{"id":2950,"date":"2026-06-05T13:35:49","date_gmt":"2026-06-05T13:35:49","guid":{"rendered":"https:\/\/blogs.mathworks.com\/finance\/?p=2950"},"modified":"2026-06-05T13:35:49","modified_gmt":"2026-06-05T13:35:49","slug":"from-eviews-to-matlab-in-one-line-reading-workfiles-directly","status":"publish","type":"post","link":"https:\/\/blogs.mathworks.com\/finance\/2026\/06\/05\/from-eviews-to-matlab-in-one-line-reading-workfiles-directly\/","title":{"rendered":"From EViews to MATLAB in One Line: Reading Workfiles Directly"},"content":{"rendered":"

Economists often keep years of work in EViews workfiles: macroeconomic series, model estimates, and curated panel data. The MATLAB Reader for EViews Workfile<\/a> reads .wf1<\/span> and .wf2<\/span> files into MATLAB, with no export step or EViews installation. This post shows how to load a workfile, inspect its metadata, plot series, and run a CAPM regression in MATLAB.<\/p>\n

Getting Started<\/h2>\n

Download the reader from the GitHub repository<\/a> (click the green Code<\/strong> button, then Download ZIP<\/strong>). Unzip it, then add the folder to your MATLAB path:<\/p>\n

addpath(\"MATLAB-Reader-for-EViews-Workfile\"<\/span>)<\/pre>\n
The reader requires MATLAB R2021a or later and no additional toolboxes.<\/p>\n
Loading a Workfile<\/h2>\n
For this walkthrough, we use two publicly available workfiles from the Principles of Econometrics (5th Edition)<\/a> companion site by Hill, Griffiths, and Lim. The first is Klein\u2019s Model I, a foundational macroeconometric dataset covering US consumption, investment, wages, and output from 1920 to 1941.<\/p>\n
Loading it takes one line:<\/p>\n
wf = wfread(\"klein.wf1\"<\/span>);<\/pre>\n
The result is a MATLAB struct. Each field is either a data series or metadata stored in the workfile:<\/p>\n
>> fieldnames(wf)\r\n  ans =\r\n      'CN'\r\n      'E'\r\n      'ELAG'\r\n      'G'\r\n      'I'\r\n      'KLAG'\r\n      'P'\r\n      'PLAG'\r\n      'TIME'\r\n      'TX'\r\n      'W'\r\n      'W1'\r\n      'W2'\r\n      'Y'\r\n      'YEAR'\r\n      'Metadata'<\/pre>\n
Fifteen series: consumption, investment, profits, wages, government spending, capital stock, and more, plus a Metadata<\/span> field containing the metadata stored in the workfile.<\/p>\n
Exploring Metadata<\/h2>\n
The Metadata<\/span> struct tells you the workfile\u2019s frequency, time span, and format without opening EViews:<\/p>\n
>> wf.Metadata.Frequency\r\n  ans = 1                        % Annual<\/i><\/span>\r\n\r\n  >> wf.Metadata.StartYear\r\n  ans = 1920\r\n\r\n  >> wf.Metadata.NumObs\r\n  ans = 22<\/pre>\n
Series Descriptions<\/h2>\n
EViews workfiles store text descriptions for each variable. The reader preserves them:<\/p>\n
>> wf.Metadata.SeriesDescriptions.CN\r\n  ans = '[Display Name] cn  [Description] consumption'<\/span>\r\n\r\n  >> wf.Metadata.SeriesDescriptions.I\r\n  ans = '[Display Name] i  [Description] net investment'<\/span>\r\n\r\n  >> wf.Metadata.SeriesDescriptions.Y\r\n  ans = '[Display Name] y  [Description] total product = cn + i + g - tx'<\/span><\/pre>\n
This embedded documentation means you can identify variables without a separate codebook.<\/p>\n
Observed vs. Derived Series<\/h2>\n
The reader classifies each series as observed<\/strong> (raw data read from the file) or derived<\/strong> (computed from other series by EViews):<\/p>\n
>> wf.Metadata.ObservedSeries\r\n  ans =\r\n      \"CN\"\r\n      \"G\"\r\n      \"I\"\r\n      \"P\"\r\n      \"TX\"\r\n      \"W1\"\r\n      \"W2\"\r\n      \"YEAR\"\r\n\r\n  >> wf.Metadata.DerivedSeries\r\n  ans =\r\n      \"E\"\r\n      \"ELAG\"\r\n      \"KLAG\"\r\n      \"PLAG\"\r\n      \"TIME\"\r\n      \"W\"\r\n      \"Y\"<\/pre>\n
For Klein\u2019s model, the observed variables are consumption (CN), government spending (G), net investment (I), profits (P), taxes (TX), and wages in the private and government sectors (W1, W2). The derived variables, total wages (W), total product (Y), lagged values (PLAG, ELAG, KLAG), were computed within EViews.<\/p>\n
Accessing Individual Series<\/h2>\n
Each series is stored as a MATLAB table with observation labels and numeric values:<\/p>\n
>> wf.CN\r\n  ans =\r\n       obs<\/strong>       CN<\/strong>\r\n      ______    ____\r\n      \"1920\"    39.8\r\n      \"1921\"    41.9\r\n      \"1922\"      45\r\n      \"1923\"    49.2\r\n      ...\r\n      \"1941\"    69.7<\/pre>\n
Output Format Options<\/h2>\n
The reader supports four output formats, selectable at load time:<\/p>\n
% Timetable \u2014 time-indexed, ready for plotting and\r\n  Econometrics Toolbox<\/i><\/span>\r\n  T = wfread(\"klein.wf1\"<\/span>, OutputType=\"timetable\"<\/span>);\r\n\r\n  % Table \u2014 flat table with all series as columns<\/i><\/span>\r\n  tbl = wfread(\"klein.wf1\"<\/span>, OutputType=\"table\"<\/span>);\r\n\r\n  % Matrix \u2014 numeric array only (for direct computation)<\/i><\/span>\r\n  M = wfread(\"klein.wf1\"<\/span>, OutputType=\"matrix\"<\/span>);<\/pre>\n
The timetable output is often the most useful. It works with MATLAB time-series functions, plotting, and toolbox workflows:<\/p>\n
>> T = wfread(\"klein.wf1\"<\/span>, OutputType=\"timetable\"<\/span>);\r\n  >> head(T, 4)\r\n       rowTimes <\/strong>      CN<\/strong>      E<\/strong>      ELAG<\/strong>     G<\/strong>      I<\/strong>      KLAG<\/strong>      P<\/strong>     ...<\/strong>\r\n      ___________    ____    ____    ____    ___    ____    _____    ____\r\n      01-Jan-1920    39.8    44.9     NaN    2.4     2.7    180.1    12.7\r\n      01-Jan-1921    41.9    45.6    44.9    3.9    -0.2    182.8    12.4\r\n      01-Jan-1922      45    50.1    45.6    3.2     1.9    182.6    16.9\r\n      01-Jan-1923    49.2    57.2    50.1    2.8     5.2    184.5    18.4<\/pre>\n
Visualization<\/h2>\n
Three lines give a picture of the US economy through the interwar period and Great Depression:<\/p>\n
T = wfread(\"klein.wf1\"<\/span>, OutputType=\"timetable\"<\/span>);\r\n\r\n  figure;\r\n  plot(T.rowTimes, T.CN, '-o'<\/span>, 'DisplayName'<\/span>, 'Consumption'<\/span>);\r\n  hold on\r\n  plot(T.rowTimes, T.I, '-s'<\/span>, 'DisplayName'<\/span>, 'Investment'<\/span>);\r\n  plot(T.rowTimes, T.Y, '-^'<\/span>, 'DisplayName'<\/span>, 'Total Product'<\/span>);\r\n  hold off\r\n  legend('Location'<\/span>,'northwest'<\/span>); grid on\r\n  ylabel('Billions of 1934 dollars'<\/span>);\r\n  title(\"Klein's Model I: US Macro Data (1920-1941)\"<\/span>);\r\n\r\n  % Shade the Great Depression (1929-1933)<\/i><\/span>\r\n  xregion(datetime(1929,1,1), datetime(1933,1,1), ...\r\n      'FaceColor'<\/span>, [0.8 0.8 0.8], 'EdgeColor'<\/span>, 'none'<\/span>);<\/pre>\n
\"Klein
Klein Model I visualization<\/figcaption><\/figure>\n
The collapse in investment during 1929-1933 is visible, and so is the asymmetry: consumption fell modestly while investment went negative.<\/p>\n
From Workfile to Estimation: CAPM in Five Lines<\/h2>\n
To show a complete load-to-estimation workflow, we switch to a second workfile: capm5.wf1<\/span>, containing monthly stock returns for six firms plus the S&P 500 market portfolio and a risk-free rate (1998-2012).<\/p>\n
The Capital Asset Pricing Model (CAPM) posits that a stock\u2019s excess return is proportional to the market\u2019s excess return, with the slope coefficient \u03b2 measuring systematic risk.<\/p>\n
% Load<\/i><\/span>\r\n  T = wfread(\"capm5.wf1\"<\/span>, OutputType=\"timetable\"<\/span>);\r\n\r\n  % Compute excess returns<\/i><\/span>\r\n  T.ExcessMSFT = T.MSFT - T.RISKFREE;\r\n  T.ExcessMKT  = T.MKT  - T.RISKFREE;\r\n\r\n  % Estimate CAPM beta<\/i><\/span>\r\n  tbl = timetable2table(T, 'ConvertRowTimes'<\/span>, false);\r\n  mdl = fitlm(tbl, 'ExcessMSFT ~ ExcessMKT'<\/span>);\r\n  disp(mdl)<\/pre>\n
Results:<\/p>\n
Linear regression model:\r\n      ExcessMSFT ~ 1 + ExcessMKT\r\n\r\n  Estimated Coefficients:\r\n                     Estimate<\/strong>       SE <\/strong>       tStat<\/strong>       pValue<\/strong>\r\n                     _________    ________    _______    __________\r\n      (Intercept)    0.0032496    0.006036    0.53838      0.59098\r\n      ExcessMKT        1.2018     0.12215     9.8389    1.63e-18\r\n\r\n  R-squared: 0.352,  F-statistic vs. constant model: 96.8, p-value = 1.63e-18<\/pre>\n
Microsoft\u2019s estimated \u03b2 of 1.20 means its excess returns were more sensitive to market excess returns than a beta-one benchmark during this sample period. The R2<\/sup> of 0.352 shows that the single-factor model explains about 35% of the variation in Microsoft excess returns.<\/p>\n
\"CAPM
CAPM scatter plot<\/figcaption><\/figure>\n
The Econometric Modeler app (Econometrics Toolbox) provides an interactive alternative. Load the timetable into the workspace, open the app with econometricModeler<\/span>, and import from the workspace to explore, test, and fit models visually, then generate code when ready to automate.<\/p>\n
What Else Can the Reader Handle?<\/h2>\n
For EViews 7+ binary workfiles, the reader can also extract:<\/p>\n