not obtainable, as is the case with PE ratios for firms with unfavorable earnings. The second is that the businesses inside a bunch can fluctuate in measurement with very small and enormous firms within the combine. Consequently, a easy common will likely be a flawed measure for an trade statistic, because it weighs the very small and the very massive firms equally, and whereas a size-weighted common could seem to be a repair, the businesses with lacking information will stay an issue. My resolution, and chances are you’ll not prefer it, it to compute aggregated values of variable, and use these aggregated values to compute the consultant statistics. Thus, my estimate the PE ratio for an trade grouping is obtained by dividing the overall market capitalization of all firms within the grouping by the overall web earnings of all firms (together with cash losers) within the grouping.

    Since my information is now international, I additionally report on these variables not solely throughout all firms globally in every trade group, however for regional sub-groupings:

I’ll admit that this breakdown could look quirky, but it surely displays the historical past of my information updates. The rationale Japan will get its personal grouping is as a result of once I began my information grouping 20 years in the past, it was a a lot bigger a part of each the worldwide financial system and markets. The rising markets grouping has turn into bigger and extra unwieldy over time, as among the international locations on this group had or have acquired developed market standing and as China and India have grown as economies and markets, I’ve began reporting statistics for them individually, along with together with them within the rising markets grouping. Europe, as a area, has turn into extra dispersed in its threat traits, with elements of Southern Europe exhibiting the volatility extra typical of rising markets.

   –   

    Within the first a part of this publish, I famous how bias can skew information evaluation, and one of many greatest sources of bias is sampling, the place you choose a subset of firms and draw the unsuitable conclusions about firms. Thus, utilizing solely the businesses within the S&P 500 or firms that market capitalizations that exceed a billion in your pattern in computing trade averages will yield outcomes that mirror what massive firms are doing or are priced at, and never your entire market. To cut back this sampling bias, I embody all publicly traded firms which have a market value that exceeds zero in my pattern, yielding a complete pattern measurement of 48,156 firms in my information universe. Word that there will likely be some sampling bias nonetheless left insofar as unlisted and privately owned companies are usually not included, however since disclosure necessities for these companies are a lot spottier, it’s unlikely that we’ll have datasets that embody these ignored firms within the pattern within the close to future. 

    By way of geography, the businesses in my pattern span the globe, and I’ll add to my earlier word on regional breakdowns, by wanting on the variety of corporations listed and market capitalizations of firms in every sub-region:

As you’ll be able to see, the USA,  with 5994 corporations and a complete market capitalization of $69.8 trillion, continues to have a dominant share of the worldwide market. Whereas US shares had a great 12 months, up virtually 16.8% within the combination, the US share of the worldwide market dipped barely from the 48.7% on the finish of 2024 to 46.8% on the finish of 2025. One of the best performing sub-region in 2025 was China, up virtually 32.5% in US greenback phrases, and the worst, once more in US greenback phrases, was India, up solely 3.31%. World equities added $26.3 trillion in market capitalization in 2025, up 21.46% for the 12 months.

    Whereas I do report averages by trade group, for 95 trade groupings, these are a part of broader sectors, and within the desk under, you’ll be able to see the breakdown of the general pattern by sector: 

Throughout all international firms, expertise is now the most important share of the market, commanding virtually 22% of total market capitalization, adopted by monetary providers with 17.51% and industrials with 12.76%. There may be vast divergence throughout sectors, by way of market efficiency in 2025, with expertise delivering the best (20.73%) and actual property and utilities the bottom. There may be clearly way more that may be on each the regional and sector analyses that may enrich this evaluation, however that must wait till the following posts

Utilization

    My information is open entry and freely obtainable, and it’s not my place to inform you the way to use it. That mentioned, it behooves me to speak about each the customers that this information is directed at, in addition to the makes use of that it’s best fitted to. 

  1. For practitioners, not educational researchers: The info that I report is for practitioners in company finance, investing and valuation, fairly than educational researchers. Thus, all the information is on the present information hyperlink is information as of the beginning of January 2026, and can be utilized in assessments and evaluation at this time. If you’re doctoral scholar or researcher, you’ll be higher served going to the uncooked information or getting access to a full information service, however should you lack that entry, and wish to obtain and use my trade averages over time, you should utilize the archived information that I’ve, with the caveat being that not all information objects have lengthy histories and my uncooked information sources have modified over time.
  2. Place to begin, not ending level: When you do resolve to make use of any of my information, please do acknowledge that it’s the place to begin in your evaluation, not a magic bullet. Thus, if you’re pricing a metal firm in Thailand, you can begin with the EV/EBITDA a number of that I report for rising market metal firms, however you need to regulate that a number of for the traits of the corporate being analyzed.
  3. Take possession: When you do use my information, whether or not or not it’s on fairness threat premiums or pricing ratios, please attempt to perceive how I compute these numbers (from my lessons or writing) and take possession of the ensuing evaluation. 

When you use my information, and acknowledge me as a supply, I thanks, however you do not want to explicitly ask me for permission. The info is within the public area for use, not for present, and I’m glad that you just had been capable of finding a use for it.

The Damodaran Bot!

       In 2024, I talked concerning the Damodaran Bot, an AI entity that had learn or watched every part that I’ve put on-line (lessons, books, writing, spreadsheets) and talked about what I may do to remain forward of its attain. I argued that AI bots is not going to solely match, however be higher than I’m, at mechanical and rule-based duties, and that my greatest pathways to making a differential benefit was find facets of my work that required multi-disciplinary (numbers plus narrative) and generalist pondering, with instinct and creativeness taking part in a key function. As I regarded on the course of that I went by way of to place my datasets collectively, I noticed that there was no side of it {that a} bot can not do higher and sooner than I can, and I plan to work on involving my bot extra in my information replace subsequent 12 months, with the top sport of getting it take over virtually your entire course of.

   I do suppose that there’s a message right here for companies which can be constructed round accumulating and processing information, and charging excessive costs for that service. Except they will discover different differentials, they’re uncovered to disruption, with AI doing a lot of what they do. Extra usually, to the extent that an excessive amount of quant investing has been constructed round sensible numbers individuals working with massive datasets to eke out extra returns, it’ll turn into tougher, not much less so, with AI within the combine. 

YouTube Video

Hyperlinks to information



Supply hyperlink

Leave a Comment

Discover more from Education for All

Subscribe now to keep reading and get access to the full archive.

Continue reading